diff options
author | Dimitrij <kvarkas@gmail.com> | 2022-10-31 00:45:23 +0300 |
---|---|---|
committer | Dimitrij <kvarkas@gmail.com> | 2022-10-31 00:45:23 +0300 |
commit | 302fb2e8ddea1c993552c9a30c02f41d01ca54a9 (patch) | |
tree | d6cf1b32664296ef2cecda33caeafbe39e6695c1 /terminal | |
parent | 59105d9b26363e47f00676bd365b2ac8d4cb536a (diff) | |
parent | 4ff82ab29a22936b78510c68f544a99e677efed3 (diff) |
Diffstat (limited to 'terminal')
-rw-r--r-- | terminal/bidi.c | 3609 | ||||
-rw-r--r-- | terminal/bidi.h | 147 | ||||
-rw-r--r-- | terminal/bidi_gettype.c | 33 | ||||
-rw-r--r-- | terminal/bidi_test.c | 372 | ||||
-rw-r--r-- | terminal/terminal.c | 7910 | ||||
-rw-r--r-- | terminal/terminal.h | 563 |
6 files changed, 12634 insertions, 0 deletions
diff --git a/terminal/bidi.c b/terminal/bidi.c new file mode 100644 index 00000000..c17671b6 --- /dev/null +++ b/terminal/bidi.c @@ -0,0 +1,3609 @@ +/* + * Implementation of the Unicode bidirectional and Arabic shaping + * algorithms for PuTTY. + * + * Original version written and kindly contributed to this code base + * by Ahmad Khalifa of Arabeyes. The bidi part was almost completely + * rewritten in 2021 by Simon Tatham to bring it up to date, but the + * shaping part is still the one by the original authors. + * + * Implementation notes: + * + * Algorithm version + * ----------------- + * + * This algorithm is up to date with Unicode Standard Annex #9 + * revision 44: + * + * https://www.unicode.org/reports/tr9/tr9-44.html + * + * and passes the full conformance test suite in Unicode 14.0.0. + * + * Paragraph and line handling + * --------------------------- + * + * The full Unicode bidi algorithm expects to receive text containing + * multiple paragraphs, together with a decision about how those + * paragraphs are broken up into lines. It calculates embedding levels + * a whole paragraph at a time without considering the line breaks, + * but then the final reordering of the text for display is done to + * each _line_ independently based on the levels computed for the text + * in that line. + * + * This algorithm omits all of that, because it's intended for use as + * a display-time transformation of a text terminal, which doesn't + * preserve enough semantic information to decide what's a paragraph + * break and what is not. So a piece of input text provided to this + * algorithm is always expected to consist of exactly one paragraph + * *and* exactly one line. + * + * Embeddings, overrides and isolates + * ---------------------------------- + * + * This implementation has full support for all the Unicode special + * control characters that modify bidi behaviour, such as + * + * U+202A LEFT-TO-RIGHT EMBEDDING + * U+202B RIGHT-TO-LEFT EMBEDDING + * U+202D LEFT-TO-RIGHT OVERRIDE + * U+202E RIGHT-TO-LEFT OVERRIDE + * U+202C POP DIRECTIONAL FORMATTING + * U+2068 FIRST STRONG ISOLATE + * U+2066 LEFT-TO-RIGHT ISOLATE + * U+2067 RIGHT-TO-LEFT ISOLATE + * U+2069 POP DIRECTIONAL ISOLATE + * + * However, at present, the terminal emulator that is a client of this + * code has no way to pass those in (because they're dropped during + * escape sequence processing and don't get stored in the terminal + * state). Nonetheless, the code is all here, so if the terminal + * emulator becomes able to record those characters at some later + * point, we'll be all set to take account of them during bidi. + * + * But the _main_ purpose of supporting the full bidi algorithm is + * simply that that's the easiest way to be sure it's correct, because + * if you support the whole thing, you can run the full conformance + * test suite. (And I don't 100% believe that restricting to the + * subset of _tests_ valid with a reduced character set will test the + * full set of _functionality_ relevant to the reduced set.) + * + * Retained formatting characters + * ------------------------------ + * + * The standard bidi algorithm, in step X9, deletes assorted + * formatting characters from the text: all the embedding and override + * section initiator characters, the Pop Directional Formatting + * character that closes one of those sections again, and any + * character labelled as Boundary Neutral. So the characters it + * returns are not a _full_ reordering of the input; some input + * characters vanish completely. + * + * This would be fine, if it were not for the fact that - as far as I + * can see - _exactly one_ Unicode code point in the discarded + * category has a wcwidth() of more than 0, namely U+00AD SOFT HYPHEN + * which is a printing character for terminal purposes but has a bidi + * class of BN. + * + * Therefore, we must implement a modified version of the algorithm, + * as described in section 5.2 of TR9, which retains those formatting + * characters so that a client can find out where they ended up in the + * reordering. + * + * Section 5.2 describes a set of modifications to the algorithm that + * are _intended_ to achieve this without changing the rest of the + * behaviour: that is, if you take the output of the modified + * algorithm and delete all the characters that the standard algorithm + * would have removed, you should end up with the remaining characters + * in the same order that the standard algorithm would have delivered. + * However, section 5.2 admits the possibility of error, and says "in + * case of any deviation the explicit algorithm is the normative + * statement for conformance". And indeed, in one or two places I + * found I had to make my own tweaks to the section 5.2 description in + * order to get the whole test suite to pass, because I think the 5.2 + * modifications if taken literally don't quite achieve that. My + * justification is that sentence of 5.2: in case of doubt, the right + * thing is to make the code behave the same as the official + * algorithm. + * + * It's possible that there might still be some undiscovered + * discrepancies between the behaviour of the standard and modified + * algorithms. So, just in case, I've kept in this code the ability to + * implement the _standard_ algorithm too! If you compile with + * -DREMOVE_FORMATTING_CHARS, this code should go back to implementing + * the literal UAX#9 bidi algorithm - so you can run your suspect + * input through both versions, making it much easier to figure out + * why they differ, and in which of the many stages of the algorithm + * the difference was introduced. + * + * However, beware that when compiling in this mode, the do_bidi + * interface to the terminal will stop working, and just abort() when + * called! The only useful thing you can do with this mode is to run + * the companion program bidi_test.c. + */ + +#include <stdlib.h> /* definition of wchar_t */ + +#include "putty.h" +#include "misc.h" +#include "bidi.h" + +typedef struct { + char type; + wchar_t form_b; +} shape_node; + +/* Kept near the actual table, for verification. */ +#define SHAPE_FIRST 0x621 +#define SHAPE_LAST (SHAPE_FIRST + lenof(shapetypes) - 1) + +static const shape_node shapetypes[] = { + /* index, Typ, Iso, Ligature Index*/ + /* 621 */ {SU, 0xFE80}, + /* 622 */ {SR, 0xFE81}, + /* 623 */ {SR, 0xFE83}, + /* 624 */ {SR, 0xFE85}, + /* 625 */ {SR, 0xFE87}, + /* 626 */ {SD, 0xFE89}, + /* 627 */ {SR, 0xFE8D}, + /* 628 */ {SD, 0xFE8F}, + /* 629 */ {SR, 0xFE93}, + /* 62A */ {SD, 0xFE95}, + /* 62B */ {SD, 0xFE99}, + /* 62C */ {SD, 0xFE9D}, + /* 62D */ {SD, 0xFEA1}, + /* 62E */ {SD, 0xFEA5}, + /* 62F */ {SR, 0xFEA9}, + /* 630 */ {SR, 0xFEAB}, + /* 631 */ {SR, 0xFEAD}, + /* 632 */ {SR, 0xFEAF}, + /* 633 */ {SD, 0xFEB1}, + /* 634 */ {SD, 0xFEB5}, + /* 635 */ {SD, 0xFEB9}, + /* 636 */ {SD, 0xFEBD}, + /* 637 */ {SD, 0xFEC1}, + /* 638 */ {SD, 0xFEC5}, + /* 639 */ {SD, 0xFEC9}, + /* 63A */ {SD, 0xFECD}, + /* 63B */ {SU, 0x0}, + /* 63C */ {SU, 0x0}, + /* 63D */ {SU, 0x0}, + /* 63E */ {SU, 0x0}, + /* 63F */ {SU, 0x0}, + /* 640 */ {SC, 0x0}, + /* 641 */ {SD, 0xFED1}, + /* 642 */ {SD, 0xFED5}, + /* 643 */ {SD, 0xFED9}, + /* 644 */ {SD, 0xFEDD}, + /* 645 */ {SD, 0xFEE1}, + /* 646 */ {SD, 0xFEE5}, + /* 647 */ {SD, 0xFEE9}, + /* 648 */ {SR, 0xFEED}, + /* 649 */ {SR, 0xFEEF}, /* SD */ + /* 64A */ {SD, 0xFEF1}, + /* 64B */ {SU, 0x0}, + /* 64C */ {SU, 0x0}, + /* 64D */ {SU, 0x0}, + /* 64E */ {SU, 0x0}, + /* 64F */ {SU, 0x0}, + /* 650 */ {SU, 0x0}, + /* 651 */ {SU, 0x0}, + /* 652 */ {SU, 0x0}, + /* 653 */ {SU, 0x0}, + /* 654 */ {SU, 0x0}, + /* 655 */ {SU, 0x0}, + /* 656 */ {SU, 0x0}, + /* 657 */ {SU, 0x0}, + /* 658 */ {SU, 0x0}, + /* 659 */ {SU, 0x0}, + /* 65A */ {SU, 0x0}, + /* 65B */ {SU, 0x0}, + /* 65C */ {SU, 0x0}, + /* 65D */ {SU, 0x0}, + /* 65E */ {SU, 0x0}, + /* 65F */ {SU, 0x0}, + /* 660 */ {SU, 0x0}, + /* 661 */ {SU, 0x0}, + /* 662 */ {SU, 0x0}, + /* 663 */ {SU, 0x0}, + /* 664 */ {SU, 0x0}, + /* 665 */ {SU, 0x0}, + /* 666 */ {SU, 0x0}, + /* 667 */ {SU, 0x0}, + /* 668 */ {SU, 0x0}, + /* 669 */ {SU, 0x0}, + /* 66A */ {SU, 0x0}, + /* 66B */ {SU, 0x0}, + /* 66C */ {SU, 0x0}, + /* 66D */ {SU, 0x0}, + /* 66E */ {SU, 0x0}, + /* 66F */ {SU, 0x0}, + /* 670 */ {SU, 0x0}, + /* 671 */ {SR, 0xFB50}, + /* 672 */ {SU, 0x0}, + /* 673 */ {SU, 0x0}, + /* 674 */ {SU, 0x0}, + /* 675 */ {SU, 0x0}, + /* 676 */ {SU, 0x0}, + /* 677 */ {SU, 0x0}, + /* 678 */ {SU, 0x0}, + /* 679 */ {SD, 0xFB66}, + /* 67A */ {SD, 0xFB5E}, + /* 67B */ {SD, 0xFB52}, + /* 67C */ {SU, 0x0}, + /* 67D */ {SU, 0x0}, + /* 67E */ {SD, 0xFB56}, + /* 67F */ {SD, 0xFB62}, + /* 680 */ {SD, 0xFB5A}, + /* 681 */ {SU, 0x0}, + /* 682 */ {SU, 0x0}, + /* 683 */ {SD, 0xFB76}, + /* 684 */ {SD, 0xFB72}, + /* 685 */ {SU, 0x0}, + /* 686 */ {SD, 0xFB7A}, + /* 687 */ {SD, 0xFB7E}, + /* 688 */ {SR, 0xFB88}, + /* 689 */ {SU, 0x0}, + /* 68A */ {SU, 0x0}, + /* 68B */ {SU, 0x0}, + /* 68C */ {SR, 0xFB84}, + /* 68D */ {SR, 0xFB82}, + /* 68E */ {SR, 0xFB86}, + /* 68F */ {SU, 0x0}, + /* 690 */ {SU, 0x0}, + /* 691 */ {SR, 0xFB8C}, + /* 692 */ {SU, 0x0}, + /* 693 */ {SU, 0x0}, + /* 694 */ {SU, 0x0}, + /* 695 */ {SU, 0x0}, + /* 696 */ {SU, 0x0}, + /* 697 */ {SU, 0x0}, + /* 698 */ {SR, 0xFB8A}, + /* 699 */ {SU, 0x0}, + /* 69A */ {SU, 0x0}, + /* 69B */ {SU, 0x0}, + /* 69C */ {SU, 0x0}, + /* 69D */ {SU, 0x0}, + /* 69E */ {SU, 0x0}, + /* 69F */ {SU, 0x0}, + /* 6A0 */ {SU, 0x0}, + /* 6A1 */ {SU, 0x0}, + /* 6A2 */ {SU, 0x0}, + /* 6A3 */ {SU, 0x0}, + /* 6A4 */ {SD, 0xFB6A}, + /* 6A5 */ {SU, 0x0}, + /* 6A6 */ {SD, 0xFB6E}, + /* 6A7 */ {SU, 0x0}, + /* 6A8 */ {SU, 0x0}, + /* 6A9 */ {SD, 0xFB8E}, + /* 6AA */ {SU, 0x0}, + /* 6AB */ {SU, 0x0}, + /* 6AC */ {SU, 0x0}, + /* 6AD */ {SD, 0xFBD3}, + /* 6AE */ {SU, 0x0}, + /* 6AF */ {SD, 0xFB92}, + /* 6B0 */ {SU, 0x0}, + /* 6B1 */ {SD, 0xFB9A}, + /* 6B2 */ {SU, 0x0}, + /* 6B3 */ {SD, 0xFB96}, + /* 6B4 */ {SU, 0x0}, + /* 6B5 */ {SU, 0x0}, + /* 6B6 */ {SU, 0x0}, + /* 6B7 */ {SU, 0x0}, + /* 6B8 */ {SU, 0x0}, + /* 6B9 */ {SU, 0x0}, + /* 6BA */ {SR, 0xFB9E}, + /* 6BB */ {SD, 0xFBA0}, + /* 6BC */ {SU, 0x0}, + /* 6BD */ {SU, 0x0}, + /* 6BE */ {SD, 0xFBAA}, + /* 6BF */ {SU, 0x0}, + /* 6C0 */ {SR, 0xFBA4}, + /* 6C1 */ {SD, 0xFBA6}, + /* 6C2 */ {SU, 0x0}, + /* 6C3 */ {SU, 0x0}, + /* 6C4 */ {SU, 0x0}, + /* 6C5 */ {SR, 0xFBE0}, + /* 6C6 */ {SR, 0xFBD9}, + /* 6C7 */ {SR, 0xFBD7}, + /* 6C8 */ {SR, 0xFBDB}, + /* 6C9 */ {SR, 0xFBE2}, + /* 6CA */ {SU, 0x0}, + /* 6CB */ {SR, 0xFBDE}, + /* 6CC */ {SD, 0xFBFC}, + /* 6CD */ {SU, 0x0}, + /* 6CE */ {SU, 0x0}, + /* 6CF */ {SU, 0x0}, + /* 6D0 */ {SU, 0x0}, + /* 6D1 */ {SU, 0x0}, + /* 6D2 */ {SR, 0xFBAE}, +}; + +/* + * Returns the bidi character type of ch. + * + * The data table in this function is constructed from the Unicode + * Character Database version 14.0.0, downloadable from unicode.org at + * the URL + * + * https://www.unicode.org/Public/14.0.0/ucd/ + * + * by the following fragment of Perl: + +perl -ne '@_=split ";"; $num = hex $_[0]; $type = $_[4];' \ + -e '$fl = ($_[1] =~ /First/ ? 1 : $_[1] =~ /Last/ ? 2 : 0);' \ + -e 'if ($type eq $runtype and ($runend == $num-1 or ' \ + -e ' ($fl==2 and $pfl==1))) {$runend = $num;} else { &reset; }' \ + -e '$pfl=$fl; END { &reset }; sub reset {' \ + -e 'printf" {0x%04x, 0x%04x, %s},\n",$runstart,$runend,$runtype' \ + -e ' if defined $runstart and $runtype ne "ON";' \ + -e '$runstart=$runend=$num; $runtype=$type;}' \ + UnicodeData.txt + + */ +unsigned char bidi_getType(int ch) +{ + static const struct { + int first, last, type; + } lookup[] = { + {0x0000, 0x0008, BN}, + {0x0009, 0x0009, S}, + {0x000a, 0x000a, B}, + {0x000b, 0x000b, S}, + {0x000c, 0x000c, WS}, + {0x000d, 0x000d, B}, + {0x000e, 0x001b, BN}, + {0x001c, 0x001e, B}, + {0x001f, 0x001f, S}, + {0x0020, 0x0020, WS}, + {0x0023, 0x0025, ET}, + {0x002b, 0x002b, ES}, + {0x002c, 0x002c, CS}, + {0x002d, 0x002d, ES}, + {0x002e, 0x002f, CS}, + {0x0030, 0x0039, EN}, + {0x003a, 0x003a, CS}, + {0x0041, 0x005a, L}, + {0x0061, 0x007a, L}, + {0x007f, 0x0084, BN}, + {0x0085, 0x0085, B}, + {0x0086, 0x009f, BN}, + {0x00a0, 0x00a0, CS}, + {0x00a2, 0x00a5, ET}, + {0x00aa, 0x00aa, L}, + {0x00ad, 0x00ad, BN}, + {0x00b0, 0x00b1, ET}, + {0x00b2, 0x00b3, EN}, + {0x00b5, 0x00b5, L}, + {0x00b9, 0x00b9, EN}, + {0x00ba, 0x00ba, L}, + {0x00c0, 0x00d6, L}, + {0x00d8, 0x00f6, L}, + {0x00f8, 0x02b8, L}, + {0x02bb, 0x02c1, L}, + {0x02d0, 0x02d1, L}, + {0x02e0, 0x02e4, L}, + {0x02ee, 0x02ee, L}, + {0x0300, 0x036f, NSM}, + {0x0370, 0x0373, L}, + {0x0376, 0x0377, L}, + {0x037a, 0x037d, L}, + {0x037f, 0x037f, L}, + {0x0386, 0x0386, L}, + {0x0388, 0x038a, L}, + {0x038c, 0x038c, L}, + {0x038e, 0x03a1, L}, + {0x03a3, 0x03f5, L}, + {0x03f7, 0x0482, L}, + {0x0483, 0x0489, NSM}, + {0x048a, 0x052f, L}, + {0x0531, 0x0556, L}, + {0x0559, 0x0589, L}, + {0x058f, 0x058f, ET}, + {0x0591, 0x05bd, NSM}, + {0x05be, 0x05be, R}, + {0x05bf, 0x05bf, NSM}, + {0x05c0, 0x05c0, R}, + {0x05c1, 0x05c2, NSM}, + {0x05c3, 0x05c3, R}, + {0x05c4, 0x05c5, NSM}, + {0x05c6, 0x05c6, R}, + {0x05c7, 0x05c7, NSM}, + {0x05d0, 0x05ea, R}, + {0x05ef, 0x05f4, R}, + {0x0600, 0x0605, AN}, + {0x0608, 0x0608, AL}, + {0x0609, 0x060a, ET}, + {0x060b, 0x060b, AL}, + {0x060c, 0x060c, CS}, + {0x060d, 0x060d, AL}, + {0x0610, 0x061a, NSM}, + {0x061b, 0x064a, AL}, + {0x064b, 0x065f, NSM}, + {0x0660, 0x0669, AN}, + {0x066a, 0x066a, ET}, + {0x066b, 0x066c, AN}, + {0x066d, 0x066f, AL}, + {0x0670, 0x0670, NSM}, + {0x0671, 0x06d5, AL}, + {0x06d6, 0x06dc, NSM}, + {0x06dd, 0x06dd, AN}, + {0x06df, 0x06e4, NSM}, + {0x06e5, 0x06e6, AL}, + {0x06e7, 0x06e8, NSM}, + {0x06ea, 0x06ed, NSM}, + {0x06ee, 0x06ef, AL}, + {0x06f0, 0x06f9, EN}, + {0x06fa, 0x070d, AL}, + {0x070f, 0x0710, AL}, + {0x0711, 0x0711, NSM}, + {0x0712, 0x072f, AL}, + {0x0730, 0x074a, NSM}, + {0x074d, 0x07a5, AL}, + {0x07a6, 0x07b0, NSM}, + {0x07b1, 0x07b1, AL}, + {0x07c0, 0x07ea, R}, + {0x07eb, 0x07f3, NSM}, + {0x07f4, 0x07f5, R}, + {0x07fa, 0x07fa, R}, + {0x07fd, 0x07fd, NSM}, + {0x07fe, 0x0815, R}, + {0x0816, 0x0819, NSM}, + {0x081a, 0x081a, R}, + {0x081b, 0x0823, NSM}, + {0x0824, 0x0824, R}, + {0x0825, 0x0827, NSM}, + {0x0828, 0x0828, R}, + {0x0829, 0x082d, NSM}, + {0x0830, 0x083e, R}, + {0x0840, 0x0858, R}, + {0x0859, 0x085b, NSM}, + {0x085e, 0x085e, R}, + {0x0860, 0x086a, AL}, + {0x0870, 0x088e, AL}, + {0x0890, 0x0891, AN}, + {0x0898, 0x089f, NSM}, + {0x08a0, 0x08c9, AL}, + {0x08ca, 0x08e1, NSM}, + {0x08e2, 0x08e2, AN}, + {0x08e3, 0x0902, NSM}, + {0x0903, 0x0939, L}, + {0x093a, 0x093a, NSM}, + {0x093b, 0x093b, L}, + {0x093c, 0x093c, NSM}, + {0x093d, 0x0940, L}, + {0x0941, 0x0948, NSM}, + {0x0949, 0x094c, L}, + {0x094d, 0x094d, NSM}, + {0x094e, 0x0950, L}, + {0x0951, 0x0957, NSM}, + {0x0958, 0x0961, L}, + {0x0962, 0x0963, NSM}, + {0x0964, 0x0980, L}, + {0x0981, 0x0981, NSM}, + {0x0982, 0x0983, L}, + {0x0985, 0x098c, L}, + {0x098f, 0x0990, L}, + {0x0993, 0x09a8, L}, + {0x09aa, 0x09b0, L}, + {0x09b2, 0x09b2, L}, + {0x09b6, 0x09b9, L}, + {0x09bc, 0x09bc, NSM}, + {0x09bd, 0x09c0, L}, + {0x09c1, 0x09c4, NSM}, + {0x09c7, 0x09c8, L}, + {0x09cb, 0x09cc, L}, + {0x09cd, 0x09cd, NSM}, + {0x09ce, 0x09ce, L}, + {0x09d7, 0x09d7, L}, + {0x09dc, 0x09dd, L}, + {0x09df, 0x09e1, L}, + {0x09e2, 0x09e3, NSM}, + {0x09e6, 0x09f1, L}, + {0x09f2, 0x09f3, ET}, + {0x09f4, 0x09fa, L}, + {0x09fb, 0x09fb, ET}, + {0x09fc, 0x09fd, L}, + {0x09fe, 0x09fe, NSM}, + {0x0a01, 0x0a02, NSM}, + {0x0a03, 0x0a03, L}, + {0x0a05, 0x0a0a, L}, + {0x0a0f, 0x0a10, L}, + {0x0a13, 0x0a28, L}, + {0x0a2a, 0x0a30, L}, + {0x0a32, 0x0a33, L}, + {0x0a35, 0x0a36, L}, + {0x0a38, 0x0a39, L}, + {0x0a3c, 0x0a3c, NSM}, + {0x0a3e, 0x0a40, L}, + {0x0a41, 0x0a42, NSM}, + {0x0a47, 0x0a48, NSM}, + {0x0a4b, 0x0a4d, NSM}, + {0x0a51, 0x0a51, NSM}, + {0x0a59, 0x0a5c, L}, + {0x0a5e, 0x0a5e, L}, + {0x0a66, 0x0a6f, L}, + {0x0a70, 0x0a71, NSM}, + {0x0a72, 0x0a74, L}, + {0x0a75, 0x0a75, NSM}, + {0x0a76, 0x0a76, L}, + {0x0a81, 0x0a82, NSM}, + {0x0a83, 0x0a83, L}, + {0x0a85, 0x0a8d, L}, + {0x0a8f, 0x0a91, L}, + {0x0a93, 0x0aa8, L}, + {0x0aaa, 0x0ab0, L}, + {0x0ab2, 0x0ab3, L}, + {0x0ab5, 0x0ab9, L}, + {0x0abc, 0x0abc, NSM}, + {0x0abd, 0x0ac0, L}, + {0x0ac1, 0x0ac5, NSM}, + {0x0ac7, 0x0ac8, NSM}, + {0x0ac9, 0x0ac9, L}, + {0x0acb, 0x0acc, L}, + {0x0acd, 0x0acd, NSM}, + {0x0ad0, 0x0ad0, L}, + {0x0ae0, 0x0ae1, L}, + {0x0ae2, 0x0ae3, NSM}, + {0x0ae6, 0x0af0, L}, + {0x0af1, 0x0af1, ET}, + {0x0af9, 0x0af9, L}, + {0x0afa, 0x0aff, NSM}, + {0x0b01, 0x0b01, NSM}, + {0x0b02, 0x0b03, L}, + {0x0b05, 0x0b0c, L}, + {0x0b0f, 0x0b10, L}, + {0x0b13, 0x0b28, L}, + {0x0b2a, 0x0b30, L}, + {0x0b32, 0x0b33, L}, + {0x0b35, 0x0b39, L}, + {0x0b3c, 0x0b3c, NSM}, + {0x0b3d, 0x0b3e, L}, + {0x0b3f, 0x0b3f, NSM}, + {0x0b40, 0x0b40, L}, + {0x0b41, 0x0b44, NSM}, + {0x0b47, 0x0b48, L}, + {0x0b4b, 0x0b4c, L}, + {0x0b4d, 0x0b4d, NSM}, + {0x0b55, 0x0b56, NSM}, + {0x0b57, 0x0b57, L}, + {0x0b5c, 0x0b5d, L}, + {0x0b5f, 0x0b61, L}, + {0x0b62, 0x0b63, NSM}, + {0x0b66, 0x0b77, L}, + {0x0b82, 0x0b82, NSM}, + {0x0b83, 0x0b83, L}, + {0x0b85, 0x0b8a, L}, + {0x0b8e, 0x0b90, L}, + {0x0b92, 0x0b95, L}, + {0x0b99, 0x0b9a, L}, + {0x0b9c, 0x0b9c, L}, + {0x0b9e, 0x0b9f, L}, + {0x0ba3, 0x0ba4, L}, + {0x0ba8, 0x0baa, L}, + {0x0bae, 0x0bb9, L}, + {0x0bbe, 0x0bbf, L}, + {0x0bc0, 0x0bc0, NSM}, + {0x0bc1, 0x0bc2, L}, + {0x0bc6, 0x0bc8, L}, + {0x0bca, 0x0bcc, L}, + {0x0bcd, 0x0bcd, NSM}, + {0x0bd0, 0x0bd0, L}, + {0x0bd7, 0x0bd7, L}, + {0x0be6, 0x0bf2, L}, + {0x0bf9, 0x0bf9, ET}, + {0x0c00, 0x0c00, NSM}, + {0x0c01, 0x0c03, L}, + {0x0c04, 0x0c04, NSM}, + {0x0c05, 0x0c0c, L}, + {0x0c0e, 0x0c10, L}, + {0x0c12, 0x0c28, L}, + {0x0c2a, 0x0c39, L}, + {0x0c3c, 0x0c3c, NSM}, + {0x0c3d, 0x0c3d, L}, + {0x0c3e, 0x0c40, NSM}, + {0x0c41, 0x0c44, L}, + {0x0c46, 0x0c48, NSM}, + {0x0c4a, 0x0c4d, NSM}, + {0x0c55, 0x0c56, NSM}, + {0x0c58, 0x0c5a, L}, + {0x0c5d, 0x0c5d, L}, + {0x0c60, 0x0c61, L}, + {0x0c62, 0x0c63, NSM}, + {0x0c66, 0x0c6f, L}, + {0x0c77, 0x0c77, L}, + {0x0c7f, 0x0c80, L}, + {0x0c81, 0x0c81, NSM}, + {0x0c82, 0x0c8c, L}, + {0x0c8e, 0x0c90, L}, + {0x0c92, 0x0ca8, L}, + {0x0caa, 0x0cb3, L}, + {0x0cb5, 0x0cb9, L}, + {0x0cbc, 0x0cbc, NSM}, + {0x0cbd, 0x0cc4, L}, + {0x0cc6, 0x0cc8, L}, + {0x0cca, 0x0ccb, L}, + {0x0ccc, 0x0ccd, NSM}, + {0x0cd5, 0x0cd6, L}, + {0x0cdd, 0x0cde, L}, + {0x0ce0, 0x0ce1, L}, + {0x0ce2, 0x0ce3, NSM}, + {0x0ce6, 0x0cef, L}, + {0x0cf1, 0x0cf2, L}, + {0x0d00, 0x0d01, NSM}, + {0x0d02, 0x0d0c, L}, + {0x0d0e, 0x0d10, L}, + {0x0d12, 0x0d3a, L}, + {0x0d3b, 0x0d3c, NSM}, + {0x0d3d, 0x0d40, L}, + {0x0d41, 0x0d44, NSM}, + {0x0d46, 0x0d48, L}, + {0x0d4a, 0x0d4c, L}, + {0x0d4d, 0x0d4d, NSM}, + {0x0d4e, 0x0d4f, L}, + {0x0d54, 0x0d61, L}, + {0x0d62, 0x0d63, NSM}, + {0x0d66, 0x0d7f, L}, + {0x0d81, 0x0d81, NSM}, + {0x0d82, 0x0d83, L}, + {0x0d85, 0x0d96, L}, + {0x0d9a, 0x0db1, L}, + {0x0db3, 0x0dbb, L}, + {0x0dbd, 0x0dbd, L}, + {0x0dc0, 0x0dc6, L}, + {0x0dca, 0x0dca, NSM}, + {0x0dcf, 0x0dd1, L}, + {0x0dd2, 0x0dd4, NSM}, + {0x0dd6, 0x0dd6, NSM}, + {0x0dd8, 0x0ddf, L}, + {0x0de6, 0x0def, L}, + {0x0df2, 0x0df4, L}, + {0x0e01, 0x0e30, L}, + {0x0e31, 0x0e31, NSM}, + {0x0e32, 0x0e33, L}, + {0x0e34, 0x0e3a, NSM}, + {0x0e3f, 0x0e3f, ET}, + {0x0e40, 0x0e46, L}, + {0x0e47, 0x0e4e, NSM}, + {0x0e4f, 0x0e5b, L}, + {0x0e81, 0x0e82, L}, + {0x0e84, 0x0e84, L}, + {0x0e86, 0x0e8a, L}, + {0x0e8c, 0x0ea3, L}, + {0x0ea5, 0x0ea5, L}, + {0x0ea7, 0x0eb0, L}, + {0x0eb1, 0x0eb1, NSM}, + {0x0eb2, 0x0eb3, L}, + {0x0eb4, 0x0ebc, NSM}, + {0x0ebd, 0x0ebd, L}, + {0x0ec0, 0x0ec4, L}, + {0x0ec6, 0x0ec6, L}, + {0x0ec8, 0x0ecd, NSM}, + {0x0ed0, 0x0ed9, L}, + {0x0edc, 0x0edf, L}, + {0x0f00, 0x0f17, L}, + {0x0f18, 0x0f19, NSM}, + {0x0f1a, 0x0f34, L}, + {0x0f35, 0x0f35, NSM}, + {0x0f36, 0x0f36, L}, + {0x0f37, 0x0f37, NSM}, + {0x0f38, 0x0f38, L}, + {0x0f39, 0x0f39, NSM}, + {0x0f3e, 0x0f47, L}, + {0x0f49, 0x0f6c, L}, + {0x0f71, 0x0f7e, NSM}, + {0x0f7f, 0x0f7f, L}, + {0x0f80, 0x0f84, NSM}, + {0x0f85, 0x0f85, L}, + {0x0f86, 0x0f87, NSM}, + {0x0f88, 0x0f8c, L}, + {0x0f8d, 0x0f97, NSM}, + {0x0f99, 0x0fbc, NSM}, + {0x0fbe, 0x0fc5, L}, + {0x0fc6, 0x0fc6, NSM}, + {0x0fc7, 0x0fcc, L}, + {0x0fce, 0x0fda, L}, + {0x1000, 0x102c, L}, + {0x102d, 0x1030, NSM}, + {0x1031, 0x1031, L}, + {0x1032, 0x1037, NSM}, + {0x1038, 0x1038, L}, + {0x1039, 0x103a, NSM}, + {0x103b, 0x103c, L}, + {0x103d, 0x103e, NSM}, + {0x103f, 0x1057, L}, + {0x1058, 0x1059, NSM}, + {0x105a, 0x105d, L}, + {0x105e, 0x1060, NSM}, + {0x1061, 0x1070, L}, + {0x1071, 0x1074, NSM}, + {0x1075, 0x1081, L}, + {0x1082, 0x1082, NSM}, + {0x1083, 0x1084, L}, + {0x1085, 0x1086, NSM}, + {0x1087, 0x108c, L}, + {0x108d, 0x108d, NSM}, + {0x108e, 0x109c, L}, + {0x109d, 0x109d, NSM}, + {0x109e, 0x10c5, L}, + {0x10c7, 0x10c7, L}, + {0x10cd, 0x10cd, L}, + {0x10d0, 0x1248, L}, + {0x124a, 0x124d, L}, + {0x1250, 0x1256, L}, + {0x1258, 0x1258, L}, + {0x125a, 0x125d, L}, + {0x1260, 0x1288, L}, + {0x128a, 0x128d, L}, + {0x1290, 0x12b0, L}, + {0x12b2, 0x12b5, L}, + {0x12b8, 0x12be, L}, + {0x12c0, 0x12c0, L}, + {0x12c2, 0x12c5, L}, + {0x12c8, 0x12d6, L}, + {0x12d8, 0x1310, L}, + {0x1312, 0x1315, L}, + {0x1318, 0x135a, L}, + {0x135d, 0x135f, NSM}, + {0x1360, 0x137c, L}, + {0x1380, 0x138f, L}, + {0x13a0, 0x13f5, L}, + {0x13f8, 0x13fd, L}, + {0x1401, 0x167f, L}, + {0x1680, 0x1680, WS}, + {0x1681, 0x169a, L}, + {0x16a0, 0x16f8, L}, + {0x1700, 0x1711, L}, + {0x1712, 0x1714, NSM}, + {0x1715, 0x1715, L}, + {0x171f, 0x1731, L}, + {0x1732, 0x1733, NSM}, + {0x1734, 0x1736, L}, + {0x1740, 0x1751, L}, + {0x1752, 0x1753, NSM}, + {0x1760, 0x176c, L}, + {0x176e, 0x1770, L}, + {0x1772, 0x1773, NSM}, + {0x1780, 0x17b3, L}, + {0x17b4, 0x17b5, NSM}, + {0x17b6, 0x17b6, L}, + {0x17b7, 0x17bd, NSM}, + {0x17be, 0x17c5, L}, + {0x17c6, 0x17c6, NSM}, + {0x17c7, 0x17c8, L}, + {0x17c9, 0x17d3, NSM}, + {0x17d4, 0x17da, L}, + {0x17db, 0x17db, ET}, + {0x17dc, 0x17dc, L}, + {0x17dd, 0x17dd, NSM}, + {0x17e0, 0x17e9, L}, + {0x180b, 0x180d, NSM}, + {0x180e, 0x180e, BN}, + {0x180f, 0x180f, NSM}, + {0x1810, 0x1819, L}, + {0x1820, 0x1878, L}, + {0x1880, 0x1884, L}, + {0x1885, 0x1886, NSM}, + {0x1887, 0x18a8, L}, + {0x18a9, 0x18a9, NSM}, + {0x18aa, 0x18aa, L}, + {0x18b0, 0x18f5, L}, + {0x1900, 0x191e, L}, + {0x1920, 0x1922, NSM}, + {0x1923, 0x1926, L}, + {0x1927, 0x1928, NSM}, + {0x1929, 0x192b, L}, + {0x1930, 0x1931, L}, + {0x1932, 0x1932, NSM}, + {0x1933, 0x1938, L}, + {0x1939, 0x193b, NSM}, + {0x1946, 0x196d, L}, + {0x1970, 0x1974, L}, + {0x1980, 0x19ab, L}, + {0x19b0, 0x19c9, L}, + {0x19d0, 0x19da, L}, + {0x1a00, 0x1a16, L}, + {0x1a17, 0x1a18, NSM}, + {0x1a19, 0x1a1a, L}, + {0x1a1b, 0x1a1b, NSM}, + {0x1a1e, 0x1a55, L}, + {0x1a56, 0x1a56, NSM}, + {0x1a57, 0x1a57, L}, + {0x1a58, 0x1a5e, NSM}, + {0x1a60, 0x1a60, NSM}, + {0x1a61, 0x1a61, L}, + {0x1a62, 0x1a62, NSM}, + {0x1a63, 0x1a64, L}, + {0x1a65, 0x1a6c, NSM}, + {0x1a6d, 0x1a72, L}, + {0x1a73, 0x1a7c, NSM}, + {0x1a7f, 0x1a7f, NSM}, + {0x1a80, 0x1a89, L}, + {0x1a90, 0x1a99, L}, + {0x1aa0, 0x1aad, L}, + {0x1ab0, 0x1ace, NSM}, + {0x1b00, 0x1b03, NSM}, + {0x1b04, 0x1b33, L}, + {0x1b34, 0x1b34, NSM}, + {0x1b35, 0x1b35, L}, + {0x1b36, 0x1b3a, NSM}, + {0x1b3b, 0x1b3b, L}, + {0x1b3c, 0x1b3c, NSM}, + {0x1b3d, 0x1b41, L}, + {0x1b42, 0x1b42, NSM}, + {0x1b43, 0x1b4c, L}, + {0x1b50, 0x1b6a, L}, + {0x1b6b, 0x1b73, NSM}, + {0x1b74, 0x1b7e, L}, + {0x1b80, 0x1b81, NSM}, + {0x1b82, 0x1ba1, L}, + {0x1ba2, 0x1ba5, NSM}, + {0x1ba6, 0x1ba7, L}, + {0x1ba8, 0x1ba9, NSM}, + {0x1baa, 0x1baa, L}, + {0x1bab, 0x1bad, NSM}, + {0x1bae, 0x1be5, L}, + {0x1be6, 0x1be6, NSM}, + {0x1be7, 0x1be7, L}, + {0x1be8, 0x1be9, NSM}, + {0x1bea, 0x1bec, L}, + {0x1bed, 0x1bed, NSM}, + {0x1bee, 0x1bee, L}, + {0x1bef, 0x1bf1, NSM}, + {0x1bf2, 0x1bf3, L}, + {0x1bfc, 0x1c2b, L}, + {0x1c2c, 0x1c33, NSM}, + {0x1c34, 0x1c35, L}, + {0x1c36, 0x1c37, NSM}, + {0x1c3b, 0x1c49, L}, + {0x1c4d, 0x1c88, L}, + {0x1c90, 0x1cba, L}, + {0x1cbd, 0x1cc7, L}, + {0x1cd0, 0x1cd2, NSM}, + {0x1cd3, 0x1cd3, L}, + {0x1cd4, 0x1ce0, NSM}, + {0x1ce1, 0x1ce1, L}, + {0x1ce2, 0x1ce8, NSM}, + {0x1ce9, 0x1cec, L}, + {0x1ced, 0x1ced, NSM}, + {0x1cee, 0x1cf3, L}, + {0x1cf4, 0x1cf4, NSM}, + {0x1cf5, 0x1cf7, L}, + {0x1cf8, 0x1cf9, NSM}, + {0x1cfa, 0x1cfa, L}, + {0x1d00, 0x1dbf, L}, + {0x1dc0, 0x1dff, NSM}, + {0x1e00, 0x1f15, L}, + {0x1f18, 0x1f1d, L}, + {0x1f20, 0x1f45, L}, + {0x1f48, 0x1f4d, L}, + {0x1f50, 0x1f57, L}, + {0x1f59, 0x1f59, L}, + {0x1f5b, 0x1f5b, L}, + {0x1f5d, 0x1f5d, L}, + {0x1f5f, 0x1f7d, L}, + {0x1f80, 0x1fb4, L}, + {0x1fb6, 0x1fbc, L}, + {0x1fbe, 0x1fbe, L}, + {0x1fc2, 0x1fc4, L}, + {0x1fc6, 0x1fcc, L}, + {0x1fd0, 0x1fd3, L}, + {0x1fd6, 0x1fdb, L}, + {0x1fe0, 0x1fec, L}, + {0x1ff2, 0x1ff4, L}, + {0x1ff6, 0x1ffc, L}, + {0x2000, 0x200a, WS}, + {0x200b, 0x200d, BN}, + {0x200e, 0x200e, L}, + {0x200f, 0x200f, R}, + {0x2028, 0x2028, WS}, + {0x2029, 0x2029, B}, + {0x202a, 0x202a, LRE}, + {0x202b, 0x202b, RLE}, + {0x202c, 0x202c, PDF}, + {0x202d, 0x202d, LRO}, + {0x202e, 0x202e, RLO}, + {0x202f, 0x202f, CS}, + {0x2030, 0x2034, ET}, + {0x2044, 0x2044, CS}, + {0x205f, 0x205f, WS}, + {0x2060, 0x2064, BN}, + {0x2066, 0x2066, LRI}, + {0x2067, 0x2067, RLI}, + {0x2068, 0x2068, FSI}, + {0x2069, 0x2069, PDI}, + {0x206a, 0x206f, BN}, + {0x2070, 0x2070, EN}, + {0x2071, 0x2071, L}, + {0x2074, 0x2079, EN}, + {0x207a, 0x207b, ES}, + {0x207f, 0x207f, L}, + {0x2080, 0x2089, EN}, + {0x208a, 0x208b, ES}, + {0x2090, 0x209c, L}, + {0x20a0, 0x20c0, ET}, + {0x20d0, 0x20f0, NSM}, + {0x2102, 0x2102, L}, + {0x2107, 0x2107, L}, + {0x210a, 0x2113, L}, + {0x2115, 0x2115, L}, + {0x2119, 0x211d, L}, + {0x2124, 0x2124, L}, + {0x2126, 0x2126, L}, + {0x2128, 0x2128, L}, + {0x212a, 0x212d, L}, + {0x212e, 0x212e, ET}, + {0x212f, 0x2139, L}, + {0x213c, 0x213f, L}, + {0x2145, 0x2149, L}, + {0x214e, 0x214f, L}, + {0x2160, 0x2188, L}, + {0x2212, 0x2212, ES}, + {0x2213, 0x2213, ET}, + {0x2336, 0x237a, L}, + {0x2395, 0x2395, L}, + {0x2488, 0x249b, EN}, + {0x249c, 0x24e9, L}, + {0x26ac, 0x26ac, L}, + {0x2800, 0x28ff, L}, + {0x2c00, 0x2ce4, L}, + {0x2ceb, 0x2cee, L}, + {0x2cef, 0x2cf1, NSM}, + {0x2cf2, 0x2cf3, L}, + {0x2d00, 0x2d25, L}, + {0x2d27, 0x2d27, L}, + {0x2d2d, 0x2d2d, L}, + {0x2d30, 0x2d67, L}, + {0x2d6f, 0x2d70, L}, + {0x2d7f, 0x2d7f, NSM}, + {0x2d80, 0x2d96, L}, + {0x2da0, 0x2da6, L}, + {0x2da8, 0x2dae, L}, + {0x2db0, 0x2db6, L}, + {0x2db8, 0x2dbe, L}, + {0x2dc0, 0x2dc6, L}, + {0x2dc8, 0x2dce, L}, + {0x2dd0, 0x2dd6, L}, + {0x2dd8, 0x2dde, L}, + {0x2de0, 0x2dff, NSM}, + {0x3000, 0x3000, WS}, + {0x3005, 0x3007, L}, + {0x3021, 0x3029, L}, + {0x302a, 0x302d, NSM}, + {0x302e, 0x302f, L}, + {0x3031, 0x3035, L}, + {0x3038, 0x303c, L}, + {0x3041, 0x3096, L}, + {0x3099, 0x309a, NSM}, + {0x309d, 0x309f, L}, + {0x30a1, 0x30fa, L}, + {0x30fc, 0x30ff, L}, + {0x3105, 0x312f, L}, + {0x3131, 0x318e, L}, + {0x3190, 0x31bf, L}, + {0x31f0, 0x321c, L}, + {0x3220, 0x324f, L}, + {0x3260, 0x327b, L}, + {0x327f, 0x32b0, L}, + {0x32c0, 0x32cb, L}, + {0x32d0, 0x3376, L}, + {0x337b, 0x33dd, L}, + {0x33e0, 0x33fe, L}, + {0x3400, 0x4dbf, L}, + {0x4e00, 0xa48c, L}, + {0xa4d0, 0xa60c, L}, + {0xa610, 0xa62b, L}, + {0xa640, 0xa66e, L}, + {0xa66f, 0xa672, NSM}, + {0xa674, 0xa67d, NSM}, + {0xa680, 0xa69d, L}, + {0xa69e, 0xa69f, NSM}, + {0xa6a0, 0xa6ef, L}, + {0xa6f0, 0xa6f1, NSM}, + {0xa6f2, 0xa6f7, L}, + {0xa722, 0xa787, L}, + {0xa789, 0xa7ca, L}, + {0xa7d0, 0xa7d1, L}, + {0xa7d3, 0xa7d3, L}, + {0xa7d5, 0xa7d9, L}, + {0xa7f2, 0xa801, L}, + {0xa802, 0xa802, NSM}, + {0xa803, 0xa805, L}, + {0xa806, 0xa806, NSM}, + {0xa807, 0xa80a, L}, + {0xa80b, 0xa80b, NSM}, + {0xa80c, 0xa824, L}, + {0xa825, 0xa826, NSM}, + {0xa827, 0xa827, L}, + {0xa82c, 0xa82c, NSM}, + {0xa830, 0xa837, L}, + {0xa838, 0xa839, ET}, + {0xa840, 0xa873, L}, + {0xa880, 0xa8c3, L}, + {0xa8c4, 0xa8c5, NSM}, + {0xa8ce, 0xa8d9, L}, + {0xa8e0, 0xa8f1, NSM}, + {0xa8f2, 0xa8fe, L}, + {0xa8ff, 0xa8ff, NSM}, + {0xa900, 0xa925, L}, + {0xa926, 0xa92d, NSM}, + {0xa92e, 0xa946, L}, + {0xa947, 0xa951, NSM}, + {0xa952, 0xa953, L}, + {0xa95f, 0xa97c, L}, + {0xa980, 0xa982, NSM}, + {0xa983, 0xa9b2, L}, + {0xa9b3, 0xa9b3, NSM}, + {0xa9b4, 0xa9b5, L}, + {0xa9b6, 0xa9b9, NSM}, + {0xa9ba, 0xa9bb, L}, + {0xa9bc, 0xa9bd, NSM}, + {0xa9be, 0xa9cd, L}, + {0xa9cf, 0xa9d9, L}, + {0xa9de, 0xa9e4, L}, + {0xa9e5, 0xa9e5, NSM}, + {0xa9e6, 0xa9fe, L}, + {0xaa00, 0xaa28, L}, + {0xaa29, 0xaa2e, NSM}, + {0xaa2f, 0xaa30, L}, + {0xaa31, 0xaa32, NSM}, + {0xaa33, 0xaa34, L}, + {0xaa35, 0xaa36, NSM}, + {0xaa40, 0xaa42, L}, + {0xaa43, 0xaa43, NSM}, + {0xaa44, 0xaa4b, L}, + {0xaa4c, 0xaa4c, NSM}, + {0xaa4d, 0xaa4d, L}, + {0xaa50, 0xaa59, L}, + {0xaa5c, 0xaa7b, L}, + {0xaa7c, 0xaa7c, NSM}, + {0xaa7d, 0xaaaf, L}, + {0xaab0, 0xaab0, NSM}, + {0xaab1, 0xaab1, L}, + {0xaab2, 0xaab4, NSM}, + {0xaab5, 0xaab6, L}, + {0xaab7, 0xaab8, NSM}, + {0xaab9, 0xaabd, L}, + {0xaabe, 0xaabf, NSM}, + {0xaac0, 0xaac0, L}, + {0xaac1, 0xaac1, NSM}, + {0xaac2, 0xaac2, L}, + {0xaadb, 0xaaeb, L}, + {0xaaec, 0xaaed, NSM}, + {0xaaee, 0xaaf5, L}, + {0xaaf6, 0xaaf6, NSM}, + {0xab01, 0xab06, L}, + {0xab09, 0xab0e, L}, + {0xab11, 0xab16, L}, + {0xab20, 0xab26, L}, + {0xab28, 0xab2e, L}, + {0xab30, 0xab69, L}, + {0xab70, 0xabe4, L}, + {0xabe5, 0xabe5, NSM}, + {0xabe6, 0xabe7, L}, + {0xabe8, 0xabe8, NSM}, + {0xabe9, 0xabec, L}, + {0xabed, 0xabed, NSM}, + {0xabf0, 0xabf9, L}, + {0xac00, 0xd7a3, L}, + {0xd7b0, 0xd7c6, L}, + {0xd7cb, 0xd7fb, L}, + {0xd800, 0xfa6d, L}, + {0xfa70, 0xfad9, L}, + {0xfb00, 0xfb06, L}, + {0xfb13, 0xfb17, L}, + {0xfb1d, 0xfb1d, R}, + {0xfb1e, 0xfb1e, NSM}, + {0xfb1f, 0xfb28, R}, + {0xfb29, 0xfb29, ES}, + {0xfb2a, 0xfb36, R}, + {0xfb38, 0xfb3c, R}, + {0xfb3e, 0xfb3e, R}, + {0xfb40, 0xfb41, R}, + {0xfb43, 0xfb44, R}, + {0xfb46, 0xfb4f, R}, + {0xfb50, 0xfbc2, AL}, + {0xfbd3, 0xfd3d, AL}, + {0xfd50, 0xfd8f, AL}, + {0xfd92, 0xfdc7, AL}, + {0xfdf0, 0xfdfc, AL}, + {0xfe00, 0xfe0f, NSM}, + {0xfe20, 0xfe2f, NSM}, + {0xfe50, 0xfe50, CS}, + {0xfe52, 0xfe52, CS}, + {0xfe55, 0xfe55, CS}, + {0xfe5f, 0xfe5f, ET}, + {0xfe62, 0xfe63, ES}, + {0xfe69, 0xfe6a, ET}, + {0xfe70, 0xfe74, AL}, + {0xfe76, 0xfefc, AL}, + {0xfeff, 0xfeff, BN}, + {0xff03, 0xff05, ET}, + {0xff0b, 0xff0b, ES}, + {0xff0c, 0xff0c, CS}, + {0xff0d, 0xff0d, ES}, + {0xff0e, 0xff0f, CS}, + {0xff10, 0xff19, EN}, + {0xff1a, 0xff1a, CS}, + {0xff21, 0xff3a, L}, + {0xff41, 0xff5a, L}, + {0xff66, 0xffbe, L}, + {0xffc2, 0xffc7, L}, + {0xffca, 0xffcf, L}, + {0xffd2, 0xffd7, L}, + {0xffda, 0xffdc, L}, + {0xffe0, 0xffe1, ET}, + {0xffe5, 0xffe6, ET}, + {0x10000, 0x1000b, L}, + {0x1000d, 0x10026, L}, + {0x10028, 0x1003a, L}, + {0x1003c, 0x1003d, L}, + {0x1003f, 0x1004d, L}, + {0x10050, 0x1005d, L}, + {0x10080, 0x100fa, L}, + {0x10100, 0x10100, L}, + {0x10102, 0x10102, L}, + {0x10107, 0x10133, L}, + {0x10137, 0x1013f, L}, + {0x1018d, 0x1018e, L}, + {0x101d0, 0x101fc, L}, + {0x101fd, 0x101fd, NSM}, + {0x10280, 0x1029c, L}, + {0x102a0, 0x102d0, L}, + {0x102e0, 0x102e0, NSM}, + {0x102e1, 0x102fb, EN}, + {0x10300, 0x10323, L}, + {0x1032d, 0x1034a, L}, + {0x10350, 0x10375, L}, + {0x10376, 0x1037a, NSM}, + {0x10380, 0x1039d, L}, + {0x1039f, 0x103c3, L}, + {0x103c8, 0x103d5, L}, + {0x10400, 0x1049d, L}, + {0x104a0, 0x104a9, L}, + {0x104b0, 0x104d3, L}, + {0x104d8, 0x104fb, L}, + {0x10500, 0x10527, L}, + {0x10530, 0x10563, L}, + {0x1056f, 0x1057a, L}, + {0x1057c, 0x1058a, L}, + {0x1058c, 0x10592, L}, + {0x10594, 0x10595, L}, + {0x10597, 0x105a1, L}, + {0x105a3, 0x105b1, L}, + {0x105b3, 0x105b9, L}, + {0x105bb, 0x105bc, L}, + {0x10600, 0x10736, L}, + {0x10740, 0x10755, L}, + {0x10760, 0x10767, L}, + {0x10780, 0x10785, L}, + {0x10787, 0x107b0, L}, + {0x107b2, 0x107ba, L}, + {0x10800, 0x10805, R}, + {0x10808, 0x10808, R}, + {0x1080a, 0x10835, R}, + {0x10837, 0x10838, R}, + {0x1083c, 0x1083c, R}, + {0x1083f, 0x10855, R}, + {0x10857, 0x1089e, R}, + {0x108a7, 0x108af, R}, + {0x108e0, 0x108f2, R}, + {0x108f4, 0x108f5, R}, + {0x108fb, 0x1091b, R}, + {0x10920, 0x10939, R}, + {0x1093f, 0x1093f, R}, + {0x10980, 0x109b7, R}, + {0x109bc, 0x109cf, R}, + {0x109d2, 0x10a00, R}, + {0x10a01, 0x10a03, NSM}, + {0x10a05, 0x10a06, NSM}, + {0x10a0c, 0x10a0f, NSM}, + {0x10a10, 0x10a13, R}, + {0x10a15, 0x10a17, R}, + {0x10a19, 0x10a35, R}, + {0x10a38, 0x10a3a, NSM}, + {0x10a3f, 0x10a3f, NSM}, + {0x10a40, 0x10a48, R}, + {0x10a50, 0x10a58, R}, + {0x10a60, 0x10a9f, R}, + {0x10ac0, 0x10ae4, R}, + {0x10ae5, 0x10ae6, NSM}, + {0x10aeb, 0x10af6, R}, + {0x10b00, 0x10b35, R}, + {0x10b40, 0x10b55, R}, + {0x10b58, 0x10b72, R}, + {0x10b78, 0x10b91, R}, + {0x10b99, 0x10b9c, R}, + {0x10ba9, 0x10baf, R}, + {0x10c00, 0x10c48, R}, + {0x10c80, 0x10cb2, R}, + {0x10cc0, 0x10cf2, R}, + {0x10cfa, 0x10cff, R}, + {0x10d00, 0x10d23, AL}, + {0x10d24, 0x10d27, NSM}, + {0x10d30, 0x10d39, AN}, + {0x10e60, 0x10e7e, AN}, + {0x10e80, 0x10ea9, R}, + {0x10eab, 0x10eac, NSM}, + {0x10ead, 0x10ead, R}, + {0x10eb0, 0x10eb1, R}, + {0x10f00, 0x10f27, R}, + {0x10f30, 0x10f45, AL}, + {0x10f46, 0x10f50, NSM}, + {0x10f51, 0x10f59, AL}, + {0x10f70, 0x10f81, R}, + {0x10f82, 0x10f85, NSM}, + {0x10f86, 0x10f89, R}, + {0x10fb0, 0x10fcb, R}, + {0x10fe0, 0x10ff6, R}, + {0x11000, 0x11000, L}, + {0x11001, 0x11001, NSM}, + {0x11002, 0x11037, L}, + {0x11038, 0x11046, NSM}, + {0x11047, 0x1104d, L}, + {0x11066, 0x1106f, L}, + {0x11070, 0x11070, NSM}, + {0x11071, 0x11072, L}, + {0x11073, 0x11074, NSM}, + {0x11075, 0x11075, L}, + {0x1107f, 0x11081, NSM}, + {0x11082, 0x110b2, L}, + {0x110b3, 0x110b6, NSM}, + {0x110b7, 0x110b8, L}, + {0x110b9, 0x110ba, NSM}, + {0x110bb, 0x110c1, L}, + {0x110c2, 0x110c2, NSM}, + {0x110cd, 0x110cd, L}, + {0x110d0, 0x110e8, L}, + {0x110f0, 0x110f9, L}, + {0x11100, 0x11102, NSM}, + {0x11103, 0x11126, L}, + {0x11127, 0x1112b, NSM}, + {0x1112c, 0x1112c, L}, + {0x1112d, 0x11134, NSM}, + {0x11136, 0x11147, L}, + {0x11150, 0x11172, L}, + {0x11173, 0x11173, NSM}, + {0x11174, 0x11176, L}, + {0x11180, 0x11181, NSM}, + {0x11182, 0x111b5, L}, + {0x111b6, 0x111be, NSM}, + {0x111bf, 0x111c8, L}, + {0x111c9, 0x111cc, NSM}, + {0x111cd, 0x111ce, L}, + {0x111cf, 0x111cf, NSM}, + {0x111d0, 0x111df, L}, + {0x111e1, 0x111f4, L}, + {0x11200, 0x11211, L}, + {0x11213, 0x1122e, L}, + {0x1122f, 0x11231, NSM}, + {0x11232, 0x11233, L}, + {0x11234, 0x11234, NSM}, + {0x11235, 0x11235, L}, + {0x11236, 0x11237, NSM}, + {0x11238, 0x1123d, L}, + {0x1123e, 0x1123e, NSM}, + {0x11280, 0x11286, L}, + {0x11288, 0x11288, L}, + {0x1128a, 0x1128d, L}, + {0x1128f, 0x1129d, L}, + {0x1129f, 0x112a9, L}, + {0x112b0, 0x112de, L}, + {0x112df, 0x112df, NSM}, + {0x112e0, 0x112e2, L}, + {0x112e3, 0x112ea, NSM}, + {0x112f0, 0x112f9, L}, + {0x11300, 0x11301, NSM}, + {0x11302, 0x11303, L}, + {0x11305, 0x1130c, L}, + {0x1130f, 0x11310, L}, + {0x11313, 0x11328, L}, + {0x1132a, 0x11330, L}, + {0x11332, 0x11333, L}, + {0x11335, 0x11339, L}, + {0x1133b, 0x1133c, NSM}, + {0x1133d, 0x1133f, L}, + {0x11340, 0x11340, NSM}, + {0x11341, 0x11344, L}, + {0x11347, 0x11348, L}, + {0x1134b, 0x1134d, L}, + {0x11350, 0x11350, L}, + {0x11357, 0x11357, L}, + {0x1135d, 0x11363, L}, + {0x11366, 0x1136c, NSM}, + {0x11370, 0x11374, NSM}, + {0x11400, 0x11437, L}, + {0x11438, 0x1143f, NSM}, + {0x11440, 0x11441, L}, + {0x11442, 0x11444, NSM}, + {0x11445, 0x11445, L}, + {0x11446, 0x11446, NSM}, + {0x11447, 0x1145b, L}, + {0x1145d, 0x1145d, L}, + {0x1145e, 0x1145e, NSM}, + {0x1145f, 0x11461, L}, + {0x11480, 0x114b2, L}, + {0x114b3, 0x114b8, NSM}, + {0x114b9, 0x114b9, L}, + {0x114ba, 0x114ba, NSM}, + {0x114bb, 0x114be, L}, + {0x114bf, 0x114c0, NSM}, + {0x114c1, 0x114c1, L}, + {0x114c2, 0x114c3, NSM}, + {0x114c4, 0x114c7, L}, + {0x114d0, 0x114d9, L}, + {0x11580, 0x115b1, L}, + {0x115b2, 0x115b5, NSM}, + {0x115b8, 0x115bb, L}, + {0x115bc, 0x115bd, NSM}, + {0x115be, 0x115be, L}, + {0x115bf, 0x115c0, NSM}, + {0x115c1, 0x115db, L}, + {0x115dc, 0x115dd, NSM}, + {0x11600, 0x11632, L}, + {0x11633, 0x1163a, NSM}, + {0x1163b, 0x1163c, L}, + {0x1163d, 0x1163d, NSM}, + {0x1163e, 0x1163e, L}, + {0x1163f, 0x11640, NSM}, + {0x11641, 0x11644, L}, + {0x11650, 0x11659, L}, + {0x11680, 0x116aa, L}, + {0x116ab, 0x116ab, NSM}, + {0x116ac, 0x116ac, L}, + {0x116ad, 0x116ad, NSM}, + {0x116ae, 0x116af, L}, + {0x116b0, 0x116b5, NSM}, + {0x116b6, 0x116b6, L}, + {0x116b7, 0x116b7, NSM}, + {0x116b8, 0x116b9, L}, + {0x116c0, 0x116c9, L}, + {0x11700, 0x1171a, L}, + {0x1171d, 0x1171f, NSM}, + {0x11720, 0x11721, L}, + {0x11722, 0x11725, NSM}, + {0x11726, 0x11726, L}, + {0x11727, 0x1172b, NSM}, + {0x11730, 0x11746, L}, + {0x11800, 0x1182e, L}, + {0x1182f, 0x11837, NSM}, + {0x11838, 0x11838, L}, + {0x11839, 0x1183a, NSM}, + {0x1183b, 0x1183b, L}, + {0x118a0, 0x118f2, L}, + {0x118ff, 0x11906, L}, + {0x11909, 0x11909, L}, + {0x1190c, 0x11913, L}, + {0x11915, 0x11916, L}, + {0x11918, 0x11935, L}, + {0x11937, 0x11938, L}, + {0x1193b, 0x1193c, NSM}, + {0x1193d, 0x1193d, L}, + {0x1193e, 0x1193e, NSM}, + {0x1193f, 0x11942, L}, + {0x11943, 0x11943, NSM}, + {0x11944, 0x11946, L}, + {0x11950, 0x11959, L}, + {0x119a0, 0x119a7, L}, + {0x119aa, 0x119d3, L}, + {0x119d4, 0x119d7, NSM}, + {0x119da, 0x119db, NSM}, + {0x119dc, 0x119df, L}, + {0x119e0, 0x119e0, NSM}, + {0x119e1, 0x119e4, L}, + {0x11a00, 0x11a00, L}, + {0x11a01, 0x11a06, NSM}, + {0x11a07, 0x11a08, L}, + {0x11a09, 0x11a0a, NSM}, + {0x11a0b, 0x11a32, L}, + {0x11a33, 0x11a38, NSM}, + {0x11a39, 0x11a3a, L}, + {0x11a3b, 0x11a3e, NSM}, + {0x11a3f, 0x11a46, L}, + {0x11a47, 0x11a47, NSM}, + {0x11a50, 0x11a50, L}, + {0x11a51, 0x11a56, NSM}, + {0x11a57, 0x11a58, L}, + {0x11a59, 0x11a5b, NSM}, + {0x11a5c, 0x11a89, L}, + {0x11a8a, 0x11a96, NSM}, + {0x11a97, 0x11a97, L}, + {0x11a98, 0x11a99, NSM}, + {0x11a9a, 0x11aa2, L}, + {0x11ab0, 0x11af8, L}, + {0x11c00, 0x11c08, L}, + {0x11c0a, 0x11c2f, L}, + {0x11c30, 0x11c36, NSM}, + {0x11c38, 0x11c3d, NSM}, + {0x11c3e, 0x11c45, L}, + {0x11c50, 0x11c6c, L}, + {0x11c70, 0x11c8f, L}, + {0x11c92, 0x11ca7, NSM}, + {0x11ca9, 0x11ca9, L}, + {0x11caa, 0x11cb0, NSM}, + {0x11cb1, 0x11cb1, L}, + {0x11cb2, 0x11cb3, NSM}, + {0x11cb4, 0x11cb4, L}, + {0x11cb5, 0x11cb6, NSM}, + {0x11d00, 0x11d06, L}, + {0x11d08, 0x11d09, L}, + {0x11d0b, 0x11d30, L}, + {0x11d31, 0x11d36, NSM}, + {0x11d3a, 0x11d3a, NSM}, + {0x11d3c, 0x11d3d, NSM}, + {0x11d3f, 0x11d45, NSM}, + {0x11d46, 0x11d46, L}, + {0x11d47, 0x11d47, NSM}, + {0x11d50, 0x11d59, L}, + {0x11d60, 0x11d65, L}, + {0x11d67, 0x11d68, L}, + {0x11d6a, 0x11d8e, L}, + {0x11d90, 0x11d91, NSM}, + {0x11d93, 0x11d94, L}, + {0x11d95, 0x11d95, NSM}, + {0x11d96, 0x11d96, L}, + {0x11d97, 0x11d97, NSM}, + {0x11d98, 0x11d98, L}, + {0x11da0, 0x11da9, L}, + {0x11ee0, 0x11ef2, L}, + {0x11ef3, 0x11ef4, NSM}, + {0x11ef5, 0x11ef8, L}, + {0x11fb0, 0x11fb0, L}, + {0x11fc0, 0x11fd4, L}, + {0x11fdd, 0x11fe0, ET}, + {0x11fff, 0x12399, L}, + {0x12400, 0x1246e, L}, + {0x12470, 0x12474, L}, + {0x12480, 0x12543, L}, + {0x12f90, 0x12ff2, L}, + {0x13000, 0x1342e, L}, + {0x13430, 0x13438, L}, + {0x14400, 0x14646, L}, + {0x16800, 0x16a38, L}, + {0x16a40, 0x16a5e, L}, + {0x16a60, 0x16a69, L}, + {0x16a6e, 0x16abe, L}, + {0x16ac0, 0x16ac9, L}, + {0x16ad0, 0x16aed, L}, + {0x16af0, 0x16af4, NSM}, + {0x16af5, 0x16af5, L}, + {0x16b00, 0x16b2f, L}, + {0x16b30, 0x16b36, NSM}, + {0x16b37, 0x16b45, L}, + {0x16b50, 0x16b59, L}, + {0x16b5b, 0x16b61, L}, + {0x16b63, 0x16b77, L}, + {0x16b7d, 0x16b8f, L}, + {0x16e40, 0x16e9a, L}, + {0x16f00, 0x16f4a, L}, + {0x16f4f, 0x16f4f, NSM}, + {0x16f50, 0x16f87, L}, + {0x16f8f, 0x16f92, NSM}, + {0x16f93, 0x16f9f, L}, + {0x16fe0, 0x16fe1, L}, + {0x16fe3, 0x16fe3, L}, + {0x16fe4, 0x16fe4, NSM}, + {0x16ff0, 0x16ff1, L}, + {0x17000, 0x187f7, L}, + {0x18800, 0x18cd5, L}, + {0x18d00, 0x18d08, L}, + {0x1aff0, 0x1aff3, L}, + {0x1aff5, 0x1affb, L}, + {0x1affd, 0x1affe, L}, + {0x1b000, 0x1b122, L}, + {0x1b150, 0x1b152, L}, + {0x1b164, 0x1b167, L}, + {0x1b170, 0x1b2fb, L}, + {0x1bc00, 0x1bc6a, L}, + {0x1bc70, 0x1bc7c, L}, + {0x1bc80, 0x1bc88, L}, + {0x1bc90, 0x1bc99, L}, + {0x1bc9c, 0x1bc9c, L}, + {0x1bc9d, 0x1bc9e, NSM}, + {0x1bc9f, 0x1bc9f, L}, + {0x1bca0, 0x1bca3, BN}, + {0x1cf00, 0x1cf2d, NSM}, + {0x1cf30, 0x1cf46, NSM}, + {0x1cf50, 0x1cfc3, L}, + {0x1d000, 0x1d0f5, L}, + {0x1d100, 0x1d126, L}, + {0x1d129, 0x1d166, L}, + {0x1d167, 0x1d169, NSM}, + {0x1d16a, 0x1d172, L}, + {0x1d173, 0x1d17a, BN}, + {0x1d17b, 0x1d182, NSM}, + {0x1d183, 0x1d184, L}, + {0x1d185, 0x1d18b, NSM}, + {0x1d18c, 0x1d1a9, L}, + {0x1d1aa, 0x1d1ad, NSM}, + {0x1d1ae, 0x1d1e8, L}, + {0x1d242, 0x1d244, NSM}, + {0x1d2e0, 0x1d2f3, L}, + {0x1d360, 0x1d378, L}, + {0x1d400, 0x1d454, L}, + {0x1d456, 0x1d49c, L}, + {0x1d49e, 0x1d49f, L}, + {0x1d4a2, 0x1d4a2, L}, + {0x1d4a5, 0x1d4a6, L}, + {0x1d4a9, 0x1d4ac, L}, + {0x1d4ae, 0x1d4b9, L}, + {0x1d4bb, 0x1d4bb, L}, + {0x1d4bd, 0x1d4c3, L}, + {0x1d4c5, 0x1d505, L}, + {0x1d507, 0x1d50a, L}, + {0x1d50d, 0x1d514, L}, + {0x1d516, 0x1d51c, L}, + {0x1d51e, 0x1d539, L}, + {0x1d53b, 0x1d53e, L}, + {0x1d540, 0x1d544, L}, + {0x1d546, 0x1d546, L}, + {0x1d54a, 0x1d550, L}, + {0x1d552, 0x1d6a5, L}, + {0x1d6a8, 0x1d6da, L}, + {0x1d6dc, 0x1d714, L}, + {0x1d716, 0x1d74e, L}, + {0x1d750, 0x1d788, L}, + {0x1d78a, 0x1d7c2, L}, + {0x1d7c4, 0x1d7cb, L}, + {0x1d7ce, 0x1d7ff, EN}, + {0x1d800, 0x1d9ff, L}, + {0x1da00, 0x1da36, NSM}, + {0x1da37, 0x1da3a, L}, + {0x1da3b, 0x1da6c, NSM}, + {0x1da6d, 0x1da74, L}, + {0x1da75, 0x1da75, NSM}, + {0x1da76, 0x1da83, L}, + {0x1da84, 0x1da84, NSM}, + {0x1da85, 0x1da8b, L}, + {0x1da9b, 0x1da9f, NSM}, + {0x1daa1, 0x1daaf, NSM}, + {0x1df00, 0x1df1e, L}, + {0x1e000, 0x1e006, NSM}, + {0x1e008, 0x1e018, NSM}, + {0x1e01b, 0x1e021, NSM}, + {0x1e023, 0x1e024, NSM}, + {0x1e026, 0x1e02a, NSM}, + {0x1e100, 0x1e12c, L}, + {0x1e130, 0x1e136, NSM}, + {0x1e137, 0x1e13d, L}, + {0x1e140, 0x1e149, L}, + {0x1e14e, 0x1e14f, L}, + {0x1e290, 0x1e2ad, L}, + {0x1e2ae, 0x1e2ae, NSM}, + {0x1e2c0, 0x1e2eb, L}, + {0x1e2ec, 0x1e2ef, NSM}, + {0x1e2f0, 0x1e2f9, L}, + {0x1e2ff, 0x1e2ff, ET}, + {0x1e7e0, 0x1e7e6, L}, + {0x1e7e8, 0x1e7eb, L}, + {0x1e7ed, 0x1e7ee, L}, + {0x1e7f0, 0x1e7fe, L}, + {0x1e800, 0x1e8c4, R}, + {0x1e8c7, 0x1e8cf, R}, + {0x1e8d0, 0x1e8d6, NSM}, + {0x1e900, 0x1e943, R}, + {0x1e944, 0x1e94a, NSM}, + {0x1e94b, 0x1e94b, R}, + {0x1e950, 0x1e959, R}, + {0x1e95e, 0x1e95f, R}, + {0x1ec71, 0x1ecb4, AL}, + {0x1ed01, 0x1ed3d, AL}, + {0x1ee00, 0x1ee03, AL}, + {0x1ee05, 0x1ee1f, AL}, + {0x1ee21, 0x1ee22, AL}, + {0x1ee24, 0x1ee24, AL}, + {0x1ee27, 0x1ee27, AL}, + {0x1ee29, 0x1ee32, AL}, + {0x1ee34, 0x1ee37, AL}, + {0x1ee39, 0x1ee39, AL}, + {0x1ee3b, 0x1ee3b, AL}, + {0x1ee42, 0x1ee42, AL}, + {0x1ee47, 0x1ee47, AL}, + {0x1ee49, 0x1ee49, AL}, + {0x1ee4b, 0x1ee4b, AL}, + {0x1ee4d, 0x1ee4f, AL}, + {0x1ee51, 0x1ee52, AL}, + {0x1ee54, 0x1ee54, AL}, + {0x1ee57, 0x1ee57, AL}, + {0x1ee59, 0x1ee59, AL}, + {0x1ee5b, 0x1ee5b, AL}, + {0x1ee5d, 0x1ee5d, AL}, + {0x1ee5f, 0x1ee5f, AL}, + {0x1ee61, 0x1ee62, AL}, + {0x1ee64, 0x1ee64, AL}, + {0x1ee67, 0x1ee6a, AL}, + {0x1ee6c, 0x1ee72, AL}, + {0x1ee74, 0x1ee77, AL}, + {0x1ee79, 0x1ee7c, AL}, + {0x1ee7e, 0x1ee7e, AL}, + {0x1ee80, 0x1ee89, AL}, + {0x1ee8b, 0x1ee9b, AL}, + {0x1eea1, 0x1eea3, AL}, + {0x1eea5, 0x1eea9, AL}, + {0x1eeab, 0x1eebb, AL}, + {0x1f100, 0x1f10a, EN}, + {0x1f110, 0x1f12e, L}, + {0x1f130, 0x1f169, L}, + {0x1f170, 0x1f1ac, L}, + {0x1f1e6, 0x1f202, L}, + {0x1f210, 0x1f23b, L}, + {0x1f240, 0x1f248, L}, + {0x1f250, 0x1f251, L}, + {0x1fbf0, 0x1fbf9, EN}, + {0x20000, 0x2a6df, L}, + {0x2a700, 0x2b738, L}, + {0x2b740, 0x2b81d, L}, + {0x2b820, 0x2cea1, L}, + {0x2ceb0, 0x2ebe0, L}, + {0x2f800, 0x2fa1d, L}, + {0x30000, 0x3134a, L}, + {0xe0001, 0xe0001, BN}, + {0xe0020, 0xe007f, BN}, + {0xe0100, 0xe01ef, NSM}, + {0xf0000, 0xffffd, L}, + {0x100000, 0x10fffd, L}, + }; + + int i, j, k; + + i = -1; + j = lenof(lookup); + + while (j - i > 1) { + k = (i + j) / 2; + if (ch < lookup[k].first) + j = k; + else if (ch > lookup[k].last) + i = k; + else + return lookup[k].type; + } + + /* + * If we reach here, the character was not in any of the + * intervals listed in the lookup table. This means we return + * ON (`Other Neutrals'). This is the appropriate code for any + * character genuinely not listed in the Unicode table, and + * also the table above has deliberately left out any + * characters _explicitly_ listed as ON (to save space!). + */ + return ON; +} + +/* + * Return the mirrored version of a glyph. + + * The data table in this function is constructed from the Unicode + * Character Database version 14.0.0, downloadable from unicode.org at + * the URL + * + * https://www.unicode.org/Public/14.0.0/ucd/ + * + * by the following fragment of Perl: + +perl -e ' + while (<<>>) { + chomp; s{\s}{}g; s{#.*$}{}; next unless /./; + @_ = split /;/, $_; + $src = hex $_[0]; $dst = hex $_[1]; + $m{$src}=$dst; $m{$dst}=$src; + } + for $src (sort {$a <=> $b} keys %m) { + printf " {0x%04x, 0x%04x},\n", $src, $m{$src}; + } +' BidiMirroring.txt + + * + * FIXME: there are also glyphs which the text rendering engine is + * supposed to display left-right reflected, since no mirrored glyph + * exists in Unicode itself to indicate the reflected form. Those are + * listed in comments in BidiMirroring.txt. Many of them are + * mathematical, e.g. the square root sign, or set difference + * operator, or integral sign. No API currently exists here to + * communicate the need for that reflected display back to the client. + */ +static unsigned mirror_glyph(unsigned int ch) +{ + static const struct { + unsigned src, dst; + } mirror_pairs[] = { + {0x0028, 0x0029}, + {0x0029, 0x0028}, + {0x003c, 0x003e}, + {0x003e, 0x003c}, + {0x005b, 0x005d}, + {0x005d, 0x005b}, + {0x007b, 0x007d}, + {0x007d, 0x007b}, + {0x00ab, 0x00bb}, + {0x00bb, 0x00ab}, + {0x0f3a, 0x0f3b}, + {0x0f3b, 0x0f3a}, + {0x0f3c, 0x0f3d}, + {0x0f3d, 0x0f3c}, + {0x169b, 0x169c}, + {0x169c, 0x169b}, + {0x2039, 0x203a}, + {0x203a, 0x2039}, + {0x2045, 0x2046}, + {0x2046, 0x2045}, + {0x207d, 0x207e}, + {0x207e, 0x207d}, + {0x208d, 0x208e}, + {0x208e, 0x208d}, + {0x2208, 0x220b}, + {0x2209, 0x220c}, + {0x220a, 0x220d}, + {0x220b, 0x2208}, + {0x220c, 0x2209}, + {0x220d, 0x220a}, + {0x2215, 0x29f5}, + {0x221f, 0x2bfe}, + {0x2220, 0x29a3}, + {0x2221, 0x299b}, + {0x2222, 0x29a0}, + {0x2224, 0x2aee}, + {0x223c, 0x223d}, + {0x223d, 0x223c}, + {0x2243, 0x22cd}, + {0x2245, 0x224c}, + {0x224c, 0x2245}, + {0x2252, 0x2253}, + {0x2253, 0x2252}, + {0x2254, 0x2255}, + {0x2255, 0x2254}, + {0x2264, 0x2265}, + {0x2265, 0x2264}, + {0x2266, 0x2267}, + {0x2267, 0x2266}, + {0x2268, 0x2269}, + {0x2269, 0x2268}, + {0x226a, 0x226b}, + {0x226b, 0x226a}, + {0x226e, 0x226f}, + {0x226f, 0x226e}, + {0x2270, 0x2271}, + {0x2271, 0x2270}, + {0x2272, 0x2273}, + {0x2273, 0x2272}, + {0x2274, 0x2275}, + {0x2275, 0x2274}, + {0x2276, 0x2277}, + {0x2277, 0x2276}, + {0x2278, 0x2279}, + {0x2279, 0x2278}, + {0x227a, 0x227b}, + {0x227b, 0x227a}, + {0x227c, 0x227d}, + {0x227d, 0x227c}, + {0x227e, 0x227f}, + {0x227f, 0x227e}, + {0x2280, 0x2281}, + {0x2281, 0x2280}, + {0x2282, 0x2283}, + {0x2283, 0x2282}, + {0x2284, 0x2285}, + {0x2285, 0x2284}, + {0x2286, 0x2287}, + {0x2287, 0x2286}, + {0x2288, 0x2289}, + {0x2289, 0x2288}, + {0x228a, 0x228b}, + {0x228b, 0x228a}, + {0x228f, 0x2290}, + {0x2290, 0x228f}, + {0x2291, 0x2292}, + {0x2292, 0x2291}, + {0x2298, 0x29b8}, + {0x22a2, 0x22a3}, + {0x22a3, 0x22a2}, + {0x22a6, 0x2ade}, + {0x22a8, 0x2ae4}, + {0x22a9, 0x2ae3}, + {0x22ab, 0x2ae5}, + {0x22b0, 0x22b1}, + {0x22b1, 0x22b0}, + {0x22b2, 0x22b3}, + {0x22b3, 0x22b2}, + {0x22b4, 0x22b5}, + {0x22b5, 0x22b4}, + {0x22b6, 0x22b7}, + {0x22b7, 0x22b6}, + {0x22b8, 0x27dc}, + {0x22c9, 0x22ca}, + {0x22ca, 0x22c9}, + {0x22cb, 0x22cc}, + {0x22cc, 0x22cb}, + {0x22cd, 0x2243}, + {0x22d0, 0x22d1}, + {0x22d1, 0x22d0}, + {0x22d6, 0x22d7}, + {0x22d7, 0x22d6}, + {0x22d8, 0x22d9}, + {0x22d9, 0x22d8}, + {0x22da, 0x22db}, + {0x22db, 0x22da}, + {0x22dc, 0x22dd}, + {0x22dd, 0x22dc}, + {0x22de, 0x22df}, + {0x22df, 0x22de}, + {0x22e0, 0x22e1}, + {0x22e1, 0x22e0}, + {0x22e2, 0x22e3}, + {0x22e3, 0x22e2}, + {0x22e4, 0x22e5}, + {0x22e5, 0x22e4}, + {0x22e6, 0x22e7}, + {0x22e7, 0x22e6}, + {0x22e8, 0x22e9}, + {0x22e9, 0x22e8}, + {0x22ea, 0x22eb}, + {0x22eb, 0x22ea}, + {0x22ec, 0x22ed}, + {0x22ed, 0x22ec}, + {0x22f0, 0x22f1}, + {0x22f1, 0x22f0}, + {0x22f2, 0x22fa}, + {0x22f3, 0x22fb}, + {0x22f4, 0x22fc}, + {0x22f6, 0x22fd}, + {0x22f7, 0x22fe}, + {0x22fa, 0x22f2}, + {0x22fb, 0x22f3}, + {0x22fc, 0x22f4}, + {0x22fd, 0x22f6}, + {0x22fe, 0x22f7}, + {0x2308, 0x2309}, + {0x2309, 0x2308}, + {0x230a, 0x230b}, + {0x230b, 0x230a}, + {0x2329, 0x232a}, + {0x232a, 0x2329}, + {0x2768, 0x2769}, + {0x2769, 0x2768}, + {0x276a, 0x276b}, + {0x276b, 0x276a}, + {0x276c, 0x276d}, + {0x276d, 0x276c}, + {0x276e, 0x276f}, + {0x276f, 0x276e}, + {0x2770, 0x2771}, + {0x2771, 0x2770}, + {0x2772, 0x2773}, + {0x2773, 0x2772}, + {0x2774, 0x2775}, + {0x2775, 0x2774}, + {0x27c3, 0x27c4}, + {0x27c4, 0x27c3}, + {0x27c5, 0x27c6}, + {0x27c6, 0x27c5}, + {0x27c8, 0x27c9}, + {0x27c9, 0x27c8}, + {0x27cb, 0x27cd}, + {0x27cd, 0x27cb}, + {0x27d5, 0x27d6}, + {0x27d6, 0x27d5}, + {0x27dc, 0x22b8}, + {0x27dd, 0x27de}, + {0x27de, 0x27dd}, + {0x27e2, 0x27e3}, + {0x27e3, 0x27e2}, + {0x27e4, 0x27e5}, + {0x27e5, 0x27e4}, + {0x27e6, 0x27e7}, + {0x27e7, 0x27e6}, + {0x27e8, 0x27e9}, + {0x27e9, 0x27e8}, + {0x27ea, 0x27eb}, + {0x27eb, 0x27ea}, + {0x27ec, 0x27ed}, + {0x27ed, 0x27ec}, + {0x27ee, 0x27ef}, + {0x27ef, 0x27ee}, + {0x2983, 0x2984}, + {0x2984, 0x2983}, + {0x2985, 0x2986}, + {0x2986, 0x2985}, + {0x2987, 0x2988}, + {0x2988, 0x2987}, + {0x2989, 0x298a}, + {0x298a, 0x2989}, + {0x298b, 0x298c}, + {0x298c, 0x298b}, + {0x298d, 0x2990}, + {0x298e, 0x298f}, + {0x298f, 0x298e}, + {0x2990, 0x298d}, + {0x2991, 0x2992}, + {0x2992, 0x2991}, + {0x2993, 0x2994}, + {0x2994, 0x2993}, + {0x2995, 0x2996}, + {0x2996, 0x2995}, + {0x2997, 0x2998}, + {0x2998, 0x2997}, + {0x299b, 0x2221}, + {0x29a0, 0x2222}, + {0x29a3, 0x2220}, + {0x29a4, 0x29a5}, + {0x29a5, 0x29a4}, + {0x29a8, 0x29a9}, + {0x29a9, 0x29a8}, + {0x29aa, 0x29ab}, + {0x29ab, 0x29aa}, + {0x29ac, 0x29ad}, + {0x29ad, 0x29ac}, + {0x29ae, 0x29af}, + {0x29af, 0x29ae}, + {0x29b8, 0x2298}, + {0x29c0, 0x29c1}, + {0x29c1, 0x29c0}, + {0x29c4, 0x29c5}, + {0x29c5, 0x29c4}, + {0x29cf, 0x29d0}, + {0x29d0, 0x29cf}, + {0x29d1, 0x29d2}, + {0x29d2, 0x29d1}, + {0x29d4, 0x29d5}, + {0x29d5, 0x29d4}, + {0x29d8, 0x29d9}, + {0x29d9, 0x29d8}, + {0x29da, 0x29db}, + {0x29db, 0x29da}, + {0x29e8, 0x29e9}, + {0x29e9, 0x29e8}, + {0x29f5, 0x2215}, + {0x29f8, 0x29f9}, + {0x29f9, 0x29f8}, + {0x29fc, 0x29fd}, + {0x29fd, 0x29fc}, + {0x2a2b, 0x2a2c}, + {0x2a2c, 0x2a2b}, + {0x2a2d, 0x2a2e}, + {0x2a2e, 0x2a2d}, + {0x2a34, 0x2a35}, + {0x2a35, 0x2a34}, + {0x2a3c, 0x2a3d}, + {0x2a3d, 0x2a3c}, + {0x2a64, 0x2a65}, + {0x2a65, 0x2a64}, + {0x2a79, 0x2a7a}, + {0x2a7a, 0x2a79}, + {0x2a7b, 0x2a7c}, + {0x2a7c, 0x2a7b}, + {0x2a7d, 0x2a7e}, + {0x2a7e, 0x2a7d}, + {0x2a7f, 0x2a80}, + {0x2a80, 0x2a7f}, + {0x2a81, 0x2a82}, + {0x2a82, 0x2a81}, + {0x2a83, 0x2a84}, + {0x2a84, 0x2a83}, + {0x2a85, 0x2a86}, + {0x2a86, 0x2a85}, + {0x2a87, 0x2a88}, + {0x2a88, 0x2a87}, + {0x2a89, 0x2a8a}, + {0x2a8a, 0x2a89}, + {0x2a8b, 0x2a8c}, + {0x2a8c, 0x2a8b}, + {0x2a8d, 0x2a8e}, + {0x2a8e, 0x2a8d}, + {0x2a8f, 0x2a90}, + {0x2a90, 0x2a8f}, + {0x2a91, 0x2a92}, + {0x2a92, 0x2a91}, + {0x2a93, 0x2a94}, + {0x2a94, 0x2a93}, + {0x2a95, 0x2a96}, + {0x2a96, 0x2a95}, + {0x2a97, 0x2a98}, + {0x2a98, 0x2a97}, + {0x2a99, 0x2a9a}, + {0x2a9a, 0x2a99}, + {0x2a9b, 0x2a9c}, + {0x2a9c, 0x2a9b}, + {0x2a9d, 0x2a9e}, + {0x2a9e, 0x2a9d}, + {0x2a9f, 0x2aa0}, + {0x2aa0, 0x2a9f}, + {0x2aa1, 0x2aa2}, + {0x2aa2, 0x2aa1}, + {0x2aa6, 0x2aa7}, + {0x2aa7, 0x2aa6}, + {0x2aa8, 0x2aa9}, + {0x2aa9, 0x2aa8}, + {0x2aaa, 0x2aab}, + {0x2aab, 0x2aaa}, + {0x2aac, 0x2aad}, + {0x2aad, 0x2aac}, + {0x2aaf, 0x2ab0}, + {0x2ab0, 0x2aaf}, + {0x2ab1, 0x2ab2}, + {0x2ab2, 0x2ab1}, + {0x2ab3, 0x2ab4}, + {0x2ab4, 0x2ab3}, + {0x2ab5, 0x2ab6}, + {0x2ab6, 0x2ab5}, + {0x2ab7, 0x2ab8}, + {0x2ab8, 0x2ab7}, + {0x2ab9, 0x2aba}, + {0x2aba, 0x2ab9}, + {0x2abb, 0x2abc}, + {0x2abc, 0x2abb}, + {0x2abd, 0x2abe}, + {0x2abe, 0x2abd}, + {0x2abf, 0x2ac0}, + {0x2ac0, 0x2abf}, + {0x2ac1, 0x2ac2}, + {0x2ac2, 0x2ac1}, + {0x2ac3, 0x2ac4}, + {0x2ac4, 0x2ac3}, + {0x2ac5, 0x2ac6}, + {0x2ac6, 0x2ac5}, + {0x2ac7, 0x2ac8}, + {0x2ac8, 0x2ac7}, + {0x2ac9, 0x2aca}, + {0x2aca, 0x2ac9}, + {0x2acb, 0x2acc}, + {0x2acc, 0x2acb}, + {0x2acd, 0x2ace}, + {0x2ace, 0x2acd}, + {0x2acf, 0x2ad0}, + {0x2ad0, 0x2acf}, + {0x2ad1, 0x2ad2}, + {0x2ad2, 0x2ad1}, + {0x2ad3, 0x2ad4}, + {0x2ad4, 0x2ad3}, + {0x2ad5, 0x2ad6}, + {0x2ad6, 0x2ad5}, + {0x2ade, 0x22a6}, + {0x2ae3, 0x22a9}, + {0x2ae4, 0x22a8}, + {0x2ae5, 0x22ab}, + {0x2aec, 0x2aed}, + {0x2aed, 0x2aec}, + {0x2aee, 0x2224}, + {0x2af7, 0x2af8}, + {0x2af8, 0x2af7}, + {0x2af9, 0x2afa}, + {0x2afa, 0x2af9}, + {0x2bfe, 0x221f}, + {0x2e02, 0x2e03}, + {0x2e03, 0x2e02}, + {0x2e04, 0x2e05}, + {0x2e05, 0x2e04}, + {0x2e09, 0x2e0a}, + {0x2e0a, 0x2e09}, + {0x2e0c, 0x2e0d}, + {0x2e0d, 0x2e0c}, + {0x2e1c, 0x2e1d}, + {0x2e1d, 0x2e1c}, + {0x2e20, 0x2e21}, + {0x2e21, 0x2e20}, + {0x2e22, 0x2e23}, + {0x2e23, 0x2e22}, + {0x2e24, 0x2e25}, + {0x2e25, 0x2e24}, + {0x2e26, 0x2e27}, + {0x2e27, 0x2e26}, + {0x2e28, 0x2e29}, + {0x2e29, 0x2e28}, + {0x2e55, 0x2e56}, + {0x2e56, 0x2e55}, + {0x2e57, 0x2e58}, + {0x2e58, 0x2e57}, + {0x2e59, 0x2e5a}, + {0x2e5a, 0x2e59}, + {0x2e5b, 0x2e5c}, + {0x2e5c, 0x2e5b}, + {0x3008, 0x3009}, + {0x3009, 0x3008}, + {0x300a, 0x300b}, + {0x300b, 0x300a}, + {0x300c, 0x300d}, + {0x300d, 0x300c}, + {0x300e, 0x300f}, + {0x300f, 0x300e}, + {0x3010, 0x3011}, + {0x3011, 0x3010}, + {0x3014, 0x3015}, + {0x3015, 0x3014}, + {0x3016, 0x3017}, + {0x3017, 0x3016}, + {0x3018, 0x3019}, + {0x3019, 0x3018}, + {0x301a, 0x301b}, + {0x301b, 0x301a}, + {0xfe59, 0xfe5a}, + {0xfe5a, 0xfe59}, + {0xfe5b, 0xfe5c}, + {0xfe5c, 0xfe5b}, + {0xfe5d, 0xfe5e}, + {0xfe5e, 0xfe5d}, + {0xfe64, 0xfe65}, + {0xfe65, 0xfe64}, + {0xff08, 0xff09}, + {0xff09, 0xff08}, + {0xff1c, 0xff1e}, + {0xff1e, 0xff1c}, + {0xff3b, 0xff3d}, + {0xff3d, 0xff3b}, + {0xff5b, 0xff5d}, + {0xff5d, 0xff5b}, + {0xff5f, 0xff60}, + {0xff60, 0xff5f}, + {0xff62, 0xff63}, + {0xff63, 0xff62}, + }; + + int i, j, k; + + i = -1; + j = lenof(mirror_pairs); + + while (j - i > 1) { + k = (i + j) / 2; + if (ch < mirror_pairs[k].src) + j = k; + else if (ch > mirror_pairs[k].src) + i = k; + else + return mirror_pairs[k].dst; + } + + return ch; +} + +/* + * Identify the bracket characters treated specially by bidi rule + * BD19, and return their paired character(s). + * + * The data table in this function is constructed from the Unicode + * Character Database version 14.0.0, downloadable from unicode.org at + * the URL + * + * https://www.unicode.org/Public/14.0.0/ucd/ + * + * by the following fragment of Perl: + +perl -e ' + open BIDIBRACKETS, "<", $ARGV[0] or die; + while (<BIDIBRACKETS>) { + chomp; s{\s}{}g; s{#.*$}{}; next unless /./; + @_ = split /;/, $_; + $src = hex $_[0]; $dst = hex $_[1]; $kind = $_[2]; + $m{$src}=[$kind, $dst]; + } + open UNICODEDATA, "<", $ARGV[1] or die; + while (<UNICODEDATA>) { + chomp; @_ = split /;/, $_; + $src = hex $_[0]; next unless defined $m{$src}; + if ($_[5] =~ /^[0-9a-f]+$/i) { + $equiv = hex $_[5]; + $e{$src} = $equiv; + $e{$equiv} = $src; + } + } + for $src (sort {$a <=> $b} keys %m) { + ($kind, $dst) = @{$m{$src}}; + $equiv = 0 + $e{$dst}; + printf " {0x%04x, {0x%04x, 0x%04x, %s}},\n", $src, $dst, $equiv, + $kind eq "c" ? "BT_CLOSE" : "BT_OPEN"; + } +' BidiBrackets.txt UnicodeData.txt + + */ +typedef enum { BT_NONE, BT_OPEN, BT_CLOSE } BracketType; +typedef struct BracketTypeData { + unsigned partner, equiv_partner; + BracketType type; +} BracketTypeData; +static BracketTypeData bracket_type(unsigned int ch) +{ + static const struct { + unsigned src; + BracketTypeData payload; + } bracket_pairs[] = { + {0x0028, {0x0029, 0x0000, BT_OPEN}}, + {0x0029, {0x0028, 0x0000, BT_CLOSE}}, + {0x005b, {0x005d, 0x0000, BT_OPEN}}, + {0x005d, {0x005b, 0x0000, BT_CLOSE}}, + {0x007b, {0x007d, 0x0000, BT_OPEN}}, + {0x007d, {0x007b, 0x0000, BT_CLOSE}}, + {0x0f3a, {0x0f3b, 0x0000, BT_OPEN}}, + {0x0f3b, {0x0f3a, 0x0000, BT_CLOSE}}, + {0x0f3c, {0x0f3d, 0x0000, BT_OPEN}}, + {0x0f3d, {0x0f3c, 0x0000, BT_CLOSE}}, + {0x169b, {0x169c, 0x0000, BT_OPEN}}, + {0x169c, {0x169b, 0x0000, BT_CLOSE}}, + {0x2045, {0x2046, 0x0000, BT_OPEN}}, + {0x2046, {0x2045, 0x0000, BT_CLOSE}}, + {0x207d, {0x207e, 0x0000, BT_OPEN}}, + {0x207e, {0x207d, 0x0000, BT_CLOSE}}, + {0x208d, {0x208e, 0x0000, BT_OPEN}}, + {0x208e, {0x208d, 0x0000, BT_CLOSE}}, + {0x2308, {0x2309, 0x0000, BT_OPEN}}, + {0x2309, {0x2308, 0x0000, BT_CLOSE}}, + {0x230a, {0x230b, 0x0000, BT_OPEN}}, + {0x230b, {0x230a, 0x0000, BT_CLOSE}}, + {0x2329, {0x232a, 0x3009, BT_OPEN}}, + {0x232a, {0x2329, 0x3008, BT_CLOSE}}, + {0x2768, {0x2769, 0x0000, BT_OPEN}}, + {0x2769, {0x2768, 0x0000, BT_CLOSE}}, + {0x276a, {0x276b, 0x0000, BT_OPEN}}, + {0x276b, {0x276a, 0x0000, BT_CLOSE}}, + {0x276c, {0x276d, 0x0000, BT_OPEN}}, + {0x276d, {0x276c, 0x0000, BT_CLOSE}}, + {0x276e, {0x276f, 0x0000, BT_OPEN}}, + {0x276f, {0x276e, 0x0000, BT_CLOSE}}, + {0x2770, {0x2771, 0x0000, BT_OPEN}}, + {0x2771, {0x2770, 0x0000, BT_CLOSE}}, + {0x2772, {0x2773, 0x0000, BT_OPEN}}, + {0x2773, {0x2772, 0x0000, BT_CLOSE}}, + {0x2774, {0x2775, 0x0000, BT_OPEN}}, + {0x2775, {0x2774, 0x0000, BT_CLOSE}}, + {0x27c5, {0x27c6, 0x0000, BT_OPEN}}, + {0x27c6, {0x27c5, 0x0000, BT_CLOSE}}, + {0x27e6, {0x27e7, 0x0000, BT_OPEN}}, + {0x27e7, {0x27e6, 0x0000, BT_CLOSE}}, + {0x27e8, {0x27e9, 0x0000, BT_OPEN}}, + {0x27e9, {0x27e8, 0x0000, BT_CLOSE}}, + {0x27ea, {0x27eb, 0x0000, BT_OPEN}}, + {0x27eb, {0x27ea, 0x0000, BT_CLOSE}}, + {0x27ec, {0x27ed, 0x0000, BT_OPEN}}, + {0x27ed, {0x27ec, 0x0000, BT_CLOSE}}, + {0x27ee, {0x27ef, 0x0000, BT_OPEN}}, + {0x27ef, {0x27ee, 0x0000, BT_CLOSE}}, + {0x2983, {0x2984, 0x0000, BT_OPEN}}, + {0x2984, {0x2983, 0x0000, BT_CLOSE}}, + {0x2985, {0x2986, 0x0000, BT_OPEN}}, + {0x2986, {0x2985, 0x0000, BT_CLOSE}}, + {0x2987, {0x2988, 0x0000, BT_OPEN}}, + {0x2988, {0x2987, 0x0000, BT_CLOSE}}, + {0x2989, {0x298a, 0x0000, BT_OPEN}}, + {0x298a, {0x2989, 0x0000, BT_CLOSE}}, + {0x298b, {0x298c, 0x0000, BT_OPEN}}, + {0x298c, {0x298b, 0x0000, BT_CLOSE}}, + {0x298d, {0x2990, 0x0000, BT_OPEN}}, + {0x298e, {0x298f, 0x0000, BT_CLOSE}}, + {0x298f, {0x298e, 0x0000, BT_OPEN}}, + {0x2990, {0x298d, 0x0000, BT_CLOSE}}, + {0x2991, {0x2992, 0x0000, BT_OPEN}}, + {0x2992, {0x2991, 0x0000, BT_CLOSE}}, + {0x2993, {0x2994, 0x0000, BT_OPEN}}, + {0x2994, {0x2993, 0x0000, BT_CLOSE}}, + {0x2995, {0x2996, 0x0000, BT_OPEN}}, + {0x2996, {0x2995, 0x0000, BT_CLOSE}}, + {0x2997, {0x2998, 0x0000, BT_OPEN}}, + {0x2998, {0x2997, 0x0000, BT_CLOSE}}, + {0x29d8, {0x29d9, 0x0000, BT_OPEN}}, + {0x29d9, {0x29d8, 0x0000, BT_CLOSE}}, + {0x29da, {0x29db, 0x0000, BT_OPEN}}, + {0x29db, {0x29da, 0x0000, BT_CLOSE}}, + {0x29fc, {0x29fd, 0x0000, BT_OPEN}}, + {0x29fd, {0x29fc, 0x0000, BT_CLOSE}}, + {0x2e22, {0x2e23, 0x0000, BT_OPEN}}, + {0x2e23, {0x2e22, 0x0000, BT_CLOSE}}, + {0x2e24, {0x2e25, 0x0000, BT_OPEN}}, + {0x2e25, {0x2e24, 0x0000, BT_CLOSE}}, + {0x2e26, {0x2e27, 0x0000, BT_OPEN}}, + {0x2e27, {0x2e26, 0x0000, BT_CLOSE}}, + {0x2e28, {0x2e29, 0x0000, BT_OPEN}}, + {0x2e29, {0x2e28, 0x0000, BT_CLOSE}}, + {0x2e55, {0x2e56, 0x0000, BT_OPEN}}, + {0x2e56, {0x2e55, 0x0000, BT_CLOSE}}, + {0x2e57, {0x2e58, 0x0000, BT_OPEN}}, + {0x2e58, {0x2e57, 0x0000, BT_CLOSE}}, + {0x2e59, {0x2e5a, 0x0000, BT_OPEN}}, + {0x2e5a, {0x2e59, 0x0000, BT_CLOSE}}, + {0x2e5b, {0x2e5c, 0x0000, BT_OPEN}}, + {0x2e5c, {0x2e5b, 0x0000, BT_CLOSE}}, + {0x3008, {0x3009, 0x232a, BT_OPEN}}, + {0x3009, {0x3008, 0x2329, BT_CLOSE}}, + {0x300a, {0x300b, 0x0000, BT_OPEN}}, + {0x300b, {0x300a, 0x0000, BT_CLOSE}}, + {0x300c, {0x300d, 0x0000, BT_OPEN}}, + {0x300d, {0x300c, 0x0000, BT_CLOSE}}, + {0x300e, {0x300f, 0x0000, BT_OPEN}}, + {0x300f, {0x300e, 0x0000, BT_CLOSE}}, + {0x3010, {0x3011, 0x0000, BT_OPEN}}, + {0x3011, {0x3010, 0x0000, BT_CLOSE}}, + {0x3014, {0x3015, 0x0000, BT_OPEN}}, + {0x3015, {0x3014, 0x0000, BT_CLOSE}}, + {0x3016, {0x3017, 0x0000, BT_OPEN}}, + {0x3017, {0x3016, 0x0000, BT_CLOSE}}, + {0x3018, {0x3019, 0x0000, BT_OPEN}}, + {0x3019, {0x3018, 0x0000, BT_CLOSE}}, + {0x301a, {0x301b, 0x0000, BT_OPEN}}, + {0x301b, {0x301a, 0x0000, BT_CLOSE}}, + {0xfe59, {0xfe5a, 0x0000, BT_OPEN}}, + {0xfe5a, {0xfe59, 0x0000, BT_CLOSE}}, + {0xfe5b, {0xfe5c, 0x0000, BT_OPEN}}, + {0xfe5c, {0xfe5b, 0x0000, BT_CLOSE}}, + {0xfe5d, {0xfe5e, 0x0000, BT_OPEN}}, + {0xfe5e, {0xfe5d, 0x0000, BT_CLOSE}}, + {0xff08, {0xff09, 0x0000, BT_OPEN}}, + {0xff09, {0xff08, 0x0000, BT_CLOSE}}, + {0xff3b, {0xff3d, 0x0000, BT_OPEN}}, + {0xff3d, {0xff3b, 0x0000, BT_CLOSE}}, + {0xff5b, {0xff5d, 0x0000, BT_OPEN}}, + {0xff5d, {0xff5b, 0x0000, BT_CLOSE}}, + {0xff5f, {0xff60, 0x0000, BT_OPEN}}, + {0xff60, {0xff5f, 0x0000, BT_CLOSE}}, + {0xff62, {0xff63, 0x0000, BT_OPEN}}, + {0xff63, {0xff62, 0x0000, BT_CLOSE}}, + }; + + int i, j, k; + + i = -1; + j = lenof(bracket_pairs); + + while (j - i > 1) { + k = (i + j) / 2; + if (ch < bracket_pairs[k].src) { + j = k; + } else if (ch > bracket_pairs[k].src) { + i = k; + } else { + return bracket_pairs[k].payload; + } + } + + static const BracketTypeData null = { 0, 0, BT_NONE }; + return null; +} + +/* + * Function exported to front ends to allow them to identify + * bidi-active characters (in case, for example, the platform's + * text display function can't conveniently be prevented from doing + * its own bidi and so special treatment is required for characters + * that would cause the bidi algorithm to activate). + * + * This function is passed a single Unicode code point, and returns + * nonzero if the presence of this code point can possibly cause + * the bidi algorithm to do any reordering. Thus, any string + * composed entirely of characters for which is_rtl() returns zero + * should be safe to pass to a bidi-active platform display + * function without fear. + * + * (is_rtl() must therefore also return true for any character + * which would be affected by Arabic shaping, but this isn't + * important because all such characters are right-to-left so it + * would have flagged them anyway.) + */ +bool is_rtl(int c) +{ + return typeIsBidiActive(bidi_getType(c)); +} + +/* The Main shaping function, and the only one to be used + * by the outside world. + * + * line: buffer to apply shaping to. this must be passed by doBidi() first + * to: output buffer for the shaped data + * count: number of characters in line + */ +int do_shape(bidi_char *line, bidi_char *to, int count) +{ + int i, tempShape; + bool ligFlag = false; + + for (i=0; i<count; i++) { + to[i] = line[i]; + tempShape = STYPE(line[i].wc); + switch (tempShape) { + case SC: + break; + + case SU: + break; + + case SR: + tempShape = (i+1 < count ? STYPE(line[i+1].wc) : SU); + if ((tempShape == SL) || (tempShape == SD) || (tempShape == SC)) + to[i].wc = SFINAL((SISOLATED(line[i].wc))); + else + to[i].wc = SISOLATED(line[i].wc); + break; + + + case SD: + /* Make Ligatures */ + tempShape = (i+1 < count ? STYPE(line[i+1].wc) : SU); + if (line[i].wc == 0x644) { + if (i > 0) switch (line[i-1].wc) { + case 0x622: + ligFlag = true; + if ((tempShape == SL) || (tempShape == SD) || (tempShape == SC)) + to[i].wc = 0xFEF6; + else + to[i].wc = 0xFEF5; + break; + case 0x623: + ligFlag = true; + if ((tempShape == SL) || (tempShape == SD) || (tempShape == SC)) + to[i].wc = 0xFEF8; + else + to[i].wc = 0xFEF7; + break; + case 0x625: + ligFlag = true; + if ((tempShape == SL) || (tempShape == SD) || (tempShape == SC)) + to[i].wc = 0xFEFA; + else + to[i].wc = 0xFEF9; + break; + case 0x627: + ligFlag = true; + if ((tempShape == SL) || (tempShape == SD) || (tempShape == SC)) + to[i].wc = 0xFEFC; + else + to[i].wc = 0xFEFB; + break; + } + if (ligFlag) { + to[i-1].wc = 0x20; + ligFlag = false; + break; + } + } + + if ((tempShape == SL) || (tempShape == SD) || (tempShape == SC)) { + tempShape = (i > 0 ? STYPE(line[i-1].wc) : SU); + if ((tempShape == SR) || (tempShape == SD) || (tempShape == SC)) + to[i].wc = SMEDIAL((SISOLATED(line[i].wc))); + else + to[i].wc = SFINAL((SISOLATED(line[i].wc))); + break; + } + + tempShape = (i > 0 ? STYPE(line[i-1].wc) : SU); + if ((tempShape == SR) || (tempShape == SD) || (tempShape == SC)) + to[i].wc = SINITIAL((SISOLATED(line[i].wc))); + else + to[i].wc = SISOLATED(line[i].wc); + break; + + + } + } + return 1; +} + +typedef enum { DO_NEUTRAL, DO_LTR, DO_RTL } DirectionalOverride; + +typedef struct DSStackEntry { + /* + * An entry in the directional status stack (rule section X). + */ + unsigned char level; + bool isolate; + DirectionalOverride override; +} DSStackEntry; + +typedef struct BracketStackEntry { + /* + * An entry in the bracket-pair-tracking stack (rule BD16). + */ + unsigned ch; + size_t c; +} BracketStackEntry; + +typedef struct IsolatingRunSequence { + size_t start, end; + BidiType sos, eos, embeddingDirection; +} IsolatingRunSequence; + +#define MAX_DEPTH 125 /* specified in the standard */ + +struct BidiContext { + /* + * Storage space preserved between runs, all allocated to the same + * length (internal_array_sizes). + */ + size_t internal_array_sizes; + BidiType *types, *origTypes; + unsigned char *levels; + size_t *irsindices, *bracketpos; + bool *irsdone; + + /* + * Separately allocated with its own size field + */ + IsolatingRunSequence *irslist; + size_t irslistsize; + + /* + * Rewritten to point to the input to the currently active run of + * the bidi algorithm + */ + bidi_char *text; + size_t textlen; + + /* + * State within a run of the algorithm + */ + BidiType paragraphOverride; + DSStackEntry dsstack[MAX_DEPTH + 2]; + size_t ds_sp; + size_t overflowIsolateCount, overflowEmbeddingCount, validIsolateCount; + unsigned char paragraphLevel; + size_t *irs; + size_t irslen; + BidiType sos, eos, embeddingDirection; + BracketStackEntry bstack[63]; /* constant size specified in rule BD16 */ +}; + +BidiContext *bidi_new_context(void) +{ + BidiContext *ctx = snew(BidiContext); + memset(ctx, 0, sizeof(BidiContext)); + return ctx; +} + +void bidi_free_context(BidiContext *ctx) +{ + sfree(ctx->types); + sfree(ctx->origTypes); + sfree(ctx->levels); + sfree(ctx->irsindices); + sfree(ctx->irsdone); + sfree(ctx->bracketpos); + sfree(ctx->irslist); + sfree(ctx); +} + +static void ensure_arrays(BidiContext *ctx, size_t textlen) +{ + if (textlen <= ctx->internal_array_sizes) + return; + ctx->internal_array_sizes = textlen; + ctx->types = sresize(ctx->types, ctx->internal_array_sizes, BidiType); + ctx->origTypes = sresize(ctx->origTypes, ctx->internal_array_sizes, + BidiType); + ctx->levels = sresize(ctx->levels, ctx->internal_array_sizes, + unsigned char); + ctx->irsindices = sresize(ctx->irsindices, ctx->internal_array_sizes, + size_t); + ctx->irsdone = sresize(ctx->irsdone, ctx->internal_array_sizes, bool); + ctx->bracketpos = sresize(ctx->bracketpos, ctx->internal_array_sizes, + size_t); +} + +static void setup_types(BidiContext *ctx) +{ + for (size_t i = 0; i < ctx->textlen; i++) + ctx->types[i] = ctx->origTypes[i] = bidi_getType(ctx->text[i].wc); +} + +static bool text_needs_bidi(BidiContext *ctx) +{ + /* + * Initial optimisation: check for any bidi-active character at + * all in an input line. If there aren't any, we can skip the + * whole algorithm. + * + * Also include the paragraph override in this check! + */ + for (size_t i = 0; i < ctx->textlen; i++) + if (typeIsBidiActive(ctx->types[i])) + return true; + return typeIsBidiActive(ctx->paragraphOverride); +} + +static size_t find_matching_pdi(const BidiType *types, size_t i, size_t size) +{ + /* Assuming that types[i] is an isolate initiator, find its + * matching PDI by rule BD9. */ + unsigned counter = 1; + i++; + for (; i < size; i++) { + BidiType t = types[i]; + if (typeIsIsolateInitiator(t)) { + counter++; + } else if (t == PDI) { + counter--; + if (counter == 0) + return i; + } + } + + /* If no PDI was found, return the length of the array. */ + return size; +} + +static unsigned char rule_p2_p3(const BidiType *types, size_t size) +{ + /* + * Rule P2. Find the first strong type (L, R or AL), ignoring + * anything inside an isolated segment. + * + * Rule P3. If that type is R or AL, choose a paragraph embeddding + * level of 1, otherwise 0. + */ + for (size_t i = 0; i < size; i++) { + BidiType t = types[i]; + if (typeIsIsolateInitiator(t)) + i = find_matching_pdi(types, i, size); + else if (typeIsStrong(t)) + return (t == L ? 0 : 1); + } + + return 0; /* default if no strong type found */ +} + +static void set_paragraph_level(BidiContext *ctx) +{ + if (ctx->paragraphOverride == L) + ctx->paragraphLevel = 0; + else if (ctx->paragraphOverride == R) + ctx->paragraphLevel = 1; + else + ctx->paragraphLevel = rule_p2_p3(ctx->types, ctx->textlen); +} + +static inline unsigned char nextOddLevel(unsigned char x) { return (x+1)|1; } +static inline unsigned char nextEvenLevel(unsigned char x) { return (x|1)+1; } + +static inline void push(BidiContext *ctx, unsigned char level, + DirectionalOverride override, bool isolate) +{ + ctx->ds_sp++; + assert(ctx->ds_sp < lenof(ctx->dsstack)); + ctx->dsstack[ctx->ds_sp].level = level; + ctx->dsstack[ctx->ds_sp].override = override; + ctx->dsstack[ctx->ds_sp].isolate = isolate; +} + +static inline void pop(BidiContext *ctx) +{ + assert(ctx->ds_sp > 0); + ctx->ds_sp--; +} + +static void process_explicit_embeddings(BidiContext *ctx) +{ + /* + * Rule X1 initialisation. + */ + ctx->ds_sp = (size_t)-1; + push(ctx, ctx->paragraphLevel, DO_NEUTRAL, false); + ctx->overflowIsolateCount = 0; + ctx->overflowEmbeddingCount = 0; + ctx->validIsolateCount = 0; + + #define stk (&ctx->dsstack[ctx->ds_sp]) + + for (size_t i = 0; i < ctx->textlen; i++) { + BidiType t = ctx->types[i]; + switch (t) { + case RLE: case LRE: case RLO: case LRO: { + /* Rules X2-X5 */ + unsigned char newLevel; + DirectionalOverride override; + +#ifndef REMOVE_FORMATTING_CHARS + ctx->levels[i] = stk->level; +#endif + + switch (t) { + case RLE: /* rule X2 */ + newLevel = nextOddLevel(stk->level); + override = DO_NEUTRAL; + break; + case LRE: /* rule X3 */ + newLevel = nextEvenLevel(stk->level); + override = DO_NEUTRAL; + break; + case RLO: /* rule X4 */ + newLevel = nextOddLevel(stk->level); + override = DO_RTL; + break; + case LRO: /* rule X5 */ + newLevel = nextEvenLevel(stk->level); + override = DO_LTR; + break; + default: + unreachable("how did this get past the outer switch?"); + } + + if (newLevel <= MAX_DEPTH && + ctx->overflowIsolateCount == 0 && + ctx->overflowEmbeddingCount == 0) { + /* Embedding code is valid. Push a stack entry. */ + push(ctx, newLevel, override, false); + } else { + /* Embedding code is an overflow one. */ + if (ctx->overflowIsolateCount == 0) + ctx->overflowEmbeddingCount++; + } + break; + } + + case RLI: case LRI: case FSI: { + /* Rules X5a, X5b, X5c */ + + if (t == FSI) { + /* Rule X5c: decide whether this should be treated + * like RLI or LRI */ + size_t pdi = find_matching_pdi(ctx->types, i, ctx->textlen); + unsigned char level = rule_p2_p3(ctx->types + (i + 1), + pdi - (i + 1)); + t = (level == 1 ? RLI : LRI); + } + + ctx->levels[i] = stk->level; + if (stk->override != DO_NEUTRAL) + ctx->types[i] = (stk->override == DO_LTR ? L : + stk->override == DO_RTL ? R : t); + + unsigned char newLevel = (t == RLI ? nextOddLevel(stk->level) : + nextEvenLevel(stk->level)); + + if (newLevel <= MAX_DEPTH && + ctx->overflowIsolateCount == 0 && + ctx->overflowEmbeddingCount == 0) { + /* Isolate code is valid. Push a stack entry. */ + push(ctx, newLevel, DO_NEUTRAL, true); + ctx->validIsolateCount++; + } else { + /* Isolate code is an overflow one. */ + ctx->overflowIsolateCount++; + } + break; + } + + case PDI: { + /* Rule X6a */ + if (ctx->overflowIsolateCount > 0) { + ctx->overflowIsolateCount--; + } else if (ctx->validIsolateCount == 0) { + /* Do nothing: spurious isolate-pop */ + } else { + /* Valid isolate-pop. We expect that the stack must + * therefore contain at least one isolate==true entry, + * so pop everything up to and including it. */ + ctx->overflowEmbeddingCount = 0; + while (!stk->isolate) + pop(ctx); + pop(ctx); + ctx->validIsolateCount--; + } + ctx->levels[i] = stk->level; + if (stk->override != DO_NEUTRAL) + ctx->types[i] = (stk->override == DO_LTR ? L : R); + break; + } + + case PDF: { + /* Rule X7 */ + if (ctx->overflowIsolateCount > 0) { + /* Do nothing if we've overflowed on isolates */ + } else if (ctx->overflowEmbeddingCount > 0) { + ctx->overflowEmbeddingCount--; + } else if (ctx->ds_sp > 0 && !stk->isolate) { + pop(ctx); + } else { + /* Do nothing: spurious embedding-pop */ + } + +#ifndef REMOVE_FORMATTING_CHARS + ctx->levels[i] = stk->level; +#endif + break; + } + + case B: { + /* Rule X8: if an explicit paragraph separator appears in + * this text at all then it does not participate in any of + * the above, and just gets assigned the paragraph level. + * + * PS, it had better be right at the end of the text, + * because we have not implemented rule P1 in this code. */ + assert(i == ctx->textlen - 1); + ctx->levels[i] = ctx->paragraphLevel; + break; + } + + case BN: { + /* + * The section 5.2 adjustment to rule X6 says that we + * apply it to BN just like any other class. But I think + * this can't possibly give the same results as the + * unmodified algorithm. + * + * Proof: adding RLO BN or LRO BN at the end of a + * paragraph should not change the output of the standard + * algorithm, because the override doesn't affect the BN + * in rule X6, and then rule X9 removes both. But with the + * modified rule X6, the BN is changed into R or L, and + * then rule X9 doesn't remove it, and then you've added a + * strong type that will set eos for the level run just + * before the override. And whatever the standard + * algorithm set eos to, _one_ of these override sequences + * will disagree with it. + * + * So I think we just set the BN's level, and don't change + * its type. + */ + ctx->levels[i] = stk->level; + break; + } + + default: { + /* Rule X6. */ + ctx->levels[i] = stk->level; + if (stk->override != DO_NEUTRAL) + ctx->types[i] = (stk->override == DO_LTR ? L : R); + break; + } + } + } + + #undef stk +} + +static void remove_embedding_characters(BidiContext *ctx) +{ +#ifndef REMOVE_FORMATTING_CHARS + /* + * Rule X9, as modified by section 5.2: turn embedding (but not + * isolate) characters into BN. + */ + for (size_t i = 0; i < ctx->textlen; i++) { + BidiType t = ctx->types[i]; + if (typeIsRemovedDuringProcessing(t)) { + ctx->types[i] = BN; + + /* + * My own adjustment to the section 5.2 mods: a sequence + * of contiguous BN generated by this setup should never + * be at different levels from each other. + * + * An example where this goes wrong is if you open two + * LREs in sequence, then close them again: + * + * ... LRE LRE PDF PDF ... + * + * The initial level assignment gives level 0 to the outer + * LRE/PDF pair, and level 2 to the inner one. The + * standard algorithm would remove all four, so this + * doesn't matter, and you end up with no break in the + * surrounding level run. But if you just rewrite the + * types of all those characters to BN and leave the + * levels in that state, then the modified algorithm will + * leave the middle two BN at level 2, dividing what + * should have been a long level run at level 0 into two + * separate ones. + */ + if (i > 0 && ctx->types[i-1] == BN) + ctx->levels[i] = ctx->levels[i-1]; + } + } +#else + /* + * Rule X9, original version: completely remove embedding + * start/end characters and also boundary neutrals. + */ + size_t outpos = 0; + for (size_t i = 0; i < ctx->textlen; i++) { + BidiType t = ctx->types[i]; + if (!typeIsRemovedDuringProcessing(t)) { + ctx->text[outpos] = ctx->text[i]; + ctx->levels[outpos] = ctx->levels[i]; + ctx->types[outpos] = ctx->types[i]; + ctx->origTypes[outpos] = ctx->origTypes[i]; + outpos++; + } + } + ctx->textlen = outpos; +#endif +} + +typedef void (*irs_fn_t)(BidiContext *ctx); + +static void find_isolating_run_sequences(BidiContext *ctx, irs_fn_t process) +{ + /* + * Rule X10 / BD13. Now that we've assigned an embedding level to + * each character in the text, we have to divide the text into + * subsequences on which to do the next stage of processing. + * + * In earlier issues of the bidi algorithm, these subsequences + * were contiguous in the original text, and each one was a 'level + * run': a maximal contiguous subsequence of characters all at the + * same embedding level. + * + * But now we have isolates, and the point of an (isolate + * initiator ... PDI) sequence is that the whole sequence should + * be treated like a single BN for the purposes of formatting + * everything outside it. As a result, we now have to recombine + * our level runs into longer sequences, on the principle that if + * a level run ends with an isolate initiator, then we bring it + * together with whatever later level run starts with the matching + * PDI. + * + * These subsequences are no longer contiguous (the whole point is + * that between the isolate initiator and the PDI is some other + * text that we've skipped over). They're called 'isolating run + * sequences'. + */ + + memset(ctx->irsdone, 0, ctx->textlen); + size_t i = 0; + size_t n_irs = 0; + size_t indexpos = 0; + while (i < ctx->textlen) { + if (ctx->irsdone[i]) { + i++; + continue; + } + + /* + * Found a character not already processed. Start a new + * sequence here. + */ + sgrowarray(ctx->irslist, ctx->irslistsize, n_irs); + IsolatingRunSequence *irs = &ctx->irslist[n_irs++]; + irs->start = indexpos; + size_t j = i; + size_t irslevel = ctx->levels[i]; + while (j < ctx->textlen) { + /* + * We expect that all level runs in this sequence will be + * at the same level as each other, by construction of how + * we set up the levels from the isolates in the first + * place. + */ + assert(ctx->levels[j] == irslevel); + + do { + ctx->irsdone[j] = true; + ctx->irsindices[indexpos++] = j++; + } while (j < ctx->textlen && ctx->levels[j] == irslevel); + if (!typeIsIsolateInitiator(ctx->types[j-1])) + break; /* this IRS is ended */ + j = find_matching_pdi(ctx->types, j-1, ctx->textlen); + } + irs->end = indexpos; + + /* + * Determine the start-of-sequence and end-of-sequence types + * for this sequence. + * + * These depend on the embedding levels of surrounding text. + * But processing each run can change those levels. That's why + * we have to use a two-pass strategy here, first identifying + * all the isolating run sequences using the input level data, + * and not processing any of them until we know where they all + * are. + */ + size_t p; + unsigned char level_inside, level_outside, level_max; + + p = i; + level_inside = ctx->levels[p]; + level_outside = ctx->paragraphLevel; + while (p > 0) { + p--; + if (ctx->types[p] != BN) { + level_outside = ctx->levels[p]; + break; + } + } + level_max = max(level_inside, level_outside); + irs->sos = (level_max % 2 ? R : L); + + p = ctx->irsindices[irs->end - 1]; + level_inside = ctx->levels[p]; + level_outside = ctx->paragraphLevel; + if (typeIsIsolateInitiator(ctx->types[p])) { + /* Special case: if an isolating run sequence ends in an + * unmatched isolate initiator, then level_outside is + * taken to be the paragraph embedding level and the + * loop below is skipped. */ + } else { + while (p+1 < ctx->textlen) { + p++; + if (ctx->types[p] != BN) { + level_outside = ctx->levels[p]; + break; + } + } + } + level_max = max(level_inside, level_outside); + irs->eos = (level_max % 2 ? R : L); + + irs->embeddingDirection = (irslevel % 2 ? R : L); + + /* + * Now we've listed in ctx->irsindices[] the index of every + * character that's part of this isolating run sequence, and + * recorded an entry in irslist containing the interval of + * indices relevant to this IRS, plus its assorted metadata. + * We've also marked those locations in the input text as done + * in ctx->irsdone, so that we'll skip over them when the + * outer iteration reaches them later. + */ + } + + for (size_t k = 0; k < n_irs; k++) { + IsolatingRunSequence *irs = &ctx->irslist[k]; + ctx->irs = ctx->irsindices + irs->start; + ctx->irslen = irs->end - irs->start; + ctx->sos = irs->sos; + ctx->eos = irs->eos; + ctx->embeddingDirection = irs->embeddingDirection; + process(ctx); + } + + /* Reset irslen to 0 when we've finished. This means any other + * functions that absentmindedly try to use irslen at all will end + * up doing nothing at all, which should be easier to detect and + * debug than if they run on subtly the wrong subset of the + * text. */ + ctx->irslen = 0; +} + +static void remove_nsm(BidiContext *ctx) +{ + /* Rule W1: NSM gains the type of the previous character, or sos + * at the start of the run, with the exception that isolation + * boundaries turn into ON. */ + BidiType prevType = ctx->sos; + for (size_t c = 0; c < ctx->irslen; c++) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + if (t == NSM) { + ctx->types[i] = prevType; + } else if (typeIsIsolateInitiatorOrPDI(t)) { + prevType = ON; +#ifndef REMOVE_FORMATTING_CHARS + } else if (t == BN) { + /* section 5.2 adjustment: these don't affect prevType */ +#endif + } else { + prevType = t; + } + } +} + +static void change_en_to_an(BidiContext *ctx) +{ + /* Rule W2: EN becomes AN if the previous strong type is AL. (The + * spec says that the 'previous strong type' is counted as sos at + * the start of the run, although it hardly matters, since sos + * can't be AL.) */ + BidiType prevStrongType = ctx->sos; + for (size_t c = 0; c < ctx->irslen; c++) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + if (t == EN && prevStrongType == AL) { + ctx->types[i] = AN; + } else if (typeIsStrong(t)) { + prevStrongType = t; + } + } +} + +static void change_al_to_r(BidiContext *ctx) +{ + /* Rule W3: AL becomes R unconditionally. (The only difference + * between the two types was their effect on nearby numbers, which + * was dealt with in rule W2, so now we're done with the + * distinction.) */ + for (size_t c = 0; c < ctx->irslen; c++) { + size_t i = ctx->irs[c]; + if (ctx->types[i] == AL) + ctx->types[i] = R; + } +} + +static void eliminate_separators_between_numbers(BidiContext *ctx) +{ + /* Rule W4: a single numeric separator between two numbers of the + * same type compatible with that separator takes the type of the + * number. ES is a separator type compatible only with EN; CS is a + * separator type compatible with either EN or AN. + * + * Section 5.2 adjustment: intervening BNs do not break this, so + * instead of simply looking at types[irs[c-1]] and types[irs[c+1]], + * we must track the last three indices we saw that were not BN. */ + size_t i1 = 0, i2 = 0; + BidiType t0 = ON, t1 = ON, t2 = ON; + for (size_t c = 0; c < ctx->irslen; c++) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + +#ifndef REMOVE_FORMATTING_CHARS + if (t == BN) + continue; +#endif + + i1 = i2; i2 = i; + t0 = t1; t1 = t2; t2 = t; + if (t0 == t2 && ((t1 == ES && t0 == EN) || + (t1 == CS && (t0 == EN || t0 == AN)))) { + ctx->types[i1] = t0; + } + } +} + +static void eliminate_et_next_to_en(BidiContext *ctx) +{ + /* Rule W5: a sequence of ET adjacent to an EN take the type EN. + * This is easiest to implement with one loop in each direction. + * + * Section 5.2 adjustment: include BN with ET. (We don't need to + * #ifdef that out, because in the standard algorithm, we won't + * have any BN left in any case.) */ + + bool modifying = false; + + for (size_t c = 0; c < ctx->irslen; c++) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + if (t == EN) { + modifying = true; + } else if (modifying && typeIsETOrBN(t)) { + ctx->types[i] = EN; + } else { + modifying = false; + } + } + + for (size_t c = ctx->irslen; c-- > 0 ;) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + if (t == EN) { + modifying = true; + } else if (modifying && typeIsETOrBN(t)) { + ctx->types[i] = EN; + } else { + modifying = false; + } + } +} + +static void eliminate_separators_and_terminators(BidiContext *ctx) +{ + /* Rule W6: all separators and terminators change to ON. + * + * (The spec is not quite clear on which bidi types are included + * in this; one assumes ES, ET and CS, but what about S? I _think_ + * the answer is that this is a rule in the W section, so it's + * implicitly supposed to only apply to types designated as weakly + * directional, so not S.) */ + +#ifndef REMOVE_FORMATTING_CHARS + /* + * Section 5.2 adjustment: this also applies to any BN adjacent on + * either side to one of these types, which is easiest to + * implement with a separate double-loop converting those to an + * arbitrary one of the affected types, say CS. + * + * This double loop can be completely skipped in the standard + * algorithm. + */ + bool modifying = false; + + for (size_t c = 0; c < ctx->irslen; c++) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + if (typeIsWeakSeparatorOrTerminator(t)) { + modifying = true; + } else if (modifying && t == BN) { + ctx->types[i] = CS; + } else { + modifying = false; + } + } + + for (size_t c = ctx->irslen; c-- > 0 ;) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + if (typeIsWeakSeparatorOrTerminator(t)) { + modifying = true; + } else if (modifying && t == BN) { + ctx->types[i] = CS; + } else { + modifying = false; + } + } +#endif + + /* Now the main part of rule W6 */ + for (size_t c = 0; c < ctx->irslen; c++) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + if (typeIsWeakSeparatorOrTerminator(t)) + ctx->types[i] = ON; + } +} + +static void change_en_to_l(BidiContext *ctx) +{ + /* Rule W7: EN becomes L if the previous strong type (or sos) is L. */ + BidiType prevStrongType = ctx->sos; + for (size_t c = 0; c < ctx->irslen; c++) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + if (t == EN && prevStrongType == L) { + ctx->types[i] = L; + } else if (typeIsStrong(t)) { + prevStrongType = t; + } + } +} + +typedef void (*bracket_pair_fn)(BidiContext *ctx, size_t copen, size_t cclose); + +static void find_bracket_pairs(BidiContext *ctx, bracket_pair_fn process) +{ + const size_t NO_BRACKET = ~(size_t)0; + + /* + * Rule BD16. + */ + size_t sp = 0; + for (size_t c = 0; c < ctx->irslen; c++) + ctx->bracketpos[c] = NO_BRACKET; + + for (size_t c = 0; c < ctx->irslen; c++) { + size_t i = ctx->irs[c]; + unsigned wc = ctx->text[i].wc; + BracketTypeData bt = bracket_type(wc); + if (bt.type == BT_OPEN) { + if (sp >= lenof(ctx->bstack)) { + /* + * Stack overflow. The spec says we simply give up at + * this point. + */ + goto found_all_pairs; + } + + ctx->bstack[sp].ch = wc; + ctx->bstack[sp].c = c; + sp++; + } else if (bt.type == BT_CLOSE) { + size_t new_sp = sp; + + /* + * Search up the stack for an entry containing a matching + * open bracket. If we find it, pop that entry and + * everything deeper, and record a matching pair. If we + * reach the bottom of the stack without finding anything, + * leave sp where it started. + */ + while (new_sp-- > 0) { + if (ctx->bstack[new_sp].ch == bt.partner || + ctx->bstack[new_sp].ch == bt.equiv_partner) { + /* Found a stack element matching this one */ + size_t cstart = ctx->bstack[new_sp].c; + ctx->bracketpos[cstart] = c; + sp = new_sp; + break; + } + } + } + } + + found_all_pairs: + for (size_t c = 0; c < ctx->irslen; c++) { + if (ctx->bracketpos[c] != NO_BRACKET) { + process(ctx, c, ctx->bracketpos[c]); + } + } +} + +static BidiType get_bracket_type(BidiContext *ctx, size_t copen, size_t cclose) +{ + /* + * Rule N0: a pair of matched brackets containing at least one + * strong type takes on the current embedding direction, unless + * all of these are true at once: + * + * (a) there are no strong types inside the brackets matching the + * current embedding direction + * (b) there _is_ at least one strong type inside the brackets + * that is _opposite_ to the current embedding direction + * (c) the strong type preceding the open bracket is also + * opposite to the current embedding direction + * + * in which case they take on the opposite direction. + * + * For these purposes, number types (EN and AN) count as R. + */ + + bool foundOppositeTypeInside = false; + for (size_t c = copen + 1; c < cclose; c++) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + if (typeIsStrongOrNumber(t)) { + t = t == L ? L : R; /* numbers count as R */ + if (t == ctx->embeddingDirection) { + /* Found something inside the brackets matching the + * current level, so (a) is violated. */ + return ctx->embeddingDirection; + } else { + foundOppositeTypeInside = true; + } + } + } + + if (!foundOppositeTypeInside) { + /* No strong types at all inside the brackets, so return ON to + * indicate that we're not messing with their type at all. */ + return ON; + } + + /* There was an opposite strong type in the brackets. Look + * backwards to the preceding strong type, and go with that, + * whichever it is. */ + for (size_t c = copen; c-- > 0 ;) { + size_t i = ctx->irs[c]; + BidiType t = ctx->types[i]; + if (typeIsStrongOrNumber(t)) { + t = t == L ? L : R; /* numbers count as R */ + return t; + } + } + + /* Fallback: if the preceding strong type was not found, go with + * sos. */ + return ctx->sos; +} + +static void reset_bracket_type(BidiContext *ctx, size_t c, BidiType t) +{ + /* Final bullet point of rule N0: when we change the type of a + * bracket, the same change applies to any contiguous sequence of + * characters after it whose _original_ bidi type was NSM. */ + do { + ctx->types[ctx->irs[c++]] = t; + +#ifndef REMOVE_FORMATTING_CHARS + while (c < ctx->irslen && ctx->origTypes[ctx->irs[c]] == BN) { + /* Section 5.2 adjustment: skip past BN in the process. */ + c++; + } +#endif + } while (c < ctx->irslen && ctx->origTypes[ctx->irs[c]] == NSM); +} + +static void resolve_brackets(BidiContext *ctx, size_t copen, size_t cclose) +{ + if (typeIsNeutral(ctx->types[ctx->irs[copen]]) && + typeIsNeutral(ctx->types[ctx->irs[cclose]])) { + BidiType t = get_bracket_type(ctx, copen, cclose); + if (t != ON) { + reset_bracket_type(ctx, copen, t); + reset_bracket_type(ctx, cclose, t); + } + } +} + +static void remove_ni(BidiContext *ctx) +{ + /* + * Rules N1 and N2 together: neutral or isolate characters take + * the direction of the surrounding strong text if the nearest + * strong characters on each side match, and otherwise, they take + * the embedding direction. + */ + const size_t NO_INDEX = ~(size_t)0; + BidiType prevStrongType = ctx->sos; + size_t c_ni_start = NO_INDEX; + for (size_t c = 0; c <= ctx->irslen; c++) { + BidiType t; + + if (c < ctx->irslen) { + size_t i = ctx->irs[c]; + t = ctx->types[i]; + } else { + /* One extra loop iteration, using eos to resolve the + * final sequence of NI if any */ + t = ctx->eos; + } + + if (typeIsStrongOrNumber(t)) { + t = t == L ? L : R; /* numbers count as R */ + if (c_ni_start != NO_INDEX) { + /* There are some NI we have to fix up */ + BidiType ni_type = (t == prevStrongType ? t : + ctx->embeddingDirection); + for (size_t c2 = c_ni_start; c2 < c; c2++) { + size_t i2 = ctx->irs[c2]; + BidiType t2 = ctx->types[i2]; + if (typeIsNeutralOrIsolate(t2)) + ctx->types[i2] = ni_type; + } + } + prevStrongType = t; + c_ni_start = NO_INDEX; + } else if (typeIsNeutralOrIsolate(t) && c_ni_start == NO_INDEX) { + c_ni_start = c; + } + } +} + +static void resolve_implicit_levels(BidiContext *ctx) +{ + /* Rules I1 and I2 */ + for (size_t c = 0; c < ctx->irslen; c++) { + size_t i = ctx->irs[c]; + unsigned char level = ctx->levels[i]; + BidiType t = ctx->types[i]; + if (level % 2 == 0) { + /* Rule I1 */ + if (t == R) + ctx->levels[i] += 1; + else if (t == AN || t == EN) + ctx->levels[i] += 2; + } else { + /* Rule I2 */ + if (t == L || t == AN || t == EN) + ctx->levels[i] += 1; + } + } +} + +static void process_isolating_run_sequence(BidiContext *ctx) +{ + /* Section W: resolve weak types */ + remove_nsm(ctx); + change_en_to_an(ctx); + change_al_to_r(ctx); + eliminate_separators_between_numbers(ctx); + eliminate_et_next_to_en(ctx); + eliminate_separators_and_terminators(ctx); + change_en_to_l(ctx); + + /* Section N: resolve neutral types (and isolates) */ + find_bracket_pairs(ctx, resolve_brackets); + remove_ni(ctx); + + /* Section I: resolve implicit levels */ + resolve_implicit_levels(ctx); +} + +static void reset_whitespace_and_separators(BidiContext *ctx) +{ + /* + * Rule L1: segment and paragraph separators, plus whitespace + * preceding them, all reset to the paragraph embedding level. + * This also applies to whitespace at the very end. + * + * This is done using the original types, not the versions that + * the rest of this algorithm has been merrily mutating. + */ + bool modifying = true; + for (size_t i = ctx->textlen; i-- > 0 ;) { + BidiType t = ctx->origTypes[i]; + if (typeIsSegmentOrParaSeparator(t)) { + ctx->levels[i] = ctx->paragraphLevel; + modifying = true; + } else if (modifying) { + if (typeIsWhitespaceOrIsolate(t)) { + ctx->levels[i] = ctx->paragraphLevel; + } else if (!typeIsRemovedDuringProcessing(t)) { + modifying = false; + } + } + } + +#ifndef REMOVE_FORMATTING_CHARS + /* + * Section 5.2 adjustment: types removed by rule X9 take the level + * of the character to their left. + */ + for (size_t i = 0; i < ctx->textlen; i++) { + BidiType t = ctx->origTypes[i]; + if (typeIsRemovedDuringProcessing(t)) { + /* Section 5.2 adjustment */ + ctx->levels[i] = (i > 0 ? ctx->levels[i-1] : ctx->paragraphLevel); + } + } +#endif /* ! REMOVE_FORMATTING_CHARS */ +} + +static void reverse(BidiContext *ctx, size_t start, size_t end) +{ + for (size_t i = start, j = end; i < j; i++, j--) { + bidi_char tmp = ctx->text[i]; + ctx->text[i] = ctx->text[j]; + ctx->text[j] = tmp; + } +} + +static void mirror_glyphs(BidiContext *ctx) +{ + /* + * Rule L3: any character with a mirror-image pair at an odd + * embedding level is replaced by its mirror image. + * + * This is specified in the standard as happening _after_ rule L2 + * (the actual reordering of the text). But it's much easier to + * implement it before, while our levels[] array still matches up + * to the text order. + */ + for (size_t i = 0; i < ctx->textlen; i++) { + if (ctx->levels[i] % 2) + ctx->text[i].wc = mirror_glyph(ctx->text[i].wc); + } +} + +static void reverse_sequences(BidiContext *ctx) +{ + /* + * Rule L2: every maximal contiguous sequence of characters at a + * given level or higher is reversed. + */ + unsigned level = 0; + for (size_t i = 0; i < ctx->textlen; i++) + level = max(level, ctx->levels[i]); + + for (; level >= 1; level--) { + for (size_t i = 0; i < ctx->textlen; i++) { + if (ctx->levels[i] >= level) { + size_t start = i; + while (i+1 < ctx->textlen && ctx->levels[i+1] >= level) + i++; + reverse(ctx, start, i); + } + } + } +} + +/* + * The Main Bidi Function. The two wrappers below it present different + * external APIs for different purposes, but everything comes through + * here. + * + * text: a buffer of size textlen containing text to apply the + * Bidirectional algorithm to. + */ +static void do_bidi_new(BidiContext *ctx, bidi_char *text, size_t textlen) +{ + ensure_arrays(ctx, textlen); + ctx->text = text; + ctx->textlen = textlen; + setup_types(ctx); + + /* Quick initial test: see if we need to bother with any work at all */ + if (!text_needs_bidi(ctx)) + return; + + set_paragraph_level(ctx); + process_explicit_embeddings(ctx); + remove_embedding_characters(ctx); + find_isolating_run_sequences(ctx, process_isolating_run_sequence); + + /* If this implementation distinguished paragraphs from lines, + * then this would be the point where we repeat the remainder of + * the algorithm once for each line in the paragraph. */ + + reset_whitespace_and_separators(ctx); + mirror_glyphs(ctx); + reverse_sequences(ctx); +} + +size_t do_bidi_test(BidiContext *ctx, bidi_char *text, size_t textlen, + int override) +{ + ctx->paragraphOverride = (override > 0 ? L : override < 0 ? R : ON); + do_bidi_new(ctx, text, textlen); + return ctx->textlen; +} + +void do_bidi(BidiContext *ctx, bidi_char *text, size_t textlen) +{ +#ifdef REMOVE_FORMATTING_CHARACTERS + abort(); /* can't use the standard algorithm in a live terminal */ +#else + ctx->paragraphOverride = ON; + do_bidi_new(ctx, text, textlen); +#endif +} diff --git a/terminal/bidi.h b/terminal/bidi.h new file mode 100644 index 00000000..90d68e5b --- /dev/null +++ b/terminal/bidi.h @@ -0,0 +1,147 @@ +/* + * Header file shared between bidi.c and its tests. Not used by + * anything outside the bidi subsystem. + */ + +#ifndef PUTTY_BIDI_H +#define PUTTY_BIDI_H + +#define LMASK 0x3F /* Embedding Level mask */ +#define OMASK 0xC0 /* Override mask */ +#define OISL 0x80 /* Override is L */ +#define OISR 0x40 /* Override is R */ + +/* Shaping Helpers */ +#define STYPE(xh) ((((xh) >= SHAPE_FIRST) && ((xh) <= SHAPE_LAST)) ? \ +shapetypes[(xh)-SHAPE_FIRST].type : SU) /*))*/ +#define SISOLATED(xh) (shapetypes[(xh)-SHAPE_FIRST].form_b) +#define SFINAL(xh) ((xh)+1) +#define SINITIAL(xh) ((xh)+2) +#define SMEDIAL(ch) ((ch)+3) + +#define leastGreaterOdd(x) ( ((x)+1) | 1 ) +#define leastGreaterEven(x) ( ((x)+2) &~ 1 ) + +/* Function declarations used outside bidi.c */ +unsigned char bidi_getType(int ch); + +/* Bidi character types */ +#define BIDI_CHAR_TYPE_LIST(X) \ + X(L) \ + X(LRE) \ + X(LRO) \ + X(LRI) \ + X(R) \ + X(AL) \ + X(RLE) \ + X(RLO) \ + X(RLI) \ + X(PDF) \ + X(PDI) \ + X(FSI) \ + X(EN) \ + X(ES) \ + X(ET) \ + X(AN) \ + X(CS) \ + X(NSM) \ + X(BN) \ + X(B) \ + X(S) \ + X(WS) \ + X(ON) \ + /* end of list */ + +/* Shaping Types */ +#define SHAPING_CHAR_TYPE_LIST(X) \ + X(SL) /* Left-Joining, doesn't exist in U+0600 - U+06FF */ \ + X(SR) /* Right-Joining, ie has Isolated, Final */ \ + X(SD) /* Dual-Joining, ie has Isolated, Final, Initial, Medial */ \ + X(SU) /* Non-Joining */ \ + X(SC) /* Join-Causing, like U+0640 (TATWEEL) */ \ + /* end of list */ + +#define ENUM_DECL(name) name, +typedef enum { BIDI_CHAR_TYPE_LIST(ENUM_DECL) N_BIDI_TYPES } BidiType; +typedef enum { SHAPING_CHAR_TYPE_LIST(ENUM_DECL) N_SHAPING_TYPES } ShapingType; +#undef ENUM_DECL + +static inline bool typeIsStrong(BidiType t) +{ + return ((1<<L) | (1<<R) | (1<<AL)) & (1 << t); +} +static inline bool typeIsWeak(BidiType t) +{ + return ((1<<EN) | (1<<ES) | (1<<ET) | (1<<AN) | + (1<<CS) | (1<<NSM) | (1<<BN)) & (1 << t); +} +static inline bool typeIsNeutral(BidiType t) +{ + return ((1<<B) | (1<<S) | (1<<WS) | (1<<ON)) & (1 << t); +} +static inline bool typeIsBidiActive(BidiType t) +{ + return ((1<<R) | (1<<AL) | (1<<AN) | (1<<RLE) | (1<<LRE) | (1<<RLO) | + (1<<LRO) | (1<<PDF) | (1<<RLI)) & (1 << t); +} +static inline bool typeIsIsolateInitiator(BidiType t) +{ + return ((1<<LRI) | (1<<RLI) | (1<<FSI)) & (1 << t); +} +static inline bool typeIsIsolateInitiatorOrPDI(BidiType t) +{ + return ((1<<LRI) | (1<<RLI) | (1<<FSI) | (1<<PDI)) & (1 << t); +} +static inline bool typeIsEmbeddingInitiator(BidiType t) +{ + return ((1<<LRE) | (1<<RLE) | (1<<LRO) | (1<<RLO)) & (1 << t); +} +static inline bool typeIsEmbeddingInitiatorOrPDF(BidiType t) +{ + return ((1<<LRE) | (1<<RLE) | (1<<LRO) | (1<<RLO) | (1<<PDF)) & (1 << t); +} +static inline bool typeIsWeakSeparatorOrTerminator(BidiType t) +{ + return ((1<<ES) | (1<<ET) | (1<<CS)) & (1 << t); +} +static inline bool typeIsNeutralOrIsolate(BidiType t) +{ + return ((1<<S) | (1<<WS) | (1<<ON) | (1<<FSI) | (1<<LRI) | (1<<RLI) | + (1<<PDI)) & (1 << t); +} +static inline bool typeIsSegmentOrParaSeparator(BidiType t) +{ + return ((1<<S) | (1<<B)) & (1 << t); +} +static inline bool typeIsWhitespaceOrIsolate(BidiType t) +{ + return ((1<<WS) | (1<<FSI) | (1<<LRI) | (1<<RLI) | (1<<PDI)) & (1 << t); +} +static inline bool typeIsRemovedDuringProcessing(BidiType t) +{ + return ((1<<RLE) | (1<<LRE) | (1<<RLO) | (1<<LRO) | (1<<PDF) | + (1<<BN)) & (1 << t); +} +static inline bool typeIsStrongOrNumber(BidiType t) +{ + return ((1<<L) | (1<<R) | (1<<AL) | (1<<EN) | (1<<AN)) & (1 << t); +} +static inline bool typeIsETOrBN(BidiType t) +{ + return ((1<<ET) | (1<<BN)) & (1 << t); +} + +/* + * More featureful interface to the bidi code, for use in bidi_test.c. + * It returns a potentially different value of textlen (in case we're + * compiling in REMOVE_FORMATTING_CHARACTERS mode), and also permits + * you to pass in an override to the paragraph direction (because many + * of the UCD conformance tests use one). + * + * 'override' is 0 for no override, +1 for left-to-right, -1 for + * right-to-left. + */ +size_t do_bidi_test(BidiContext *ctx, bidi_char *text, size_t textlen, + int override); + +#endif /* PUTTY_BIDI_H */ diff --git a/terminal/bidi_gettype.c b/terminal/bidi_gettype.c new file mode 100644 index 00000000..f3f5338e --- /dev/null +++ b/terminal/bidi_gettype.c @@ -0,0 +1,33 @@ +/* + * Standalone test program that exposes the minibidi getType function. + */ + +#include <stdio.h> +#include <assert.h> + +#include "putty.h" +#include "misc.h" +#include "bidi.h" + +void out_of_memory(void) +{ + fprintf(stderr, "out of memory!\n"); + exit(2); +} + +#define TYPETONAME(X) #X, +static const char *const typenames[] = { BIDI_CHAR_TYPE_LIST(TYPETONAME) }; +#undef TYPETONAME + +int main(int argc, char **argv) +{ + int i; + + for (i = 1; i < argc; i++) { + unsigned long chr = strtoul(argv[i], NULL, 0); + int type = bidi_getType(chr); + printf("U+%04x: %s\n", (unsigned)chr, typenames[type]); + } + + return 0; +} diff --git a/terminal/bidi_test.c b/terminal/bidi_test.c new file mode 100644 index 00000000..1acd1d68 --- /dev/null +++ b/terminal/bidi_test.c @@ -0,0 +1,372 @@ +/* + * Test program that reads the Unicode bidi algorithm test case lists + * that form part of the Unicode Character Database: + * + * https://www.unicode.org/Public/UCD/latest/ucd/BidiTest.txt + * https://www.unicode.org/Public/UCD/latest/ucd/BidiCharacterTest.txt + */ + +#include <ctype.h> + +#include "putty.h" +#include "misc.h" +#include "bidi.h" + +static int pass = 0, fail = 0; + +static BidiContext *ctx; + +static const char *extract_word(char **ptr) +{ + char *p = *ptr; + while (*p && isspace((unsigned char)*p)) p++; + + char *start = p; + while (*p && !isspace((unsigned char)*p)) p++; + + if (*p) { + *p++ = '\0'; + while (*p && isspace((unsigned char)*p)) p++; + } + + *ptr = p; + return start; +} + +#define TYPETONAME(X) #X, +static const char *const typenames[] = { BIDI_CHAR_TYPE_LIST(TYPETONAME) }; +#undef TYPETONAME + +static void run_test(const char *filename, unsigned lineno, + bidi_char *bcs, size_t bcs_len, + const unsigned *order, size_t order_len, + int override) +{ + size_t bcs_orig_len = bcs_len; + bidi_char *bcs_orig = snewn(bcs_orig_len, bidi_char); + if (bcs_orig_len) + memcpy(bcs_orig, bcs, bcs_orig_len * sizeof(bidi_char)); + + bcs_len = do_bidi_test(ctx, bcs, bcs_len, override); + + /* + * TR9 revision 44 rule X9 says we remove explicit embedding + * controls and BN characters. So the test cases don't list them + * in the expected outputs. Do the same to our own output - unless + * we're testing the standard version of the algorithm, in which + * case, we expect the output to be exactly as the test cases say. + */ + unsigned *our_order = snewn(bcs_len, unsigned); + size_t our_order_len = 0; + for (size_t i = 0; i < bcs_len; i++) { + BidiType t = bidi_getType(bcs[i].wc); +#ifndef REMOVE_FORMATTING_CHARS + if (typeIsRemovedDuringProcessing(t)) + continue; +#endif + our_order[our_order_len++] = bcs[i].index; + } + + bool ok = false; + if (our_order_len == order_len) { + ok = true; + for (size_t i = 0; i < our_order_len; i++) + if (our_order[i] != order[i]) + ok = false; + } + if (ok) { + pass++; + } else { + fail++; + printf("%s:%u: failed order\n", filename, lineno); + printf(" input chars:"); + for (size_t i = 0; i < bcs_orig_len; i++) + printf(" %04x", bcs_orig[i].wc); + printf("\n"); + printf(" classes: "); + for (size_t i = 0; i < bcs_orig_len; i++) + printf(" %-4s", typenames[bidi_getType(bcs_orig[i].wc)]); + printf("\n"); + printf(" para level = %s\n", + override > 0 ? "LTR" : override < 0 ? "RTL" : "auto"); + printf(" expected:"); + for (size_t i = 0; i < order_len; i++) + printf(" %u", order[i]); + printf("\n"); + printf(" got: "); + for (size_t i = 0; i < our_order_len; i++) + printf(" %u", our_order[i]); + printf("\n"); + } + + /* Put the original data back so we can re-test with another override */ + memcpy(bcs, bcs_orig, bcs_orig_len * sizeof(bidi_char)); + + sfree(bcs_orig); + sfree(our_order); +} + +static void class_test(const char *filename, FILE *fp) +{ + unsigned lineno = 0; + size_t bcs_size = 0, bcs_len = 0; + bidi_char *bcs = NULL; + size_t order_size = 0, order_len = 0; + unsigned *order = NULL; + + /* Preliminary: find a representative character of every bidi + * type. Prefer positive-width ones if available. */ + unsigned representatives[N_BIDI_TYPES]; + for (size_t i = 0; i < N_BIDI_TYPES; i++) + representatives[i] = 0; + for (unsigned uc = 1; uc < 0x110000; uc++) { + unsigned type = bidi_getType(uc); + if (!representatives[type] || + (mk_wcwidth(representatives[type]) <= 0 && mk_wcwidth(uc) > 0)) + representatives[type] = uc; + } + + while (true) { + lineno++; + char *line = chomp(fgetline(fp)); + if (!line) + break; + + /* Skip blank lines and comments */ + if (!line[0] || line[0] == '#') { + sfree(line); + continue; + } + + /* Parse @Reorder lines, which tell us the expected output + * order for all following test cases (until superseded) */ + if (strstartswith(line, "@Reorder:")) { + char *p = line; + extract_word(&p); /* eat the "@Reorder:" header itself */ + order_len = 0; + while (1) { + const char *word = extract_word(&p); + if (!*word) + break; + sgrowarray(order, order_size, order_len); + order[order_len++] = strtoul(word, NULL, 0); + } + + sfree(line); + continue; + } + + /* Skip @Levels lines, which we don't (yet?) do anything with */ + if (strstartswith(line, "@Levels:")) { + sfree(line); + continue; + } + + /* Everything remaining should be an actual test */ + char *semicolon = strchr(line, ';'); + if (!semicolon) { + printf("%s:%u: bad test line': no bitmap\n", filename, lineno); + sfree(line); + continue; + } + *semicolon++ = '\0'; + unsigned bitmask = strtoul(semicolon, NULL, 0); + char *p = line; + bcs_len = 0; + bool test_ok = true; + while (1) { + const char *word = extract_word(&p); + if (!*word) + break; + unsigned type; + for (type = 0; type < N_BIDI_TYPES; type++) + if (!strcmp(word, typenames[type])) + break; + if (type == N_BIDI_TYPES) { + printf("%s:%u: bad test line: bad bidi type '%s'\n", + filename, lineno, word); + test_ok = false; + break; + } + sgrowarray(bcs, bcs_size, bcs_len); + bcs[bcs_len].wc = representatives[type]; + bcs[bcs_len].origwc = bcs[bcs_len].wc; + bcs[bcs_len].index = bcs_len; + bcs[bcs_len].nchars = 1; + bcs_len++; + } + + if (!test_ok) { + sfree(line); + continue; + } + + if (bitmask & 1) + run_test(filename, lineno, bcs, bcs_len, order, order_len, 0); + if (bitmask & 2) + run_test(filename, lineno, bcs, bcs_len, order, order_len, +1); + if (bitmask & 4) + run_test(filename, lineno, bcs, bcs_len, order, order_len, -1); + + sfree(line); + } + + sfree(bcs); + sfree(order); +} + +static void char_test(const char *filename, FILE *fp) +{ + unsigned lineno = 0; + size_t bcs_size = 0, bcs_len = 0; + bidi_char *bcs = NULL; + size_t order_size = 0, order_len = 0; + unsigned *order = NULL; + + while (true) { + lineno++; + char *line = chomp(fgetline(fp)); + if (!line) + break; + + /* Skip blank lines and comments */ + if (!line[0] || line[0] == '#') { + sfree(line); + continue; + } + + /* Break each test line up into its main fields */ + ptrlen input_pl, para_dir_pl, order_pl; + { + ptrlen pl = ptrlen_from_asciz(line); + input_pl = ptrlen_get_word(&pl, ";"); + para_dir_pl = ptrlen_get_word(&pl, ";"); + ptrlen_get_word(&pl, ";"); /* paragraph level, which we ignore */ + ptrlen_get_word(&pl, ";"); /* embedding levels, which we ignore */ + order_pl = ptrlen_get_word(&pl, ";"); + } + + int override; + { + char *para_dir_str = mkstr(para_dir_pl); + unsigned para_dir = strtoul(para_dir_str, NULL, 0); + sfree(para_dir_str); + + override = (para_dir == 0 ? +1 : para_dir == 1 ? -1 : 0); + } + + /* Break up the input into Unicode characters */ + bcs_len = 0; + { + ptrlen pl = input_pl; + while (pl.len) { + ptrlen chr = ptrlen_get_word(&pl, " "); + char *chrstr = mkstr(chr); + sgrowarray(bcs, bcs_size, bcs_len); + bcs[bcs_len].wc = strtoul(chrstr, NULL, 16); + bcs[bcs_len].origwc = bcs[bcs_len].wc; + bcs[bcs_len].index = bcs_len; + bcs[bcs_len].nchars = 1; + bcs_len++; + sfree(chrstr); + } + } + + /* Ditto the expected output order */ + order_len = 0; + { + ptrlen pl = order_pl; + while (pl.len) { + ptrlen chr = ptrlen_get_word(&pl, " "); + char *chrstr = mkstr(chr); + sgrowarray(order, order_size, order_len); + order[order_len++] = strtoul(chrstr, NULL, 0); + sfree(chrstr); + } + } + + run_test(filename, lineno, bcs, bcs_len, order, order_len, override); + sfree(line); + } + + sfree(bcs); + sfree(order); +} + +void out_of_memory(void) +{ + fprintf(stderr, "out of memory!\n"); + exit(2); +} + +static void usage(FILE *fp) +{ + fprintf(fp, "\ +usage: bidi_test ( ( --class | --char ) infile... )...\n\ +e.g.: bidi_test --class BidiTest.txt --char BidiCharacterTest.txt\n\ +also: --help display this text\n\ +"); +} + +int main(int argc, char **argv) +{ + void (*testfn)(const char *, FILE *) = NULL; + bool doing_opts = true; + const char *filename = NULL; + bool done_something = false; + + ctx = bidi_new_context(); + + while (--argc > 0) { + const char *arg = *++argv; + if (doing_opts && arg[0] == '-' && arg[1]) { + if (!strcmp(arg, "--")) { + doing_opts = false; + } else if (!strcmp(arg, "--class")) { + testfn = class_test; + } else if (!strcmp(arg, "--char")) { + testfn = char_test; + } else if (!strcmp(arg, "--help")) { + usage(stdout); + return 0; + } else { + fprintf(stderr, "unrecognised option '%s'\n", arg); + return 1; + } + } else { + const char *filename = arg; + + if (!testfn) { + fprintf(stderr, "no mode argument provided before filename " + "'%s'\n", filename); + return 1; + } + + if (!strcmp(filename, "-")) { + testfn("<standard input>", stdin); + } else { + FILE *fp = fopen(filename, "r"); + if (!fp) { + fprintf(stderr, "unable to open '%s'\n", filename); + return 1; + } + testfn(filename, fp); + fclose(fp); + } + done_something = true; + } + } + + if (!done_something) { + usage(stderr); + return 1; + } + + if (!filename) + filename = "-"; + + printf("pass %d fail %d total %d\n", pass, fail, pass + fail); + + bidi_free_context(ctx); + return fail != 0; +} diff --git a/terminal/terminal.c b/terminal/terminal.c new file mode 100644 index 00000000..37fa3513 --- /dev/null +++ b/terminal/terminal.c @@ -0,0 +1,7910 @@ +/* + * Terminal emulator. + */ + +#include <stdio.h> +#include <stdlib.h> +#include <ctype.h> +#include <limits.h> +#include <wchar.h> + +#include <time.h> +#include <assert.h> +#include "putty.h" +#include "terminal.h" + +#define VT52_PLUS + +#define CL_ANSIMIN 0x0001 /* Codes in all ANSI like terminals. */ +#define CL_VT100 0x0002 /* VT100 */ +#define CL_VT100AVO 0x0004 /* VT100 +AVO; 132x24 (not 132x14) & attrs */ +#define CL_VT102 0x0008 /* VT102 */ +#define CL_VT220 0x0010 /* VT220 */ +#define CL_VT320 0x0020 /* VT320 */ +#define CL_VT420 0x0040 /* VT420 */ +#define CL_VT510 0x0080 /* VT510, NB VT510 includes ANSI */ +#define CL_VT340TEXT 0x0100 /* VT340 extensions that appear in the VT420 */ +#define CL_SCOANSI 0x1000 /* SCOANSI not in ANSIMIN. */ +#define CL_ANSI 0x2000 /* ANSI ECMA-48 not in the VT100..VT420 */ +#define CL_OTHER 0x4000 /* Others, Xterm, linux, putty, dunno, etc */ + +#define TM_VT100 (CL_ANSIMIN|CL_VT100) +#define TM_VT100AVO (TM_VT100|CL_VT100AVO) +#define TM_VT102 (TM_VT100AVO|CL_VT102) +#define TM_VT220 (TM_VT102|CL_VT220) +#define TM_VTXXX (TM_VT220|CL_VT340TEXT|CL_VT510|CL_VT420|CL_VT320) +#define TM_SCOANSI (CL_ANSIMIN|CL_SCOANSI) + +#define TM_PUTTY (0xFFFF) + +#define UPDATE_DELAY ((TICKSPERSEC+49)/50)/* ticks to defer window update */ +#define TBLINK_DELAY ((TICKSPERSEC*9+19)/20)/* ticks between text blinks*/ +#define CBLINK_DELAY (CURSORBLINK) /* ticks between cursor blinks */ +#define VBELL_DELAY (VBELL_TIMEOUT) /* visual bell timeout in ticks */ + +#define compatibility(x) \ + if ( ((CL_##x)&term->compatibility_level) == 0 ) { \ + term->termstate=TOPLEVEL; \ + break; \ + } +#define compatibility2(x,y) \ + if ( ((CL_##x|CL_##y)&term->compatibility_level) == 0 ) { \ + term->termstate=TOPLEVEL; \ + break; \ + } + +#define has_compat(x) ( ((CL_##x)&term->compatibility_level) != 0 ) + +static const char *const EMPTY_WINDOW_TITLE = ""; + +static const char sco2ansicolour[] = { 0, 4, 2, 6, 1, 5, 3, 7 }; + +#define sel_nl_sz (sizeof(sel_nl)/sizeof(wchar_t)) +static const wchar_t sel_nl[] = SEL_NL; + +/* + * Fetch the character at a particular position in a line array, + * for purposes of `wordtype'. The reason this isn't just a simple + * array reference is that if the character we find is UCSWIDE, + * then we must look one space further to the left. + */ +#define UCSGET(a, x) \ + ( (x)>0 && (a)[(x)].chr == UCSWIDE ? (a)[(x)-1].chr : (a)[(x)].chr ) + +/* + * Detect the various aliases of U+0020 SPACE. + */ +#define IS_SPACE_CHR(chr) \ + ((chr) == 0x20 || (DIRECT_CHAR(chr) && ((chr) & 0xFF) == 0x20)) + +/* + * Spot magic CSETs. + */ +#define CSET_OF(chr) (DIRECT_CHAR(chr)||DIRECT_FONT(chr) ? (chr)&CSET_MASK : 0) + +/* + * Internal prototypes. + */ +static void resizeline(Terminal *, termline *, int); +static termline *lineptr(Terminal *, int, int, int); +static void unlineptr(termline *); +static void check_line_size(Terminal *, termline *); +static void do_paint(Terminal *); +static void erase_lots(Terminal *, bool, bool, bool); +static int find_last_nonempty_line(Terminal *, tree234 *); +static void swap_screen(Terminal *, int, bool, bool); +static void update_sbar(Terminal *); +static void deselect(Terminal *); +static void term_print_finish(Terminal *); +static void scroll(Terminal *, int, int, int, bool); +static void parse_optionalrgb(optionalrgb *out, unsigned *values); +static void term_added_data(Terminal *term, bool); +static void term_update_raw_mouse_mode(Terminal *term); +static void term_out_cb(void *); + +static termline *newtermline(Terminal *term, int cols, bool bce) +{ + termline *line; + int j; + + line = snew(termline); + line->chars = snewn(cols, termchar); + for (j = 0; j < cols; j++) + line->chars[j] = (bce ? term->erase_char : term->basic_erase_char); + line->cols = line->size = cols; + line->lattr = LATTR_NORM; + line->trusted = false; + line->temporary = false; + line->cc_free = 0; + + return line; +} + +static void freetermline(termline *line) +{ + if (line) { + sfree(line->chars); + sfree(line); + } +} + +static void unlineptr(termline *line) +{ + if (line->temporary) + freetermline(line); +} + +const int colour_indices_conf_to_oscp[CONF_NCOLOURS] = { + #define COLOUR_ENTRY(id,name) OSCP_COLOUR_##id, + CONF_COLOUR_LIST(COLOUR_ENTRY) + #undef COLOUR_ENTRY +}; + +const int colour_indices_conf_to_osc4[CONF_NCOLOURS] = { + #define COLOUR_ENTRY(id,name) OSC4_COLOUR_##id, + CONF_COLOUR_LIST(COLOUR_ENTRY) + #undef COLOUR_ENTRY +}; + +const int colour_indices_oscp_to_osc4[OSCP_NCOLOURS] = { + #define COLOUR_ENTRY(id) OSC4_COLOUR_##id, + OSCP_COLOUR_LIST(COLOUR_ENTRY) + #undef COLOUR_ENTRY +}; + +#ifdef TERM_CC_DIAGS +/* + * Diagnostic function: verify that a termline has a correct + * combining character structure. + * + * This is a performance-intensive check, so it's no longer enabled + * by default. + */ +static void cc_check(termline *line) +{ + unsigned char *flags; + int i, j; + + assert(line->size >= line->cols); + + flags = snewn(line->size, unsigned char); + + for (i = 0; i < line->size; i++) + flags[i] = (i < line->cols); + + for (i = 0; i < line->cols; i++) { + j = i; + while (line->chars[j].cc_next) { + j += line->chars[j].cc_next; + assert(j >= line->cols && j < line->size); + assert(!flags[j]); + flags[j] = true; + } + } + + j = line->cc_free; + if (j) { + while (1) { + assert(j >= line->cols && j < line->size); + assert(!flags[j]); + flags[j] = true; + if (line->chars[j].cc_next) + j += line->chars[j].cc_next; + else + break; + } + } + + j = 0; + for (i = 0; i < line->size; i++) + j += (flags[i] != 0); + + assert(j == line->size); + + sfree(flags); +} +#endif + +static void clear_cc(termline *line, int col); + +/* + * Add a combining character to a character cell. + */ +static void add_cc(termline *line, int col, unsigned long chr) +{ + int newcc; + + assert(col >= 0 && col < line->cols); + + /* + * Don't add combining characters at all to U+FFFD REPLACEMENT + * CHARACTER. (Partly it's a slightly incoherent idea in the first + * place; mostly, U+FFFD is what we generate if a cell already has + * too many ccs, in which case we want it to be a fixed point when + * further ccs are added.) + */ + if (line->chars[col].chr == 0xFFFD) + return; + + /* + * Walk the cc list of the cell in question to find its current + * end point. + */ + size_t ncc = 0; + int origcol = col; + while (line->chars[col].cc_next) { + col += line->chars[col].cc_next; + if (++ncc >= CC_LIMIT) { + /* + * There are already too many combining characters in this + * character cell. Change strategy: throw out the entire + * chain and replace the main character with U+FFFD. + * + * (Rationale: extrapolating from UTR #36 section 3.6.2 + * suggests the principle that it's better to substitute + * U+FFFD than to _ignore_ input completely. Also, if the + * user copies and pastes an overcombined character cell, + * this way it will clearly indicate that we haven't + * reproduced the writer's original intentions, instead of + * looking as if it was the _writer's_ fault that the 33rd + * cc is missing.) + * + * Per the code above, this will also prevent any further + * ccs from being added to this cell. + */ + clear_cc(line, origcol); + line->chars[origcol].chr = 0xFFFD; + return; + } + } + + /* + * Extend the cols array if the free list is empty. + */ + if (!line->cc_free) { + int n = line->size; + + size_t tmpsize = line->size; + sgrowarray(line->chars, tmpsize, tmpsize); + assert(tmpsize <= INT_MAX); + line->size = tmpsize; + + line->cc_free = n; + while (n < line->size) { + if (n+1 < line->size) + line->chars[n].cc_next = 1; + else + line->chars[n].cc_next = 0; + n++; + } + } + + /* + * `col' now points at the last cc currently in this cell; so + * we simply add another one. + */ + newcc = line->cc_free; + if (line->chars[newcc].cc_next) + line->cc_free = newcc + line->chars[newcc].cc_next; + else + line->cc_free = 0; + line->chars[newcc].cc_next = 0; + line->chars[newcc].chr = chr; + line->chars[col].cc_next = newcc - col; + +#ifdef TERM_CC_DIAGS + cc_check(line); +#endif +} + +/* + * Clear the combining character list in a character cell. + */ +static void clear_cc(termline *line, int col) +{ + int oldfree, origcol = col; + + assert(col >= 0 && col < line->cols); + + if (!line->chars[col].cc_next) + return; /* nothing needs doing */ + + oldfree = line->cc_free; + line->cc_free = col + line->chars[col].cc_next; + while (line->chars[col].cc_next) + col += line->chars[col].cc_next; + if (oldfree) + line->chars[col].cc_next = oldfree - col; + else + line->chars[col].cc_next = 0; + + line->chars[origcol].cc_next = 0; + +#ifdef TERM_CC_DIAGS + cc_check(line); +#endif +} + +/* + * Compare two character cells for equality. Special case required + * in do_paint() where we override what we expect the chr and attr + * fields to be. + */ +static bool termchars_equal_override(termchar *a, termchar *b, + unsigned long bchr, unsigned long battr) +{ + /* FULL-TERMCHAR */ + if (!truecolour_equal(a->truecolour, b->truecolour)) + return false; + if (a->chr != bchr) + return false; + if ((a->attr &~ DATTR_MASK) != (battr &~ DATTR_MASK)) + return false; + while (a->cc_next || b->cc_next) { + if (!a->cc_next || !b->cc_next) + return false; /* one cc-list ends, other does not */ + a += a->cc_next; + b += b->cc_next; + if (a->chr != b->chr) + return false; + } + return true; +} + +static bool termchars_equal(termchar *a, termchar *b) +{ + return termchars_equal_override(a, b, b->chr, b->attr); +} + +/* + * Copy a character cell. (Requires a pointer to the destination + * termline, so as to access its free list.) + */ +static void copy_termchar(termline *destline, int x, termchar *src) +{ + clear_cc(destline, x); + + destline->chars[x] = *src; /* copy everything except cc-list */ + destline->chars[x].cc_next = 0; /* and make sure this is zero */ + + while (src->cc_next) { + src += src->cc_next; + add_cc(destline, x, src->chr); + } + +#ifdef TERM_CC_DIAGS + cc_check(destline); +#endif +} + +/* + * Move a character cell within its termline. + */ +static void move_termchar(termline *line, termchar *dest, termchar *src) +{ + /* First clear the cc list from the original char, just in case. */ + clear_cc(line, dest - line->chars); + + /* Move the character cell and adjust its cc_next. */ + *dest = *src; /* copy everything except cc-list */ + if (src->cc_next) + dest->cc_next = src->cc_next - (dest-src); + + /* Ensure the original cell doesn't have a cc list. */ + src->cc_next = 0; + +#ifdef TERM_CC_DIAGS + cc_check(line); +#endif +} + +/* + * Compress and decompress a termline into an RLE-based format for + * storing in scrollback. (Since scrollback almost never needs to + * be modified and exists in huge quantities, this is a sensible + * tradeoff, particularly since it allows us to continue adding + * features to the main termchar structure without proportionally + * bloating the terminal emulator's memory footprint unless those + * features are in constant use.) + */ +static void makerle(strbuf *b, termline *ldata, + void (*makeliteral)(strbuf *b, termchar *c, + unsigned long *state)) +{ + int hdrpos, hdrsize, n, prevlen, prevpos, thislen, thispos; + bool prev2; + termchar *c = ldata->chars; + unsigned long state = 0, oldstate; + + n = ldata->cols; + + hdrpos = b->len; + hdrsize = 0; + put_byte(b, 0); + prevlen = prevpos = 0; + prev2 = false; + + while (n-- > 0) { + thispos = b->len; + makeliteral(b, c++, &state); + thislen = b->len - thispos; + if (thislen == prevlen && + !memcmp(b->u + prevpos, b->u + thispos, thislen)) { + /* + * This literal precisely matches the previous one. + * Turn it into a run if it's worthwhile. + * + * With one-byte literals, it costs us two bytes to + * encode a run, plus another byte to write the header + * to resume normal output; so a three-element run is + * neutral, and anything beyond that is unconditionally + * worthwhile. With two-byte literals or more, even a + * 2-run is a win. + */ + if (thislen > 1 || prev2) { + int runpos, runlen; + + /* + * It's worth encoding a run. Start at prevpos, + * unless hdrsize==0 in which case we can back up + * another one and start by overwriting hdrpos. + */ + + hdrsize--; /* remove the literal at prevpos */ + if (prev2) { + assert(hdrsize > 0); + hdrsize--; + prevpos -= prevlen;/* and possibly another one */ + } + + if (hdrsize == 0) { + assert(prevpos == hdrpos + 1); + runpos = hdrpos; + strbuf_shrink_to(b, prevpos+prevlen); + } else { + memmove(b->u + prevpos+1, b->u + prevpos, prevlen); + runpos = prevpos; + strbuf_shrink_to(b, prevpos+prevlen+1); + /* + * Terminate the previous run of ordinary + * literals. + */ + assert(hdrsize >= 1 && hdrsize <= 128); + b->u[hdrpos] = hdrsize - 1; + } + + runlen = prev2 ? 3 : 2; + + while (n > 0 && runlen < 129) { + int tmppos, tmplen; + tmppos = b->len; + oldstate = state; + makeliteral(b, c, &state); + tmplen = b->len - tmppos; + bool match = tmplen == thislen && + !memcmp(b->u + runpos+1, b->u + tmppos, tmplen); + strbuf_shrink_to(b, tmppos); + if (!match) { + state = oldstate; + break; /* run over */ + } + n--, c++, runlen++; + } + + assert(runlen >= 2 && runlen <= 129); + b->u[runpos] = runlen + 0x80 - 2; + + hdrpos = b->len; + hdrsize = 0; + put_byte(b, 0); + /* And ensure this run doesn't interfere with the next. */ + prevlen = prevpos = 0; + prev2 = false; + + continue; + } else { + /* + * Just flag that the previous two literals were + * identical, in case we find a third identical one + * we want to turn into a run. + */ + prev2 = true; + prevlen = thislen; + prevpos = thispos; + } + } else { + prev2 = false; + prevlen = thislen; + prevpos = thispos; + } + + /* + * This character isn't (yet) part of a run. Add it to + * hdrsize. + */ + hdrsize++; + if (hdrsize == 128) { + b->u[hdrpos] = hdrsize - 1; + hdrpos = b->len; + hdrsize = 0; + put_byte(b, 0); + prevlen = prevpos = 0; + prev2 = false; + } + } + + /* + * Clean up. + */ + if (hdrsize > 0) { + assert(hdrsize <= 128); + b->u[hdrpos] = hdrsize - 1; + } else { + strbuf_shrink_to(b, hdrpos); + } +} +static void makeliteral_chr(strbuf *b, termchar *c, unsigned long *state) +{ + /* + * My encoding for characters is UTF-8-like, in that it stores + * 7-bit ASCII in one byte and uses high-bit-set bytes as + * introducers to indicate a longer sequence. However, it's + * unlike UTF-8 in that it doesn't need to be able to + * resynchronise, and therefore I don't want to waste two bits + * per byte on having recognisable continuation characters. + * Also I don't want to rule out the possibility that I may one + * day use values 0x80000000-0xFFFFFFFF for interesting + * purposes, so unlike UTF-8 I need a full 32-bit range. + * Accordingly, here is my encoding: + * + * 00000000-0000007F: 0xxxxxxx (but see below) + * 00000080-00003FFF: 10xxxxxx xxxxxxxx + * 00004000-001FFFFF: 110xxxxx xxxxxxxx xxxxxxxx + * 00200000-0FFFFFFF: 1110xxxx xxxxxxxx xxxxxxxx xxxxxxxx + * 10000000-FFFFFFFF: 11110ZZZ xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx + * + * (`Z' is like `x' but is always going to be zero since the + * values I'm encoding don't go above 2^32. In principle the + * five-byte form of the encoding could extend to 2^35, and + * there could be six-, seven-, eight- and nine-byte forms as + * well to allow up to 64-bit values to be encoded. But that's + * completely unnecessary for these purposes!) + * + * The encoding as written above would be very simple, except + * that 7-bit ASCII can occur in several different ways in the + * terminal data; sometimes it crops up in the D800 page + * (CSET_ASCII) but at other times it's in the 0000 page (real + * Unicode). Therefore, this encoding is actually _stateful_: + * the one-byte encoding of 00-7F actually indicates `reuse the + * upper three bytes of the last character', and to encode an + * absolute value of 00-7F you need to use the two-byte form + * instead. + */ + if ((c->chr & ~0x7F) == *state) { + put_byte(b, (unsigned char)(c->chr & 0x7F)); + } else if (c->chr < 0x4000) { + put_byte(b, (unsigned char)(((c->chr >> 8) & 0x3F) | 0x80)); + put_byte(b, (unsigned char)(c->chr & 0xFF)); + } else if (c->chr < 0x200000) { + put_byte(b, (unsigned char)(((c->chr >> 16) & 0x1F) | 0xC0)); + put_uint16(b, c->chr & 0xFFFF); + } else if (c->chr < 0x10000000) { + put_byte(b, (unsigned char)(((c->chr >> 24) & 0x0F) | 0xE0)); + put_byte(b, (unsigned char)((c->chr >> 16) & 0xFF)); + put_uint16(b, c->chr & 0xFFFF); + } else { + put_byte(b, 0xF0); + put_uint32(b, c->chr); + } + *state = c->chr & ~0xFF; +} +static void makeliteral_attr(strbuf *b, termchar *c, unsigned long *state) +{ + /* + * My encoding for attributes is 16-bit-granular and assumes + * that the top bit of the word is never required. I either + * store a two-byte value with the top bit clear (indicating + * just that value), or a four-byte value with the top bit set + * (indicating the same value with its top bit clear). + * + * However, first I permute the bits of the attribute value, so + * that the eight bits of colour (four in each of fg and bg) + * which are never non-zero unless xterm 256-colour mode is in + * use are placed higher up the word than everything else. This + * ensures that attribute values remain 16-bit _unless_ the + * user uses extended colour. + */ + unsigned attr, colourbits; + + attr = c->attr; + + assert(ATTR_BGSHIFT > ATTR_FGSHIFT); + + colourbits = (attr >> (ATTR_BGSHIFT + 4)) & 0xF; + colourbits <<= 4; + colourbits |= (attr >> (ATTR_FGSHIFT + 4)) & 0xF; + + attr = (((attr >> (ATTR_BGSHIFT + 8)) << (ATTR_BGSHIFT + 4)) | + (attr & ((1 << (ATTR_BGSHIFT + 4))-1))); + attr = (((attr >> (ATTR_FGSHIFT + 8)) << (ATTR_FGSHIFT + 4)) | + (attr & ((1 << (ATTR_FGSHIFT + 4))-1))); + + attr |= (colourbits << (32-9)); + + if (attr < 0x8000) { + put_byte(b, (unsigned char)((attr >> 8) & 0xFF)); + put_byte(b, (unsigned char)(attr & 0xFF)); + } else { + put_byte(b, (unsigned char)(((attr >> 24) & 0x7F) | 0x80)); + put_byte(b, (unsigned char)((attr >> 16) & 0xFF)); + put_byte(b, (unsigned char)((attr >> 8) & 0xFF)); + put_byte(b, (unsigned char)(attr & 0xFF)); + } +} +static void makeliteral_truecolour(strbuf *b, termchar *c, unsigned long *state) +{ + /* + * Put the used parts of the colour info into the buffer. + */ + put_byte(b, ((c->truecolour.fg.enabled ? 1 : 0) | + (c->truecolour.bg.enabled ? 2 : 0))); + if (c->truecolour.fg.enabled) { + put_byte(b, c->truecolour.fg.r); + put_byte(b, c->truecolour.fg.g); + put_byte(b, c->truecolour.fg.b); + } + if (c->truecolour.bg.enabled) { + put_byte(b, c->truecolour.bg.r); + put_byte(b, c->truecolour.bg.g); + put_byte(b, c->truecolour.bg.b); + } +} +static void makeliteral_cc(strbuf *b, termchar *c, unsigned long *state) +{ + /* + * For combining characters, I just encode a bunch of ordinary + * chars using makeliteral_chr, and terminate with a \0 + * character (which I know won't come up as a combining char + * itself). + * + * I don't use the stateful encoding in makeliteral_chr. + */ + unsigned long zstate; + termchar z; + + while (c->cc_next) { + c += c->cc_next; + + assert(c->chr != 0); + + zstate = 0; + makeliteral_chr(b, c, &zstate); + } + + z.chr = 0; + zstate = 0; + makeliteral_chr(b, &z, &zstate); +} + +typedef struct compressed_scrollback_line { + size_t len; +} compressed_scrollback_line; + +static termline *decompressline(compressed_scrollback_line *line); + +static compressed_scrollback_line *compressline(termline *ldata) +{ + strbuf *b = strbuf_new(); + + /* Leave space for the header structure */ + strbuf_append(b, sizeof(compressed_scrollback_line)); + + /* + * First, store the column count, 7 bits at a time, least + * significant `digit' first, with the high bit set on all but + * the last. + */ + { + int n = ldata->cols; + while (n >= 128) { + put_byte(b, (unsigned char)((n & 0x7F) | 0x80)); + n >>= 7; + } + put_byte(b, (unsigned char)(n)); + } + + /* + * Next store the lattrs; same principle. We add one extra bit to + * this to indicate the trust state of the line. + */ + { + int n = ldata->lattr | (ldata->trusted ? 0x10000 : 0); + while (n >= 128) { + put_byte(b, (unsigned char)((n & 0x7F) | 0x80)); + n >>= 7; + } + put_byte(b, (unsigned char)(n)); + } + + /* + * Now we store a sequence of separate run-length encoded + * fragments, each containing exactly as many symbols as there + * are columns in the ldata. + * + * All of these have a common basic format: + * + * - a byte 00-7F indicates that X+1 literals follow it + * - a byte 80-FF indicates that a single literal follows it + * and expects to be repeated (X-0x80)+2 times. + * + * The format of the `literals' varies between the fragments. + */ + makerle(b, ldata, makeliteral_chr); + makerle(b, ldata, makeliteral_attr); + makerle(b, ldata, makeliteral_truecolour); + makerle(b, ldata, makeliteral_cc); + + size_t linelen = b->len - sizeof(compressed_scrollback_line); + compressed_scrollback_line *line = + (compressed_scrollback_line *)strbuf_to_str(b); + line->len = linelen; + + /* + * Diagnostics: ensure that the compressed data really does + * decompress to the right thing. + * + * This is a bit performance-heavy for production code. + */ +#ifdef TERM_CC_DIAGS +#ifndef CHECK_SB_COMPRESSION + { + termline *dcl; + int i; + +#ifdef DIAGNOSTIC_SB_COMPRESSION + for (i = 0; i < b->len; i++) { + printf(" %02x ", b->data[i]); + } + printf("\n"); +#endif + + dcl = decompressline(line); + assert(ldata->cols == dcl->cols); + assert(ldata->lattr == dcl->lattr); + for (i = 0; i < ldata->cols; i++) + assert(termchars_equal(&ldata->chars[i], &dcl->chars[i])); + +#ifdef DIAGNOSTIC_SB_COMPRESSION + printf("%d cols (%d bytes) -> %d bytes (factor of %g)\n", + ldata->cols, 4 * ldata->cols, dused, + (double)dused / (4 * ldata->cols)); +#endif + + freetermline(dcl); + } +#endif +#endif /* TERM_CC_DIAGS */ + + return line; +} + +static void readrle(BinarySource *bs, termline *ldata, + void (*readliteral)(BinarySource *bs, termchar *c, + termline *ldata, unsigned long *state)) +{ + int n = 0; + unsigned long state = 0; + + while (n < ldata->cols) { + int hdr = get_byte(bs); + + if (hdr >= 0x80) { + /* A run. */ + + size_t pos = bs->pos, count = hdr + 2 - 0x80; + while (count--) { + assert(n < ldata->cols); + bs->pos = pos; + readliteral(bs, ldata->chars + n, ldata, &state); + n++; + } + } else { + /* Just a sequence of consecutive literals. */ + + int count = hdr + 1; + while (count--) { + assert(n < ldata->cols); + readliteral(bs, ldata->chars + n, ldata, &state); + n++; + } + } + } + + assert(n == ldata->cols); +} +static void readliteral_chr(BinarySource *bs, termchar *c, termline *ldata, + unsigned long *state) +{ + int byte; + + /* + * 00000000-0000007F: 0xxxxxxx + * 00000080-00003FFF: 10xxxxxx xxxxxxxx + * 00004000-001FFFFF: 110xxxxx xxxxxxxx xxxxxxxx + * 00200000-0FFFFFFF: 1110xxxx xxxxxxxx xxxxxxxx xxxxxxxx + * 10000000-FFFFFFFF: 11110ZZZ xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx + */ + + byte = get_byte(bs); + if (byte < 0x80) { + c->chr = byte | *state; + } else if (byte < 0xC0) { + c->chr = (byte &~ 0xC0) << 8; + c->chr |= get_byte(bs); + } else if (byte < 0xE0) { + c->chr = (byte &~ 0xE0) << 16; + c->chr |= get_uint16(bs); + } else if (byte < 0xF0) { + c->chr = (byte &~ 0xF0) << 24; + c->chr |= get_byte(bs) << 16; + c->chr |= get_uint16(bs); + } else { + assert(byte == 0xF0); + c->chr = get_uint32(bs); + } + *state = c->chr & ~0xFF; +} +static void readliteral_attr(BinarySource *bs, termchar *c, termline *ldata, + unsigned long *state) +{ + unsigned val, attr, colourbits; + + val = get_uint16(bs); + + if (val >= 0x8000) { + val &= ~0x8000; + val <<= 16; + val |= get_uint16(bs); + } + + colourbits = (val >> (32-9)) & 0xFF; + attr = (val & ((1<<(32-9))-1)); + + attr = (((attr >> (ATTR_FGSHIFT + 4)) << (ATTR_FGSHIFT + 8)) | + (attr & ((1 << (ATTR_FGSHIFT + 4))-1))); + attr = (((attr >> (ATTR_BGSHIFT + 4)) << (ATTR_BGSHIFT + 8)) | + (attr & ((1 << (ATTR_BGSHIFT + 4))-1))); + + attr |= (colourbits >> 4) << (ATTR_BGSHIFT + 4); + attr |= (colourbits & 0xF) << (ATTR_FGSHIFT + 4); + + c->attr = attr; +} +static void readliteral_truecolour( + BinarySource *bs, termchar *c, termline *ldata, unsigned long *state) +{ + int flags = get_byte(bs); + + if (flags & 1) { + c->truecolour.fg.enabled = true; + c->truecolour.fg.r = get_byte(bs); + c->truecolour.fg.g = get_byte(bs); + c->truecolour.fg.b = get_byte(bs); + } else { + c->truecolour.fg = optionalrgb_none; + } + + if (flags & 2) { + c->truecolour.bg.enabled = true; + c->truecolour.bg.r = get_byte(bs); + c->truecolour.bg.g = get_byte(bs); + c->truecolour.bg.b = get_byte(bs); + } else { + c->truecolour.bg = optionalrgb_none; + } +} +static void readliteral_cc(BinarySource *bs, termchar *c, termline *ldata, + unsigned long *state) +{ + termchar n; + unsigned long zstate; + int x = c - ldata->chars; + + c->cc_next = 0; + + while (1) { + zstate = 0; + readliteral_chr(bs, &n, ldata, &zstate); + if (!n.chr) + break; + add_cc(ldata, x, n.chr); + } +} + +static termline *decompressline(compressed_scrollback_line *line) +{ + int ncols, byte, shift; + BinarySource bs[1]; + termline *ldata; + + BinarySource_BARE_INIT(bs, line+1, line->len); + + /* + * First read in the column count. + */ + ncols = shift = 0; + do { + byte = get_byte(bs); + ncols |= (byte & 0x7F) << shift; + shift += 7; + } while (byte & 0x80); + + /* + * Now create the output termline. + */ + ldata = snew(termline); + ldata->chars = snewn(ncols, termchar); + ldata->cols = ldata->size = ncols; + ldata->temporary = true; + ldata->cc_free = 0; + + /* + * We must set all the cc pointers in ldata->chars to 0 right + * now, so that cc diagnostics that verify the integrity of the + * whole line will make sense while we're in the middle of + * building it up. + */ + { + int i; + for (i = 0; i < ldata->cols; i++) + ldata->chars[i].cc_next = 0; + } + + /* + * Now read in the lattr. + */ + int lattr = shift = 0; + do { + byte = get_byte(bs); + lattr |= (byte & 0x7F) << shift; + shift += 7; + } while (byte & 0x80); + ldata->lattr = lattr & 0xFFFF; + ldata->trusted = (lattr & 0x10000) != 0; + + /* + * Now we read in each of the RLE streams in turn. + */ + readrle(bs, ldata, readliteral_chr); + readrle(bs, ldata, readliteral_attr); + readrle(bs, ldata, readliteral_truecolour); + readrle(bs, ldata, readliteral_cc); + + /* And we always expect that we ended up exactly at the end of the + * compressed data. */ + assert(!get_err(bs)); + assert(get_avail(bs) == 0); + + return ldata; +} + +/* + * Resize a line to make it `cols' columns wide. + */ +static void resizeline(Terminal *term, termline *line, int cols) +{ + int i, oldcols; + + if (line->cols != cols) { + + oldcols = line->cols; + + /* + * This line is the wrong length, which probably means it + * hasn't been accessed since a resize. Resize it now. + * + * First, go through all the characters that will be thrown + * out in the resize (if we're shrinking the line) and + * return their cc lists to the cc free list. + */ + for (i = cols; i < oldcols; i++) + clear_cc(line, i); + + /* + * If we're shrinking the line, we now bodily move the + * entire cc section from where it started to where it now + * needs to be. (We have to do this before the resize, so + * that the data we're copying is still there. However, if + * we're expanding, we have to wait until _after_ the + * resize so that the space we're copying into is there.) + */ + if (cols < oldcols) + memmove(line->chars + cols, line->chars + oldcols, + (line->size - line->cols) * TSIZE); + + /* + * Now do the actual resize, leaving the _same_ amount of + * cc space as there was to begin with. + */ + line->size += cols - oldcols; + line->chars = sresize(line->chars, line->size, TTYPE); + line->cols = cols; + + /* + * If we're expanding the line, _now_ we move the cc + * section. + */ + if (cols > oldcols) + memmove(line->chars + cols, line->chars + oldcols, + (line->size - line->cols) * TSIZE); + + /* + * Go through what's left of the original line, and adjust + * the first cc_next pointer in each list. (All the + * subsequent ones are still valid because they are + * relative offsets within the cc block.) Also do the same + * to the head of the cc_free list. + */ + for (i = 0; i < oldcols && i < cols; i++) + if (line->chars[i].cc_next) + line->chars[i].cc_next += cols - oldcols; + if (line->cc_free) + line->cc_free += cols - oldcols; + + /* + * And finally fill in the new space with erase chars. (We + * don't have to worry about cc lists here, because we + * _know_ the erase char doesn't have one.) + */ + for (i = oldcols; i < cols; i++) + line->chars[i] = term->basic_erase_char; + +#ifdef TERM_CC_DIAGS + cc_check(line); +#endif + } +} + +/* + * Get the number of lines in the scrollback. + */ +static int sblines(Terminal *term) +{ + int sblines = count234(term->scrollback); + if (term->erase_to_scrollback && + term->alt_which && term->alt_screen) { + sblines += term->alt_sblines; + } + return sblines; +} + +static void null_line_error(Terminal *term, int y, int lineno, + tree234 *whichtree, int treeindex, + const char *varname) +{ + modalfatalbox("%s==NULL in terminal.c\n" + "lineno=%d y=%d w=%d h=%d\n" + "count(scrollback=%p)=%d\n" + "count(screen=%p)=%d\n" + "count(alt=%p)=%d alt_sblines=%d\n" + "whichtree=%p treeindex=%d\n" + "commitid=%s\n\n" + "Please contact <putty@projects.tartarus.org> " + "and pass on the above information.", + varname, lineno, y, term->cols, term->rows, + term->scrollback, count234(term->scrollback), + term->screen, count234(term->screen), + term->alt_screen, count234(term->alt_screen), + term->alt_sblines, whichtree, treeindex, commitid); +} + +/* + * Retrieve a line of the screen or of the scrollback, according to + * whether the y coordinate is non-negative or negative + * (respectively). + */ +static termline *lineptr(Terminal *term, int y, int lineno, int screen) +{ + termline *line; + tree234 *whichtree; + int treeindex; + + if (y >= 0) { + whichtree = term->screen; + treeindex = y; + } else { + int altlines = 0; + + assert(!screen); + + if (term->erase_to_scrollback && + term->alt_which && term->alt_screen) { + altlines = term->alt_sblines; + } + if (y < -altlines) { + whichtree = term->scrollback; + treeindex = y + altlines + count234(term->scrollback); + } else { + whichtree = term->alt_screen; + treeindex = y + term->alt_sblines; + /* treeindex = y + count234(term->alt_screen); */ + } + } + if (whichtree == term->scrollback) { + compressed_scrollback_line *cline = index234(whichtree, treeindex); + if (!cline) + null_line_error(term, y, lineno, whichtree, treeindex, "cline"); + line = decompressline(cline); + } else { + line = index234(whichtree, treeindex); + } + + /* We assume that we don't screw up and retrieve something out of range. */ + if (line == NULL) + null_line_error(term, y, lineno, whichtree, treeindex, "line"); + assert(line != NULL); + + /* + * Here we resize lines to _at least_ the right length, but we + * don't truncate them. Truncation is done as a side effect of + * modifying the line. + * + * The point of this policy is to try to arrange that resizing the + * terminal window repeatedly - e.g. successive steps in an X11 + * opaque window-resize drag, or resizing as a side effect of + * retiling by tiling WMs such as xmonad - does not throw away + * data gratuitously. Specifically, we want a sequence of resize + * operations with no terminal output between them to have the + * same effect as a single resize to the ultimate terminal size, + * and also (for the case in which xmonad narrows a window that's + * scrolling things) we want scrolling up new text at the bottom + * of a narrowed window to avoid truncating lines further up when + * the window is re-widened. + */ + if (term->cols > line->cols) + resizeline(term, line, term->cols); + + return line; +} + +#define lineptr(x) (lineptr)(term,x,__LINE__,0) +#define scrlineptr(x) (lineptr)(term,x,__LINE__,1) + +/* + * Coerce a termline to the terminal's current width. Unlike the + * optional resize in lineptr() above, this is potentially destructive + * of text, since it can shrink as well as grow the line. + * + * We call this whenever a termline is actually going to be modified. + * Helpfully, putting a single call to this function in check_boundary + * deals with _nearly_ all such cases, leaving only a few things like + * bulk erase and ESC#8 to handle separately. + */ +static void check_line_size(Terminal *term, termline *line) +{ + if (term->cols != line->cols) /* trivial optimisation */ + resizeline(term, line, term->cols); +} + +static void term_schedule_tblink(Terminal *term); +static void term_schedule_cblink(Terminal *term); +static void term_update_callback(void *ctx); + +static void term_timer(void *ctx, unsigned long now) +{ + Terminal *term = (Terminal *)ctx; + + if (term->tblink_pending && now == term->next_tblink) { + term->tblinker = !term->tblinker; + term->tblink_pending = false; + term_schedule_tblink(term); + term->window_update_pending = true; + } + + if (term->cblink_pending && now == term->next_cblink) { + term->cblinker = !term->cblinker; + term->cblink_pending = false; + term_schedule_cblink(term); + term->window_update_pending = true; + } + + if (term->in_vbell && now == term->vbell_end) { + term->in_vbell = false; + term->window_update_pending = true; + } + + if (term->window_update_cooldown && + now == term->window_update_cooldown_end) { + term->window_update_cooldown = false; + } + + if (term->window_update_pending) + term_update_callback(term); +} + +static void term_update_callback(void *ctx) +{ + Terminal *term = (Terminal *)ctx; + if (!term->window_update_pending) + return; + if (!term->window_update_cooldown) { + term_update(term); + term->window_update_cooldown = true; + term->window_update_cooldown_end = schedule_timer( + UPDATE_DELAY, term_timer, term); + } +} + +static void term_schedule_update(Terminal *term) +{ + if (!term->window_update_pending) { + term->window_update_pending = true; + queue_toplevel_callback(term_update_callback, term); + } +} + +/* + * Call this whenever the terminal window state changes, to queue + * an update. + */ +static void seen_disp_event(Terminal *term) +{ + term->seen_disp_event = true; /* for scrollback-reset-on-activity */ + term_schedule_update(term); +} + +/* + * Call when the terminal's blinking-text settings change, or when + * a text blink has just occurred. + */ +static void term_schedule_tblink(Terminal *term) +{ + if (term->blink_is_real) { + if (!term->tblink_pending) + term->next_tblink = schedule_timer(TBLINK_DELAY, term_timer, term); + term->tblink_pending = true; + } else { + term->tblinker = true; /* reset when not in use */ + term->tblink_pending = false; + } +} + +/* + * Likewise with cursor blinks. + */ +static void term_schedule_cblink(Terminal *term) +{ + if (term->blink_cur && term->has_focus) { + if (!term->cblink_pending) + term->next_cblink = schedule_timer(CBLINK_DELAY, term_timer, term); + term->cblink_pending = true; + } else { + term->cblinker = true; /* reset when not in use */ + term->cblink_pending = false; + } +} + +/* + * Call to reset cursor blinking on new output. + */ +static void term_reset_cblink(Terminal *term) +{ + seen_disp_event(term); + term->cblinker = true; + term->cblink_pending = false; + term_schedule_cblink(term); +} + +/* + * Call to begin a visual bell. + */ +static void term_schedule_vbell(Terminal *term, bool already_started, + long startpoint) +{ + long ticks_already_gone; + + if (already_started) + ticks_already_gone = GETTICKCOUNT() - startpoint; + else + ticks_already_gone = 0; + + if (ticks_already_gone < VBELL_DELAY) { + term->in_vbell = true; + term->vbell_end = schedule_timer(VBELL_DELAY - ticks_already_gone, + term_timer, term); + } else { + term->in_vbell = false; + } +} + +/* + * Set up power-on settings for the terminal. + * If 'clear' is false, don't actually clear the primary screen, and + * position the cursor below the last non-blank line (scrolling if + * necessary). + */ +static void power_on(Terminal *term, bool clear) +{ + term->alt_x = term->alt_y = 0; + term->savecurs.x = term->savecurs.y = 0; + term->alt_savecurs.x = term->alt_savecurs.y = 0; + term->alt_t = term->marg_t = 0; + if (term->rows != -1) + term->alt_b = term->marg_b = term->rows - 1; + else + term->alt_b = term->marg_b = 0; + if (term->cols != -1) { + int i; + for (i = 0; i < term->cols; i++) + term->tabs[i] = (i % 8 == 0 ? true : false); + } + term->alt_om = term->dec_om = conf_get_bool(term->conf, CONF_dec_om); + term->alt_ins = false; + term->insert = false; + term->alt_wnext = false; + term->wrapnext = false; + term->save_wnext = false; + term->alt_save_wnext = false; + term->alt_wrap = term->wrap = conf_get_bool(term->conf, CONF_wrap_mode); + term->alt_cset = term->cset = term->save_cset = term->alt_save_cset = 0; + term->alt_utf = false; + term->utf = false; + term->save_utf = false; + term->alt_save_utf = false; + term->utf8.state = 0; + term->alt_sco_acs = term->sco_acs = + term->save_sco_acs = term->alt_save_sco_acs = 0; + term->cset_attr[0] = term->cset_attr[1] = + term->save_csattr = term->alt_save_csattr = CSET_ASCII; + term->rvideo = false; + term->in_vbell = false; + term->cursor_on = true; + term->big_cursor = false; + term->default_attr = term->save_attr = + term->alt_save_attr = term->curr_attr = ATTR_DEFAULT; + term->curr_truecolour.fg = term->curr_truecolour.bg = optionalrgb_none; + term->save_truecolour = term->alt_save_truecolour = term->curr_truecolour; + term->app_cursor_keys = conf_get_bool(term->conf, CONF_app_cursor); + term->app_keypad_keys = conf_get_bool(term->conf, CONF_app_keypad); + term->use_bce = conf_get_bool(term->conf, CONF_bce); + term->blink_is_real = conf_get_bool(term->conf, CONF_blinktext); + term->erase_char = term->basic_erase_char; + term->alt_which = 0; + term_print_finish(term); + term->xterm_mouse = 0; + term->xterm_extended_mouse = false; + term->urxvt_extended_mouse = false; + win_set_raw_mouse_mode(term->win, false); + term->win_pointer_shape_pending = true; + term->win_pointer_shape_raw = false; + term->bracketed_paste = false; + term->srm_echo = false; + { + int i; + for (i = 0; i < 256; i++) + term->wordness[i] = conf_get_int_int(term->conf, CONF_wordness, i); + } + if (term->screen) { + swap_screen(term, 1, false, false); + erase_lots(term, false, true, true); + swap_screen(term, 0, false, false); + if (clear) + erase_lots(term, false, true, true); + term->curs.y = find_last_nonempty_line(term, term->screen) + 1; + if (term->curs.y == term->rows) { + term->curs.y--; + scroll(term, 0, term->rows - 1, 1, true); + } + } else { + term->curs.y = 0; + } + term->curs.x = 0; + term_schedule_tblink(term); + term_schedule_cblink(term); + term_schedule_update(term); +} + +/* + * Force a screen update. + */ +void term_update(Terminal *term) +{ + term->window_update_pending = false; + + if (term->win_move_pending) { + win_move(term->win, term->win_move_pending_x, + term->win_move_pending_y); + term->win_move_pending = false; + } + if (term->win_resize_pending == WIN_RESIZE_NEED_SEND) { + term->win_resize_pending = WIN_RESIZE_AWAIT_REPLY; + win_request_resize(term->win, term->win_resize_pending_w, + term->win_resize_pending_h); + } + if (term->win_zorder_pending) { + win_set_zorder(term->win, term->win_zorder_top); + term->win_zorder_pending = false; + } + if (term->win_minimise_pending) { + win_set_minimised(term->win, term->win_minimise_enable); + term->win_minimise_pending = false; + } + if (term->win_maximise_pending) { + win_set_maximised(term->win, term->win_maximise_enable); + term->win_maximise_pending = false; + } + if (term->win_title_pending) { + win_set_title(term->win, term->window_title, + term->wintitle_codepage); + term->win_title_pending = false; + } + if (term->win_icon_title_pending) { + win_set_icon_title(term->win, term->icon_title, + term->icontitle_codepage); + term->win_icon_title_pending = false; + } + if (term->win_pointer_shape_pending) { + win_set_raw_mouse_mode_pointer(term->win, term->win_pointer_shape_raw); + term->win_pointer_shape_pending = false; + } + if (term->win_refresh_pending) { + win_refresh(term->win); + term->win_refresh_pending = false; + } + if (term->win_palette_pending) { + unsigned start = term->win_palette_pending_min; + unsigned ncolours = term->win_palette_pending_limit - start; + win_palette_set(term->win, start, ncolours, term->palette + start); + term->win_palette_pending = false; + } + + if (win_setup_draw_ctx(term->win)) { + bool need_sbar_update = term->seen_disp_event || + term->win_scrollbar_update_pending; + term->win_scrollbar_update_pending = false; + if (term->seen_disp_event && term->scroll_on_disp) { + term->disptop = 0; /* return to main screen */ + term->seen_disp_event = false; + need_sbar_update = true; + } + + if (need_sbar_update) + update_sbar(term); + do_paint(term); + win_set_cursor_pos( + term->win, term->curs.x, term->curs.y - term->disptop); + win_free_draw_ctx(term->win); + } +} + +/* + * Called from front end when a keypress occurs, to trigger + * anything magical that needs to happen in that situation. + */ +void term_seen_key_event(Terminal *term) +{ + /* + * On any keypress, clear the bell overload mechanism + * completely, on the grounds that large numbers of + * beeps coming from deliberate key action are likely + * to be intended (e.g. beeps from filename completion + * blocking repeatedly). + */ + term->beep_overloaded = false; + while (term->beephead) { + struct beeptime *tmp = term->beephead; + term->beephead = tmp->next; + sfree(tmp); + } + term->beeptail = NULL; + term->nbeeps = 0; + + /* + * Reset the scrollback on keypress, if we're doing that. + */ + if (term->scroll_on_key) { + term->disptop = 0; /* return to main screen */ + seen_disp_event(term); + } +} + +/* + * Same as power_on(), but an external function. + */ +void term_pwron(Terminal *term, bool clear) +{ + power_on(term, clear); + if (term->ldisc) /* cause ldisc to notice changes */ + ldisc_echoedit_update(term->ldisc); + term->disptop = 0; + deselect(term); + term_update(term); +} + +static void set_erase_char(Terminal *term) +{ + term->erase_char = term->basic_erase_char; + if (term->use_bce) { + term->erase_char.attr = (term->curr_attr & + (ATTR_FGMASK | ATTR_BGMASK)); + term->erase_char.truecolour.bg = term->curr_truecolour.bg; + } +} + +/* + * We copy a bunch of stuff out of the Conf structure into local + * fields in the Terminal structure, to avoid the repeated tree234 + * lookups which would be involved in fetching them from the former + * every time. + */ +static void term_copy_stuff_from_conf(Terminal *term) +{ + term->ansi_colour = conf_get_bool(term->conf, CONF_ansi_colour); + term->no_arabicshaping = conf_get_bool(term->conf, CONF_no_arabicshaping); + term->beep = conf_get_int(term->conf, CONF_beep); + term->bellovl = conf_get_bool(term->conf, CONF_bellovl); + term->bellovl_n = conf_get_int(term->conf, CONF_bellovl_n); + term->bellovl_s = conf_get_int(term->conf, CONF_bellovl_s); + term->bellovl_t = conf_get_int(term->conf, CONF_bellovl_t); + term->no_bidi = conf_get_bool(term->conf, CONF_no_bidi); + term->bksp_is_delete = conf_get_bool(term->conf, CONF_bksp_is_delete); + term->blink_cur = conf_get_bool(term->conf, CONF_blink_cur); + term->blinktext = conf_get_bool(term->conf, CONF_blinktext); + term->cjk_ambig_wide = conf_get_bool(term->conf, CONF_cjk_ambig_wide); + term->conf_height = conf_get_int(term->conf, CONF_height); + term->conf_width = conf_get_int(term->conf, CONF_width); + term->crhaslf = conf_get_bool(term->conf, CONF_crhaslf); + term->erase_to_scrollback = conf_get_bool(term->conf, CONF_erase_to_scrollback); + term->funky_type = conf_get_int(term->conf, CONF_funky_type); + term->sharrow_type = conf_get_int(term->conf, CONF_sharrow_type); + term->lfhascr = conf_get_bool(term->conf, CONF_lfhascr); + term->logflush = conf_get_bool(term->conf, CONF_logflush); + term->logtype = conf_get_int(term->conf, CONF_logtype); + term->mouse_override = conf_get_bool(term->conf, CONF_mouse_override); + term->nethack_keypad = conf_get_bool(term->conf, CONF_nethack_keypad); + term->no_alt_screen = conf_get_bool(term->conf, CONF_no_alt_screen); + term->no_applic_c = conf_get_bool(term->conf, CONF_no_applic_c); + term->no_applic_k = conf_get_bool(term->conf, CONF_no_applic_k); + term->no_dbackspace = conf_get_bool(term->conf, CONF_no_dbackspace); + term->no_mouse_rep = conf_get_bool(term->conf, CONF_no_mouse_rep); + term->no_remote_charset = conf_get_bool(term->conf, CONF_no_remote_charset); + term->no_remote_resize = conf_get_bool(term->conf, CONF_no_remote_resize); + term->no_remote_wintitle = conf_get_bool(term->conf, CONF_no_remote_wintitle); + term->no_remote_clearscroll = conf_get_bool(term->conf, CONF_no_remote_clearscroll); + term->rawcnp = conf_get_bool(term->conf, CONF_rawcnp); + term->utf8linedraw = conf_get_bool(term->conf, CONF_utf8linedraw); + term->rect_select = conf_get_bool(term->conf, CONF_rect_select); + term->remote_qtitle_action = conf_get_int(term->conf, CONF_remote_qtitle_action); + term->rxvt_homeend = conf_get_bool(term->conf, CONF_rxvt_homeend); + term->scroll_on_disp = conf_get_bool(term->conf, CONF_scroll_on_disp); + term->scroll_on_key = conf_get_bool(term->conf, CONF_scroll_on_key); + term->xterm_mouse_forbidden = conf_get_bool(term->conf, CONF_no_mouse_rep); + term->xterm_256_colour = conf_get_bool(term->conf, CONF_xterm_256_colour); + term->true_colour = conf_get_bool(term->conf, CONF_true_colour); + + /* + * Parse the control-character escapes in the configured + * answerback string. + */ + { + char *answerback = conf_get_str(term->conf, CONF_answerback); + int maxlen = strlen(answerback); + + term->answerback = snewn(maxlen, char); + term->answerbacklen = 0; + + while (*answerback) { + char *n; + char c = ctrlparse(answerback, &n); + if (n) { + term->answerback[term->answerbacklen++] = c; + answerback = n; + } else { + term->answerback[term->answerbacklen++] = *answerback++; + } + } + } +} + +void term_pre_reconfig(Terminal *term, Conf *conf) +{ + + /* + * Copy the current window title into the stored previous + * configuration, so that doing nothing to the window title field + * in the config box doesn't reset the title to its startup state. + */ + conf_set_str(conf, CONF_wintitle, term->window_title); +} + +/* + * When the user reconfigures us, we need to check the forbidden- + * alternate-screen config option, disable raw mouse mode if the + * user has disabled mouse reporting, and abandon a print job if + * the user has disabled printing. + */ +void term_reconfig(Terminal *term, Conf *conf) +{ + /* + * Before adopting the new config, check all those terminal + * settings which control power-on defaults; and if they've + * changed, we will modify the current state as well as the + * default one. The full list is: Auto wrap mode, DEC Origin + * Mode, BCE, blinking text, character classes. + */ + bool reset_wrap, reset_decom, reset_bce, reset_tblink, reset_charclass; + bool palette_changed = false; + int i; + + reset_wrap = (conf_get_bool(term->conf, CONF_wrap_mode) != + conf_get_bool(conf, CONF_wrap_mode)); + reset_decom = (conf_get_bool(term->conf, CONF_dec_om) != + conf_get_bool(conf, CONF_dec_om)); + reset_bce = (conf_get_bool(term->conf, CONF_bce) != + conf_get_bool(conf, CONF_bce)); + reset_tblink = (conf_get_bool(term->conf, CONF_blinktext) != + conf_get_bool(conf, CONF_blinktext)); + reset_charclass = false; + for (i = 0; i < 256; i++) + if (conf_get_int_int(term->conf, CONF_wordness, i) != + conf_get_int_int(conf, CONF_wordness, i)) + reset_charclass = true; + + /* + * If the bidi or shaping settings have changed, flush the bidi + * cache completely. + */ + if (conf_get_bool(term->conf, CONF_no_arabicshaping) != + conf_get_bool(conf, CONF_no_arabicshaping) || + conf_get_bool(term->conf, CONF_no_bidi) != + conf_get_bool(conf, CONF_no_bidi)) { + for (i = 0; i < term->bidi_cache_size; i++) { + sfree(term->pre_bidi_cache[i].chars); + sfree(term->post_bidi_cache[i].chars); + term->pre_bidi_cache[i].width = -1; + term->pre_bidi_cache[i].chars = NULL; + term->post_bidi_cache[i].width = -1; + term->post_bidi_cache[i].chars = NULL; + } + } + + { + const char *old_title = conf_get_str(term->conf, CONF_wintitle); + const char *new_title = conf_get_str(conf, CONF_wintitle); + if (strcmp(old_title, new_title)) { + sfree(term->window_title); + term->window_title = dupstr(new_title); + term->wintitle_codepage = DEFAULT_CODEPAGE; + term->win_title_pending = true; + term_schedule_update(term); + } + } + + /* + * Just setting conf is sufficient to cause colour setting changes + * to appear on the next ESC]R palette reset. But we should also + * check whether any colour settings have been changed, so that + * they can be updated immediately if they haven't been overridden + * by some escape sequence. + */ + { + int i, j; + for (i = 0; i < CONF_NCOLOURS; i++) { + for (j = 0; j < 3; j++) + if (conf_get_int_int(term->conf, CONF_colours, i*3+j) != + conf_get_int_int(conf, CONF_colours, i*3+j)) + break; + if (j < 3) { + /* Actually enacting the change has to be deferred + * until the new conf is installed. */ + palette_changed = true; + break; + } + } + } + + conf_free(term->conf); + term->conf = conf_copy(conf); + + if (reset_wrap) + term->alt_wrap = term->wrap = conf_get_bool(term->conf, CONF_wrap_mode); + if (reset_decom) + term->alt_om = term->dec_om = conf_get_bool(term->conf, CONF_dec_om); + if (reset_bce) { + term->use_bce = conf_get_bool(term->conf, CONF_bce); + set_erase_char(term); + } + if (reset_tblink) { + term->blink_is_real = conf_get_bool(term->conf, CONF_blinktext); + } + if (reset_charclass) + for (i = 0; i < 256; i++) + term->wordness[i] = conf_get_int_int(term->conf, CONF_wordness, i); + + if (conf_get_bool(term->conf, CONF_no_alt_screen)) + swap_screen(term, 0, false, false); + if (conf_get_bool(term->conf, CONF_no_remote_charset)) { + term->cset_attr[0] = term->cset_attr[1] = CSET_ASCII; + term->sco_acs = term->alt_sco_acs = 0; + term->utf = false; + } + if (!conf_get_str(term->conf, CONF_printer)) { + term_print_finish(term); + } + if (palette_changed) + term_notify_palette_changed(term); + term_schedule_tblink(term); + term_schedule_cblink(term); + term_copy_stuff_from_conf(term); + term_update_raw_mouse_mode(term); +} + +/* + * Clear the scrollback. + */ +void term_clrsb(Terminal *term) +{ + unsigned char *line; + int i; + + /* + * Scroll forward to the current screen, if we were back in the + * scrollback somewhere until now. + */ + term->disptop = 0; + + /* + * Clear the actual scrollback. + */ + while ((line = delpos234(term->scrollback, 0)) != NULL) { + sfree(line); /* this is compressed data, not a termline */ + } + + /* + * When clearing the scrollback, we also truncate any termlines on + * the current screen which have remembered data from a previous + * larger window size. Rationale: clearing the scrollback is + * sometimes done to protect privacy, so the user intention is + * specifically that we should not retain evidence of what + * previously happened in the terminal, and that ought to include + * evidence to the right as well as evidence above. + */ + for (i = 0; i < term->rows; i++) + check_line_size(term, scrlineptr(i)); + + /* + * That operation has invalidated the selection, if it overlapped + * the scrollback at all. + */ + if (term->selstate != NO_SELECTION && term->selstart.y < 0) + deselect(term); + + /* + * There are now no lines of real scrollback which can be pulled + * back into the screen by a resize, and no lines of the alternate + * screen which should be displayed as if part of the scrollback. + */ + term->tempsblines = 0; + term->alt_sblines = 0; + + /* + * The scrollbar will need updating to reflect the new state of + * the world. + */ + term->win_scrollbar_update_pending = true; + term_schedule_update(term); +} + +const optionalrgb optionalrgb_none = {0, 0, 0, 0}; + +void term_setup_window_titles(Terminal *term, const char *title_hostname) +{ + const char *conf_title = conf_get_str(term->conf, CONF_wintitle); + sfree(term->window_title); + sfree(term->icon_title); + if (*conf_title) { + term->window_title = dupstr(conf_title); + term->icon_title = dupstr(conf_title); + } else { + if (title_hostname && *title_hostname) + term->window_title = dupcat(title_hostname, " - ", appname); + else + term->window_title = dupstr(appname); + term->icon_title = dupstr(term->window_title); + } + term->wintitle_codepage = term->icontitle_codepage = DEFAULT_CODEPAGE; + term->win_title_pending = true; + term->win_icon_title_pending = true; +} + +static void palette_rebuild(Terminal *term) +{ + unsigned min_changed = OSC4_NCOLOURS, max_changed = 0; + + if (term->win_palette_pending) { + /* Possibly extend existing range. */ + min_changed = term->win_palette_pending_min; + max_changed = term->win_palette_pending_limit - 1; + } else { + /* Start with empty range. */ + min_changed = OSC4_NCOLOURS; + max_changed = 0; + } + + for (unsigned i = 0; i < OSC4_NCOLOURS; i++) { + rgb new_value; + bool found = false; + + for (unsigned j = lenof(term->subpalettes); j-- > 0 ;) { + if (term->subpalettes[j].present[i]) { + new_value = term->subpalettes[j].values[i]; + found = true; + break; + } + } + + assert(found); /* we expect SUBPAL_CONF to always be set */ + + if (new_value.r != term->palette[i].r || + new_value.g != term->palette[i].g || + new_value.b != term->palette[i].b) { + term->palette[i] = new_value; + if (min_changed > i) + min_changed = i; + if (max_changed < i) + max_changed = i; + } + } + + if (min_changed <= max_changed) { + /* + * At least one colour changed (or we had an update scheduled + * already). Schedule a redraw event to pass the result back + * to the TermWin. This also requires invalidating the rest + * of the window, because usually all the text will need + * redrawing in the new colours. + * (If there was an update pending and this palette rebuild + * didn't actually change anything, we'll harmlessly reinforce + * the existing update request.) + */ + term->win_palette_pending = true; + term->win_palette_pending_min = min_changed; + term->win_palette_pending_limit = max_changed + 1; + term_invalidate(term); + } +} + +/* + * Rebuild the palette from configuration and platform colours. + * If 'keep_overrides' set, any escape-sequence-specified overrides will + * remain in place. + */ +static void palette_reset(Terminal *term, bool keep_overrides) +{ + for (unsigned i = 0; i < OSC4_NCOLOURS; i++) + term->subpalettes[SUBPAL_CONF].present[i] = true; + + /* + * Copy all the palette information out of the Conf. + */ + for (unsigned i = 0; i < CONF_NCOLOURS; i++) { + rgb *col = &term->subpalettes[SUBPAL_CONF].values[ + colour_indices_conf_to_osc4[i]]; + col->r = conf_get_int_int(term->conf, CONF_colours, i*3+0); + col->g = conf_get_int_int(term->conf, CONF_colours, i*3+1); + col->b = conf_get_int_int(term->conf, CONF_colours, i*3+2); + } + + /* + * Directly invent the rest of the xterm-256 colours. + */ + for (unsigned i = 0; i < 216; i++) { + rgb *col = &term->subpalettes[SUBPAL_CONF].values[i + 16]; + int r = i / 36, g = (i / 6) % 6, b = i % 6; + col->r = r ? r * 40 + 55 : 0; + col->g = g ? g * 40 + 55 : 0; + col->b = b ? b * 40 + 55 : 0; + } + for (unsigned i = 0; i < 24; i++) { + rgb *col = &term->subpalettes[SUBPAL_CONF].values[i + 232]; + int shade = i * 10 + 8; + col->r = col->g = col->b = shade; + } + + /* + * Re-fetch any OS-local overrides. + */ + for (unsigned i = 0; i < OSC4_NCOLOURS; i++) + term->subpalettes[SUBPAL_PLATFORM].present[i] = false; + win_palette_get_overrides(term->win, term); + + if (!keep_overrides) { + /* + * Get rid of all escape-sequence configuration. + */ + for (unsigned i = 0; i < OSC4_NCOLOURS; i++) + term->subpalettes[SUBPAL_SESSION].present[i] = false; + } + + /* + * Rebuild the composite palette. + */ + palette_rebuild(term); +} + +void term_palette_override(Terminal *term, unsigned osc4_index, rgb rgb) +{ + /* + * We never expect to be called except as re-entry from our own + * call to win_palette_get_overrides above, so we need not mess + * about calling palette_rebuild. + */ + term->subpalettes[SUBPAL_PLATFORM].present[osc4_index] = true; + term->subpalettes[SUBPAL_PLATFORM].values[osc4_index] = rgb; +} + +/* + * Initialise the terminal. + */ +Terminal *term_init(Conf *myconf, struct unicode_data *ucsdata, TermWin *win) +{ + Terminal *term; + + /* + * Allocate a new Terminal structure and initialise the fields + * that need it. + */ + term = snew(Terminal); + term->win = win; + term->ucsdata = ucsdata; + term->conf = conf_copy(myconf); + term->logctx = NULL; + term->compatibility_level = TM_PUTTY; + strcpy(term->id_string, "\033[?6c"); + term->cblink_pending = term->tblink_pending = false; + term->paste_buffer = NULL; + term->paste_len = 0; + bufchain_init(&term->inbuf); + bufchain_init(&term->printer_buf); + term->printing = term->only_printing = false; + term->print_job = NULL; + term->vt52_mode = false; + term->cr_lf_return = false; + term->seen_disp_event = false; + term->mouse_is_down = 0; + term->reset_132 = false; + term->cblinker = false; + term->tblinker = false; + term->has_focus = true; + term->repeat_off = false; + term->termstate = TOPLEVEL; + term->selstate = NO_SELECTION; + term->curstype = 0; + + term_copy_stuff_from_conf(term); + + term->screen = term->alt_screen = term->scrollback = NULL; + term->tempsblines = 0; + term->alt_sblines = 0; + term->disptop = 0; + term->disptext = NULL; + term->dispcursx = term->dispcursy = -1; + term->tabs = NULL; + deselect(term); + term->rows = term->cols = -1; + power_on(term, true); + term->beephead = term->beeptail = NULL; + term->nbeeps = 0; + term->lastbeep = false; + term->beep_overloaded = false; + term->attr_mask = 0xffffffff; + term->backend = NULL; + term->in_term_out = false; + term->ltemp = NULL; + term->ltemp_size = 0; + term->wcFrom = NULL; + term->wcTo = NULL; + term->wcFromTo_size = 0; + + term->window_update_pending = false; + term->window_update_cooldown = false; + + term->bidi_cache_size = 0; + term->pre_bidi_cache = term->post_bidi_cache = NULL; + + /* FULL-TERMCHAR */ + term->basic_erase_char.chr = CSET_ASCII | ' '; + term->basic_erase_char.attr = ATTR_DEFAULT; + term->basic_erase_char.cc_next = 0; + term->basic_erase_char.truecolour.fg = optionalrgb_none; + term->basic_erase_char.truecolour.bg = optionalrgb_none; + term->erase_char = term->basic_erase_char; + + term->last_selected_text = NULL; + term->last_selected_attr = NULL; + term->last_selected_tc = NULL; + term->last_selected_len = 0; + /* TermWin implementations will typically extend these with + * clipboard ids they know about */ + term->mouse_select_clipboards[0] = CLIP_LOCAL; + term->n_mouse_select_clipboards = 1; + term->mouse_paste_clipboard = CLIP_NULL; + + term->last_graphic_char = 0; + + term->trusted = true; + + term->bracketed_paste_active = false; + + term->window_title = dupstr(""); + term->icon_title = dupstr(""); + term->wintitle_codepage = term->icontitle_codepage = DEFAULT_CODEPAGE; + term->minimised = false; + term->winpos_x = term->winpos_y = 0; + term->winpixsize_x = term->winpixsize_y = 0; + + term->win_move_pending = false; + term->win_resize_pending = WIN_RESIZE_NO; + term->win_zorder_pending = false; + term->win_minimise_pending = false; + term->win_maximise_pending = false; + term->win_title_pending = false; + term->win_icon_title_pending = false; + term->win_pointer_shape_pending = false; + term->win_refresh_pending = false; + term->win_scrollbar_update_pending = false; + term->win_palette_pending = false; + + term->bidi_ctx = bidi_new_context(); + + palette_reset(term, false); + + return term; +} + +void term_free(Terminal *term) +{ + termline *line; + struct beeptime *beep; + int i; + + while ((line = delpos234(term->scrollback, 0)) != NULL) + sfree(line); /* compressed data, not a termline */ + freetree234(term->scrollback); + while ((line = delpos234(term->screen, 0)) != NULL) + freetermline(line); + freetree234(term->screen); + while ((line = delpos234(term->alt_screen, 0)) != NULL) + freetermline(line); + freetree234(term->alt_screen); + if (term->disptext) { + for (i = 0; i < term->rows; i++) + freetermline(term->disptext[i]); + } + sfree(term->disptext); + while (term->beephead) { + beep = term->beephead; + term->beephead = beep->next; + sfree(beep); + } + bufchain_clear(&term->inbuf); + if(term->print_job) + printer_finish_job(term->print_job); + bufchain_clear(&term->printer_buf); + sfree(term->paste_buffer); + sfree(term->ltemp); + sfree(term->wcFrom); + sfree(term->wcTo); + sfree(term->answerback); + + for (i = 0; i < term->bidi_cache_size; i++) { + sfree(term->pre_bidi_cache[i].chars); + sfree(term->post_bidi_cache[i].chars); + sfree(term->post_bidi_cache[i].forward); + sfree(term->post_bidi_cache[i].backward); + } + sfree(term->pre_bidi_cache); + sfree(term->post_bidi_cache); + + sfree(term->tabs); + + expire_timer_context(term); + delete_callbacks_for_context(term); + + conf_free(term->conf); + + sfree(term->window_title); + sfree(term->icon_title); + + bidi_free_context(term->bidi_ctx); + + sfree(term); +} + +void term_set_trust_status(Terminal *term, bool trusted) +{ + term->trusted = trusted; +} + +void term_get_cursor_position(Terminal *term, int *x, int *y) +{ + *x = term->curs.x; + *y = term->curs.y; +} + +/* + * Set up the terminal for a given size. + */ +void term_size(Terminal *term, int newrows, int newcols, int newsavelines) +{ + tree234 *newalt; + termline **newdisp, *line; + int i, j, oldrows = term->rows; + int sblen; + int save_alt_which = term->alt_which; + + if (newrows == term->rows && newcols == term->cols && + newsavelines == term->savelines) + return; /* nothing to do */ + + /* Behave sensibly if we're given zero (or negative) rows/cols */ + + if (newrows < 1) newrows = 1; + if (newcols < 1) newcols = 1; + + deselect(term); + swap_screen(term, 0, false, false); + + term->alt_t = term->marg_t = 0; + term->alt_b = term->marg_b = newrows - 1; + + if (term->rows == -1) { + term->scrollback = newtree234(NULL); + term->screen = newtree234(NULL); + term->tempsblines = 0; + term->rows = 0; + } + + /* + * Resize the screen and scrollback. We only need to shift + * lines around within our data structures, because lineptr() + * will take care of resizing each individual line if + * necessary. So: + * + * - If the new screen is longer, we shunt lines in from temporary + * scrollback if possible, otherwise we add new blank lines at + * the bottom. + * + * - If the new screen is shorter, we remove any blank lines at + * the bottom if possible, otherwise shunt lines above the cursor + * to scrollback if possible, otherwise delete lines below the + * cursor. + * + * - Then, if the new scrollback length is less than the + * amount of scrollback we actually have, we must throw some + * away. + */ + sblen = count234(term->scrollback); + /* Do this loop to expand the screen if newrows > rows */ + assert(term->rows == count234(term->screen)); + while (term->rows < newrows) { + if (term->tempsblines > 0) { + compressed_scrollback_line *cline; + /* Insert a line from the scrollback at the top of the screen. */ + assert(sblen >= term->tempsblines); + cline = delpos234(term->scrollback, --sblen); + line = decompressline(cline); + sfree(cline); + line->temporary = false; /* reconstituted line is now real */ + term->tempsblines -= 1; + addpos234(term->screen, line, 0); + term->curs.y += 1; + term->savecurs.y += 1; + term->alt_y += 1; + term->alt_savecurs.y += 1; + } else { + /* Add a new blank line at the bottom of the screen. */ + line = newtermline(term, newcols, false); + addpos234(term->screen, line, count234(term->screen)); + } + term->rows += 1; + } + /* Do this loop to shrink the screen if newrows < rows */ + while (term->rows > newrows) { + if (term->curs.y < term->rows - 1) { + /* delete bottom row, unless it contains the cursor */ + line = delpos234(term->screen, term->rows - 1); + freetermline(line); + } else { + /* push top row to scrollback */ + line = delpos234(term->screen, 0); + addpos234(term->scrollback, compressline(line), sblen++); + freetermline(line); + term->tempsblines += 1; + term->curs.y -= 1; + term->savecurs.y -= 1; + term->alt_y -= 1; + term->alt_savecurs.y -= 1; + } + term->rows -= 1; + } + assert(term->rows == newrows); + assert(count234(term->screen) == newrows); + + /* Delete any excess lines from the scrollback. */ + while (sblen > newsavelines) { + line = delpos234(term->scrollback, 0); + sfree(line); + sblen--; + } + if (sblen < term->tempsblines) + term->tempsblines = sblen; + assert(count234(term->scrollback) <= newsavelines); + assert(count234(term->scrollback) >= term->tempsblines); + term->disptop = 0; + + /* Make a new displayed text buffer. */ + newdisp = snewn(newrows, termline *); + for (i = 0; i < newrows; i++) { + newdisp[i] = newtermline(term, newcols, false); + for (j = 0; j < newcols; j++) + newdisp[i]->chars[j].attr = ATTR_INVALID; + } + if (term->disptext) { + for (i = 0; i < oldrows; i++) + freetermline(term->disptext[i]); + } + sfree(term->disptext); + term->disptext = newdisp; + term->dispcursx = term->dispcursy = -1; + + /* Make a new alternate screen. */ + newalt = newtree234(NULL); + for (i = 0; i < newrows; i++) { + line = newtermline(term, newcols, true); + addpos234(newalt, line, i); + } + if (term->alt_screen) { + while (NULL != (line = delpos234(term->alt_screen, 0))) + freetermline(line); + freetree234(term->alt_screen); + } + term->alt_screen = newalt; + term->alt_sblines = 0; + + term->tabs = sresize(term->tabs, newcols, unsigned char); + { + int i; + for (i = (term->cols > 0 ? term->cols : 0); i < newcols; i++) + term->tabs[i] = (i % 8 == 0 ? true : false); + } + + /* Check that the cursor positions are still valid. */ + if (term->savecurs.y < 0) + term->savecurs.y = 0; + if (term->savecurs.y >= newrows) + term->savecurs.y = newrows - 1; + if (term->savecurs.x >= newcols) + term->savecurs.x = newcols - 1; + if (term->alt_savecurs.y < 0) + term->alt_savecurs.y = 0; + if (term->alt_savecurs.y >= newrows) + term->alt_savecurs.y = newrows - 1; + if (term->alt_savecurs.x >= newcols) + term->alt_savecurs.x = newcols - 1; + if (term->curs.y < 0) + term->curs.y = 0; + if (term->curs.y >= newrows) + term->curs.y = newrows - 1; + if (term->curs.x >= newcols) + term->curs.x = newcols - 1; + if (term->alt_y < 0) + term->alt_y = 0; + if (term->alt_y >= newrows) + term->alt_y = newrows - 1; + if (term->alt_x >= newcols) + term->alt_x = newcols - 1; + term->alt_x = term->alt_y = 0; + term->wrapnext = false; + term->alt_wnext = false; + + term->rows = newrows; + term->cols = newcols; + term->savelines = newsavelines; + + swap_screen(term, save_alt_which, false, false); + + term->win_scrollbar_update_pending = true; + term_schedule_update(term); + if (term->backend) + backend_size(term->backend, term->cols, term->rows); +} + +void term_resize_request_completed(Terminal *term) +{ + assert(term->win_resize_pending == WIN_RESIZE_AWAIT_REPLY); + term->win_resize_pending = WIN_RESIZE_NO; + queue_toplevel_callback(term_out_cb, term); +} + +/* + * Hand a backend to the terminal, so it can be notified of resizes. + */ +void term_provide_backend(Terminal *term, Backend *backend) +{ + term->backend = backend; + if (term->backend && term->cols > 0 && term->rows > 0) + backend_size(term->backend, term->cols, term->rows); +} + +/* Find the bottom line on the screen that has any content. + * If only the top line has content, returns 0. + * If no lines have content, return -1. + */ +static int find_last_nonempty_line(Terminal *term, tree234 *screen) +{ + int i; + for (i = count234(screen) - 1; i >= 0; i--) { + termline *line = index234(screen, i); + int j; + for (j = 0; j < line->cols; j++) + if (!termchars_equal(&line->chars[j], &term->erase_char)) + break; + if (j != line->cols) break; + } + return i; +} + +/* + * Swap screens. If `reset' is true and we have been asked to + * switch to the alternate screen, we must bring most of its + * configuration from the main screen and erase the contents of the + * alternate screen completely. (This is even true if we're already + * on it! Blame xterm.) + */ +static void swap_screen(Terminal *term, int which, + bool reset, bool keep_cur_pos) +{ + int t; + bool bt; + pos tp; + truecolour ttc; + tree234 *ttr; + + if (!which) + reset = false; /* do no weird resetting if which==0 */ + + if (which != term->alt_which) { + if (term->erase_to_scrollback && term->alt_screen && + term->alt_which && term->disptop < 0) { + /* + * We're swapping away from the alternate screen, so some + * lines are about to vanish from the virtual scrollback. + * Adjust disptop by that much, so that (if we're not + * resetting the scrollback anyway on a display event) the + * current scroll position still ends up pointing at the + * same text. + */ + term->disptop += term->alt_sblines; + if (term->disptop > 0) + term->disptop = 0; + } + + term->alt_which = which; + + ttr = term->alt_screen; + term->alt_screen = term->screen; + term->screen = ttr; + term->alt_sblines = ( + term->alt_screen ? + find_last_nonempty_line(term, term->alt_screen) + 1 : 0); + t = term->curs.x; + if (!reset && !keep_cur_pos) + term->curs.x = term->alt_x; + term->alt_x = t; + t = term->curs.y; + if (!reset && !keep_cur_pos) + term->curs.y = term->alt_y; + term->alt_y = t; + t = term->marg_t; + if (!reset) term->marg_t = term->alt_t; + term->alt_t = t; + t = term->marg_b; + if (!reset) term->marg_b = term->alt_b; + term->alt_b = t; + bt = term->dec_om; + if (!reset) term->dec_om = term->alt_om; + term->alt_om = bt; + bt = term->wrap; + if (!reset) term->wrap = term->alt_wrap; + term->alt_wrap = bt; + bt = term->wrapnext; + if (!reset) term->wrapnext = term->alt_wnext; + term->alt_wnext = bt; + bt = term->insert; + if (!reset) term->insert = term->alt_ins; + term->alt_ins = bt; + t = term->cset; + if (!reset) term->cset = term->alt_cset; + term->alt_cset = t; + bt = term->utf; + if (!reset) term->utf = term->alt_utf; + term->alt_utf = bt; + t = term->sco_acs; + if (!reset) term->sco_acs = term->alt_sco_acs; + term->alt_sco_acs = t; + + tp = term->savecurs; + if (!reset) + term->savecurs = term->alt_savecurs; + term->alt_savecurs = tp; + t = term->save_cset; + if (!reset) + term->save_cset = term->alt_save_cset; + term->alt_save_cset = t; + t = term->save_csattr; + if (!reset) + term->save_csattr = term->alt_save_csattr; + term->alt_save_csattr = t; + t = term->save_attr; + if (!reset) + term->save_attr = term->alt_save_attr; + term->alt_save_attr = t; + ttc = term->save_truecolour; + if (!reset) + term->save_truecolour = term->alt_save_truecolour; + term->alt_save_truecolour = ttc; + bt = term->save_utf; + if (!reset) + term->save_utf = term->alt_save_utf; + term->alt_save_utf = bt; + bt = term->save_wnext; + if (!reset) + term->save_wnext = term->alt_save_wnext; + term->alt_save_wnext = bt; + t = term->save_sco_acs; + if (!reset) + term->save_sco_acs = term->alt_save_sco_acs; + term->alt_save_sco_acs = t; + + if (term->erase_to_scrollback && term->alt_screen && + term->alt_which && term->disptop < 0) { + /* + * Inverse of the adjustment at the top of this function. + * This time, we're swapping _to_ the alternate screen, so + * some lines are about to _appear_ in the virtual + * scrollback, and we adjust disptop in the other + * direction. + * + * Both these adjustments depend on the value stored in + * term->alt_sblines while the alt screen is selected, + * which is why we had to do one _before_ switching away + * from it and the other _after_ switching to it. + */ + term->disptop -= term->alt_sblines; + int limit = -sblines(term); + if (term->disptop < limit) + term->disptop = limit; + } + } + + if (reset && term->screen) { + /* + * Yes, this _is_ supposed to honour background-colour-erase. + */ + erase_lots(term, false, true, true); + } +} + +/* + * Update the scroll bar. + */ +static void update_sbar(Terminal *term) +{ + int nscroll = sblines(term); + win_set_scrollbar(term->win, nscroll + term->rows, + nscroll + term->disptop, term->rows); +} + +/* + * Check whether the region bounded by the two pointers intersects + * the scroll region, and de-select the on-screen selection if so. + */ +static void check_selection(Terminal *term, pos from, pos to) +{ + if (poslt(from, term->selend) && poslt(term->selstart, to)) + deselect(term); +} + +static void clear_line(Terminal *term, termline *line) +{ + resizeline(term, line, term->cols); + for (int i = 0; i < term->cols; i++) + copy_termchar(line, i, &term->erase_char); + line->lattr = LATTR_NORM; +} + +static void check_trust_status(Terminal *term, termline *line) +{ + if (line->trusted != term->trusted) { + /* + * If we're displaying trusted output on a previously + * untrusted line, or vice versa, we need to switch the + * 'trusted' attribute on this terminal line, and also clear + * all its previous contents. + */ + clear_line(term, line); + line->trusted = term->trusted; + } +} + +/* + * Scroll the screen. (`lines' is +ve for scrolling forward, -ve + * for backward.) `sb' is true if the scrolling is permitted to + * affect the scrollback buffer. + */ +static void scroll(Terminal *term, int topline, int botline, + int lines, bool sb) +{ + termline *line; + int seltop, scrollwinsize; + + if (topline != 0 || term->alt_which != 0) + sb = false; + + scrollwinsize = botline - topline + 1; + + if (lines < 0) { + lines = -lines; + if (lines > scrollwinsize) + lines = scrollwinsize; + while (lines-- > 0) { + line = delpos234(term->screen, botline); + resizeline(term, line, term->cols); + clear_line(term, line); + addpos234(term->screen, line, topline); + + if (term->selstart.y >= topline && term->selstart.y <= botline) { + term->selstart.y++; + if (term->selstart.y > botline) { + term->selstart.y = botline + 1; + term->selstart.x = 0; + } + } + if (term->selend.y >= topline && term->selend.y <= botline) { + term->selend.y++; + if (term->selend.y > botline) { + term->selend.y = botline + 1; + term->selend.x = 0; + } + } + } + } else { + if (lines > scrollwinsize) + lines = scrollwinsize; + while (lines-- > 0) { + line = delpos234(term->screen, topline); +#ifdef TERM_CC_DIAGS + cc_check(line); +#endif + if (sb && term->savelines > 0) { + int sblen = count234(term->scrollback); + /* + * We must add this line to the scrollback. We'll + * remove a line from the top of the scrollback if + * the scrollback is full. + */ + if (sblen == term->savelines) { + unsigned char *cline; + + sblen--; + cline = delpos234(term->scrollback, 0); + sfree(cline); + } else + term->tempsblines += 1; + + addpos234(term->scrollback, compressline(line), sblen); + + /* now `line' itself can be reused as the bottom line */ + + /* + * If the user is currently looking at part of the + * scrollback, and they haven't enabled any options + * that are going to reset the scrollback as a + * result of this movement, then the chances are + * they'd like to keep looking at the same line. So + * we move their viewpoint at the same rate as the + * scroll, at least until their viewpoint hits the + * top end of the scrollback buffer, at which point + * we don't have the choice any more. + * + * Thanks to Jan Holmen Holsten for the idea and + * initial implementation. + */ + if (term->disptop > -term->savelines && term->disptop < 0) + term->disptop--; + } + resizeline(term, line, term->cols); + clear_line(term, line); + line->trusted = false; + addpos234(term->screen, line, botline); + + /* + * If the selection endpoints move into the scrollback, + * we keep them moving until they hit the top. However, + * of course, if the line _hasn't_ moved into the + * scrollback then we don't do this, and cut them off + * at the top of the scroll region. + * + * This applies to selstart and selend (for an existing + * selection), and also selanchor (for one being + * selected as we speak). + */ + seltop = sb ? -term->savelines : topline; + + if (term->selstate != NO_SELECTION) { + if (term->selstart.y >= seltop && + term->selstart.y <= botline) { + term->selstart.y--; + if (term->selstart.y < seltop) { + term->selstart.y = seltop; + term->selstart.x = 0; + } + } + if (term->selend.y >= seltop && term->selend.y <= botline) { + term->selend.y--; + if (term->selend.y < seltop) { + term->selend.y = seltop; + term->selend.x = 0; + } + } + if (term->selanchor.y >= seltop && + term->selanchor.y <= botline) { + term->selanchor.y--; + if (term->selanchor.y < seltop) { + term->selanchor.y = seltop; + term->selanchor.x = 0; + } + } + } + } + } +} + +/* + * Move the cursor to a given position, clipping at boundaries. We + * may or may not want to clip at the scroll margin: marg_clip is 0 + * not to, 1 to disallow _passing_ the margins, and 2 to disallow + * even _being_ outside the margins. + */ +static void move(Terminal *term, int x, int y, int marg_clip) +{ + if (x < 0) + x = 0; + if (x >= term->cols) + x = term->cols - 1; + if (marg_clip) { + if ((term->curs.y >= term->marg_t || marg_clip == 2) && + y < term->marg_t) + y = term->marg_t; + if ((term->curs.y <= term->marg_b || marg_clip == 2) && + y > term->marg_b) + y = term->marg_b; + } + if (y < 0) + y = 0; + if (y >= term->rows) + y = term->rows - 1; + term->curs.x = x; + term->curs.y = y; + term->wrapnext = false; +} + +/* + * Save or restore the cursor and SGR mode. + */ +static void save_cursor(Terminal *term, bool save) +{ + if (save) { + term->savecurs = term->curs; + term->save_attr = term->curr_attr; + term->save_truecolour = term->curr_truecolour; + term->save_cset = term->cset; + term->save_utf = term->utf; + term->save_wnext = term->wrapnext; + term->save_csattr = term->cset_attr[term->cset]; + term->save_sco_acs = term->sco_acs; + } else { + term->curs = term->savecurs; + /* Make sure the window hasn't shrunk since the save */ + if (term->curs.x >= term->cols) + term->curs.x = term->cols - 1; + if (term->curs.y >= term->rows) + term->curs.y = term->rows - 1; + + term->curr_attr = term->save_attr; + term->curr_truecolour = term->save_truecolour; + term->cset = term->save_cset; + term->utf = term->save_utf; + term->wrapnext = term->save_wnext; + /* + * wrapnext might reset to False if the x position is no + * longer at the rightmost edge. + */ + if (term->wrapnext && term->curs.x < term->cols-1) + term->wrapnext = false; + term->cset_attr[term->cset] = term->save_csattr; + term->sco_acs = term->save_sco_acs; + set_erase_char(term); + } +} + +/* + * This function is called before doing _anything_ which affects + * only part of a line of text. It is used to mark the boundary + * between two character positions, and it indicates that some sort + * of effect is going to happen on only one side of that boundary. + * + * The effect of this function is to check whether a CJK + * double-width character is straddling the boundary, and to remove + * it and replace it with two spaces if so. (Of course, one or + * other of those spaces is then likely to be replaced with + * something else again, as a result of whatever happens next.) + * + * Also, if the boundary is at the right-hand _edge_ of the screen, + * it implies something deliberate is being done to the rightmost + * column position; hence we must clear LATTR_WRAPPED2. + * + * The input to the function is the coordinates of the _second_ + * character of the pair. + */ +static void check_boundary(Terminal *term, int x, int y) +{ + termline *ldata; + + /* Validate input coordinates, just in case. */ + if (x <= 0 || x > term->cols) + return; + + ldata = scrlineptr(y); + check_trust_status(term, ldata); + check_line_size(term, ldata); + if (x == term->cols) { + ldata->lattr &= ~LATTR_WRAPPED2; + } else { + if (ldata->chars[x].chr == UCSWIDE) { + clear_cc(ldata, x-1); + clear_cc(ldata, x); + ldata->chars[x-1].chr = ' ' | CSET_ASCII; + ldata->chars[x] = ldata->chars[x-1]; + } + } +} + +/* + * Erase a large portion of the screen: the whole screen, or the + * whole line, or parts thereof. + */ +static void erase_lots(Terminal *term, + bool line_only, bool from_begin, bool to_end) +{ + pos start, end; + bool erase_lattr; + bool erasing_lines_from_top = false; + + if (line_only) { + start.y = term->curs.y; + start.x = 0; + end.y = term->curs.y + 1; + end.x = 0; + erase_lattr = false; + } else { + start.y = 0; + start.x = 0; + end.y = term->rows; + end.x = 0; + erase_lattr = true; + } + + /* This is the endpoint of the clearing operation that is not + * either the start or end of the line / screen. */ + pos boundary = term->curs; + + if (!from_begin) { + /* + * If we're erasing from the current char to the end of + * line/screen, then we take account of wrapnext, so as to + * maintain the invariant that writing a printing character + * followed by ESC[K should not overwrite the character you + * _just wrote_. That is, when wrapnext says the cursor is + * 'logically' at the very rightmost edge of the screen + * instead of just before the last printing char, ESC[K should + * do nothing at all, and ESC[J should clear the next line but + * leave this one unchanged. + * + * This adjusted position will also be the position we use for + * check_boundary (i.e. the thing we ensure isn't in the + * middle of a double-width printing char). + */ + if (term->wrapnext) + incpos(boundary); + + start = boundary; + } + if (!to_end) { + /* + * If we're erasing from the start of (at least) the line _to_ + * the current position, then that is taken to mean 'inclusive + * of the cell under the cursor', which means we don't + * consider wrapnext at all: whether it's set or not, we still + * clear the cell under the cursor. + * + * Again, that incremented boundary position is where we + * should be careful of a straddling wide character. + */ + incpos(boundary); + end = boundary; + } + if (!from_begin || !to_end) + check_boundary(term, boundary.x, boundary.y); + check_selection(term, start, end); + + /* Clear screen also forces a full window redraw, just in case. */ + if (start.y == 0 && start.x == 0 && end.y == term->rows) + term_invalidate(term); + + /* Lines scrolled away shouldn't be brought back on if the terminal + * resizes. */ + if (start.y == 0 && start.x == 0 && end.x == 0 && erase_lattr) + erasing_lines_from_top = true; + + if (term->erase_to_scrollback && erasing_lines_from_top) { + /* If it's a whole number of lines, starting at the top, and + * we're fully erasing them, erase by scrolling and keep the + * lines in the scrollback. */ + int scrolllines = end.y; + if (end.y == term->rows) { + /* Shrink until we find a non-empty row.*/ + scrolllines = find_last_nonempty_line(term, term->screen) + 1; + } + if (scrolllines > 0) + scroll(term, 0, scrolllines - 1, scrolllines, true); + } else { + termline *ldata = scrlineptr(start.y); + check_trust_status(term, ldata); + while (poslt(start, end)) { + check_line_size(term, ldata); + if (start.x == term->cols) { + if (!erase_lattr) + ldata->lattr &= ~(LATTR_WRAPPED | LATTR_WRAPPED2); + else + ldata->lattr = LATTR_NORM; + } else { + copy_termchar(ldata, start.x, &term->erase_char); + } + if (incpos(start) && start.y < term->rows) { + ldata = scrlineptr(start.y); + check_trust_status(term, ldata); + } + } + } + + /* After an erase of lines from the top of the screen, we shouldn't + * bring the lines back again if the terminal enlarges (since the user or + * application has explicitly thrown them away). */ + if (erasing_lines_from_top && !(term->alt_which)) + term->tempsblines = 0; +} + +/* + * Insert or delete characters within the current line. n is +ve if + * insertion is desired, and -ve for deletion. + */ +static void insch(Terminal *term, int n) +{ + int dir = (n < 0 ? -1 : +1); + int m, j; + pos eol; + termline *ldata; + + n = (n < 0 ? -n : n); + if (n > term->cols - term->curs.x) + n = term->cols - term->curs.x; + m = term->cols - term->curs.x - n; + + /* + * We must de-highlight the selection if it overlaps any part of + * the region affected by this operation, i.e. the region from the + * current cursor position to end-of-line, _unless_ the entirety + * of the selection is going to be moved to the left or right by + * this operation but otherwise unchanged, in which case we can + * simply move the highlight with the text. + */ + eol.y = term->curs.y; + eol.x = term->cols; + if (poslt(term->curs, term->selend) && poslt(term->selstart, eol)) { + pos okstart = term->curs; + pos okend = eol; + if (dir > 0) { + /* Insertion: n characters at EOL will be splatted. */ + okend.x -= n; + } else { + /* Deletion: n characters at cursor position will be splatted. */ + okstart.x += n; + } + if (posle(okstart, term->selstart) && posle(term->selend, okend)) { + /* Selection is contained entirely in the interval + * [okstart,okend), so we need only adjust the selection + * bounds. */ + term->selstart.x += dir * n; + term->selend.x += dir * n; + assert(term->selstart.x >= term->curs.x); + assert(term->selstart.x < term->cols); + assert(term->selend.x > term->curs.x); + assert(term->selend.x <= term->cols); + } else { + /* Selection is not wholly contained in that interval, so + * we must unhighlight it. */ + deselect(term); + } + } + + check_boundary(term, term->curs.x, term->curs.y); + if (dir < 0) + check_boundary(term, term->curs.x + n, term->curs.y); + ldata = scrlineptr(term->curs.y); + check_trust_status(term, ldata); + if (dir < 0) { + for (j = 0; j < m; j++) + move_termchar(ldata, + ldata->chars + term->curs.x + j, + ldata->chars + term->curs.x + j + n); + while (n--) + copy_termchar(ldata, term->curs.x + m++, &term->erase_char); + } else { + for (j = m; j-- ;) + move_termchar(ldata, + ldata->chars + term->curs.x + j + n, + ldata->chars + term->curs.x + j); + while (n--) + copy_termchar(ldata, term->curs.x + n, &term->erase_char); + } +} + +static void term_update_raw_mouse_mode(Terminal *term) +{ + bool want_raw = (term->xterm_mouse != 0 && !term->xterm_mouse_forbidden); + win_set_raw_mouse_mode(term->win, want_raw); + term->win_pointer_shape_pending = true; + term->win_pointer_shape_raw = want_raw; + term_schedule_update(term); +} + +static void term_request_resize(Terminal *term, int cols, int rows) +{ + if (term->cols == cols && term->rows == rows) + return; /* don't need to do anything */ + + term->win_resize_pending = WIN_RESIZE_NEED_SEND; + term->win_resize_pending_w = cols; + term->win_resize_pending_h = rows; + term_schedule_update(term); +} + +/* + * Toggle terminal mode `mode' to state `state'. (`query' indicates + * whether the mode is a DEC private one or a normal one.) + */ +static void toggle_mode(Terminal *term, int mode, int query, bool state) +{ + if (query == 1) { + switch (mode) { + case 1: /* DECCKM: application cursor keys */ + term->app_cursor_keys = state; + break; + case 2: /* DECANM: VT52 mode */ + term->vt52_mode = !state; + if (term->vt52_mode) { + term->blink_is_real = false; + term->vt52_bold = false; + } else { + term->blink_is_real = term->blinktext; + } + term_schedule_tblink(term); + break; + case 3: /* DECCOLM: 80/132 columns */ + deselect(term); + if (!term->no_remote_resize) + term_request_resize(term, state ? 132 : 80, term->rows); + term->reset_132 = state; + term->alt_t = term->marg_t = 0; + term->alt_b = term->marg_b = term->rows - 1; + move(term, 0, 0, 0); + erase_lots(term, false, true, true); + break; + case 5: /* DECSCNM: reverse video */ + /* + * Toggle reverse video. If we receive an OFF within the + * visual bell timeout period after an ON, we trigger an + * effective visual bell, so that ESC[?5hESC[?5l will + * always be an actually _visible_ visual bell. + */ + if (term->rvideo && !state) { + /* This is an OFF, so set up a vbell */ + term_schedule_vbell(term, true, term->rvbell_startpoint); + } else if (!term->rvideo && state) { + /* This is an ON, so we notice the time and save it. */ + term->rvbell_startpoint = GETTICKCOUNT(); + } + term->rvideo = state; + seen_disp_event(term); + break; + case 6: /* DECOM: DEC origin mode */ + term->dec_om = state; + break; + case 7: /* DECAWM: auto wrap */ + term->wrap = state; + break; + case 8: /* DECARM: auto key repeat */ + term->repeat_off = !state; + break; + case 25: /* DECTCEM: enable/disable cursor */ + compatibility2(OTHER, VT220); + term->cursor_on = state; + seen_disp_event(term); + break; + case 47: /* alternate screen */ + compatibility(OTHER); + deselect(term); + swap_screen(term, term->no_alt_screen ? 0 : state, false, false); + if (term->scroll_on_disp) + term->disptop = 0; + break; + case 1000: /* xterm mouse 1 (normal) */ + term->xterm_mouse = state ? 1 : 0; + term_update_raw_mouse_mode(term); + break; + case 1002: /* xterm mouse 2 (inc. button drags) */ + term->xterm_mouse = state ? 2 : 0; + term_update_raw_mouse_mode(term); + break; + case 1006: /* xterm extended mouse */ + term->xterm_extended_mouse = state; + break; + case 1015: /* urxvt extended mouse */ + term->urxvt_extended_mouse = state; + break; + case 1047: /* alternate screen */ + compatibility(OTHER); + deselect(term); + swap_screen(term, term->no_alt_screen ? 0 : state, true, true); + if (term->scroll_on_disp) + term->disptop = 0; + break; + case 1048: /* save/restore cursor */ + if (!term->no_alt_screen) + save_cursor(term, state); + if (!state) seen_disp_event(term); + break; + case 1049: /* cursor & alternate screen */ + if (state && !term->no_alt_screen) + save_cursor(term, state); + if (!state) seen_disp_event(term); + compatibility(OTHER); + deselect(term); + swap_screen(term, term->no_alt_screen ? 0 : state, true, false); + if (!state && !term->no_alt_screen) + save_cursor(term, state); + if (term->scroll_on_disp) + term->disptop = 0; + break; + case 2004: /* xterm bracketed paste */ + term->bracketed_paste = state ? true : false; + break; + } + } else if (query == 0) { + switch (mode) { + case 4: /* IRM: set insert mode */ + compatibility(VT102); + term->insert = state; + break; + case 12: /* SRM: set echo mode */ + term->srm_echo = !state; + break; + case 20: /* LNM: Return sends ... */ + term->cr_lf_return = state; + break; + case 34: /* WYULCURM: Make cursor BIG */ + compatibility2(OTHER, VT220); + term->big_cursor = !state; + } + } +} + +/* + * Process an OSC sequence: set window title or icon name. + */ +static void do_osc(Terminal *term) +{ + if (term->osc_w) { + while (term->osc_strlen--) + term->wordness[(unsigned char)term->osc_string[term->osc_strlen]] = + term->esc_args[0]; + } else { + term->osc_string[term->osc_strlen] = '\0'; + switch (term->esc_args[0]) { + case 0: + case 1: + if (!term->no_remote_wintitle) { + sfree(term->icon_title); + term->icon_title = dupstr(term->osc_string); + term->icontitle_codepage = term->ucsdata->line_codepage; + term->win_icon_title_pending = true; + term_schedule_update(term); + } + if (term->esc_args[0] == 1) + break; + /* fall through: parameter 0 means set both */ + case 2: + case 21: + if (!term->no_remote_wintitle) { + sfree(term->window_title); + term->window_title = dupstr(term->osc_string); + term->wintitle_codepage = term->ucsdata->line_codepage; + term->win_title_pending = true; + term_schedule_update(term); + } + break; + case 4: + if (term->ldisc && !strcmp(term->osc_string, "?")) { + unsigned index = term->esc_args[1]; + if (index < OSC4_NCOLOURS) { + rgb colour = term->palette[index]; + char *reply_buf = dupprintf( + "\033]4;%u;rgb:%04x/%04x/%04x\007", index, + (unsigned)colour.r * 0x0101, + (unsigned)colour.g * 0x0101, + (unsigned)colour.b * 0x0101); + ldisc_send(term->ldisc, reply_buf, strlen(reply_buf), + false); + sfree(reply_buf); + } + } + break; + } + } +} + +/* + * ANSI printing routines. + */ +static void term_print_setup(Terminal *term, char *printer) +{ + bufchain_clear(&term->printer_buf); + term->print_job = printer_start_job(printer); +} +static void term_print_flush(Terminal *term) +{ + size_t size; + while ((size = bufchain_size(&term->printer_buf)) > 5) { + ptrlen data = bufchain_prefix(&term->printer_buf); + if (data.len > size-5) + data.len = size-5; + printer_job_data(term->print_job, data.ptr, data.len); + bufchain_consume(&term->printer_buf, data.len); + } +} +static void term_print_finish(Terminal *term) +{ + size_t size; + char c; + + if (!term->printing && !term->only_printing) + return; /* we need do nothing */ + + term_print_flush(term); + while ((size = bufchain_size(&term->printer_buf)) > 0) { + ptrlen data = bufchain_prefix(&term->printer_buf); + c = *(char *)data.ptr; + if (c == '\033' || c == '\233') { + bufchain_consume(&term->printer_buf, size); + break; + } else { + printer_job_data(term->print_job, &c, 1); + bufchain_consume(&term->printer_buf, 1); + } + } + printer_finish_job(term->print_job); + term->print_job = NULL; + term->printing = term->only_printing = false; +} + +static void term_display_graphic_char(Terminal *term, unsigned long c) +{ + termline *cline = scrlineptr(term->curs.y); + int width = 0; + if (DIRECT_CHAR(c)) + width = 1; + if (!width) + width = term_char_width(term, c); + + if (term->wrapnext && term->wrap && width > 0) { + cline->lattr |= LATTR_WRAPPED; + if (term->curs.y == term->marg_b) + scroll(term, term->marg_t, term->marg_b, 1, true); + else if (term->curs.y < term->rows - 1) + term->curs.y++; + term->curs.x = 0; + term->wrapnext = false; + cline = scrlineptr(term->curs.y); + } + if (term->insert && width > 0) + insch(term, width); + if (term->selstate != NO_SELECTION) { + pos cursplus = term->curs; + incpos(cursplus); + check_selection(term, term->curs, cursplus); + } + if (((c & CSET_MASK) == CSET_ASCII || + (c & CSET_MASK) == 0) && term->logctx) + logtraffic(term->logctx, (unsigned char) c, LGTYP_ASCII); + + check_trust_status(term, cline); + + int linecols = term->cols; + if (cline->trusted) + linecols -= TRUST_SIGIL_WIDTH; + + /* + * Preliminary check: if the terminal is only one character cell + * wide, then we cannot display any double-width character at all. + * Substitute single-width REPLACEMENT CHARACTER instead. + */ + if (width == 2 && linecols < 2) { + width = 1; + c = 0xFFFD; + } + + switch (width) { + case 2: + /* + * If we're about to display a double-width character starting + * in the rightmost column, then we do something special + * instead. We must print a space in the last column of the + * screen, then wrap; and we also set LATTR_WRAPPED2 which + * instructs subsequent cut-and-pasting not only to splice + * this line to the one after it, but to ignore the space in + * the last character position as well. (Because what was + * actually output to the terminal was presumably just a + * sequence of CJK characters, and we don't want a space to be + * pasted in the middle of those just because they had the + * misfortune to start in the wrong parity column. xterm + * concurs.) + */ + check_boundary(term, term->curs.x, term->curs.y); + check_boundary(term, term->curs.x+2, term->curs.y); + if (term->curs.x >= linecols-1) { + copy_termchar(cline, term->curs.x, + &term->erase_char); + cline->lattr |= LATTR_WRAPPED | LATTR_WRAPPED2; + if (term->curs.y == term->marg_b) + scroll(term, term->marg_t, term->marg_b, + 1, true); + else if (term->curs.y < term->rows - 1) + term->curs.y++; + term->curs.x = 0; + cline = scrlineptr(term->curs.y); + /* Now we must check_boundary again, of course. */ + check_boundary(term, term->curs.x, term->curs.y); + check_boundary(term, term->curs.x+2, term->curs.y); + } + + /* FULL-TERMCHAR */ + clear_cc(cline, term->curs.x); + cline->chars[term->curs.x].chr = c; + cline->chars[term->curs.x].attr = term->curr_attr; + cline->chars[term->curs.x].truecolour = + term->curr_truecolour; + + term->curs.x++; + + /* FULL-TERMCHAR */ + clear_cc(cline, term->curs.x); + cline->chars[term->curs.x].chr = UCSWIDE; + cline->chars[term->curs.x].attr = term->curr_attr; + cline->chars[term->curs.x].truecolour = + term->curr_truecolour; + + break; + case 1: + check_boundary(term, term->curs.x, term->curs.y); + check_boundary(term, term->curs.x+1, term->curs.y); + + /* FULL-TERMCHAR */ + clear_cc(cline, term->curs.x); + cline->chars[term->curs.x].chr = c; + cline->chars[term->curs.x].attr = term->curr_attr; + cline->chars[term->curs.x].truecolour = + term->curr_truecolour; + + break; + case 0: + if (term->curs.x > 0) { + int x = term->curs.x - 1; + + /* If we're in wrapnext state, the character to combine + * with is _here_, not to our left. */ + if (term->wrapnext) + x++; + + /* + * If the previous character is UCSWIDE, back up another + * one. + */ + if (cline->chars[x].chr == UCSWIDE) { + assert(x > 0); + x--; + } + + add_cc(cline, x, c); + seen_disp_event(term); + } + return; + default: + return; + } + term->curs.x++; + if (term->curs.x >= linecols) { + term->curs.x = linecols - 1; + term->wrapnext = true; + if (term->wrap && term->vt52_mode) { + cline->lattr |= LATTR_WRAPPED; + if (term->curs.y == term->marg_b) + scroll(term, term->marg_t, term->marg_b, 1, true); + else if (term->curs.y < term->rows - 1) + term->curs.y++; + term->curs.x = 0; + term->wrapnext = false; + } + } + seen_disp_event(term); +} + +static strbuf *term_input_data_from_unicode( + Terminal *term, const wchar_t *widebuf, int len) +{ + strbuf *buf = strbuf_new(); + + if (in_utf(term)) { + /* + * Translate input wide characters into UTF-8 to go in the + * terminal's input data queue. + */ + for (int i = 0; i < len; i++) { + unsigned long ch = widebuf[i]; + + if (IS_SURROGATE(ch)) { +#ifdef PLATFORM_IS_UTF16 + if (i+1 < len) { + unsigned long ch2 = widebuf[i+1]; + if (IS_SURROGATE_PAIR(ch, ch2)) { + ch = FROM_SURROGATES(ch, ch2); + i++; + } + } else +#endif + { + /* Unrecognised UTF-16 sequence */ + ch = '.'; + } + } + + char utf8_chr[6]; + put_data(buf, utf8_chr, encode_utf8(utf8_chr, ch)); + } + } else { + /* + * Call to the character-set subsystem to translate into + * whatever charset the terminal is currently configured in. + * + * Since the terminal doesn't currently support any multibyte + * character set other than UTF-8, we can assume here that + * there will be at most one output byte per input wchar_t. + * (But also we must allow space for the trailing NUL that + * wc_to_mb will write.) + */ + char *bufptr = strbuf_append(buf, len + 1); + int rv; + rv = wc_to_mb(term->ucsdata->line_codepage, 0, widebuf, len, + bufptr, len + 1, NULL); + strbuf_shrink_to(buf, rv < 0 ? 0 : rv); + } + + return buf; +} + +static strbuf *term_input_data_from_charset( + Terminal *term, int codepage, const char *str, int len) +{ + strbuf *buf; + + if (codepage < 0) { + buf = strbuf_new(); + put_data(buf, str, len); + } else { + int widesize = len * 2; /* allow for UTF-16 surrogates */ + wchar_t *widebuf = snewn(widesize, wchar_t); + int widelen = mb_to_wc(codepage, 0, str, len, widebuf, widesize); + buf = term_input_data_from_unicode(term, widebuf, widelen); + sfree(widebuf); + } + + return buf; +} + +static inline void term_bracketed_paste_start(Terminal *term) +{ + ptrlen seq = PTRLEN_LITERAL("\033[200~"); + if (term->ldisc) + ldisc_send(term->ldisc, seq.ptr, seq.len, false); + term->bracketed_paste_active = true; +} + +static inline void term_bracketed_paste_stop(Terminal *term) +{ + if (!term->bracketed_paste_active) + return; + + ptrlen seq = PTRLEN_LITERAL("\033[201~"); + if (term->ldisc) + ldisc_send(term->ldisc, seq.ptr, seq.len, false); + term->bracketed_paste_active = false; +} + +static inline void term_keyinput_internal( + Terminal *term, const void *buf, int len, bool interactive) +{ + if (term->srm_echo) { + /* + * Implement the terminal-level local echo behaviour that + * ECMA-48 specifies when terminal mode 12 is configured off + * (ESC[12l). In this mode, data input to the terminal via the + * keyboard is also added to the output buffer. But this + * doesn't apply to escape sequences generated as session + * input _within_ the terminal, e.g. in response to terminal + * query sequences, or the bracketing sequences of bracketed + * paste mode. Those will be sent directly via + * ldisc_send(term->ldisc, ...) and won't go through this + * function. + */ + + /* Mimic the special case of negative length in ldisc_send */ + int true_len = len >= 0 ? len : strlen(buf); + + bufchain_add(&term->inbuf, buf, true_len); + term_added_data(term, false); + } + if (interactive) + term_bracketed_paste_stop(term); + if (term->ldisc) + ldisc_send(term->ldisc, buf, len, interactive); + term_seen_key_event(term); +} + +unsigned long term_translate( + Terminal *term, struct term_utf8_decode *utf8, unsigned char c) +{ + if (in_utf(term)) { + switch (utf8->state) { + case 0: + if (c < 0x80) { + /* UTF-8 must be stateless so we ignore iso2022. */ + if (term->ucsdata->unitab_ctrl[c] != 0xFF) { + return term->ucsdata->unitab_ctrl[c]; + } else if ((term->utf8linedraw) && + (term->cset_attr[term->cset] == CSET_LINEDRW)) { + /* Linedraw characters are explicitly enabled */ + return c | CSET_LINEDRW; + } else { + return c | CSET_ASCII; + } + } else if ((c & 0xe0) == 0xc0) { + utf8->size = utf8->state = 1; + utf8->chr = (c & 0x1f); + } else if ((c & 0xf0) == 0xe0) { + utf8->size = utf8->state = 2; + utf8->chr = (c & 0x0f); + } else if ((c & 0xf8) == 0xf0) { + utf8->size = utf8->state = 3; + utf8->chr = (c & 0x07); + } else if ((c & 0xfc) == 0xf8) { + utf8->size = utf8->state = 4; + utf8->chr = (c & 0x03); + } else if ((c & 0xfe) == 0xfc) { + utf8->size = utf8->state = 5; + utf8->chr = (c & 0x01); + } else { + return UCSINVALID; + } + return UCSINCOMPLETE; + case 1: + case 2: + case 3: + case 4: + case 5: + if ((c & 0xC0) != 0x80) { + utf8->state = 0; + return UCSTRUNCATED; /* caller will then give us the + * same byte again */ + } + utf8->chr = (utf8->chr << 6) | (c & 0x3f); + if (--utf8->state) + return UCSINCOMPLETE; + + unsigned long t = utf8->chr; + + /* Is somebody trying to be evil! */ + if (t < 0x80 || + (t < 0x800 && utf8->size >= 2) || + (t < 0x10000 && utf8->size >= 3) || + (t < 0x200000 && utf8->size >= 4) || + (t < 0x4000000 && utf8->size >= 5)) + return UCSINVALID; + + /* Unicode line separator and paragraph separator are CR-LF */ + if (t == 0x2028 || t == 0x2029) + return 0x85; + + /* High controls are probably a Baaad idea too. */ + if (t < 0xA0) + return 0xFFFD; + + /* The UTF-16 surrogates are not nice either. */ + /* The standard give the option of decoding these: + * I don't want to! */ + if (t >= 0xD800 && t < 0xE000) + return UCSINVALID; + + /* ISO 10646 characters now limited to UTF-16 range. */ + if (t > 0x10FFFF) + return UCSINVALID; + + /* This is currently a TagPhobic application.. */ + if (t >= 0xE0000 && t <= 0xE007F) + return UCSINCOMPLETE; + + /* U+FEFF is best seen as a null. */ + if (t == 0xFEFF) + return UCSINCOMPLETE; + /* But U+FFFE is an error. */ + if (t == 0xFFFE || t == 0xFFFF) + return UCSINVALID; + + return t; + } + } else if (term->sco_acs && + (c!='\033' && c!='\012' && c!='\015' && c!='\b')) { + /* Are we in the nasty ACS mode? Note: no sco in utf mode. */ + if (term->sco_acs == 2) + c |= 0x80; + + return c | CSET_SCOACS; + } else { + switch (term->cset_attr[term->cset]) { + /* + * Linedraw characters are different from 'ESC ( B' + * only for a small range. For ones outside that + * range, make sure we use the same font as well as + * the same encoding. + */ + case CSET_LINEDRW: + if (term->ucsdata->unitab_ctrl[c] != 0xFF) + return term->ucsdata->unitab_ctrl[c]; + else + return c | CSET_LINEDRW; + break; + + case CSET_GBCHR: + /* If UK-ASCII, make the '#' a LineDraw Pound */ + if (c == '#') + return '}' | CSET_LINEDRW; + /* fall through */ + + case CSET_ASCII: + if (term->ucsdata->unitab_ctrl[c] != 0xFF) + return term->ucsdata->unitab_ctrl[c]; + else + return c | CSET_ASCII; + break; + case CSET_SCOACS: + if (c >= ' ') + return c | CSET_SCOACS; + break; + } + } + return c; +} + +/* + * Remove everything currently in `inbuf' and stick it up on the + * in-memory display. There's a big state machine in here to + * process escape sequences... + */ +static void term_out(Terminal *term, bool called_from_term_data) +{ + unsigned long c; + int unget; + const unsigned char *chars; + size_t nchars_got = 0, nchars_used = 0; + + /* + * During drag-selects, we do not process terminal input, because + * the user will want the screen to hold still to be selected. + */ + if (term->selstate == DRAGGING) + return; + + unget = -1; + + chars = NULL; /* placate compiler warnings */ + while (nchars_got < nchars_used || + unget != -1 || + bufchain_size(&term->inbuf) > 0) { + if (unget != -1) { + /* + * Handle a character we left in 'unget' the last time + * round this loop. This happens if a UTF-8 sequence is + * aborted early, by containing fewer continuation bytes + * than its introducer expected: the non-continuation byte + * that interrupted the sequence must now be processed + * as a fresh piece of input in its own right. + */ + c = unget; + unget = -1; + } else { + /* + * If we're waiting for a terminal resize triggered by an + * escape sequence, we defer processing the terminal + * output until we receive acknowledgment from the front + * end that the resize has happened, so that further + * output will be processed in the context of the new + * size. + * + * This test goes inside the main while-loop, so that we + * exit early if we encounter a resize escape sequence + * part way through term->inbuf. + * + * It's also in the branch of this if statement that + * doesn't deal with a character left in 'unget' by the + * previous loop iteration, because if we break out of + * this loop with an ungot character still pending, we'll + * lose it. (And in any case, if the previous thing that + * happened was a truncated UTF-8 sequence, then it won't + * have scheduled a pending resize.) + */ + if (term->win_resize_pending != WIN_RESIZE_NO) + break; + + if (nchars_got == nchars_used) { + /* Delete the previous chunk from the bufchain */ + bufchain_consume(&term->inbuf, nchars_used); + nchars_used = 0; + + if (bufchain_size(&term->inbuf) == 0) + break; /* no more data */ + + ptrlen data = bufchain_prefix(&term->inbuf); + chars = data.ptr; + nchars_got = data.len; + assert(chars != NULL); + assert(nchars_used < nchars_got); + } + c = chars[nchars_used++]; + + /* + * Optionally log the session traffic to a file. Useful for + * debugging and possibly also useful for actual logging. + */ + if (term->logtype == LGTYP_DEBUG && term->logctx) + logtraffic(term->logctx, (unsigned char) c, LGTYP_DEBUG); + } + + /* Note only VT220+ are 8-bit VT102 is seven bit, it shouldn't even + * be able to display 8-bit characters, but I'll let that go 'cause + * of i18n. + */ + + /* + * If we're printing, add the character to the printer + * buffer. + */ + if (term->printing) { + bufchain_add(&term->printer_buf, &c, 1); + + /* + * If we're in print-only mode, we use a much simpler + * state machine designed only to recognise the ESC[4i + * termination sequence. + */ + if (term->only_printing) { + if (c == '\033') + term->print_state = 1; + else if (c == (unsigned char)'\233') + term->print_state = 2; + else if (c == '[' && term->print_state == 1) + term->print_state = 2; + else if (c == '4' && term->print_state == 2) + term->print_state = 3; + else if (c == 'i' && term->print_state == 3) + term->print_state = 4; + else + term->print_state = 0; + if (term->print_state == 4) { + term_print_finish(term); + } + continue; + } + } + + /* Do character-set translation. */ + if (term->termstate == TOPLEVEL) { + unsigned long t = term_translate(term, &term->utf8, c); + switch (t) { + case UCSINCOMPLETE: + continue; /* didn't complete a multibyte char */ + case UCSTRUNCATED: + unget = c; + /* fall through */ + case UCSINVALID: + c = UCSERR; + break; + default: + c = t; + break; + } + } + + /* + * How about C1 controls? + * Explicitly ignore SCI (0x9a), which we don't translate to DECID. + */ + if ((c & -32) == 0x80 && term->termstate < DO_CTRLS && + !term->vt52_mode && has_compat(VT220)) { + if (c == 0x9a) + c = 0; + else { + term->termstate = SEEN_ESC; + term->esc_query = 0; + c = '@' + (c & 0x1F); + } + } + + /* Or the GL control. */ + if (c == '\177' && term->termstate < DO_CTRLS && has_compat(OTHER)) { + if (term->curs.x && !term->wrapnext) + term->curs.x--; + term->wrapnext = false; + /* destructive backspace might be disabled */ + if (!term->no_dbackspace) { + check_boundary(term, term->curs.x, term->curs.y); + check_boundary(term, term->curs.x+1, term->curs.y); + copy_termchar(scrlineptr(term->curs.y), + term->curs.x, &term->erase_char); + } + } else + /* Or normal C0 controls. */ + if ((c & ~0x1F) == 0 && term->termstate < DO_CTRLS) { + switch (c) { + case '\005': /* ENQ: terminal type query */ + /* + * Strictly speaking this is VT100 but a VT100 defaults to + * no response. Other terminals respond at their option. + * + * Don't put a CR in the default string as this tends to + * upset some weird software. + */ + compatibility(ANSIMIN); + if (term->ldisc) { + strbuf *buf = term_input_data_from_charset( + term, DEFAULT_CODEPAGE, + term->answerback, term->answerbacklen); + ldisc_send(term->ldisc, buf->s, buf->len, false); + strbuf_free(buf); + } + break; + case '\007': { /* BEL: Bell */ + if (term->termstate == SEEN_OSC || + term->termstate == SEEN_OSC_W) { + /* + * In an OSC context, BEL is one of the ways to terminate + * the whole sequence. We process it as such even if we + * haven't got into the final OSC_STRING state yet, so that + * OSC sequences without a string will be handled cleanly. + */ + do_osc(term); + term->termstate = TOPLEVEL; + break; + } + + struct beeptime *newbeep; + unsigned long ticks; + + ticks = GETTICKCOUNT(); + + if (!term->beep_overloaded) { + newbeep = snew(struct beeptime); + newbeep->ticks = ticks; + newbeep->next = NULL; + if (!term->beephead) + term->beephead = newbeep; + else + term->beeptail->next = newbeep; + term->beeptail = newbeep; + term->nbeeps++; + } + + /* + * Throw out any beeps that happened more than + * t seconds ago. + */ + while (term->beephead && + term->beephead->ticks < ticks - term->bellovl_t) { + struct beeptime *tmp = term->beephead; + term->beephead = tmp->next; + sfree(tmp); + if (!term->beephead) + term->beeptail = NULL; + term->nbeeps--; + } + + if (term->bellovl && term->beep_overloaded && + ticks - term->lastbeep >= (unsigned)term->bellovl_s) { + /* + * If we're currently overloaded and the + * last beep was more than s seconds ago, + * leave overload mode. + */ + term->beep_overloaded = false; + } else if (term->bellovl && !term->beep_overloaded && + term->nbeeps >= term->bellovl_n) { + /* + * Now, if we have n or more beeps + * remaining in the queue, go into overload + * mode. + */ + term->beep_overloaded = true; + } + term->lastbeep = ticks; + + /* + * Perform an actual beep if we're not overloaded. + */ + if (!term->bellovl || !term->beep_overloaded) { + win_bell(term->win, term->beep); + + if (term->beep == BELL_VISUAL) { + term_schedule_vbell(term, false, 0); + } + } + seen_disp_event(term); + break; + } + case '\b': /* BS: Back space */ + if (term->curs.x == 0 && (term->curs.y == 0 || !term->wrap)) + /* do nothing */ ; + else if (term->curs.x == 0 && term->curs.y > 0) + term->curs.x = term->cols - 1, term->curs.y--; + else if (term->wrapnext) + term->wrapnext = false; + else + term->curs.x--; + seen_disp_event(term); + break; + case '\016': /* LS1: Locking-shift one */ + compatibility(VT100); + term->cset = 1; + break; + case '\017': /* LS0: Locking-shift zero */ + compatibility(VT100); + term->cset = 0; + break; + case '\033': /* ESC: Escape */ + if (term->vt52_mode) + term->termstate = VT52_ESC; + else if (term->termstate == SEEN_OSC || + term->termstate == SEEN_OSC_W) { + /* Be prepared to terminate an OSC early */ + term->termstate = OSC_MAYBE_ST; + } else { + compatibility(ANSIMIN); + term->termstate = SEEN_ESC; + term->esc_query = 0; + } + break; + case '\015': /* CR: Carriage return */ + term->curs.x = 0; + term->wrapnext = false; + seen_disp_event(term); + + if (term->crhaslf) { + if (term->curs.y == term->marg_b) + scroll(term, term->marg_t, term->marg_b, 1, true); + else if (term->curs.y < term->rows - 1) + term->curs.y++; + } + if (term->logctx) + logtraffic(term->logctx, (unsigned char) c, LGTYP_ASCII); + break; + case '\014': /* FF: Form feed */ + if (has_compat(SCOANSI)) { + move(term, 0, 0, 0); + erase_lots(term, false, false, true); + if (term->scroll_on_disp) + term->disptop = 0; + term->wrapnext = false; + seen_disp_event(term); + break; + } + case '\013': /* VT: Line tabulation */ + compatibility(VT100); + case '\012': /* LF: Line feed */ + if (term->curs.y == term->marg_b) + scroll(term, term->marg_t, term->marg_b, 1, true); + else if (term->curs.y < term->rows - 1) + term->curs.y++; + if (term->lfhascr) + term->curs.x = 0; + term->wrapnext = false; + seen_disp_event(term); + if (term->logctx) + logtraffic(term->logctx, (unsigned char) c, LGTYP_ASCII); + break; + case '\t': { /* HT: Character tabulation */ + pos old_curs = term->curs; + termline *ldata = scrlineptr(term->curs.y); + + do { + term->curs.x++; + } while (term->curs.x < term->cols - 1 && + !term->tabs[term->curs.x]); + + if ((ldata->lattr & LATTR_MODE) != LATTR_NORM) { + if (term->curs.x >= term->cols / 2) + term->curs.x = term->cols / 2 - 1; + } else { + if (term->curs.x >= term->cols) + term->curs.x = term->cols - 1; + } + + check_selection(term, old_curs, term->curs); + seen_disp_event(term); + break; + } + } + } else + switch (term->termstate) { + case TOPLEVEL: + /* Only graphic characters get this far; + * ctrls are stripped above */ + term_display_graphic_char(term, c); + term->last_graphic_char = c; + break; + + case OSC_MAYBE_ST: + /* + * This state is virtually identical to SEEN_ESC, with the + * exception that we have an OSC sequence in the pipeline, + * and _if_ we see a backslash, we process it. + */ + if (c == '\\') { + do_osc(term); + term->termstate = TOPLEVEL; + break; + } + /* else fall through */ + case SEEN_ESC: + if (c >= ' ' && c <= '/') { + if (term->esc_query) + term->esc_query = -1; + else + term->esc_query = c; + break; + } + term->termstate = TOPLEVEL; + switch (ANSI(c, term->esc_query)) { + case '[': /* enter CSI mode */ + term->termstate = SEEN_CSI; + term->esc_nargs = 1; + term->esc_args[0] = ARG_DEFAULT; + term->esc_query = 0; + break; + case ']': /* OSC: xterm escape sequences */ + /* Compatibility is nasty here, xterm, linux, decterm yuk! */ + compatibility(OTHER); + term->termstate = SEEN_OSC; + term->osc_strlen = 0; + term->esc_args[0] = 0; + term->esc_nargs = 1; + break; + case '7': /* DECSC: save cursor */ + compatibility(VT100); + save_cursor(term, true); + break; + case '8': /* DECRC: restore cursor */ + compatibility(VT100); + save_cursor(term, false); + seen_disp_event(term); + break; + case '=': /* DECKPAM: Keypad application mode */ + compatibility(VT100); + term->app_keypad_keys = true; + break; + case '>': /* DECKPNM: Keypad numeric mode */ + compatibility(VT100); + term->app_keypad_keys = false; + break; + case 'D': /* IND: exactly equivalent to LF */ + compatibility(VT100); + if (term->curs.y == term->marg_b) + scroll(term, term->marg_t, term->marg_b, 1, true); + else if (term->curs.y < term->rows - 1) + term->curs.y++; + term->wrapnext = false; + seen_disp_event(term); + break; + case 'E': /* NEL: exactly equivalent to CR-LF */ + compatibility(VT100); + term->curs.x = 0; + if (term->curs.y == term->marg_b) + scroll(term, term->marg_t, term->marg_b, 1, true); + else if (term->curs.y < term->rows - 1) + term->curs.y++; + term->wrapnext = false; + seen_disp_event(term); + break; + case 'M': /* RI: reverse index - backwards LF */ + compatibility(VT100); + if (term->curs.y == term->marg_t) + scroll(term, term->marg_t, term->marg_b, -1, true); + else if (term->curs.y > 0) + term->curs.y--; + term->wrapnext = false; + seen_disp_event(term); + break; + case 'Z': /* DECID: terminal type query */ + compatibility(VT100); + if (term->ldisc) + ldisc_send(term->ldisc, term->id_string, + strlen(term->id_string), false); + break; + case 'c': /* RIS: restore power-on settings */ + compatibility(VT100); + power_on(term, true); + if (term->ldisc) /* cause ldisc to notice changes */ + ldisc_echoedit_update(term->ldisc); + if (term->reset_132) { + if (!term->no_remote_resize) + term_request_resize(term, 80, term->rows); + term->reset_132 = false; + } + if (term->scroll_on_disp) + term->disptop = 0; + seen_disp_event(term); + break; + case 'H': /* HTS: set a tab */ + compatibility(VT100); + term->tabs[term->curs.x] = true; + break; + + case ANSI('8', '#'): { /* DECALN: fills screen with Es :-) */ + compatibility(VT100); + termline *ldata; + int i, j; + pos scrtop, scrbot; + + for (i = 0; i < term->rows; i++) { + ldata = scrlineptr(i); + check_line_size(term, ldata); + for (j = 0; j < term->cols; j++) { + copy_termchar(ldata, j, + &term->basic_erase_char); + ldata->chars[j].chr = 'E'; + } + ldata->lattr = LATTR_NORM; + } + if (term->scroll_on_disp) + term->disptop = 0; + seen_disp_event(term); + scrtop.x = scrtop.y = 0; + scrbot.x = 0; + scrbot.y = term->rows; + check_selection(term, scrtop, scrbot); + break; + } + + case ANSI('3', '#'): + case ANSI('4', '#'): + case ANSI('5', '#'): + case ANSI('6', '#'): { + compatibility(VT100); + int nlattr; + termline *ldata; + + switch (ANSI(c, term->esc_query)) { + case ANSI('3', '#'): /* DECDHL: 2*height, top */ + nlattr = LATTR_TOP; + break; + case ANSI('4', '#'): /* DECDHL: 2*height, bottom */ + nlattr = LATTR_BOT; + break; + case ANSI('5', '#'): /* DECSWL: normal */ + nlattr = LATTR_NORM; + break; + default: /* case ANSI('6', '#'): DECDWL: 2*width */ + nlattr = LATTR_WIDE; + break; + } + ldata = scrlineptr(term->curs.y); + check_line_size(term, ldata); + check_trust_status(term, ldata); + ldata->lattr = nlattr; + break; + } + /* GZD4: G0 designate 94-set */ + case ANSI('A', '('): + compatibility(VT100); + if (!term->no_remote_charset) + term->cset_attr[0] = CSET_GBCHR; + break; + case ANSI('B', '('): + compatibility(VT100); + if (!term->no_remote_charset) + term->cset_attr[0] = CSET_ASCII; + break; + case ANSI('0', '('): + compatibility(VT100); + if (!term->no_remote_charset) + term->cset_attr[0] = CSET_LINEDRW; + break; + case ANSI('U', '('): + compatibility(OTHER); + if (!term->no_remote_charset) + term->cset_attr[0] = CSET_SCOACS; + break; + /* G1D4: G1-designate 94-set */ + case ANSI('A', ')'): + compatibility(VT100); + if (!term->no_remote_charset) + term->cset_attr[1] = CSET_GBCHR; + break; + case ANSI('B', ')'): + compatibility(VT100); + if (!term->no_remote_charset) + term->cset_attr[1] = CSET_ASCII; + break; + case ANSI('0', ')'): + compatibility(VT100); + if (!term->no_remote_charset) + term->cset_attr[1] = CSET_LINEDRW; + break; + case ANSI('U', ')'): + compatibility(OTHER); + if (!term->no_remote_charset) + term->cset_attr[1] = CSET_SCOACS; + break; + /* DOCS: Designate other coding system */ + case ANSI('8', '%'): /* Old Linux code */ + case ANSI('G', '%'): + compatibility(OTHER); + if (!term->no_remote_charset) + term->utf = true; + break; + case ANSI('@', '%'): + compatibility(OTHER); + if (!term->no_remote_charset) + term->utf = false; + break; + } + break; + case SEEN_CSI: + term->termstate = TOPLEVEL; /* default */ + if (isdigit(c)) { + if (term->esc_nargs <= ARGS_MAX) { + if (term->esc_args[term->esc_nargs - 1] == ARG_DEFAULT) + term->esc_args[term->esc_nargs - 1] = 0; + if (term->esc_args[term->esc_nargs - 1] <= + UINT_MAX / 10 && + term->esc_args[term->esc_nargs - 1] * 10 <= + UINT_MAX - c - '0') + term->esc_args[term->esc_nargs - 1] = + 10 * term->esc_args[term->esc_nargs - 1] + + c - '0'; + else + term->esc_args[term->esc_nargs - 1] = UINT_MAX; + } + term->termstate = SEEN_CSI; + } else if (c == ';') { + if (term->esc_nargs < ARGS_MAX) + term->esc_args[term->esc_nargs++] = ARG_DEFAULT; + term->termstate = SEEN_CSI; + } else if (c < '@') { + if (term->esc_query) + term->esc_query = -1; + else if (c == '?') + term->esc_query = 1; + else + term->esc_query = c; + term->termstate = SEEN_CSI; + } else +#define CLAMP(arg, lim) ((arg) = ((arg) > (lim)) ? (lim) : (arg)) + switch (ANSI(c, term->esc_query)) { + case 'A': /* CUU: move up N lines */ + CLAMP(term->esc_args[0], term->rows); + move(term, term->curs.x, + term->curs.y - def(term->esc_args[0], 1), 1); + seen_disp_event(term); + break; + case 'e': /* VPR: move down N lines */ + compatibility(ANSI); + /* FALLTHROUGH */ + case 'B': /* CUD: Cursor down */ + CLAMP(term->esc_args[0], term->rows); + move(term, term->curs.x, + term->curs.y + def(term->esc_args[0], 1), 1); + seen_disp_event(term); + break; + case 'b': /* REP: repeat previous grap */ + CLAMP(term->esc_args[0], term->rows * term->cols); + if (term->last_graphic_char) { + unsigned i; + for (i = 0; i < term->esc_args[0]; i++) + term_display_graphic_char( + term, term->last_graphic_char); + } + break; + case ANSI('c', '>'): /* DA: report xterm version */ + compatibility(OTHER); + /* this reports xterm version 136 so that VIM can + use the drag messages from the mouse reporting */ + if (term->ldisc) + ldisc_send(term->ldisc, "\033[>0;136;0c", 11, + false); + break; + case 'a': /* HPR: move right N cols */ + compatibility(ANSI); + /* FALLTHROUGH */ + case 'C': /* CUF: Cursor right */ + CLAMP(term->esc_args[0], term->cols); + move(term, term->curs.x + def(term->esc_args[0], 1), + term->curs.y, 1); + seen_disp_event(term); + break; + case 'D': /* CUB: move left N cols */ + CLAMP(term->esc_args[0], term->cols); + move(term, term->curs.x - def(term->esc_args[0], 1), + term->curs.y, 1); + seen_disp_event(term); + break; + case 'E': /* CNL: move down N lines and CR */ + compatibility(ANSI); + CLAMP(term->esc_args[0], term->rows); + move(term, 0, + term->curs.y + def(term->esc_args[0], 1), 1); + seen_disp_event(term); + break; + case 'F': /* CPL: move up N lines and CR */ + compatibility(ANSI); + CLAMP(term->esc_args[0], term->rows); + move(term, 0, + term->curs.y - def(term->esc_args[0], 1), 1); + seen_disp_event(term); + break; + case 'G': /* CHA */ + case '`': /* HPA: set horizontal posn */ + compatibility(ANSI); + CLAMP(term->esc_args[0], term->cols); + move(term, def(term->esc_args[0], 1) - 1, + term->curs.y, 0); + seen_disp_event(term); + break; + case 'd': /* VPA: set vertical posn */ + compatibility(ANSI); + CLAMP(term->esc_args[0], term->rows); + move(term, term->curs.x, + ((term->dec_om ? term->marg_t : 0) + + def(term->esc_args[0], 1) - 1), + (term->dec_om ? 2 : 0)); + seen_disp_event(term); + break; + case 'H': /* CUP */ + case 'f': /* HVP: set horz and vert posns at once */ + if (term->esc_nargs < 2) + term->esc_args[1] = ARG_DEFAULT; + CLAMP(term->esc_args[0], term->rows); + CLAMP(term->esc_args[1], term->cols); + move(term, def(term->esc_args[1], 1) - 1, + ((term->dec_om ? term->marg_t : 0) + + def(term->esc_args[0], 1) - 1), + (term->dec_om ? 2 : 0)); + seen_disp_event(term); + break; + case 'J': { /* ED: erase screen or parts of it */ + unsigned int i = def(term->esc_args[0], 0); + if (i == 3) { + /* Erase Saved Lines (xterm) + * This follows Thomas Dickey's xterm. */ + if (!term->no_remote_clearscroll) + term_clrsb(term); + } else { + i++; + if (i > 3) + i = 0; + erase_lots(term, false, !!(i & 2), !!(i & 1)); + } + if (term->scroll_on_disp) + term->disptop = 0; + seen_disp_event(term); + break; + } + case 'K': { /* EL: erase line or parts of it */ + unsigned int i = def(term->esc_args[0], 0) + 1; + if (i > 3) + i = 0; + erase_lots(term, true, !!(i & 2), !!(i & 1)); + seen_disp_event(term); + break; + } + case 'L': /* IL: insert lines */ + compatibility(VT102); + CLAMP(term->esc_args[0], term->rows); + if (term->curs.y <= term->marg_b) + scroll(term, term->curs.y, term->marg_b, + -def(term->esc_args[0], 1), false); + seen_disp_event(term); + break; + case 'M': /* DL: delete lines */ + compatibility(VT102); + CLAMP(term->esc_args[0], term->rows); + if (term->curs.y <= term->marg_b) + scroll(term, term->curs.y, term->marg_b, + def(term->esc_args[0], 1), + true); + seen_disp_event(term); + break; + case '@': /* ICH: insert chars */ + /* XXX VTTEST says this is vt220, vt510 manual says vt102 */ + compatibility(VT102); + CLAMP(term->esc_args[0], term->cols); + insch(term, def(term->esc_args[0], 1)); + seen_disp_event(term); + break; + case 'P': /* DCH: delete chars */ + compatibility(VT102); + CLAMP(term->esc_args[0], term->cols); + insch(term, -def(term->esc_args[0], 1)); + seen_disp_event(term); + break; + case 'c': /* DA: terminal type query */ + compatibility(VT100); + /* This is the response for a VT102 */ + if (term->ldisc) + ldisc_send(term->ldisc, term->id_string, + strlen(term->id_string), false); + break; + case 'n': /* DSR: cursor position query */ + if (term->ldisc) { + if (term->esc_args[0] == 6) { + char buf[32]; + sprintf(buf, "\033[%d;%dR", term->curs.y + 1, + term->curs.x + 1); + ldisc_send(term->ldisc, buf, strlen(buf), + false); + } else if (term->esc_args[0] == 5) { + ldisc_send(term->ldisc, "\033[0n", 4, false); + } + } + break; + case 'h': /* SM: toggle modes to high */ + case ANSI_QUE('h'): + compatibility(VT100); + for (int i = 0; i < term->esc_nargs; i++) + toggle_mode(term, term->esc_args[i], + term->esc_query, true); + break; + case 'i': /* MC: Media copy */ + case ANSI_QUE('i'): { + compatibility(VT100); + char *printer; + if (term->esc_nargs != 1) break; + if (term->esc_args[0] == 5 && + (printer = conf_get_str(term->conf, + CONF_printer))[0]) { + term->printing = true; + term->only_printing = !term->esc_query; + term->print_state = 0; + term_print_setup(term, printer); + } else if (term->esc_args[0] == 4 && + term->printing) { + term_print_finish(term); + } + break; + } + case 'l': /* RM: toggle modes to low */ + case ANSI_QUE('l'): + compatibility(VT100); + for (int i = 0; i < term->esc_nargs; i++) + toggle_mode(term, term->esc_args[i], + term->esc_query, false); + break; + case 'g': /* TBC: clear tabs */ + compatibility(VT100); + if (term->esc_nargs == 1) { + if (term->esc_args[0] == 0) { + term->tabs[term->curs.x] = false; + } else if (term->esc_args[0] == 3) { + int i; + for (i = 0; i < term->cols; i++) + term->tabs[i] = false; + } + } + break; + case 'r': /* DECSTBM: set scroll margins */ + compatibility(VT100); + if (term->esc_nargs <= 2) { + int top, bot; + CLAMP(term->esc_args[0], term->rows); + CLAMP(term->esc_args[1], term->rows); + top = def(term->esc_args[0], 1) - 1; + bot = (term->esc_nargs <= 1 + || term->esc_args[1] == 0 ? + term->rows : + def(term->esc_args[1], term->rows)) - 1; + if (bot >= term->rows) + bot = term->rows - 1; + /* VTTEST Bug 9 - if region is less than 2 lines + * don't change region. + */ + if (bot - top > 0) { + term->marg_t = top; + term->marg_b = bot; + term->curs.x = 0; + /* + * I used to think the cursor should be + * placed at the top of the newly marginned + * area. Apparently not: VMS TPU falls over + * if so. + * + * Well actually it should for + * Origin mode - RDB + */ + term->curs.y = (term->dec_om ? + term->marg_t : 0); + seen_disp_event(term); + } + } + break; + case 'm': /* SGR: set graphics rendition */ + /* + * A VT100 without the AVO only had one + * attribute, either underline or reverse + * video depending on the cursor type, this + * was selected by CSI 7m. + * + * case 2: + * This is sometimes DIM, eg on the GIGI and + * Linux + * case 8: + * This is sometimes INVIS various ANSI. + * case 21: + * This like 22 disables BOLD, DIM and INVIS + * + * The ANSI colours appear on any terminal + * that has colour (obviously) but the + * interaction between sgr0 and the colours + * varies but is usually related to the + * background colour erase item. The + * interaction between colour attributes and + * the mono ones is also very implementation + * dependent. + * + * The 39 and 49 attributes are likely to be + * unimplemented. + */ + for (int i = 0; i < term->esc_nargs; i++) + switch (def(term->esc_args[i], 0)) { + case 0: /* restore defaults */ + term->curr_attr = term->default_attr; + term->curr_truecolour = + term->basic_erase_char.truecolour; + break; + case 1: /* enable bold */ + compatibility(VT100AVO); + term->curr_attr |= ATTR_BOLD; + break; + case 2: /* enable dim */ + compatibility(OTHER); + term->curr_attr |= ATTR_DIM; + break; + case 21: /* (enable double underline) */ + compatibility(OTHER); + case 4: /* enable underline */ + compatibility(VT100AVO); + term->curr_attr |= ATTR_UNDER; + break; + case 5: /* enable blink */ + compatibility(VT100AVO); + term->curr_attr |= ATTR_BLINK; + break; + case 6: /* SCO light bkgrd */ + compatibility(SCOANSI); + term->blink_is_real = false; + term->curr_attr |= ATTR_BLINK; + term_schedule_tblink(term); + break; + case 7: /* enable reverse video */ + term->curr_attr |= ATTR_REVERSE; + break; + case 9: /* enable strikethrough */ + term->curr_attr |= ATTR_STRIKE; + break; + case 10: /* SCO acs off */ + compatibility(SCOANSI); + if (term->no_remote_charset) break; + term->sco_acs = 0; break; + case 11: /* SCO acs on */ + compatibility(SCOANSI); + if (term->no_remote_charset) break; + term->sco_acs = 1; break; + case 12: /* SCO acs on, |0x80 */ + compatibility(SCOANSI); + if (term->no_remote_charset) break; + term->sco_acs = 2; break; + case 22: /* disable bold and dim */ + compatibility2(OTHER, VT220); + term->curr_attr &= ~(ATTR_BOLD | ATTR_DIM); + break; + case 24: /* disable underline */ + compatibility2(OTHER, VT220); + term->curr_attr &= ~ATTR_UNDER; + break; + case 25: /* disable blink */ + compatibility2(OTHER, VT220); + term->curr_attr &= ~ATTR_BLINK; + break; + case 27: /* disable reverse video */ + compatibility2(OTHER, VT220); + term->curr_attr &= ~ATTR_REVERSE; + break; + case 29: /* disable strikethrough */ + term->curr_attr &= ~ATTR_STRIKE; + break; + case 30: + case 31: + case 32: + case 33: + case 34: + case 35: + case 36: + case 37: + /* foreground */ + term->curr_truecolour.fg.enabled = false; + term->curr_attr &= ~ATTR_FGMASK; + term->curr_attr |= + (term->esc_args[i] - 30)<<ATTR_FGSHIFT; + break; + case 90: + case 91: + case 92: + case 93: + case 94: + case 95: + case 96: + case 97: + /* aixterm-style bright foreground */ + term->curr_truecolour.fg.enabled = false; + term->curr_attr &= ~ATTR_FGMASK; + term->curr_attr |= + ((term->esc_args[i] - 90 + 8) + << ATTR_FGSHIFT); + break; + case 39: /* default-foreground */ + term->curr_truecolour.fg.enabled = false; + term->curr_attr &= ~ATTR_FGMASK; + term->curr_attr |= ATTR_DEFFG; + break; + case 40: + case 41: + case 42: + case 43: + case 44: + case 45: + case 46: + case 47: + /* background */ + term->curr_truecolour.bg.enabled = false; + term->curr_attr &= ~ATTR_BGMASK; + term->curr_attr |= + (term->esc_args[i] - 40)<<ATTR_BGSHIFT; + break; + case 100: + case 101: + case 102: + case 103: + case 104: + case 105: + case 106: + case 107: + /* aixterm-style bright background */ + term->curr_truecolour.bg.enabled = false; + term->curr_attr &= ~ATTR_BGMASK; + term->curr_attr |= + ((term->esc_args[i] - 100 + 8) + << ATTR_BGSHIFT); + break; + case 49: /* default-background */ + term->curr_truecolour.bg.enabled = false; + term->curr_attr &= ~ATTR_BGMASK; + term->curr_attr |= ATTR_DEFBG; + break; + + /* + * 256-colour and true-colour + * sequences. A 256-colour + * foreground is selected by a + * sequence of 3 arguments in the + * form 38;5;n, where n is in the + * range 0-255. A true-colour RGB + * triple is selected by 5 args of + * the form 38;2;r;g;b. Replacing + * the initial 38 with 48 in both + * cases selects the same colour + * as the background. + */ + case 38: + if (i+2 < term->esc_nargs && + term->esc_args[i+1] == 5) { + term->curr_attr &= ~ATTR_FGMASK; + term->curr_attr |= + ((term->esc_args[i+2] & 0xFF) + << ATTR_FGSHIFT); + term->curr_truecolour.fg = + optionalrgb_none; + i += 2; + } + if (i + 4 < term->esc_nargs && + term->esc_args[i + 1] == 2) { + parse_optionalrgb( + &term->curr_truecolour.fg, + term->esc_args + (i+2)); + i += 4; + } + break; + case 48: + if (i+2 < term->esc_nargs && + term->esc_args[i+1] == 5) { + term->curr_attr &= ~ATTR_BGMASK; + term->curr_attr |= + ((term->esc_args[i+2] & 0xFF) + << ATTR_BGSHIFT); + term->curr_truecolour.bg = + optionalrgb_none; + i += 2; + } + if (i + 4 < term->esc_nargs && + term->esc_args[i+1] == 2) { + parse_optionalrgb( + &term->curr_truecolour.bg, + term->esc_args + (i+2)); + i += 4; + } + break; + } + set_erase_char(term); + break; + case 's': /* save cursor */ + save_cursor(term, true); + break; + case 'u': /* restore cursor */ + save_cursor(term, false); + seen_disp_event(term); + break; + case 't': /* DECSLPP: set page size - ie window height */ + /* + * VT340/VT420 sequence DECSLPP, DEC only allows values + * 24/25/36/48/72/144 other emulators (eg dtterm) use + * illegal values (eg first arg 1..9) for window changing + * and reports. + */ + if (term->esc_nargs <= 1 + && (term->esc_args[0] < 1 || + term->esc_args[0] >= 24)) { + compatibility(VT340TEXT); + if (!term->no_remote_resize) + term_request_resize(term, term->cols, 24); + deselect(term); + } else if (term->esc_nargs >= 1 && + term->esc_args[0] >= 1 && + term->esc_args[0] < 24) { + compatibility(OTHER); + + switch (term->esc_args[0]) { + int len; + char buf[80]; + const char *p; + case 1: + term->win_minimise_pending = true; + term->win_minimise_enable = false; + term_schedule_update(term); + break; + case 2: + term->win_minimise_pending = true; + term->win_minimise_enable = true; + term_schedule_update(term); + break; + case 3: + if (term->esc_nargs >= 3) { + if (!term->no_remote_resize) { + term->win_move_pending = true; + term->win_move_pending_x = + def(term->esc_args[1], 0); + term->win_move_pending_y = + def(term->esc_args[2], 0); + term_schedule_update(term); + } + } + break; + case 4: + /* We should resize the window to a given + * size in pixels here, but currently our + * resizing code isn't healthy enough to + * manage it. */ + break; + case 5: + /* move to top */ + term->win_zorder_pending = true; + term->win_zorder_top = true; + term_schedule_update(term); + break; + case 6: + /* move to bottom */ + term->win_zorder_pending = true; + term->win_zorder_top = false; + term_schedule_update(term); + break; + case 7: + term->win_refresh_pending = true; + term_schedule_update(term); + break; + case 8: + if (term->esc_nargs >= 3 && + !term->no_remote_resize) { + term_request_resize( + term, + def(term->esc_args[2], + term->conf_width), + def(term->esc_args[1], + term->conf_height)); + } + break; + case 9: + if (term->esc_nargs >= 2) { + term->win_maximise_pending = true; + term->win_maximise_enable = + term->esc_args[1]; + term_schedule_update(term); + } + break; + case 11: + if (term->ldisc) + ldisc_send(term->ldisc, term->minimised ? + "\033[2t" : "\033[1t", 4, + false); + break; + case 13: + if (term->ldisc) { + len = sprintf(buf, "\033[3;%u;%ut", + term->winpos_x, + term->winpos_y); + ldisc_send(term->ldisc, buf, len, false); + } + break; + case 14: + if (term->ldisc) { + len = sprintf(buf, "\033[4;%u;%ut", + term->winpixsize_y, + term->winpixsize_x); + ldisc_send(term->ldisc, buf, len, false); + } + break; + case 18: + if (term->ldisc) { + len = sprintf(buf, "\033[8;%d;%dt", + term->rows, term->cols); + ldisc_send(term->ldisc, buf, len, false); + } + break; + case 19: + /* + * Hmmm. Strictly speaking we + * should return `the size of the + * screen in characters', but + * that's not easy: (a) window + * furniture being what it is it's + * hard to compute, and (b) in + * resize-font mode maximising the + * window wouldn't change the + * number of characters. *shrug*. I + * think we'll ignore it for the + * moment and see if anyone + * complains, and then ask them + * what they would like it to do. + */ + break; + case 20: + if (term->ldisc && + term->remote_qtitle_action != TITLE_NONE) { + if(term->remote_qtitle_action == TITLE_REAL) + p = term->icon_title; + else + p = EMPTY_WINDOW_TITLE; + len = strlen(p); + ldisc_send(term->ldisc, "\033]L", 3, + false); + ldisc_send(term->ldisc, p, len, false); + ldisc_send(term->ldisc, "\033\\", 2, + false); + } + break; + case 21: + if (term->ldisc && + term->remote_qtitle_action != TITLE_NONE) { + if(term->remote_qtitle_action == TITLE_REAL) + p = term->window_title; + else + p = EMPTY_WINDOW_TITLE; + len = strlen(p); + ldisc_send(term->ldisc, "\033]l", 3, + false); + ldisc_send(term->ldisc, p, len, false); + ldisc_send(term->ldisc, "\033\\", 2, + false); + } + break; + } + } + break; + case 'S': /* SU: Scroll up */ + CLAMP(term->esc_args[0], term->rows); + compatibility(SCOANSI); + scroll(term, term->marg_t, term->marg_b, + def(term->esc_args[0], 1), true); + term->wrapnext = false; + seen_disp_event(term); + break; + case 'T': /* SD: Scroll down */ + CLAMP(term->esc_args[0], term->rows); + compatibility(SCOANSI); + scroll(term, term->marg_t, term->marg_b, + -def(term->esc_args[0], 1), true); + term->wrapnext = false; + seen_disp_event(term); + break; + case ANSI('|', '*'): /* DECSNLS */ + /* + * Set number of lines on screen + * VT420 uses VGA like hardware and can + * support any size in reasonable range + * (24..49 AIUI) with no default specified. + */ + compatibility(VT420); + if (term->esc_nargs == 1 && term->esc_args[0] > 0) { + if (!term->no_remote_resize) + term_request_resize( + term, + term->cols, + def(term->esc_args[0], term->conf_height)); + deselect(term); + } + break; + case ANSI('|', '$'): /* DECSCPP */ + /* + * Set number of columns per page + * Docs imply range is only 80 or 132, but + * I'll allow any. + */ + compatibility(VT340TEXT); + if (term->esc_nargs <= 1) { + if (!term->no_remote_resize) + term_request_resize( + term, + def(term->esc_args[0], term->conf_width), + term->rows); + deselect(term); + } + break; + case 'X': { /* ECH: write N spaces w/o moving cursor */ + /* XXX VTTEST says this is vt220, vt510 manual + * says vt100 */ + compatibility(ANSIMIN); + CLAMP(term->esc_args[0], term->cols); + int n = def(term->esc_args[0], 1); + pos cursplus; + int p = term->curs.x; + termline *cline = scrlineptr(term->curs.y); + + check_trust_status(term, cline); + if (n > term->cols - term->curs.x) + n = term->cols - term->curs.x; + cursplus = term->curs; + cursplus.x += n; + check_boundary(term, term->curs.x, term->curs.y); + check_boundary(term, term->curs.x+n, term->curs.y); + check_selection(term, term->curs, cursplus); + while (n--) + copy_termchar(cline, p++, + &term->erase_char); + seen_disp_event(term); + break; + } + case 'x': /* DECREQTPARM: report terminal characteristics */ + compatibility(VT100); + if (term->ldisc) { + char buf[32]; + int i = def(term->esc_args[0], 0); + if (i == 0 || i == 1) { + strcpy(buf, "\033[2;1;1;112;112;1;0x"); + buf[2] += i; + ldisc_send(term->ldisc, buf, 20, false); + } + } + break; + case 'Z': { /* CBT */ + compatibility(OTHER); + CLAMP(term->esc_args[0], term->cols); + int i = def(term->esc_args[0], 1); + pos old_curs = term->curs; + + for(;i>0 && term->curs.x>0; i--) { + do { + term->curs.x--; + } while (term->curs.x >0 && + !term->tabs[term->curs.x]); + } + check_selection(term, old_curs, term->curs); + break; + } + case ANSI('c', '='): /* Hide or Show Cursor */ + compatibility(SCOANSI); + switch(term->esc_args[0]) { + case 0: /* hide cursor */ + term->cursor_on = false; + break; + case 1: /* restore cursor */ + term->big_cursor = false; + term->cursor_on = true; + break; + case 2: /* block cursor */ + term->big_cursor = true; + term->cursor_on = true; + break; + } + break; + case ANSI('C', '='): + /* + * set cursor start on scanline esc_args[0] and + * end on scanline esc_args[1].If you set + * the bottom scan line to a value less than + * the top scan line, the cursor will disappear. + */ + compatibility(SCOANSI); + if (term->esc_nargs >= 2) { + if (term->esc_args[0] > term->esc_args[1]) + term->cursor_on = false; + else + term->cursor_on = true; + } + break; + case ANSI('D', '='): + compatibility(SCOANSI); + term->blink_is_real = false; + term_schedule_tblink(term); + if (term->esc_args[0]>=1) + term->curr_attr |= ATTR_BLINK; + else + term->curr_attr &= ~ATTR_BLINK; + break; + case ANSI('E', '='): + compatibility(SCOANSI); + term->blink_is_real = (term->esc_args[0] >= 1); + term_schedule_tblink(term); + break; + case ANSI('F', '='): /* set normal foreground */ + compatibility(SCOANSI); + if (term->esc_args[0] < 16) { + long colour = + (sco2ansicolour[term->esc_args[0] & 0x7] | + (term->esc_args[0] & 0x8)) << + ATTR_FGSHIFT; + term->curr_attr &= ~ATTR_FGMASK; + term->curr_attr |= colour; + term->curr_truecolour.fg = optionalrgb_none; + term->default_attr &= ~ATTR_FGMASK; + term->default_attr |= colour; + set_erase_char(term); + } + break; + case ANSI('G', '='): /* set normal background */ + compatibility(SCOANSI); + if (term->esc_args[0] < 16) { + long colour = + (sco2ansicolour[term->esc_args[0] & 0x7] | + (term->esc_args[0] & 0x8)) << + ATTR_BGSHIFT; + term->curr_attr &= ~ATTR_BGMASK; + term->curr_attr |= colour; + term->curr_truecolour.bg = optionalrgb_none; + term->default_attr &= ~ATTR_BGMASK; + term->default_attr |= colour; + set_erase_char(term); + } + break; + case ANSI('L', '='): + compatibility(SCOANSI); + term->use_bce = (term->esc_args[0] <= 0); + set_erase_char(term); + break; + case ANSI('p', '"'): /* DECSCL: set compat level */ + /* + * Allow the host to make this emulator a + * 'perfect' VT102. This first appeared in + * the VT220, but we do need to get back to + * PuTTY mode so I won't check it. + * + * The arg in 40..42,50 are a PuTTY extension. + * The 2nd arg, 8bit vs 7bit is not checked. + * + * Setting VT102 mode should also change + * the Fkeys to generate PF* codes as a + * real VT102 has no Fkeys. The VT220 does + * this, F11..F13 become ESC,BS,LF other + * Fkeys send nothing. + * + * Note ESC c will NOT change this! + */ + + switch (term->esc_args[0]) { + case 61: + term->compatibility_level &= ~TM_VTXXX; + term->compatibility_level |= TM_VT102; + break; + case 62: + term->compatibility_level &= ~TM_VTXXX; + term->compatibility_level |= TM_VT220; + break; + + default: + if (term->esc_args[0] > 60 && + term->esc_args[0] < 70) + term->compatibility_level |= TM_VTXXX; + break; + + case 40: + term->compatibility_level &= TM_VTXXX; + break; + case 41: + term->compatibility_level = TM_PUTTY; + break; + case 42: + term->compatibility_level = TM_SCOANSI; + break; + + case ARG_DEFAULT: + term->compatibility_level = TM_PUTTY; + break; + case 50: + break; + } + + /* Change the response to CSI c */ + if (term->esc_args[0] == 50) { + int i; + char lbuf[64]; + strcpy(term->id_string, "\033[?"); + for (i = 1; i < term->esc_nargs; i++) { + if (i != 1) + strcat(term->id_string, ";"); + sprintf(lbuf, "%u", term->esc_args[i]); + strcat(term->id_string, lbuf); + } + strcat(term->id_string, "c"); + } +#if 0 + /* Is this a good idea ? + * Well we should do a soft reset at this point ... + */ + if (!has_compat(VT420) && has_compat(VT100)) { + if (!term->no_remote_resize) + term_request_resize(term, + term->reset_132 ? 132 : 80, + 24); + } +#endif + break; + } + break; + case SEEN_OSC: + term->osc_w = false; + switch (c) { + case 'P': /* Linux palette sequence */ + term->termstate = SEEN_OSC_P; + term->osc_strlen = 0; + break; + case 'R': /* Linux palette reset */ + palette_reset(term, false); + term_invalidate(term); + term->termstate = TOPLEVEL; + break; + case 'W': /* word-set */ + term->termstate = SEEN_OSC_W; + term->osc_w = true; + break; + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (term->esc_args[term->esc_nargs-1] <= UINT_MAX / 10 && + term->esc_args[term->esc_nargs-1] * 10 <= UINT_MAX - c - '0') + term->esc_args[term->esc_nargs-1] = + 10 * term->esc_args[term->esc_nargs-1] + c - '0'; + else + term->esc_args[term->esc_nargs-1] = UINT_MAX; + break; + case 0x9C: + /* Terminate even though we aren't in OSC_STRING yet */ + do_osc(term); + term->termstate = TOPLEVEL; + break; + case 0xC2: + if (in_utf(term)) { + /* Or be prepared for the UTF-8 version of that */ + term->termstate = OSC_MAYBE_ST_UTF8; + } + break; + default: + /* + * _Most_ other characters here terminate the + * immediate parsing of the OSC sequence and go + * into OSC_STRING state, but we deal with a + * couple of exceptions first. + */ + if (c == 'L' && term->esc_args[0] == 2) { + /* + * Grotty hack to support xterm and DECterm title + * sequences concurrently. + */ + term->esc_args[0] = 1; + } else if (c == ';' && term->esc_nargs == 1 && + term->esc_args[0] == 4) { + /* + * xterm's OSC 4 sequence to query the current + * RGB value of a colour takes a second + * numeric argument which is easiest to parse + * using the existing system rather than in + * do_osc. + */ + term->esc_args[term->esc_nargs++] = 0; + } else { + term->termstate = OSC_STRING; + term->osc_strlen = 0; + } + } + break; + case OSC_STRING: + /* + * OSC sequences can be terminated or aborted in + * various ways. + * + * The official way to terminate an OSC, per written + * standards, is the String Terminator, SC. That can + * appear in a 7-bit two-character form ESC \, or as + * an 8-bit C1 control 0x9C. + * + * We only accept 0x9C in circumstances where it + * doesn't interfere with our main character set + * processing: so in ISO 8859-1, for example, the byte + * 0x9C is interpreted as ST, but in CP437 it's + * interpreted as an ordinary printing character (as + * it happens, the pound sign), because you might + * perfectly well want to put it in the window title + * like any other printing character. + * + * In particular, in UTF-8 mode, 0x9C is a perfectly + * valid continuation byte for an ordinary printing + * character, so we don't accept the C1 control form + * of ST unless it appears as a full UTF-8 character + * in its own right, i.e. bytes 0xC2 0x9C. + * + * BEL is also treated as a clean termination of OSC, + * which I believe was a behaviour introduced by + * xterm. + * + * To prevent run-on storage of OSC data forever if + * emission of a control sequence is interrupted, we + * also treat various control characters as illegal, + * so that they abort the OSC without processing it + * and return to TOPLEVEL state. These are CR, LF, and + * any ESC that is *not* followed by \. + */ + + if (c == '\012' || c == '\015') { + /* CR or LF aborts */ + term->termstate = TOPLEVEL; + break; + } + + if (c == '\033') { + /* ESC goes into a state where we wait to see if + * the next character is \ */ + term->termstate = OSC_MAYBE_ST; + break; + } + + if (c == '\007' || (c == 0x9C && !in_utf(term) && + term->ucsdata->unitab_ctrl[c] != 0xFF)) { + /* BEL, or the C1 ST appearing as a one-byte + * encoding, cleanly terminates the OSC right here */ + do_osc(term); + term->termstate = TOPLEVEL; + break; + } + + if (c == 0xC2 && in_utf(term)) { + /* 0xC2 is the UTF-8 character that might + * introduce the encoding of C1 ST */ + term->termstate = OSC_MAYBE_ST_UTF8; + break; + } + + /* Anything else gets added to the string */ + if (term->osc_strlen < OSC_STR_MAX) + term->osc_string[term->osc_strlen++] = (char)c; + break; + case OSC_MAYBE_ST_UTF8: + /* In UTF-8 mode, we've seen C2, so are we now seeing + * 9C? */ + if (c == 0x9C) { + /* Yes, so cleanly terminate the OSC */ + do_osc(term); + term->termstate = TOPLEVEL; + break; + } + /* No, so append the pending C2 byte to the OSC string + * followed by the current character, and go back to + * OSC string accumulation */ + if (term->osc_strlen < OSC_STR_MAX) + term->osc_string[term->osc_strlen++] = 0xC2; + if (term->osc_strlen < OSC_STR_MAX) + term->osc_string[term->osc_strlen++] = (char)c; + term->termstate = OSC_STRING; + break; + case SEEN_OSC_P: { + int max = (term->osc_strlen == 0 ? 21 : 15); + int val; + if ((int)c >= '0' && (int)c <= '9') + val = c - '0'; + else if ((int)c >= 'A' && (int)c <= 'A' + max - 10) + val = c - 'A' + 10; + else if ((int)c >= 'a' && (int)c <= 'a' + max - 10) + val = c - 'a' + 10; + else { + term->termstate = TOPLEVEL; + break; + } + term->osc_string[term->osc_strlen++] = val; + if (term->osc_strlen >= 7) { + unsigned oscp_index = term->osc_string[0]; + assert(oscp_index < OSCP_NCOLOURS); + unsigned osc4_index = + colour_indices_oscp_to_osc4[oscp_index]; + + rgb *value = &term->subpalettes[SUBPAL_SESSION].values[ + osc4_index]; + value->r = term->osc_string[1] * 16 + term->osc_string[2]; + value->g = term->osc_string[3] * 16 + term->osc_string[4]; + value->b = term->osc_string[5] * 16 + term->osc_string[6]; + term->subpalettes[SUBPAL_SESSION].present[ + osc4_index] = true; + + palette_rebuild(term); + + term->termstate = TOPLEVEL; + } + break; + } + case SEEN_OSC_W: + switch (c) { + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (term->esc_args[0] <= UINT_MAX / 10 && + term->esc_args[0] * 10 <= UINT_MAX - c - '0') + term->esc_args[0] = 10 * term->esc_args[0] + c - '0'; + else + term->esc_args[0] = UINT_MAX; + break; + case 0x9C: + /* Terminate even though we aren't in OSC_STRING yet */ + do_osc(term); + term->termstate = TOPLEVEL; + break; + case 0xC2: + if (in_utf(term)) { + /* Or be prepared for the UTF-8 version of that */ + term->termstate = OSC_MAYBE_ST_UTF8; + } + break; + default: + term->termstate = OSC_STRING; + term->osc_strlen = 0; + } + break; + case VT52_ESC: + term->termstate = TOPLEVEL; + seen_disp_event(term); + switch (c) { + case 'A': + move(term, term->curs.x, term->curs.y - 1, 1); + break; + case 'B': + move(term, term->curs.x, term->curs.y + 1, 1); + break; + case 'C': + move(term, term->curs.x + 1, term->curs.y, 1); + break; + case 'D': + move(term, term->curs.x - 1, term->curs.y, 1); + break; + /* + * From the VT100 Manual + * NOTE: The special graphics characters in the VT100 + * are different from those in the VT52 + * + * From VT102 manual: + * 137 _ Blank - Same + * 140 ` Reserved - Humm. + * 141 a Solid rectangle - Similar + * 142 b 1/ - Top half of fraction for the + * 143 c 3/ - subscript numbers below. + * 144 d 5/ + * 145 e 7/ + * 146 f Degrees - Same + * 147 g Plus or minus - Same + * 150 h Right arrow + * 151 i Ellipsis (dots) + * 152 j Divide by + * 153 k Down arrow + * 154 l Bar at scan 0 + * 155 m Bar at scan 1 + * 156 n Bar at scan 2 + * 157 o Bar at scan 3 - Similar + * 160 p Bar at scan 4 - Similar + * 161 q Bar at scan 5 - Similar + * 162 r Bar at scan 6 - Same + * 163 s Bar at scan 7 - Similar + * 164 t Subscript 0 + * 165 u Subscript 1 + * 166 v Subscript 2 + * 167 w Subscript 3 + * 170 x Subscript 4 + * 171 y Subscript 5 + * 172 z Subscript 6 + * 173 { Subscript 7 + * 174 | Subscript 8 + * 175 } Subscript 9 + * 176 ~ Paragraph + * + */ + case 'F': + term->cset_attr[term->cset = 0] = CSET_LINEDRW; + break; + case 'G': + term->cset_attr[term->cset = 0] = CSET_ASCII; + break; + case 'H': + move(term, 0, 0, 0); + break; + case 'I': + if (term->curs.y == 0) + scroll(term, 0, term->rows - 1, -1, true); + else if (term->curs.y > 0) + term->curs.y--; + term->wrapnext = false; + break; + case 'J': + erase_lots(term, false, false, true); + if (term->scroll_on_disp) + term->disptop = 0; + break; + case 'K': + erase_lots(term, true, false, true); + break; +#if 0 + case 'V': + /* XXX Print cursor line */ + break; + case 'W': + /* XXX Start controller mode */ + break; + case 'X': + /* XXX Stop controller mode */ + break; +#endif + case 'Y': + term->termstate = VT52_Y1; + break; + case 'Z': + if (term->ldisc) + ldisc_send(term->ldisc, "\033/Z", 3, false); + break; + case '=': + term->app_keypad_keys = true; + break; + case '>': + term->app_keypad_keys = false; + break; + case '<': + /* XXX This should switch to VT100 mode not current or default + * VT mode. But this will only have effect in a VT220+ + * emulation. + */ + term->vt52_mode = false; + term->blink_is_real = term->blinktext; + term_schedule_tblink(term); + break; +#if 0 + case '^': + /* XXX Enter auto print mode */ + break; + case '_': + /* XXX Exit auto print mode */ + break; + case ']': + /* XXX Print screen */ + break; +#endif + +#ifdef VT52_PLUS + case 'E': + /* compatibility(ATARI) */ + move(term, 0, 0, 0); + erase_lots(term, false, false, true); + if (term->scroll_on_disp) + term->disptop = 0; + break; + case 'L': + /* compatibility(ATARI) */ + if (term->curs.y <= term->marg_b) + scroll(term, term->curs.y, term->marg_b, -1, false); + break; + case 'M': + /* compatibility(ATARI) */ + if (term->curs.y <= term->marg_b) + scroll(term, term->curs.y, term->marg_b, 1, true); + break; + case 'b': + /* compatibility(ATARI) */ + term->termstate = VT52_FG; + break; + case 'c': + /* compatibility(ATARI) */ + term->termstate = VT52_BG; + break; + case 'd': + /* compatibility(ATARI) */ + erase_lots(term, false, true, false); + if (term->scroll_on_disp) + term->disptop = 0; + break; + case 'e': + /* compatibility(ATARI) */ + term->cursor_on = true; + break; + case 'f': + /* compatibility(ATARI) */ + term->cursor_on = false; + break; + /* case 'j': Save cursor position - broken on ST */ + /* case 'k': Restore cursor position */ + case 'l': + /* compatibility(ATARI) */ + erase_lots(term, true, true, true); + term->curs.x = 0; + term->wrapnext = false; + break; + case 'o': + /* compatibility(ATARI) */ + erase_lots(term, true, true, false); + break; + case 'p': + /* compatibility(ATARI) */ + term->curr_attr |= ATTR_REVERSE; + break; + case 'q': + /* compatibility(ATARI) */ + term->curr_attr &= ~ATTR_REVERSE; + break; + case 'v': /* wrap Autowrap on - Wyse style */ + /* compatibility(ATARI) */ + term->wrap = true; + break; + case 'w': /* Autowrap off */ + /* compatibility(ATARI) */ + term->wrap = false; + break; + + case 'R': + /* compatibility(OTHER) */ + term->vt52_bold = false; + term->curr_attr = ATTR_DEFAULT; + term->curr_truecolour.fg = optionalrgb_none; + term->curr_truecolour.bg = optionalrgb_none; + set_erase_char(term); + break; + case 'S': + /* compatibility(VI50) */ + term->curr_attr |= ATTR_UNDER; + break; + case 'W': + /* compatibility(VI50) */ + term->curr_attr &= ~ATTR_UNDER; + break; + case 'U': + /* compatibility(VI50) */ + term->vt52_bold = true; + term->curr_attr |= ATTR_BOLD; + break; + case 'T': + /* compatibility(VI50) */ + term->vt52_bold = false; + term->curr_attr &= ~ATTR_BOLD; + break; +#endif + } + break; + case VT52_Y1: + term->termstate = VT52_Y2; + move(term, term->curs.x, c - ' ', 0); + break; + case VT52_Y2: + term->termstate = TOPLEVEL; + move(term, c - ' ', term->curs.y, 0); + break; + +#ifdef VT52_PLUS + case VT52_FG: + term->termstate = TOPLEVEL; + term->curr_attr &= ~ATTR_FGMASK; + term->curr_attr &= ~ATTR_BOLD; + term->curr_attr |= (c & 0xF) << ATTR_FGSHIFT; + set_erase_char(term); + break; + case VT52_BG: + term->termstate = TOPLEVEL; + term->curr_attr &= ~ATTR_BGMASK; + term->curr_attr &= ~ATTR_BLINK; + term->curr_attr |= (c & 0xF) << ATTR_BGSHIFT; + set_erase_char(term); + break; +#endif + default: break; /* placate gcc warning about enum use */ + } + if (term->selstate != NO_SELECTION) { + pos cursplus = term->curs; + incpos(cursplus); + check_selection(term, term->curs, cursplus); + } + } + + bufchain_consume(&term->inbuf, nchars_used); + + if (!called_from_term_data) + win_unthrottle(term->win, bufchain_size(&term->inbuf)); + + term_print_flush(term); + if (term->logflush && term->logctx) + logflush(term->logctx); +} + +/* Wrapper on term_out with the right prototype to be a toplevel callback */ +void term_out_cb(void *ctx) +{ + term_out((Terminal *)ctx, false); +} + +/* + * Small subroutine to parse three consecutive escape-sequence + * arguments representing a true-colour RGB triple into an + * optionalrgb. + */ +static void parse_optionalrgb(optionalrgb *out, unsigned *values) +{ + out->enabled = true; + out->r = values[0] < 256 ? values[0] : 0; + out->g = values[1] < 256 ? values[1] : 0; + out->b = values[2] < 256 ? values[2] : 0; +} + +/* + * To prevent having to run the reasonably tricky bidi algorithm + * too many times, we maintain a cache of the last lineful of data + * fed to the algorithm on each line of the display. + */ +static bool term_bidi_cache_hit(Terminal *term, int line, + termchar *lbefore, int width, bool trusted) +{ + int i; + + if (!term->pre_bidi_cache) + return false; /* cache doesn't even exist yet! */ + + if (line >= term->bidi_cache_size) + return false; /* cache doesn't have this many lines */ + + if (!term->pre_bidi_cache[line].chars) + return false; /* cache doesn't contain _this_ line */ + + if (term->pre_bidi_cache[line].width != width) + return false; /* line is wrong width */ + + if (term->pre_bidi_cache[line].trusted != trusted) + return false; /* line has wrong trust state */ + + for (i = 0; i < width; i++) + if (!termchars_equal(term->pre_bidi_cache[line].chars+i, lbefore+i)) + return false; /* line doesn't match cache */ + + return true; /* it didn't match. */ +} + +static void term_bidi_cache_store(Terminal *term, int line, termchar *lbefore, + termchar *lafter, bidi_char *wcTo, + int width, int size, bool trusted) +{ + size_t i, j; + + if (!term->pre_bidi_cache || term->bidi_cache_size <= line) { + j = term->bidi_cache_size; + sgrowarray(term->pre_bidi_cache, term->bidi_cache_size, line); + term->post_bidi_cache = sresize(term->post_bidi_cache, + term->bidi_cache_size, + struct bidi_cache_entry); + while (j < term->bidi_cache_size) { + term->pre_bidi_cache[j].chars = + term->post_bidi_cache[j].chars = NULL; + term->pre_bidi_cache[j].width = + term->post_bidi_cache[j].width = -1; + term->pre_bidi_cache[j].trusted = false; + term->post_bidi_cache[j].trusted = false; + term->pre_bidi_cache[j].forward = + term->post_bidi_cache[j].forward = NULL; + term->pre_bidi_cache[j].backward = + term->post_bidi_cache[j].backward = NULL; + j++; + } + } + + sfree(term->pre_bidi_cache[line].chars); + sfree(term->post_bidi_cache[line].chars); + sfree(term->post_bidi_cache[line].forward); + sfree(term->post_bidi_cache[line].backward); + + term->pre_bidi_cache[line].width = width; + term->pre_bidi_cache[line].trusted = trusted; + term->pre_bidi_cache[line].chars = snewn(size, termchar); + term->post_bidi_cache[line].width = width; + term->post_bidi_cache[line].trusted = trusted; + term->post_bidi_cache[line].chars = snewn(size, termchar); + term->post_bidi_cache[line].forward = snewn(width, int); + term->post_bidi_cache[line].backward = snewn(width, int); + + memcpy(term->pre_bidi_cache[line].chars, lbefore, size * TSIZE); + memcpy(term->post_bidi_cache[line].chars, lafter, size * TSIZE); + memset(term->post_bidi_cache[line].forward, 0, width * sizeof(int)); + memset(term->post_bidi_cache[line].backward, 0, width * sizeof(int)); + + for (i = j = 0; j < width; j += wcTo[i].nchars, i++) { + int p = wcTo[i].index; + + if (p != BIDI_CHAR_INDEX_NONE) { + assert(0 <= p && p < width); + + for (int x = 0; x < wcTo[i].nchars; x++) { + term->post_bidi_cache[line].backward[j+x] = p+x; + term->post_bidi_cache[line].forward[p+x] = j+x; + } + } + } +} + +/* + * Prepare the bidi information for a screen line. Returns the + * transformed list of termchars, or NULL if no transformation at + * all took place (because bidi is disabled). If return was + * non-NULL, auxiliary information such as the forward and reverse + * mappings of permutation position are available in + * term->post_bidi_cache[scr_y].*. + */ +static termchar *term_bidi_line(Terminal *term, struct termline *ldata, + int scr_y) +{ + termchar *lchars; + int it; + + /* Do Arabic shaping and bidi. */ + if (!term->no_bidi || !term->no_arabicshaping || + (ldata->trusted && term->cols > TRUST_SIGIL_WIDTH)) { + + if (!term_bidi_cache_hit(term, scr_y, ldata->chars, term->cols, + ldata->trusted)) { + + if (term->wcFromTo_size < term->cols) { + term->wcFromTo_size = term->cols; + term->wcFrom = sresize(term->wcFrom, term->wcFromTo_size, + bidi_char); + term->wcTo = sresize(term->wcTo, term->wcFromTo_size, + bidi_char); + } + + for(it=0; it<term->cols ; it++) + { + unsigned long uc = (ldata->chars[it].chr); + + switch (uc & CSET_MASK) { + case CSET_LINEDRW: + if (!term->rawcnp) { + uc = term->ucsdata->unitab_xterm[uc & 0xFF]; + break; + } + case CSET_ASCII: + uc = term->ucsdata->unitab_line[uc & 0xFF]; + break; + case CSET_SCOACS: + uc = term->ucsdata->unitab_scoacs[uc&0xFF]; + break; + } + switch (uc & CSET_MASK) { + case CSET_ACP: + uc = term->ucsdata->unitab_font[uc & 0xFF]; + break; + case CSET_OEMCP: + uc = term->ucsdata->unitab_oemcp[uc & 0xFF]; + break; + } + + term->wcFrom[it].origwc = term->wcFrom[it].wc = + (unsigned int)uc; + term->wcFrom[it].index = it; + term->wcFrom[it].nchars = 1; + } + + if (ldata->trusted && term->cols > TRUST_SIGIL_WIDTH) { + memmove( + term->wcFrom + TRUST_SIGIL_WIDTH, term->wcFrom, + (term->cols - TRUST_SIGIL_WIDTH) * sizeof(*term->wcFrom)); + for (it = 0; it < TRUST_SIGIL_WIDTH; it++) { + term->wcFrom[it].origwc = term->wcFrom[it].wc = + (it == 0 ? TRUST_SIGIL_CHAR : + it == 1 ? UCSWIDE : ' '); + term->wcFrom[it].index = BIDI_CHAR_INDEX_NONE; + term->wcFrom[it].nchars = 1; + } + } + + int nbc = 0; + for (it = 0; it < term->cols; it++) { + term->wcFrom[nbc] = term->wcFrom[it]; + if (it+1 < term->cols && term->wcFrom[it+1].wc == UCSWIDE) { + term->wcFrom[nbc].nchars++; + it++; + } + nbc++; + } + + if(!term->no_bidi) + do_bidi(term->bidi_ctx, term->wcFrom, nbc); + + if(!term->no_arabicshaping) { + do_shape(term->wcFrom, term->wcTo, nbc); + } else { + /* If we're not calling do_shape, we must copy the + * data into wcTo anyway, unchanged */ + memcpy(term->wcTo, term->wcFrom, nbc * sizeof(*term->wcTo)); + } + + if (term->ltemp_size < ldata->size) { + term->ltemp_size = ldata->size; + term->ltemp = sresize(term->ltemp, term->ltemp_size, + termchar); + } + + memcpy(term->ltemp, ldata->chars, ldata->size * TSIZE); + + int opos = 0; + for (it=0; it<nbc; it++) { + int ipos = term->wcTo[it].index; + for (int j = 0; j < term->wcTo[it].nchars; j++) { + if (ipos != BIDI_CHAR_INDEX_NONE) { + term->ltemp[opos] = ldata->chars[ipos]; + if (term->ltemp[opos].cc_next) + term->ltemp[opos].cc_next -= opos - ipos; + + if (j > 0) + term->ltemp[opos].chr = UCSWIDE; + else if (term->wcTo[it].origwc != term->wcTo[it].wc) + term->ltemp[opos].chr = term->wcTo[it].wc; + } else { + term->ltemp[opos] = term->basic_erase_char; + term->ltemp[opos].chr = + j > 0 ? UCSWIDE : term->wcTo[it].origwc; + } + opos++; + } + } + assert(opos == term->cols); + term_bidi_cache_store(term, scr_y, ldata->chars, + term->ltemp, term->wcTo, + term->cols, ldata->size, ldata->trusted); + + lchars = term->ltemp; + } else { + lchars = term->post_bidi_cache[scr_y].chars; + } + } else { + lchars = NULL; + } + + return lchars; +} + +static void do_paint_draw(Terminal *term, termline *ldata, int x, int y, + wchar_t *ch, int ccount, + unsigned long attr, truecolour tc) +{ + if (ch[0] == TRUST_SIGIL_CHAR) { + assert(ldata->trusted); + assert(ccount == 1); + assert(attr & ATTR_WIDE); + wchar_t tch[2]; + tch[0] = tch[1] = L' '; + win_draw_text(term->win, x, y, tch, 2, term->basic_erase_char.attr, + ldata->lattr, term->basic_erase_char.truecolour); + win_draw_trust_sigil(term->win, x, y); + } else { + win_draw_text(term->win, x, y, ch, ccount, attr, ldata->lattr, tc); + if (attr & (TATTR_ACTCURS | TATTR_PASCURS)) + win_draw_cursor(term->win, x, y, ch, ccount, + attr, ldata->lattr, tc); + } +} + +/* + * Given a context, update the window. + */ +static void do_paint(Terminal *term) +{ + int i, j, our_curs_y, our_curs_x; + int rv, cursor; + pos scrpos; + wchar_t *ch; + size_t chlen; + termchar *newline; + + chlen = 1024; + ch = snewn(chlen, wchar_t); + + newline = snewn(term->cols, termchar); + + rv = (!term->rvideo ^ !term->in_vbell ? ATTR_REVERSE : 0); + + /* Depends on: + * screen array, disptop, scrtop, + * selection, rv, + * blinkpc, blink_is_real, tblinker, + * curs.y, curs.x, cblinker, blink_cur, cursor_on, has_focus, wrapnext + */ + + /* Has the cursor position or type changed ? */ + if (term->cursor_on) { + if (term->has_focus) { + if (term->cblinker || !term->blink_cur) + cursor = TATTR_ACTCURS; + else + cursor = 0; + } else + cursor = TATTR_PASCURS; + if (term->wrapnext) + cursor |= TATTR_RIGHTCURS; + } else + cursor = 0; + our_curs_y = term->curs.y - term->disptop; + { + /* + * Adjust the cursor position: + * - for bidi + * - in the case where it's resting on the right-hand half + * of a CJK wide character. xterm's behaviour here, + * which seems adequate to me, is to display the cursor + * covering the _whole_ character, exactly as if it were + * one space to the left. + */ + termline *ldata = lineptr(term->curs.y); + termchar *lchars; + + our_curs_x = term->curs.x; + + if ( (lchars = term_bidi_line(term, ldata, our_curs_y)) != NULL) { + our_curs_x = term->post_bidi_cache[our_curs_y].forward[our_curs_x]; + } else + lchars = ldata->chars; + + if (our_curs_x > 0 && + lchars[our_curs_x].chr == UCSWIDE) + our_curs_x--; + + unlineptr(ldata); + } + + /* + * If the cursor is not where it was last time we painted, and + * its previous position is visible on screen, invalidate its + * previous position. + */ + if (term->dispcursy >= 0 && + (term->curstype != cursor || + term->dispcursy != our_curs_y || + term->dispcursx != our_curs_x)) { + termchar *dispcurs = term->disptext[term->dispcursy]->chars + + term->dispcursx; + + if (term->dispcursx > 0 && dispcurs->chr == UCSWIDE) + dispcurs[-1].attr |= ATTR_INVALID; + if (term->dispcursx < term->cols-1 && dispcurs[1].chr == UCSWIDE) + dispcurs[1].attr |= ATTR_INVALID; + dispcurs->attr |= ATTR_INVALID; + + term->curstype = 0; + } + term->dispcursx = term->dispcursy = -1; + + /* The normal screen data */ + for (i = 0; i < term->rows; i++) { + termline *ldata; + termchar *lchars; + bool dirty_line, dirty_run, selected; + unsigned long attr = 0, cset = 0; + int start = 0; + int ccount = 0; + bool last_run_dirty = false; + int laststart; + bool dirtyrect; + int *backward; + truecolour tc; + + scrpos.y = i + term->disptop; + ldata = lineptr(scrpos.y); + + /* Do Arabic shaping and bidi. */ + lchars = term_bidi_line(term, ldata, i); + if (lchars) { + backward = term->post_bidi_cache[i].backward; + } else { + lchars = ldata->chars; + backward = NULL; + } + + /* + * First loop: work along the line deciding what we want + * each character cell to look like. + */ + for (j = 0; j < term->cols; j++) { + unsigned long tattr, tchar; + termchar *d = lchars + j; + scrpos.x = backward ? backward[j] : j; + + tchar = d->chr; + tattr = d->attr; + + if (!term->ansi_colour) + tattr = (tattr & ~(ATTR_FGMASK | ATTR_BGMASK)) | + ATTR_DEFFG | ATTR_DEFBG; + + if (!term->xterm_256_colour) { + int colour; + colour = (tattr & ATTR_FGMASK) >> ATTR_FGSHIFT; + if (colour >= 16 && colour < 256) + tattr = (tattr &~ ATTR_FGMASK) | ATTR_DEFFG; + colour = (tattr & ATTR_BGMASK) >> ATTR_BGSHIFT; + if (colour >= 16 && colour < 256) + tattr = (tattr &~ ATTR_BGMASK) | ATTR_DEFBG; + } + + if (term->true_colour) { + tc = d->truecolour; + } else { + tc.fg = tc.bg = optionalrgb_none; + } + + switch (tchar & CSET_MASK) { + case CSET_ASCII: + tchar = term->ucsdata->unitab_line[tchar & 0xFF]; + break; + case CSET_LINEDRW: + tchar = term->ucsdata->unitab_xterm[tchar & 0xFF]; + break; + case CSET_SCOACS: + tchar = term->ucsdata->unitab_scoacs[tchar&0xFF]; + break; + } + if (j < term->cols-1 && d[1].chr == UCSWIDE) + tattr |= ATTR_WIDE; + + /* Video reversing things */ + if (term->selstate == DRAGGING || term->selstate == SELECTED) { + if (term->seltype == LEXICOGRAPHIC) + selected = (posle(term->selstart, scrpos) && + poslt(scrpos, term->selend)); + else + selected = (posPle(term->selstart, scrpos) && + posPle_left(scrpos, term->selend)); + } else + selected = false; + tattr = (tattr ^ rv + ^ (selected ? ATTR_REVERSE : 0)); + + /* 'Real' blinking ? */ + if (term->blink_is_real && (tattr & ATTR_BLINK)) { + if (term->has_focus && term->tblinker) { + tchar = term->ucsdata->unitab_line[(unsigned char)' ']; + } + tattr &= ~ATTR_BLINK; + } + + /* + * Check the font we'll _probably_ be using to see if + * the character is wide when we don't want it to be. + */ + if (tchar != term->disptext[i]->chars[j].chr || + tattr != (term->disptext[i]->chars[j].attr &~ + (ATTR_NARROW | DATTR_MASK))) { + if ((tattr & ATTR_WIDE) == 0 && + win_char_width(term->win, tchar) == 2) + tattr |= ATTR_NARROW; + } else if (term->disptext[i]->chars[j].attr & ATTR_NARROW) + tattr |= ATTR_NARROW; + + if (i == our_curs_y && j == our_curs_x) { + tattr |= cursor; + term->curstype = cursor; + term->dispcursx = j; + term->dispcursy = i; + } + + /* FULL-TERMCHAR */ + newline[j].attr = tattr; + newline[j].chr = tchar; + newline[j].truecolour = tc; + /* Combining characters are still read from lchars */ + newline[j].cc_next = 0; + } + + /* + * Now loop over the line again, noting where things have + * changed. + * + * During this loop, we keep track of where we last saw + * DATTR_STARTRUN. Any mismatch automatically invalidates + * _all_ of the containing run that was last printed: that + * is, any rectangle that was drawn in one go in the + * previous update should be either left completely alone + * or overwritten in its entirety. This, along with the + * expectation that front ends clip all text runs to their + * bounding rectangle, should solve any possible problems + * with fonts that overflow their character cells. + */ + laststart = 0; + dirtyrect = false; + for (j = 0; j < term->cols; j++) { + if (term->disptext[i]->chars[j].attr & DATTR_STARTRUN) { + laststart = j; + dirtyrect = false; + } + + if (term->disptext[i]->chars[j].chr != newline[j].chr || + (term->disptext[i]->chars[j].attr &~ DATTR_MASK) + != newline[j].attr) { + int k; + + if (!dirtyrect) { + for (k = laststart; k < j; k++) + term->disptext[i]->chars[k].attr |= ATTR_INVALID; + + dirtyrect = true; + } + } + + if (dirtyrect) + term->disptext[i]->chars[j].attr |= ATTR_INVALID; + } + + /* + * Finally, loop once more and actually do the drawing. + */ + dirty_run = dirty_line = (ldata->lattr != + term->disptext[i]->lattr); + term->disptext[i]->lattr = ldata->lattr; + + tc = term->erase_char.truecolour; + for (j = 0; j < term->cols; j++) { + unsigned long tattr, tchar; + bool break_run, do_copy; + termchar *d = lchars + j; + + tattr = newline[j].attr; + tchar = newline[j].chr; + + if ((term->disptext[i]->chars[j].attr ^ tattr) & ATTR_WIDE) + dirty_line = true; + + break_run = ((tattr ^ attr) & term->attr_mask) != 0; + + if (!truecolour_equal(newline[j].truecolour, tc)) + break_run = true; + +#ifdef USES_VTLINE_HACK + /* Special hack for VT100 Linedraw glyphs */ + if ((tchar >= 0x23BA && tchar <= 0x23BD) || + (j > 0 && (newline[j-1].chr >= 0x23BA && + newline[j-1].chr <= 0x23BD))) + break_run = true; +#endif + + /* + * Separate out sequences of characters that have the + * same CSET, if that CSET is a magic one. + */ + if (CSET_OF(tchar) != cset) + break_run = true; + + /* + * Break on both sides of any combined-character cell. + */ + if (d->cc_next != 0 || + (j > 0 && d[-1].cc_next != 0)) + break_run = true; + + /* + * Break on both sides of a trust sigil. + */ + if (d->chr == TRUST_SIGIL_CHAR || + (j >= 2 && d[-1].chr == UCSWIDE && + d[-2].chr == TRUST_SIGIL_CHAR)) + break_run = true; + + if (!term->ucsdata->dbcs_screenfont && !dirty_line) { + if (term->disptext[i]->chars[j].chr == tchar && + (term->disptext[i]->chars[j].attr &~ DATTR_MASK)==tattr && + truecolour_equal( + term->disptext[i]->chars[j].truecolour, tc)) + break_run = true; + else if (!dirty_run && ccount == 1) + break_run = true; + } + + if (break_run) { + if ((dirty_run || last_run_dirty) && ccount > 0) + do_paint_draw(term, ldata, start, i, ch, ccount, attr, tc); + start = j; + ccount = 0; + attr = tattr; + tc = newline[j].truecolour; + cset = CSET_OF(tchar); + if (term->ucsdata->dbcs_screenfont) + last_run_dirty = dirty_run; + dirty_run = dirty_line; + } + + do_copy = false; + if (!termchars_equal_override(&term->disptext[i]->chars[j], + d, tchar, tattr)) { + do_copy = true; + dirty_run = true; + } + + sgrowarrayn(ch, chlen, ccount, 2); + +#ifdef PLATFORM_IS_UTF16 + if (tchar > 0x10000 && tchar < 0x110000) { + ch[ccount++] = (wchar_t) HIGH_SURROGATE_OF(tchar); + ch[ccount++] = (wchar_t) LOW_SURROGATE_OF(tchar); + } else +#endif /* PLATFORM_IS_UTF16 */ + ch[ccount++] = (wchar_t) tchar; + + if (d->cc_next) { + termchar *dd = d; + + while (dd->cc_next) { + unsigned long schar; + + dd += dd->cc_next; + + schar = dd->chr; + switch (schar & CSET_MASK) { + case CSET_ASCII: + schar = term->ucsdata->unitab_line[schar & 0xFF]; + break; + case CSET_LINEDRW: + schar = term->ucsdata->unitab_xterm[schar & 0xFF]; + break; + case CSET_SCOACS: + schar = term->ucsdata->unitab_scoacs[schar&0xFF]; + break; + } + + sgrowarrayn(ch, chlen, ccount, 2); + +#ifdef PLATFORM_IS_UTF16 + if (schar > 0x10000 && schar < 0x110000) { + ch[ccount++] = (wchar_t) HIGH_SURROGATE_OF(schar); + ch[ccount++] = (wchar_t) LOW_SURROGATE_OF(schar); + } else +#endif /* PLATFORM_IS_UTF16 */ + ch[ccount++] = (wchar_t) schar; + } + + attr |= TATTR_COMBINING; + } + + if (do_copy) { + copy_termchar(term->disptext[i], j, d); + term->disptext[i]->chars[j].chr = tchar; + term->disptext[i]->chars[j].attr = tattr; + term->disptext[i]->chars[j].truecolour = tc; + if (start == j) + term->disptext[i]->chars[j].attr |= DATTR_STARTRUN; + } + + /* If it's a wide char step along to the next one. */ + if (tattr & ATTR_WIDE) { + if (++j < term->cols) { + d++; + /* + * By construction above, the cursor should not + * be on the right-hand half of this character. + * Ever. + */ + assert(!(i == our_curs_y && j == our_curs_x)); + if (!termchars_equal(&term->disptext[i]->chars[j], d)) + dirty_run = true; + copy_termchar(term->disptext[i], j, d); + } + } + } + if (dirty_run && ccount > 0) + do_paint_draw(term, ldata, start, i, ch, ccount, attr, tc); + + unlineptr(ldata); + } + + sfree(newline); + sfree(ch); +} + +/* + * Invalidate the whole screen so it will be repainted in full. + */ +void term_invalidate(Terminal *term) +{ + int i, j; + + for (i = 0; i < term->rows; i++) + for (j = 0; j < term->cols; j++) + term->disptext[i]->chars[j].attr |= ATTR_INVALID; + + term_schedule_update(term); +} + +/* + * Paint the window in response to a WM_PAINT message. + */ +void term_paint(Terminal *term, + int left, int top, int right, int bottom, bool immediately) +{ + int i, j; + if (left < 0) left = 0; + if (top < 0) top = 0; + if (right >= term->cols) right = term->cols-1; + if (bottom >= term->rows) bottom = term->rows-1; + + for (i = top; i <= bottom && i < term->rows; i++) { + if ((term->disptext[i]->lattr & LATTR_MODE) == LATTR_NORM) + for (j = left; j <= right && j < term->cols; j++) + term->disptext[i]->chars[j].attr |= ATTR_INVALID; + else + for (j = left / 2; j <= right / 2 + 1 && j < term->cols; j++) + term->disptext[i]->chars[j].attr |= ATTR_INVALID; + } + + if (immediately) { + do_paint(term); + } else { + term_schedule_update(term); + } +} + +/* + * Attempt to scroll the scrollback. The second parameter gives the + * position we want to scroll to; the first is +1 to denote that + * this position is relative to the beginning of the scrollback, -1 + * to denote it is relative to the end, and 0 to denote that it is + * relative to the current position. + */ +void term_scroll(Terminal *term, int rel, int where) +{ + int sbtop = -sblines(term); + + term->disptop = (rel < 0 ? 0 : rel > 0 ? sbtop : term->disptop) + where; + if (term->disptop < sbtop) + term->disptop = sbtop; + if (term->disptop > 0) + term->disptop = 0; + term->win_scrollbar_update_pending = true; + term_schedule_update(term); +} + +/* + * Scroll the scrollback to centre it on the beginning or end of the + * current selection, if any. + */ +void term_scroll_to_selection(Terminal *term, int which_end) +{ + pos target; + int y; + int sbtop = -sblines(term); + + if (term->selstate != SELECTED) + return; + if (which_end) + target = term->selend; + else + target = term->selstart; + + y = target.y - term->rows/2; + if (y < sbtop) + y = sbtop; + else if (y > 0) + y = 0; + term_scroll(term, -1, y); +} + +/* + * Helper routine for clipme(): growing buffer. + */ +typedef struct { + size_t bufsize; /* amount of allocated space in textbuf/attrbuf */ + size_t bufpos; /* amount of actual data */ + wchar_t *textbuf; /* buffer for copied text */ + wchar_t *textptr; /* = textbuf + bufpos (current insertion point) */ + int *attrbuf; /* buffer for copied attributes */ + int *attrptr; /* = attrbuf + bufpos */ + truecolour *tcbuf; /* buffer for copied colours */ + truecolour *tcptr; /* = tcbuf + bufpos */ +} clip_workbuf; + +static void clip_addchar(clip_workbuf *b, wchar_t chr, int attr, truecolour tc) +{ + if (b->bufpos >= b->bufsize) { + sgrowarray(b->textbuf, b->bufsize, b->bufpos); + b->textptr = b->textbuf + b->bufpos; + b->attrbuf = sresize(b->attrbuf, b->bufsize, int); + b->attrptr = b->attrbuf + b->bufpos; + b->tcbuf = sresize(b->tcbuf, b->bufsize, truecolour); + b->tcptr = b->tcbuf + b->bufpos; + } + *b->textptr++ = chr; + *b->attrptr++ = attr; + *b->tcptr++ = tc; + b->bufpos++; +} + +static void clipme(Terminal *term, pos top, pos bottom, bool rect, bool desel, + const int *clipboards, int n_clipboards) +{ + clip_workbuf buf; + int old_top_x; + int attr; + truecolour tc; + + buf.bufsize = 5120; + buf.bufpos = 0; + buf.textptr = buf.textbuf = snewn(buf.bufsize, wchar_t); + buf.attrptr = buf.attrbuf = snewn(buf.bufsize, int); + buf.tcptr = buf.tcbuf = snewn(buf.bufsize, truecolour); + + old_top_x = top.x; /* needed for rect==1 */ + + while (poslt(top, bottom)) { + bool nl = false; + termline *ldata = lineptr(top.y); + pos nlpos; + + /* + * nlpos will point at the maximum position on this line we + * should copy up to. So we start it at the end of the + * line... + */ + nlpos.y = top.y; + nlpos.x = term->cols; + + /* + * ... move it backwards if there's unused space at the end + * of the line (and also set `nl' if this is the case, + * because in normal selection mode this means we need a + * newline at the end)... + */ + if (!(ldata->lattr & LATTR_WRAPPED)) { + while (nlpos.x && + IS_SPACE_CHR(ldata->chars[nlpos.x - 1].chr) && + !ldata->chars[nlpos.x - 1].cc_next && + poslt(top, nlpos)) + decpos(nlpos); + if (poslt(nlpos, bottom)) + nl = true; + } else { + if (ldata->trusted) { + /* A wrapped line with a trust sigil on it terminates + * a few characters earlier. */ + nlpos.x = (nlpos.x < TRUST_SIGIL_WIDTH ? 0 : + nlpos.x - TRUST_SIGIL_WIDTH); + } + if (ldata->lattr & LATTR_WRAPPED2) { + /* Ignore the last char on the line in a WRAPPED2 line. */ + decpos(nlpos); + } + } + + /* + * ... and then clip it to the terminal x coordinate if + * we're doing rectangular selection. (In this case we + * still did the above, so that copying e.g. the right-hand + * column from a table doesn't fill with spaces on the + * right.) + */ + if (rect) { + if (nlpos.x > bottom.x) + nlpos.x = bottom.x; + nl = (top.y < bottom.y); + } + + while (poslt(top, bottom) && poslt(top, nlpos)) { +#if 0 + char cbuf[16], *p; + sprintf(cbuf, "<U+%04x>", (ldata[top.x] & 0xFFFF)); +#else + wchar_t cbuf[16], *p; + int c; + int x = top.x; + + if (ldata->chars[x].chr == UCSWIDE) { + top.x++; + continue; + } + + while (1) { + int uc = ldata->chars[x].chr; + attr = ldata->chars[x].attr; + tc = ldata->chars[x].truecolour; + + switch (uc & CSET_MASK) { + case CSET_LINEDRW: + if (!term->rawcnp) { + uc = term->ucsdata->unitab_xterm[uc & 0xFF]; + break; + } + case CSET_ASCII: + uc = term->ucsdata->unitab_line[uc & 0xFF]; + break; + case CSET_SCOACS: + uc = term->ucsdata->unitab_scoacs[uc&0xFF]; + break; + } + switch (uc & CSET_MASK) { + case CSET_ACP: + uc = term->ucsdata->unitab_font[uc & 0xFF]; + break; + case CSET_OEMCP: + uc = term->ucsdata->unitab_oemcp[uc & 0xFF]; + break; + } + + c = (uc & ~CSET_MASK); +#ifdef PLATFORM_IS_UTF16 + if (uc > 0x10000 && uc < 0x110000) { + cbuf[0] = 0xD800 | ((uc - 0x10000) >> 10); + cbuf[1] = 0xDC00 | ((uc - 0x10000) & 0x3FF); + cbuf[2] = 0; + } else +#endif + { + cbuf[0] = uc; + cbuf[1] = 0; + } + + if (DIRECT_FONT(uc)) { + if (c >= ' ' && c != 0x7F) { + char buf[4]; + WCHAR wbuf[4]; + int rv; + if (is_dbcs_leadbyte(term->ucsdata->font_codepage, (BYTE) c)) { + buf[0] = c; + buf[1] = (char) (0xFF & ldata->chars[top.x + 1].chr); + rv = mb_to_wc(term->ucsdata->font_codepage, 0, buf, 2, wbuf, 4); + top.x++; + } else { + buf[0] = c; + rv = mb_to_wc(term->ucsdata->font_codepage, 0, buf, 1, wbuf, 4); + } + + if (rv > 0) { + memcpy(cbuf, wbuf, rv * sizeof(wchar_t)); + cbuf[rv] = 0; + } + } + } +#endif + + for (p = cbuf; *p; p++) + clip_addchar(&buf, *p, attr, tc); + + if (ldata->chars[x].cc_next) + x += ldata->chars[x].cc_next; + else + break; + } + top.x++; + } + if (nl) { + int i; + for (i = 0; i < sel_nl_sz; i++) + clip_addchar(&buf, sel_nl[i], 0, term->basic_erase_char.truecolour); + } + top.y++; + top.x = rect ? old_top_x : 0; + + unlineptr(ldata); + } +#if SELECTION_NUL_TERMINATED + clip_addchar(&buf, 0, 0, term->basic_erase_char.truecolour); +#endif + /* Finally, transfer all that to the clipboard(s). */ + { + int i; + bool clip_local = false; + for (i = 0; i < n_clipboards; i++) { + if (clipboards[i] == CLIP_LOCAL) { + clip_local = true; + } else if (clipboards[i] != CLIP_NULL) { + win_clip_write( + term->win, clipboards[i], buf.textbuf, buf.attrbuf, + buf.tcbuf, buf.bufpos, desel); + } + } + if (clip_local) { + sfree(term->last_selected_text); + sfree(term->last_selected_attr); + sfree(term->last_selected_tc); + term->last_selected_text = buf.textbuf; + term->last_selected_attr = buf.attrbuf; + term->last_selected_tc = buf.tcbuf; + term->last_selected_len = buf.bufpos; + } else { + sfree(buf.textbuf); + sfree(buf.attrbuf); + sfree(buf.tcbuf); + } + } +} + +void term_copyall(Terminal *term, const int *clipboards, int n_clipboards) +{ + pos top; + pos bottom; + tree234 *screen = term->screen; + top.y = -sblines(term); + top.x = 0; + bottom.y = find_last_nonempty_line(term, screen); + bottom.x = term->cols; + clipme(term, top, bottom, false, true, clipboards, n_clipboards); +} + +static void paste_from_clip_local(void *vterm) +{ + Terminal *term = (Terminal *)vterm; + term_do_paste(term, term->last_selected_text, term->last_selected_len); +} + +void term_request_copy(Terminal *term, const int *clipboards, int n_clipboards) +{ + int i; + for (i = 0; i < n_clipboards; i++) { + assert(clipboards[i] != CLIP_LOCAL); + if (clipboards[i] != CLIP_NULL) { + win_clip_write(term->win, clipboards[i], + term->last_selected_text, term->last_selected_attr, + term->last_selected_tc, term->last_selected_len, + false); + } + } +} + +void term_request_paste(Terminal *term, int clipboard) +{ + switch (clipboard) { + case CLIP_NULL: + /* Do nothing: CLIP_NULL never has data in it. */ + break; + case CLIP_LOCAL: + queue_toplevel_callback(paste_from_clip_local, term); + break; + default: + win_clip_request_paste(term->win, clipboard); + break; + } +} + +/* + * The wordness array is mainly for deciding the disposition of the + * US-ASCII characters. + */ +static int wordtype(Terminal *term, int uc) +{ + struct ucsword { + int start, end, ctype; + }; + static const struct ucsword ucs_words[] = { + { + 128, 160, 0}, { + 161, 191, 1}, { + 215, 215, 1}, { + 247, 247, 1}, { + 0x037e, 0x037e, 1}, /* Greek question mark */ + { + 0x0387, 0x0387, 1}, /* Greek ano teleia */ + { + 0x055a, 0x055f, 1}, /* Armenian punctuation */ + { + 0x0589, 0x0589, 1}, /* Armenian full stop */ + { + 0x0700, 0x070d, 1}, /* Syriac punctuation */ + { + 0x104a, 0x104f, 1}, /* Myanmar punctuation */ + { + 0x10fb, 0x10fb, 1}, /* Georgian punctuation */ + { + 0x1361, 0x1368, 1}, /* Ethiopic punctuation */ + { + 0x166d, 0x166e, 1}, /* Canadian Syl. punctuation */ + { + 0x17d4, 0x17dc, 1}, /* Khmer punctuation */ + { + 0x1800, 0x180a, 1}, /* Mongolian punctuation */ + { + 0x2000, 0x200a, 0}, /* Various spaces */ + { + 0x2070, 0x207f, 2}, /* superscript */ + { + 0x2080, 0x208f, 2}, /* subscript */ + { + 0x200b, 0x27ff, 1}, /* punctuation and symbols */ + { + 0x3000, 0x3000, 0}, /* ideographic space */ + { + 0x3001, 0x3020, 1}, /* ideographic punctuation */ + { + 0x303f, 0x309f, 3}, /* Hiragana */ + { + 0x30a0, 0x30ff, 3}, /* Katakana */ + { + 0x3300, 0x9fff, 3}, /* CJK Ideographs */ + { + 0xac00, 0xd7a3, 3}, /* Hangul Syllables */ + { + 0xf900, 0xfaff, 3}, /* CJK Ideographs */ + { + 0xfe30, 0xfe6b, 1}, /* punctuation forms */ + { + 0xff00, 0xff0f, 1}, /* half/fullwidth ASCII */ + { + 0xff1a, 0xff20, 1}, /* half/fullwidth ASCII */ + { + 0xff3b, 0xff40, 1}, /* half/fullwidth ASCII */ + { + 0xff5b, 0xff64, 1}, /* half/fullwidth ASCII */ + { + 0xfff0, 0xffff, 0}, /* half/fullwidth ASCII */ + { + 0, 0, 0} + }; + const struct ucsword *wptr; + + switch (uc & CSET_MASK) { + case CSET_LINEDRW: + uc = term->ucsdata->unitab_xterm[uc & 0xFF]; + break; + case CSET_ASCII: + uc = term->ucsdata->unitab_line[uc & 0xFF]; + break; + case CSET_SCOACS: + uc = term->ucsdata->unitab_scoacs[uc&0xFF]; + break; + } + switch (uc & CSET_MASK) { + case CSET_ACP: + uc = term->ucsdata->unitab_font[uc & 0xFF]; + break; + case CSET_OEMCP: + uc = term->ucsdata->unitab_oemcp[uc & 0xFF]; + break; + } + + /* For DBCS fonts I can't do anything useful. Even this will sometimes + * fail as there's such a thing as a double width space. :-( + */ + if (term->ucsdata->dbcs_screenfont && + term->ucsdata->font_codepage == term->ucsdata->line_codepage) + return (uc != ' '); + + if (uc < 0x80) + return term->wordness[uc]; + + for (wptr = ucs_words; wptr->start; wptr++) { + if (uc >= wptr->start && uc <= wptr->end) + return wptr->ctype; + } + + return 2; +} + +static int line_cols(Terminal *term, termline *ldata) +{ + int cols = term->cols; + if (ldata->trusted) { + cols -= TRUST_SIGIL_WIDTH; + } + if (ldata->lattr & LATTR_WRAPPED2) + cols--; + if (cols < 0) + cols = 0; + return cols; +} + +/* + * Spread the selection outwards according to the selection mode. + */ +static pos sel_spread_half(Terminal *term, pos p, int dir) +{ + termline *ldata; + short wvalue; + int topy = -sblines(term); + + ldata = lineptr(p.y); + + switch (term->selmode) { + case SM_CHAR: + /* + * In this mode, every character is a separate unit, except + * for runs of spaces at the end of a non-wrapping line. + */ + if (!(ldata->lattr & LATTR_WRAPPED)) { + termchar *q = ldata->chars + line_cols(term, ldata); + while (q > ldata->chars && + IS_SPACE_CHR(q[-1].chr) && !q[-1].cc_next) + q--; + if (q == ldata->chars + term->cols) + q--; + if (p.x >= q - ldata->chars) + p.x = (dir == -1 ? q - ldata->chars : term->cols - 1); + } + break; + case SM_WORD: + /* + * In this mode, the units are maximal runs of characters + * whose `wordness' has the same value. + */ + wvalue = wordtype(term, UCSGET(ldata->chars, p.x)); + if (dir == +1) { + while (1) { + int maxcols = line_cols(term, ldata); + if (p.x < maxcols-1) { + if (wordtype(term, UCSGET(ldata->chars, p.x+1)) == wvalue) + p.x++; + else + break; + } else { + if (p.y+1 < term->rows && + (ldata->lattr & LATTR_WRAPPED)) { + termline *ldata2; + ldata2 = lineptr(p.y+1); + if (wordtype(term, UCSGET(ldata2->chars, 0)) + == wvalue) { + p.x = 0; + p.y++; + unlineptr(ldata); + ldata = ldata2; + } else { + unlineptr(ldata2); + break; + } + } else + break; + } + } + } else { + while (1) { + if (p.x > 0) { + if (wordtype(term, UCSGET(ldata->chars, p.x-1)) == wvalue) + p.x--; + else + break; + } else { + termline *ldata2; + int maxcols; + if (p.y <= topy) + break; + ldata2 = lineptr(p.y-1); + maxcols = line_cols(term, ldata2); + if (ldata2->lattr & LATTR_WRAPPED) { + if (wordtype(term, UCSGET(ldata2->chars, maxcols-1)) + == wvalue) { + p.x = maxcols-1; + p.y--; + unlineptr(ldata); + ldata = ldata2; + } else { + unlineptr(ldata2); + break; + } + } else + break; + } + } + } + break; + case SM_LINE: + /* + * In this mode, every line is a unit. + */ + p.x = (dir == -1 ? 0 : term->cols - 1); + break; + } + + unlineptr(ldata); + return p; +} + +static void sel_spread(Terminal *term) +{ + if (term->seltype == LEXICOGRAPHIC) { + term->selstart = sel_spread_half(term, term->selstart, -1); + decpos(term->selend); + term->selend = sel_spread_half(term, term->selend, +1); + incpos(term->selend); + } +} + +static void term_paste_callback(void *vterm) +{ + Terminal *term = (Terminal *)vterm; + + if (term->paste_len == 0) + return; + + while (term->paste_pos < term->paste_len) { + int n = 0; + while (n + term->paste_pos < term->paste_len) { + if (term->paste_buffer[term->paste_pos + n++] == '\015') + break; + } + if (term->ldisc) { + strbuf *buf = term_input_data_from_unicode( + term, term->paste_buffer + term->paste_pos, n); + term_keyinput_internal(term, buf->s, buf->len, false); + strbuf_free(buf); + } + term->paste_pos += n; + + if (term->paste_pos < term->paste_len) { + queue_toplevel_callback(term_paste_callback, term); + return; + } + } + term_bracketed_paste_stop(term); + sfree(term->paste_buffer); + term->paste_buffer = NULL; + term->paste_len = 0; +} + +/* + * Specialist string compare function. Returns true if the buffer of + * alen wide characters starting at a has as a prefix the buffer of + * blen characters starting at b. + */ +static bool wstartswith(const wchar_t *a, size_t alen, + const wchar_t *b, size_t blen) +{ + return alen >= blen && !wcsncmp(a, b, blen); +} + +void term_do_paste(Terminal *term, const wchar_t *data, int len) +{ + const wchar_t *p; + bool paste_controls = conf_get_bool(term->conf, CONF_paste_controls); + + /* + * Pasting data into the terminal counts as a keyboard event (for + * purposes of the 'Reset scrollback on keypress' config option), + * unless the paste is zero-length. + */ + if (len == 0) + return; + term_seen_key_event(term); + + if (term->paste_buffer) + sfree(term->paste_buffer); + term->paste_pos = term->paste_len = 0; + term->paste_buffer = snewn(len + 12, wchar_t); + + if (term->bracketed_paste) + term_bracketed_paste_start(term); + + p = data; + while (p < data + len) { + wchar_t wc = *p++; + + if (wc == sel_nl[0] && + wstartswith(p-1, data+len-(p-1), sel_nl, sel_nl_sz)) { + /* + * This is the (platform-dependent) sequence that the host + * OS uses to represent newlines in clipboard data. + * Normalise it to a press of CR. + */ + p += sel_nl_sz - 1; + wc = '\015'; + } + + if ((wc & ~(wint_t)0x9F) == 0) { + /* + * This is a control code, either in the range 0x00-0x1F + * or 0x80-0x9F. We reject all of these in pastecontrols + * mode, except for a small set of permitted ones. + */ + if (!paste_controls) { + /* In line with xterm 292, accepted control chars are: + * CR, LF, tab, backspace. (And DEL, i.e. 0x7F, but + * that's permitted by virtue of not matching the bit + * mask that got us into this if statement, so we + * don't have to permit it here. */ + static const unsigned mask = + (1<<13) | (1<<10) | (1<<9) | (1<<8); + + if (wc > 15 || !((mask >> wc) & 1)) + continue; + } + + if (wc == '\033' && term->bracketed_paste && + wstartswith(p-1, data+len-(p-1), L"\033[201~", 6)) { + /* + * Also, in bracketed-paste mode, reject the ESC + * character that begins the end-of-paste sequence. + */ + continue; + } + } + + term->paste_buffer[term->paste_len++] = wc; + } + + /* Assume a small paste will be OK in one go. */ + if (term->paste_len < 256) { + if (term->ldisc) { + strbuf *buf = term_input_data_from_unicode( + term, term->paste_buffer, term->paste_len); + term_keyinput_internal(term, buf->s, buf->len, false); + strbuf_free(buf); + } + if (term->paste_buffer) + sfree(term->paste_buffer); + term_bracketed_paste_stop(term); + term->paste_buffer = NULL; + term->paste_pos = term->paste_len = 0; + } + + queue_toplevel_callback(term_paste_callback, term); +} + +void term_mouse(Terminal *term, Mouse_Button braw, Mouse_Button bcooked, + Mouse_Action a, int x, int y, bool shift, bool ctrl, bool alt) +{ + pos selpoint; + termline *ldata; + bool raw_mouse = (term->xterm_mouse && + !term->no_mouse_rep && + !(term->mouse_override && shift)); + int default_seltype; + + if (y < 0) { + y = 0; + if (a == MA_DRAG && !raw_mouse) + term_scroll(term, 0, -1); + } + if (y >= term->rows) { + y = term->rows - 1; + if (a == MA_DRAG && !raw_mouse) + term_scroll(term, 0, +1); + } + if (x < 0) { + if (y > 0 && !raw_mouse && term->seltype != RECTANGULAR) { + /* + * When we're using the mouse for normal raster-based + * selection, dragging off the left edge of a terminal row + * is treated the same as the right-hand end of the + * previous row, in that it's considered to identify a + * point _before_ the first character on row y. + * + * But if the mouse action is going to be used for + * anything else - rectangular selection, or xterm mouse + * tracking - then we disable this special treatment. + */ + x = term->cols - 1; + y--; + } else + x = 0; + } + if (x >= term->cols) + x = term->cols - 1; + + selpoint.y = y + term->disptop; + ldata = lineptr(selpoint.y); + + if ((ldata->lattr & LATTR_MODE) != LATTR_NORM) + x /= 2; + + /* + * Transform x through the bidi algorithm to find the _logical_ + * click point from the physical one. + */ + if (term_bidi_line(term, ldata, y) != NULL) { + x = term->post_bidi_cache[y].backward[x]; + } + + selpoint.x = x; + unlineptr(ldata); + + /* + * If we're in the middle of a selection operation, we ignore raw + * mouse mode until it's done (we must have been not in raw mouse + * mode when it started). + * This makes use of Shift for selection reliable, and avoids the + * host seeing mouse releases for which they never saw corresponding + * presses. + */ + if (raw_mouse && + (term->selstate != ABOUT_TO) && (term->selstate != DRAGGING)) { + int encstate = 0, r, c; + bool wheel; + char abuf[32]; + int len = 0; + + if (term->ldisc) { + + switch (braw) { + case MBT_LEFT: + encstate = 0x00; /* left button down */ + wheel = false; + break; + case MBT_MIDDLE: + encstate = 0x01; + wheel = false; + break; + case MBT_RIGHT: + encstate = 0x02; + wheel = false; + break; + case MBT_WHEEL_UP: + encstate = 0x40; + wheel = true; + break; + case MBT_WHEEL_DOWN: + encstate = 0x41; + wheel = true; + break; + default: + return; + } + if (wheel) { + /* For mouse wheel buttons, we only ever expect to see + * MA_CLICK actions, and we don't try to keep track of + * the buttons being 'pressed' (since without matching + * click/release pairs that's pointless). */ + if (a != MA_CLICK) + return; + } else switch (a) { + case MA_DRAG: + if (term->xterm_mouse == 1) + return; + encstate += 0x20; + break; + case MA_RELEASE: + /* If multiple extensions are enabled, the xterm 1006 is used, so it's okay to check for only that */ + if (!term->xterm_extended_mouse) + encstate = 0x03; + term->mouse_is_down = 0; + break; + case MA_CLICK: + if (term->mouse_is_down == braw) + return; + term->mouse_is_down = braw; + break; + default: + return; + } + if (shift) + encstate += 0x04; + if (ctrl) + encstate += 0x10; + r = y + 1; + c = x + 1; + + /* Check the extensions in decreasing order of preference. Encoding the release event above assumes that 1006 comes first. */ + if (term->xterm_extended_mouse) { + len = sprintf(abuf, "\033[<%d;%d;%d%c", encstate, c, r, a == MA_RELEASE ? 'm' : 'M'); + } else if (term->urxvt_extended_mouse) { + len = sprintf(abuf, "\033[%d;%d;%dM", encstate + 32, c, r); + } else if (c <= 223 && r <= 223) { + len = sprintf(abuf, "\033[M%c%c%c", encstate + 32, c + 32, r + 32); + } + if (len > 0) + ldisc_send(term->ldisc, abuf, len, false); + } + return; + } + + /* + * Set the selection type (rectangular or normal) at the start + * of a selection attempt, from the state of Alt. + */ + if (!alt ^ !term->rect_select) + default_seltype = RECTANGULAR; + else + default_seltype = LEXICOGRAPHIC; + + if (term->selstate == NO_SELECTION) { + term->seltype = default_seltype; + } + + if (bcooked == MBT_SELECT && a == MA_CLICK) { + deselect(term); + term->selstate = ABOUT_TO; + term->seltype = default_seltype; + term->selanchor = selpoint; + term->selmode = SM_CHAR; + } else if (bcooked == MBT_SELECT && (a == MA_2CLK || a == MA_3CLK)) { + deselect(term); + term->selmode = (a == MA_2CLK ? SM_WORD : SM_LINE); + term->selstate = DRAGGING; + term->selstart = term->selanchor = selpoint; + term->selend = term->selstart; + incpos(term->selend); + sel_spread(term); + } else if ((bcooked == MBT_SELECT && a == MA_DRAG) || + (bcooked == MBT_EXTEND && a != MA_RELEASE)) { + if (a == MA_DRAG && + (term->selstate == NO_SELECTION || term->selstate == SELECTED)) { + /* + * This can happen if a front end has passed us a MA_DRAG + * without a prior MA_CLICK. OS X GTK does so, for + * example, if the initial button press was eaten by the + * WM when it activated the window in the first place. The + * nicest thing to do in this situation is to ignore + * further drags, and wait for the user to click in the + * window again properly if they want to select. + */ + return; + } + if (term->selstate == ABOUT_TO && poseq(term->selanchor, selpoint)) + return; + if (bcooked == MBT_EXTEND && a != MA_DRAG && + term->selstate == SELECTED) { + if (term->seltype == LEXICOGRAPHIC) { + /* + * For normal selection, we extend by moving + * whichever end of the current selection is closer + * to the mouse. + */ + if (posdiff(selpoint, term->selstart) < + posdiff(term->selend, term->selstart) / 2) { + term->selanchor = term->selend; + decpos(term->selanchor); + } else { + term->selanchor = term->selstart; + } + } else { + /* + * For rectangular selection, we have a choice of + * _four_ places to put selanchor and selpoint: the + * four corners of the selection. + */ + if (2*selpoint.x < term->selstart.x + term->selend.x) + term->selanchor.x = term->selend.x-1; + else + term->selanchor.x = term->selstart.x; + + if (2*selpoint.y < term->selstart.y + term->selend.y) + term->selanchor.y = term->selend.y; + else + term->selanchor.y = term->selstart.y; + } + term->selstate = DRAGGING; + } + if (term->selstate != ABOUT_TO && term->selstate != DRAGGING) + term->selanchor = selpoint; + term->selstate = DRAGGING; + if (term->seltype == LEXICOGRAPHIC) { + /* + * For normal selection, we set (selstart,selend) to + * (selpoint,selanchor) in some order. + */ + if (poslt(selpoint, term->selanchor)) { + term->selstart = selpoint; + term->selend = term->selanchor; + incpos(term->selend); + } else { + term->selstart = term->selanchor; + term->selend = selpoint; + incpos(term->selend); + } + } else { + /* + * For rectangular selection, we may need to + * interchange x and y coordinates (if the user has + * dragged in the -x and +y directions, or vice versa). + */ + term->selstart.x = min(term->selanchor.x, selpoint.x); + term->selend.x = 1+max(term->selanchor.x, selpoint.x); + term->selstart.y = min(term->selanchor.y, selpoint.y); + term->selend.y = max(term->selanchor.y, selpoint.y); + } + sel_spread(term); + } else if ((bcooked == MBT_SELECT || bcooked == MBT_EXTEND) && + a == MA_RELEASE) { + if (term->selstate == DRAGGING) { + /* + * We've completed a selection. We now transfer the + * data to the clipboard. + */ + clipme(term, term->selstart, term->selend, + (term->seltype == RECTANGULAR), false, + term->mouse_select_clipboards, + term->n_mouse_select_clipboards); + term->selstate = SELECTED; + } else + term->selstate = NO_SELECTION; + } else if (bcooked == MBT_PASTE + && (a == MA_CLICK +#if MULTICLICK_ONLY_EVENT + || a == MA_2CLK || a == MA_3CLK +#endif + )) { + term_request_paste(term, term->mouse_paste_clipboard); + } + + /* + * Since terminal output is suppressed during drag-selects, we + * should make sure to write any pending output if one has just + * finished. + */ + term_out(term, false); + term_schedule_update(term); +} + +void term_cancel_selection_drag(Terminal *term) +{ + /* + * In unusual circumstances, a mouse drag might be interrupted by + * something that steals the rest of the mouse gesture. An example + * is the GTK popup menu appearing. In that situation, we'll never + * receive the MA_RELEASE that finishes the DRAGGING state, which + * means terminal output could be suppressed indefinitely. Call + * this function from the front end in such situations to restore + * sensibleness. + */ + if (term->selstate == DRAGGING) + term->selstate = NO_SELECTION; + term_out(term, false); + term_schedule_update(term); +} + +static int shift_bitmap(bool shift, bool ctrl, bool alt, bool *consumed_alt) +{ + int bitmap = (shift ? 1 : 0) + (alt ? 2 : 0) + (ctrl ? 4 : 0); + if (bitmap) + bitmap++; + if (alt && consumed_alt) + *consumed_alt = true; + return bitmap; +} + +int format_arrow_key(char *buf, Terminal *term, int xkey, + bool shift, bool ctrl, bool alt, bool *consumed_alt) +{ + char *p = buf; + + if (term->vt52_mode) + p += sprintf(p, "\x1B%c", xkey); + else { + bool app_flg = (term->app_cursor_keys && !term->no_applic_c); +#if 0 + /* + * RDB: VT100 & VT102 manuals both state the app cursor + * keys only work if the app keypad is on. + * + * SGT: That may well be true, but xterm disagrees and so + * does at least one application, so I've #if'ed this out + * and the behaviour is back to PuTTY's original: app + * cursor and app keypad are independently switchable + * modes. If anyone complains about _this_ I'll have to + * put in a configurable option. + */ + if (!term->app_keypad_keys) + app_flg = 0; +#endif + + int bitmap = 0; + + /* Adjustment based on Shift, Ctrl and/or Alt */ + switch (term->sharrow_type) { + case SHARROW_APPLICATION: + if (ctrl) + app_flg = !app_flg; + break; + case SHARROW_BITMAP: + bitmap = shift_bitmap(shift, ctrl, alt, consumed_alt); + break; + } + + if (app_flg) + p += sprintf(p, "\x1BO%c", xkey); + else if (bitmap) + p += sprintf(p, "\x1B[1;%d%c", bitmap, xkey); + else + p += sprintf(p, "\x1B[%c", xkey); + } + + return p - buf; +} + +int format_function_key(char *buf, Terminal *term, int key_number, + bool shift, bool ctrl, bool alt, bool *consumed_alt) +{ + char *p = buf; + + static const int key_number_to_tilde_code[] = { + -1, /* no such key as F0 */ + 11, 12, 13, 14, 15, /*gap*/ 17, 18, 19, 20, 21, /*gap*/ + 23, 24, 25, 26, /*gap*/ 28, 29, /*gap*/ 31, 32, 33, 34, + }; + + assert(key_number > 0); + assert(key_number < lenof(key_number_to_tilde_code)); + + int index = key_number; + if (term->funky_type != FUNKY_XTERM_216 && term->funky_type != FUNKY_SCO) { + if (shift && index <= 10) { + shift = false; + index += 10; + } + } + + int code = key_number_to_tilde_code[index]; + + if (term->funky_type == FUNKY_SCO) { + /* SCO function keys */ + static const char sco_codes[] = + "MNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz@[\\]^_`{"; + index = (key_number >= 1 && key_number <= 12) ? key_number - 1 : 0; + if (shift) index += 12; + if (ctrl) index += 24; + p += sprintf(p, "\x1B[%c", sco_codes[index]); + } else if ((term->vt52_mode || term->funky_type == FUNKY_VT100P) && + code >= 11 && code <= 24) { + int offt = 0; + if (code > 15) + offt++; + if (code > 21) + offt++; + if (term->vt52_mode) + p += sprintf(p, "\x1B%c", code + 'P' - 11 - offt); + else + p += sprintf(p, "\x1BO%c", code + 'P' - 11 - offt); + } else if (term->funky_type == FUNKY_LINUX && code >= 11 && code <= 15) { + p += sprintf(p, "\x1B[[%c", code + 'A' - 11); + } else if ((term->funky_type == FUNKY_XTERM || + term->funky_type == FUNKY_XTERM_216) && + code >= 11 && code <= 14) { + if (term->vt52_mode) + p += sprintf(p, "\x1B%c", code + 'P' - 11); + else { + int bitmap = 0; + if (term->funky_type == FUNKY_XTERM_216) + bitmap = shift_bitmap(shift, ctrl, alt, consumed_alt); + if (bitmap) + p += sprintf(p, "\x1B[1;%d%c", bitmap, code + 'P' - 11); + else + p += sprintf(p, "\x1BO%c", code + 'P' - 11); + } + } else { + int bitmap = 0; + if (term->funky_type == FUNKY_XTERM_216) + bitmap = shift_bitmap(shift, ctrl, alt, consumed_alt); + if (bitmap) + p += sprintf(p, "\x1B[%d;%d~", code, bitmap); + else + p += sprintf(p, "\x1B[%d~", code); + } + + return p - buf; +} + +int format_small_keypad_key(char *buf, Terminal *term, SmallKeypadKey key, + bool shift, bool ctrl, bool alt, + bool *consumed_alt) +{ + char *p = buf; + + int code; + switch (key) { + case SKK_HOME: code = 1; break; + case SKK_INSERT: code = 2; break; + case SKK_DELETE: code = 3; break; + case SKK_END: code = 4; break; + case SKK_PGUP: code = 5; break; + case SKK_PGDN: code = 6; break; + default: unreachable("bad small keypad key enum value"); + } + + /* Reorder edit keys to physical order */ + if (term->funky_type == FUNKY_VT400 && code <= 6) + code = "\0\2\1\4\5\3\6"[code]; + + if (term->vt52_mode && code > 0 && code <= 6) { + p += sprintf(p, "\x1B%c", " HLMEIG"[code]); + } else if (term->funky_type == FUNKY_SCO) { + static const char codes[] = "HL.FIG"; + if (code == 3) { + *p++ = '\x7F'; + } else { + p += sprintf(p, "\x1B[%c", codes[code-1]); + } + } else if ((code == 1 || code == 4) && term->rxvt_homeend) { + p += sprintf(p, code == 1 ? "\x1B[H" : "\x1BOw"); + } else { + if (term->vt52_mode) { + p += sprintf(p, "\x1B[%d~", code); + } else { + int bitmap = 0; + if (term->funky_type == FUNKY_XTERM_216) + bitmap = shift_bitmap(shift, ctrl, alt, consumed_alt); + if (bitmap) + p += sprintf(p, "\x1B[%d;%d~", code, bitmap); + else + p += sprintf(p, "\x1B[%d~", code); + } + } + + return p - buf; +} + +int format_numeric_keypad_key(char *buf, Terminal *term, char key, + bool shift, bool ctrl) +{ + char *p = buf; + bool app_keypad = (term->app_keypad_keys && !term->no_applic_k); + + if (term->nethack_keypad && (key >= '1' && key <= '9')) { + static const char nh_base[] = "bjnh.lyku"; + char c = nh_base[key - '1']; + if (ctrl && c != '.') + c &= 0x1F; + else if (shift && c != '.') + c += 'A'-'a'; + *p++ = c; + } else { + int xkey = 0; + + if (term->funky_type == FUNKY_VT400 || + (term->funky_type <= FUNKY_LINUX && app_keypad)) { + switch (key) { + case 'G': xkey = 'P'; break; + case '/': xkey = 'Q'; break; + case '*': xkey = 'R'; break; + case '-': xkey = 'S'; break; + } + } + + if (app_keypad) { + switch (key) { + case '0': xkey = 'p'; break; + case '1': xkey = 'q'; break; + case '2': xkey = 'r'; break; + case '3': xkey = 's'; break; + case '4': xkey = 't'; break; + case '5': xkey = 'u'; break; + case '6': xkey = 'v'; break; + case '7': xkey = 'w'; break; + case '8': xkey = 'x'; break; + case '9': xkey = 'y'; break; + case '.': xkey = 'n'; break; + case '\r': xkey = 'M'; break; + + case '+': + /* + * Keypad + is tricky. It covers a space that would + * be taken up on the VT100 by _two_ keys; so we + * let Shift select between the two. Worse still, + * in xterm function key mode we change which two... + */ + if (term->funky_type == FUNKY_XTERM) + xkey = shift ? 'l' : 'k'; + else + xkey = shift ? 'm' : 'l'; + break; + + case '/': + if (term->funky_type == FUNKY_XTERM) + xkey = 'o'; + break; + case '*': + if (term->funky_type == FUNKY_XTERM) + xkey = 'j'; + break; + case '-': + if (term->funky_type == FUNKY_XTERM) + xkey = 'm'; + break; + } + } + + if (xkey) { + if (term->vt52_mode) { + if (xkey >= 'P' && xkey <= 'S') + p += sprintf(p, "\x1B%c", xkey); + else + p += sprintf(p, "\x1B?%c", xkey); + } else + p += sprintf(p, "\x1BO%c", xkey); + } + } + + return p - buf; +} + +void term_keyinputw(Terminal *term, const wchar_t *widebuf, int len) +{ + strbuf *buf = term_input_data_from_unicode(term, widebuf, len); + if (buf->len) + term_keyinput_internal(term, buf->s, buf->len, true); + strbuf_free(buf); +} + +void term_keyinput(Terminal *term, int codepage, const void *str, int len) +{ + if (codepage < 0 || codepage == term->ucsdata->line_codepage) { + /* + * This text needs no translation, either because it's already + * in the right character set, or because we got the special + * codepage value -1 from our caller which means 'this data + * should be charset-agnostic, just send it raw' (for really + * simple things like control characters). + */ + term_keyinput_internal(term, str, len, true); + } else { + strbuf *buf = term_input_data_from_charset(term, codepage, str, len); + if (buf->len) + term_keyinput_internal(term, buf->s, buf->len, true); + strbuf_free(buf); + } +} + +void term_nopaste(Terminal *term) +{ + if (term->paste_len == 0) + return; + sfree(term->paste_buffer); + term_bracketed_paste_stop(term); + term->paste_buffer = NULL; + term->paste_len = 0; +} + +static void deselect(Terminal *term) +{ + term->selstate = NO_SELECTION; + term->selstart.x = term->selstart.y = term->selend.x = term->selend.y = 0; +} + +void term_lost_clipboard_ownership(Terminal *term, int clipboard) +{ + if (!(term->n_mouse_select_clipboards > 1 && + clipboard == term->mouse_select_clipboards[1])) + return; + + deselect(term); + term_update(term); + + /* + * Since terminal output is suppressed during drag-selects, we + * should make sure to write any pending output if one has just + * finished. + */ + term_out(term, false); +} + +static void term_added_data(Terminal *term, bool called_from_term_data) +{ + if (!term->in_term_out) { + term->in_term_out = true; + term_reset_cblink(term); + term_out(term, called_from_term_data); + term->in_term_out = false; + } +} + +size_t term_data(Terminal *term, const void *data, size_t len) +{ + bufchain_add(&term->inbuf, data, len); + term_added_data(term, true); + return bufchain_size(&term->inbuf); +} + +void term_provide_logctx(Terminal *term, LogContext *logctx) +{ + term->logctx = logctx; +} + +void term_set_focus(Terminal *term, bool has_focus) +{ + term->has_focus = has_focus; + term_schedule_cblink(term); +} + +/* + * Provide "auto" settings for remote tty modes, suitable for an + * application with a terminal window. + */ +char *term_get_ttymode(Terminal *term, const char *mode) +{ + const char *val = NULL; + if (strcmp(mode, "ERASE") == 0) { + val = term->bksp_is_delete ? "^?" : "^H"; + } else if (strcmp(mode, "IUTF8") == 0) { + val = (term->ucsdata->line_codepage == CP_UTF8) ? "yes" : "no"; + } + /* FIXME: perhaps we should set ONLCR based on lfhascr as well? */ + /* FIXME: or ECHO and friends based on local echo state? */ + return dupstr(val); +} + +struct term_userpass_state { + size_t curr_prompt; + bool done_prompt; /* printed out prompt yet? */ +}; + +/* Tiny wrapper to make it easier to write lots of little strings */ +static inline void term_write(Terminal *term, ptrlen data) +{ + term_data(term, data.ptr, data.len); +} + +/* + * Signal that a prompts_t is done. This involves sending a + * notification to the caller, and also turning off our own callback + * that listens for more data arriving in the ldisc's input queue. + */ +static inline SeatPromptResult signal_prompts_t(Terminal *term, prompts_t *p, + SeatPromptResult spr) +{ + assert(p->callback && "Asynchronous userpass input requires a callback"); + queue_toplevel_callback(p->callback, p->callback_ctx); + if (term->ldisc) + ldisc_enable_prompt_callback(term->ldisc, NULL); + p->spr = spr; + return spr; +} + +/* + * Process some terminal data in the course of username/password + * input. + */ +SeatPromptResult term_get_userpass_input(Terminal *term, prompts_t *p) +{ + if (!term->ldisc) { + /* Can't handle interactive prompts without an ldisc */ + return signal_prompts_t(term, p, SPR_SW_ABORT( + "Terminal not prepared for interactive prompts")); + } + + if (p->spr.kind != SPRK_INCOMPLETE) { + /* We've already finished these prompts, so return the same + * result again */ + return p->spr; + } + + struct term_userpass_state *s = (struct term_userpass_state *)p->data; + + if (!s) { + /* + * First call. Set some stuff up. + */ + p->data = s = snew(struct term_userpass_state); + p->spr = SPR_INCOMPLETE; + s->curr_prompt = 0; + s->done_prompt = false; + /* We only print the `name' caption if we have to... */ + if (p->name_reqd && p->name) { + ptrlen plname = ptrlen_from_asciz(p->name); + term_write(term, plname); + if (!ptrlen_endswith(plname, PTRLEN_LITERAL("\n"), NULL)) + term_write(term, PTRLEN_LITERAL("\r\n")); + } + /* ...but we always print any `instruction'. */ + if (p->instruction) { + ptrlen plinst = ptrlen_from_asciz(p->instruction); + term_write(term, plinst); + if (!ptrlen_endswith(plinst, PTRLEN_LITERAL("\n"), NULL)) + term_write(term, PTRLEN_LITERAL("\r\n")); + } + /* + * Zero all the results, in case we abort half-way through. + */ + { + int i; + for (i = 0; i < (int)p->n_prompts; i++) + prompt_set_result(p->prompts[i], ""); + } + } + + while (s->curr_prompt < p->n_prompts) { + + prompt_t *pr = p->prompts[s->curr_prompt]; + bool finished_prompt = false; + + if (!s->done_prompt) { + term_write(term, ptrlen_from_asciz(pr->prompt)); + s->done_prompt = true; + } + + /* Breaking out here ensures that the prompt is printed even + * if we're now waiting for user data. */ + if (!ldisc_has_input_buffered(term->ldisc)) + break; + + /* FIXME: should we be using local-line-editing code instead? */ + while (!finished_prompt && ldisc_has_input_buffered(term->ldisc)) { + LdiscInputToken tok = ldisc_get_input_token(term->ldisc); + + char c; + if (tok.is_special) { + switch (tok.code) { + case SS_EOL: c = 13; break; + case SS_EC: c = 8; break; + case SS_IP: c = 3; break; + case SS_EOF: c = 3; break; + default: continue; + } + } else { + c = tok.chr; + } + + switch (c) { + case 10: + case 13: + term_write(term, PTRLEN_LITERAL("\r\n")); + /* go to next prompt, if any */ + s->curr_prompt++; + s->done_prompt = false; + finished_prompt = true; /* break out */ + break; + case 8: + case 127: + if (pr->result->len > 0) { + if (pr->echo) + term_write(term, PTRLEN_LITERAL("\b \b")); + strbuf_shrink_by(pr->result, 1); + } + break; + case 21: + case 27: + while (pr->result->len > 0) { + if (pr->echo) + term_write(term, PTRLEN_LITERAL("\b \b")); + strbuf_shrink_by(pr->result, 1); + } + break; + case 3: + case 4: + /* Immediate abort. */ + term_write(term, PTRLEN_LITERAL("\r\n")); + sfree(s); + p->data = NULL; + return signal_prompts_t(term, p, SPR_USER_ABORT); + default: + /* + * This simplistic check for printability is disabled + * when we're doing password input, because some people + * have control characters in their passwords. + */ + if (!pr->echo || (c >= ' ' && c <= '~') || + ((unsigned char) c >= 160)) { + put_byte(pr->result, c); + if (pr->echo) + term_write(term, make_ptrlen(&c, 1)); + } + break; + } + } + + } + + if (s->curr_prompt < p->n_prompts) { + ldisc_enable_prompt_callback(term->ldisc, p); + return SPR_INCOMPLETE; + } else { + sfree(s); + p->data = NULL; + return signal_prompts_t(term, p, SPR_OK); + } +} + +void term_notify_minimised(Terminal *term, bool minimised) +{ + term->minimised = minimised; +} + +void term_notify_palette_changed(Terminal *term) +{ + palette_reset(term, true); +} + +void term_notify_window_pos(Terminal *term, int x, int y) +{ + term->winpos_x = x; + term->winpos_y = y; +} + +void term_notify_window_size_pixels(Terminal *term, int x, int y) +{ + term->winpixsize_x = x; + term->winpixsize_y = y; +} diff --git a/terminal/terminal.h b/terminal/terminal.h new file mode 100644 index 00000000..3f918b22 --- /dev/null +++ b/terminal/terminal.h @@ -0,0 +1,563 @@ +/* + * Internals of the Terminal structure, for those other modules + * which need to look inside it. It would be nice if this could be + * folded back into terminal.c in future, with an abstraction layer + * to handle everything that other modules need to know about it; + * but for the moment, this will do. + */ + +#ifndef PUTTY_TERMINAL_H +#define PUTTY_TERMINAL_H + +#include "tree234.h" + +struct beeptime { + struct beeptime *next; + unsigned long ticks; +}; + +#define TRUST_SIGIL_WIDTH 3 +#define TRUST_SIGIL_CHAR 0xDFFE + +typedef struct { + int y, x; +} pos; + +typedef struct termchar termchar; +typedef struct termline termline; + +struct termchar { + /* + * Any code in terminal.c which definitely needs to be changed + * when extra fields are added here is labelled with a comment + * saying FULL-TERMCHAR. + */ + unsigned long chr; + unsigned long attr; + truecolour truecolour; + + /* + * The cc_next field is used to link multiple termchars + * together into a list, so as to fit more than one character + * into a character cell (Unicode combining characters). + * + * cc_next is a relative offset into the current array of + * termchars. I.e. to advance to the next character in a list, + * one does `tc += tc->next'. + * + * Zero means end of list. + */ + int cc_next; +}; + +struct termline { + unsigned short lattr; + int cols; /* number of real columns on the line */ + int size; /* number of allocated termchars + * (cc-lists may make this > cols) */ + bool temporary; /* true if decompressed from scrollback */ + int cc_free; /* offset to first cc in free list */ + struct termchar *chars; + bool trusted; +}; + +struct bidi_cache_entry { + int width; + bool trusted; + struct termchar *chars; + int *forward, *backward; /* the permutations of line positions */ +}; + +struct term_utf8_decode { + int state; /* Is there a pending UTF-8 character */ + int chr; /* and what is it so far? */ + int size; /* The size of the UTF character. */ +}; + +struct terminal_tag { + + int compatibility_level; + + tree234 *scrollback; /* lines scrolled off top of screen */ + tree234 *screen; /* lines on primary screen */ + tree234 *alt_screen; /* lines on alternate screen */ + int disptop; /* distance scrolled back (0 or -ve) */ + int tempsblines; /* number of lines of .scrollback that + can be retrieved onto the terminal + ("temporary scrollback") */ + + termline **disptext; /* buffer of text on real screen */ + int dispcursx, dispcursy; /* location of cursor on real screen */ + int curstype; /* type of cursor on real screen */ + +#define VBELL_TIMEOUT (TICKSPERSEC/10) /* visual bell lasts 1/10 sec */ + + struct beeptime *beephead, *beeptail; + int nbeeps; + bool beep_overloaded; + long lastbeep; + +#define TTYPE termchar +#define TSIZE (sizeof(TTYPE)) + + int default_attr, curr_attr, save_attr; + truecolour curr_truecolour, save_truecolour; + termchar basic_erase_char, erase_char; + + bufchain inbuf; /* terminal input buffer */ + + pos curs; /* cursor */ + pos savecurs; /* saved cursor position */ + int marg_t, marg_b; /* scroll margins */ + bool dec_om; /* DEC origin mode flag */ + bool wrap, wrapnext; /* wrap flags */ + bool insert; /* insert-mode flag */ + int cset; /* 0 or 1: which char set */ + int save_cset, save_csattr; /* saved with cursor position */ + bool save_utf, save_wnext; /* saved with cursor position */ + bool rvideo; /* global reverse video flag */ + unsigned long rvbell_startpoint; /* for ESC[?5hESC[?5l vbell */ + bool cursor_on; /* cursor enabled flag */ + bool reset_132; /* Flag ESC c resets to 80 cols */ + bool use_bce; /* Use Background coloured erase */ + bool cblinker; /* When blinking is the cursor on ? */ + bool tblinker; /* When the blinking text is on */ + bool blink_is_real; /* Actually blink blinking text */ + int sco_acs, save_sco_acs; /* CSI 10,11,12m -> OEM charset */ + bool vt52_bold; /* Force bold on non-bold colours */ + bool utf; /* Are we in toggleable UTF-8 mode? */ + term_utf8_decode utf8; /* If so, here's our decoding state */ + bool printing, only_printing; /* Are we doing ANSI printing? */ + int print_state; /* state of print-end-sequence scan */ + bufchain printer_buf; /* buffered data for printer */ + printer_job *print_job; + + /* ESC 7 saved state for the alternate screen */ + pos alt_savecurs; + int alt_save_attr; + truecolour alt_save_truecolour; + int alt_save_cset, alt_save_csattr; + bool alt_save_utf; + bool alt_save_wnext; + int alt_save_sco_acs; + + int rows, cols, savelines; + bool has_focus; + bool in_vbell; + long vbell_end; + bool app_cursor_keys, app_keypad_keys, vt52_mode; + bool repeat_off, srm_echo, cr_lf_return; + bool seen_disp_event; + bool big_cursor; + + bool xterm_mouse_forbidden; + int xterm_mouse; /* send mouse messages to host */ + bool xterm_extended_mouse; + bool urxvt_extended_mouse; + int mouse_is_down; /* used while tracking mouse buttons */ + + bool bracketed_paste, bracketed_paste_active; + + int cset_attr[2]; + +/* + * Saved settings on the alternate screen. + */ + int alt_x, alt_y; + bool alt_wnext, alt_ins; + bool alt_om, alt_wrap; + int alt_cset, alt_sco_acs; + bool alt_utf; + int alt_t, alt_b; + int alt_which; + int alt_sblines; /* # of lines on alternate screen that should be used for scrollback. */ + +#define ARGS_MAX 32 /* max # of esc sequence arguments */ +#define ARG_DEFAULT 0 /* if an arg isn't specified */ +#define def(a,d) ( (a) == ARG_DEFAULT ? (d) : (a) ) + unsigned esc_args[ARGS_MAX]; + int esc_nargs; + int esc_query; +#define ANSI(x,y) ((x)+((y)*256)) +#define ANSI_QUE(x) ANSI(x,1) + +#define OSC_STR_MAX 2048 + int osc_strlen; + char osc_string[OSC_STR_MAX + 1]; + bool osc_w; + + char id_string[1024]; + + unsigned char *tabs; + + enum { + TOPLEVEL, + SEEN_ESC, + SEEN_CSI, + SEEN_OSC, + SEEN_OSC_W, + + DO_CTRLS, + + SEEN_OSC_P, + OSC_STRING, OSC_MAYBE_ST, OSC_MAYBE_ST_UTF8, + VT52_ESC, + VT52_Y1, + VT52_Y2, + VT52_FG, + VT52_BG + } termstate; + + enum { + NO_SELECTION, ABOUT_TO, DRAGGING, SELECTED + } selstate; + enum { + LEXICOGRAPHIC, RECTANGULAR + } seltype; + enum { + SM_CHAR, SM_WORD, SM_LINE + } selmode; + pos selstart, selend, selanchor; + + short wordness[256]; + + /* Mask of attributes to pay attention to when painting. */ + int attr_mask; + + wchar_t *paste_buffer; + int paste_len, paste_pos; + + Backend *backend; + + Ldisc *ldisc; + + TermWin *win; + + LogContext *logctx; + + struct unicode_data *ucsdata; + + unsigned long last_graphic_char; + + /* + * We maintain a full copy of a Conf here, not merely a pointer + * to it. That way, when we're passed a new one for + * reconfiguration, we can check the differences and adjust the + * _current_ setting of (e.g.) auto wrap mode rather than only + * the default. + */ + Conf *conf; + + /* + * GUI implementations of seat_output call term_out, but it can + * also be called from the ldisc if the ldisc is called _within_ + * term_out. So we have to guard against re-entrancy - if + * seat_output is called recursively like this, it will simply add + * data to the end of the buffer term_out is in the process of + * working through. + */ + bool in_term_out; + + /* + * We don't permit window updates too close together, to avoid CPU + * churn pointlessly redrawing the window faster than the user can + * read. So after an update, we set window_update_cooldown = true + * and schedule a timer to reset it to false. In between those + * times, window updates are not performed, and instead we set + * window_update_pending = true, which will remind us to perform + * the deferred redraw when the cooldown period ends and + * window_update_cooldown is reset to false. + */ + bool window_update_pending, window_update_cooldown; + long window_update_cooldown_end; + + /* + * Track pending blinks and tblinks. + */ + bool tblink_pending, cblink_pending; + long next_tblink, next_cblink; + + /* + * These are buffers used by the bidi and Arabic shaping code. + */ + termchar *ltemp; + int ltemp_size; + bidi_char *wcFrom, *wcTo; + int wcFromTo_size; + struct bidi_cache_entry *pre_bidi_cache, *post_bidi_cache; + size_t bidi_cache_size; + + /* + * Current trust state, used to annotate every line of the + * terminal that a graphic character is output to. + */ + bool trusted; + + /* + * We copy a bunch of stuff out of the Conf structure into local + * fields in the Terminal structure, to avoid the repeated + * tree234 lookups which would be involved in fetching them from + * the former every time. + */ + bool ansi_colour; + char *answerback; + int answerbacklen; + bool no_arabicshaping; + int beep; + bool bellovl; + int bellovl_n; + int bellovl_s; + int bellovl_t; + bool no_bidi; + bool bksp_is_delete; + bool blink_cur; + bool blinktext; + bool cjk_ambig_wide; + int conf_height; + int conf_width; + bool crhaslf; + bool erase_to_scrollback; + int funky_type, sharrow_type; + bool lfhascr; + bool logflush; + int logtype; + bool mouse_override; + bool nethack_keypad; + bool no_alt_screen; + bool no_applic_c; + bool no_applic_k; + bool no_dbackspace; + bool no_mouse_rep; + bool no_remote_charset; + bool no_remote_resize; + bool no_remote_wintitle; + bool no_remote_clearscroll; + bool rawcnp; + bool utf8linedraw; + bool rect_select; + int remote_qtitle_action; + bool rxvt_homeend; + bool scroll_on_disp; + bool scroll_on_key; + bool xterm_256_colour; + bool true_colour; + + wchar_t *last_selected_text; + int *last_selected_attr; + truecolour *last_selected_tc; + size_t last_selected_len; + int mouse_select_clipboards[N_CLIPBOARDS]; + int n_mouse_select_clipboards; + int mouse_paste_clipboard; + + char *window_title, *icon_title; + int wintitle_codepage, icontitle_codepage; + bool minimised; + + BidiContext *bidi_ctx; + + /* Multi-layered colour palette. The colours from Conf (plus the + * default xterm-256 ones that don't have Conf ids at all) have + * lowest priority, followed by platform overrides if any, + * followed by escape-sequence overrides during the session. */ + struct term_subpalette { + rgb values[OSC4_NCOLOURS]; + bool present[OSC4_NCOLOURS]; + } subpalettes[3]; +#define SUBPAL_CONF 0 +#define SUBPAL_PLATFORM 1 +#define SUBPAL_SESSION 2 + + /* The composite palette that we make out of the above */ + rgb palette[OSC4_NCOLOURS]; + + unsigned winpos_x, winpos_y, winpixsize_x, winpixsize_y; + + /* + * Assorted 'pending' flags for ancillary window changes performed + * in term_update. Generally, to trigger one of these operations, + * you set the pending flag and/or the parameters here, then call + * term_schedule_update. + */ + bool win_move_pending; + int win_move_pending_x, win_move_pending_y; + bool win_zorder_pending; + bool win_zorder_top; + bool win_minimise_pending; + bool win_minimise_enable; + bool win_maximise_pending; + bool win_maximise_enable; + bool win_title_pending, win_icon_title_pending; + bool win_pointer_shape_pending; + bool win_pointer_shape_raw; + bool win_refresh_pending; + bool win_scrollbar_update_pending; + bool win_palette_pending; + unsigned win_palette_pending_min, win_palette_pending_limit; + + /* + * Unlike the rest of the above 'pending' flags, the one for + * window resizing has to be more complicated, because it's very + * likely that a server sending a window-resize escape sequence is + * going to follow it up immediately with further terminal output + * that draws a full-screen application expecting the terminal to + * be the new size. + * + * So, once we've requested a window resize from the TermWin, we + * have to stop processing terminal data until we get back the + * notification that our window really has changed size (or until + * we find out that it's not going to). + * + * Hence, window resizes go through a small state machine with two + * different kinds of 'pending'. NEED_SEND is the state where + * we've received an escape sequence asking for a new size but not + * yet sent it to the TermWin via win_request_resize; AWAIT_REPLY + * is the state where we've sent it to the TermWin and are + * expecting a call back to term_size(). + * + * So _both_ of those 'pending' states inhibit terminal output + * processing. + * + * (Hence, once we're in either state, we should never handle + * another resize sequence, so the only possible path through this + * state machine is to get all the way back to the ground state + * before doing anything else interesting.) + */ + enum { + WIN_RESIZE_NO, WIN_RESIZE_NEED_SEND, WIN_RESIZE_AWAIT_REPLY + } win_resize_pending; + int win_resize_pending_w, win_resize_pending_h; +}; + +static inline bool in_utf(Terminal *term) +{ + return term->utf || term->ucsdata->line_codepage == CP_UTF8; +} + +unsigned long term_translate( + Terminal *term, term_utf8_decode *utf8, unsigned char c); +static inline int term_char_width(Terminal *term, unsigned int c) +{ + return term->cjk_ambig_wide ? mk_wcwidth_cjk(c) : mk_wcwidth(c); +} + +/* + * UCSINCOMPLETE is returned from term_translate if it's successfully + * absorbed a byte but not emitted a complete character yet. + * UCSTRUNCATED indicates a truncated multibyte sequence (so the + * caller emits an error character and then calls term_translate again + * with the same input byte). UCSINVALID indicates some other invalid + * multibyte sequence, such as an overlong synonym, or a standalone + * continuation byte, or a completely illegal thing like 0xFE. These + * values are not stored in the terminal data structures at all. + */ +#define UCSINCOMPLETE 0x8000003FU /* '?' */ +#define UCSTRUNCATED 0x80000021U /* '!' */ +#define UCSINVALID 0x8000002AU /* '*' */ + +/* + * Maximum number of combining characters we're willing to store in a + * character cell. Our linked-list data representation permits an + * unlimited number of these in principle, but if we allowed that in + * practice then it would be an easy DoS to just squirt a squillion + * identical combining characters to someone's terminal and cause + * their PuTTY or pterm to consume lots of memory and CPU pointlessly. + * + * The precise figure of 32 is more or less arbitrary, but one point + * supporting it is UAX #15's comment that 30 combining characters is + * "significantly beyond what is required for any linguistic or + * technical usage". + */ +#define CC_LIMIT 32 + +/* ---------------------------------------------------------------------- + * Helper functions for dealing with the small 'pos' structure. + */ + +static inline bool poslt(pos p1, pos p2) +{ + if (p1.y != p2.y) + return p1.y < p2.y; + return p1.x < p2.x; +} + +static inline bool posle(pos p1, pos p2) +{ + if (p1.y != p2.y) + return p1.y < p2.y; + return p1.x <= p2.x; +} + +static inline bool poseq(pos p1, pos p2) +{ + return p1.y == p2.y && p1.x == p2.x; +} + +static inline int posdiff_fn(pos p1, pos p2, int cols) +{ + return (p1.y - p2.y) * (cols+1) + (p1.x - p2.x); +} + +/* Convenience wrapper on posdiff_fn which uses the 'Terminal *term' + * that more or less every function in terminal.c will have in scope. + * For safety's sake I include a TYPECHECK that ensures it really is a + * structure pointer of the right type. */ +#define GET_TERM_COLS TYPECHECK(term == (Terminal *)0, term->cols) +#define posdiff(p1,p2) posdiff_fn(p1, p2, GET_TERM_COLS) + +/* Product-order comparisons for rectangular block selection. */ + +static inline bool posPle(pos p1, pos p2) +{ + return p1.y <= p2.y && p1.x <= p2.x; +} + +static inline bool posPle_left(pos p1, pos p2) +{ + /* + * This function is used for checking whether a given character + * cell of the terminal ought to be highlighted as part of the + * selection, by comparing with term->selend. term->selend stores + * the location one space to the right of the last highlighted + * character. So we want to highlight the characters that are + * less-or-equal (in the product order) to the character just left + * of p2. + * + * (Setting up term->selend that way was the easiest way to get + * rectangular selection working at all, in a code base that had + * done lexicographic selection the way I happened to have done + * it.) + */ + return p1.y <= p2.y && p1.x < p2.x; +} + +static inline bool incpos_fn(pos *p, int cols) +{ + if (p->x == cols) { + p->x = 0; + p->y++; + return true; + } + p->x++; + return false; +} + +static inline bool decpos_fn(pos *p, int cols) +{ + if (p->x == 0) { + p->x = cols; + p->y--; + return true; + } + p->x--; + return false; +} + +/* Convenience wrappers on incpos and decpos which use term->cols + * (similarly to posdiff above), and also (for mild convenience and + * mostly historical inertia) let you leave off the & at every call + * site. */ +#define incpos(p) incpos_fn(&(p), GET_TERM_COLS) +#define decpos(p) decpos_fn(&(p), GET_TERM_COLS) + +#endif |