From 12fdf9069abe3cd2250a9efec6e059eb85ec59d8 Mon Sep 17 00:00:00 2001
From: Harley Acheson <harley.acheson@gmail.com>
Date: Tue, 27 Sep 2022 08:39:24 -0700
Subject: BLF: Editing Text with Combining Characters

Corrections for caret insertion & movement and deletion for text
strings that include non-precomposed diacritical marks (Unicode
combining characters).

See D15659 for more details and examples.

Differential Revision: https://developer.blender.org/D15659

Reviewed by Campbell Barton
---
 source/blender/blenlib/intern/string_cursor_utf8.c | 59 +++++++++++++---------
 1 file changed, 35 insertions(+), 24 deletions(-)

(limited to 'source/blender/blenlib/intern')

diff --git a/source/blender/blenlib/intern/string_cursor_utf8.c b/source/blender/blenlib/intern/string_cursor_utf8.c
index 6b7151be969..2405b134428 100644
--- a/source/blender/blenlib/intern/string_cursor_utf8.c
+++ b/source/blender/blenlib/intern/string_cursor_utf8.c
@@ -96,27 +96,35 @@ static eStrCursorDelimType cursor_delim_type_utf8(const char *ch_utf8,
   return cursor_delim_type_unicode(uch);
 }
 
+/* Keep in sync with BLI_str_cursor_step_next_utf32. */
 bool BLI_str_cursor_step_next_utf8(const char *str, size_t maxlen, int *pos)
 {
+  if ((*pos) >= (int)maxlen) {
+    return false;
+  }
   const char *str_end = str + (maxlen + 1);
   const char *str_pos = str + (*pos);
-  const char *str_next = BLI_str_find_next_char_utf8(str_pos, str_end);
-  if (str_next != str_end) {
-    (*pos) += (str_next - str_pos);
-    if ((*pos) > (int)maxlen) {
-      (*pos) = (int)maxlen;
-    }
-    return true;
+  const char *str_next = str_pos;
+  do {
+    str_next = BLI_str_find_next_char_utf8(str_next, str_end);
+  } while (str_next < str_end && str_next[0] != 0 && BLI_str_utf8_char_width(str_next) < 1);
+  (*pos) += (str_next - str_pos);
+  if ((*pos) > (int)maxlen) {
+    (*pos) = (int)maxlen;
   }
 
-  return false;
+  return true;
 }
 
-bool BLI_str_cursor_step_prev_utf8(const char *str, size_t UNUSED(maxlen), int *pos)
+/* Keep in sync with BLI_str_cursor_step_prev_utf32. */
+bool BLI_str_cursor_step_prev_utf8(const char *str, size_t maxlen, int *pos)
 {
-  if ((*pos) > 0) {
+  if ((*pos) > 0 && (*pos) <= maxlen) {
     const char *str_pos = str + (*pos);
-    const char *str_prev = BLI_str_find_prev_char_utf8(str_pos, str);
+    const char *str_prev = str_pos;
+    do {
+      str_prev = BLI_str_find_prev_char_utf8(str_prev, str);
+    } while (str_prev > str && BLI_str_utf8_char_width(str_prev) == 0);
     (*pos) -= (str_pos - str_prev);
     return true;
   }
@@ -202,26 +210,29 @@ void BLI_str_cursor_step_utf8(const char *str,
   }
 }
 
-/* UTF32 version of BLI_str_cursor_step_utf8 (keep in sync!)
- * less complex since it doesn't need to do multi-byte stepping.
- */
-
-/* Helper functions so we can match #BLI_str_cursor_step_utf8. */
-static bool cursor_step_next_utf32(const char32_t *UNUSED(str), size_t maxlen, int *pos)
+/* Keep in sync with BLI_str_cursor_step_next_utf8. */
+bool BLI_str_cursor_step_next_utf32(const char32_t *str, size_t maxlen, int *pos)
 {
   if ((*pos) >= (int)maxlen) {
     return false;
   }
-  (*pos)++;
+  do {
+    (*pos)++;
+  } while (*pos < (int)maxlen && str[*pos] != 0 && BLI_wcwidth(str[*pos]) == 0);
+
   return true;
 }
 
-static bool cursor_step_prev_utf32(const char32_t *UNUSED(str), size_t UNUSED(maxlen), int *pos)
+/* Keep in sync with BLI_str_cursor_step_prev_utf8. */
+bool BLI_str_cursor_step_prev_utf32(const char32_t *str, size_t UNUSED(maxlen), int *pos)
 {
   if ((*pos) <= 0) {
     return false;
   }
-  (*pos)--;
+  do {
+    (*pos)--;
+  } while (*pos > 0 && BLI_wcwidth(str[*pos]) == 0);
+
   return true;
 }
 
@@ -236,7 +247,7 @@ void BLI_str_cursor_step_utf32(const char32_t *str,
 
   if (direction == STRCUR_DIR_NEXT) {
     if (use_init_step) {
-      cursor_step_next_utf32(str, maxlen, pos);
+      BLI_str_cursor_step_next_utf32(str, maxlen, pos);
     }
     else {
       BLI_assert(jump == STRCUR_JUMP_DELIM);
@@ -250,7 +261,7 @@ void BLI_str_cursor_step_utf32(const char32_t *str,
        * look at function cursor_delim_type_unicode() for complete
        * list of special character, ctr -> */
       while ((*pos) < maxlen) {
-        if (cursor_step_next_utf32(str, maxlen, pos)) {
+        if (BLI_str_cursor_step_next_utf32(str, maxlen, pos)) {
           if ((jump != STRCUR_JUMP_ALL) &&
               (delim_type != cursor_delim_type_unicode((uint)str[*pos]))) {
             break;
@@ -264,7 +275,7 @@ void BLI_str_cursor_step_utf32(const char32_t *str,
   }
   else if (direction == STRCUR_DIR_PREV) {
     if (use_init_step) {
-      cursor_step_prev_utf32(str, maxlen, pos);
+      BLI_str_cursor_step_prev_utf32(str, maxlen, pos);
     }
     else {
       BLI_assert(jump == STRCUR_JUMP_DELIM);
@@ -279,7 +290,7 @@ void BLI_str_cursor_step_utf32(const char32_t *str,
        * list of special character, ctr -> */
       while ((*pos) > 0) {
         const int pos_prev = *pos;
-        if (cursor_step_prev_utf32(str, maxlen, pos)) {
+        if (BLI_str_cursor_step_prev_utf32(str, maxlen, pos)) {
           if ((jump != STRCUR_JUMP_ALL) &&
               (delim_type != cursor_delim_type_unicode((uint)str[*pos]))) {
             /* left only: compensate for index/change in direction */
-- 
cgit v1.2.3