* dcrt0.cc (dll_crt0_1): Use GetCommandLineW and convert to current

	codepage.
	* environ.cc (set_file_api_mode): Always set file api to ANSI if not
	using the OEM codepage.
	(codepage_init): Allow "utf8" codepage.
	* fhandler_clipboard.cc (set_clipboard): Convert clipbuf to void and
	cast as needed.  Always convert input to wide char and write
	CF_UNICODETEXT to clipboard.
	(fhandler_dev_clipboard::read): Read CF_UNICODETEXT from clipboard and
	convert to current codepage if CYGWIN_NATIVE format is not available.
	* fhandler_console.cc: Drop redundant undef.
	* smallprint.cc (__small_vsprintf): Convert PWCHAR and UNICODE_STRING
	to current codepage for printing.
	* strfuncs.cc: Use PWCHAR throughout.
	(get_cp): Return CP_UTF8 for utf8_cp codepage setting.
	(sys_wcstombs): Allow NULL target buffer.
	(sys_wcstombs_alloc): New function.
	(sys_mbstowcs_alloc): Ditto.
	* winsup.h (codepage_type): Add utf8_cp.
	(HEAP_NOTHEAP): Define.
	(sys_wcstombs_alloc): Declare.
	(sys_mbstowcs_alloc): Declare.

1 files changed, 67 insertions, 4 deletions

diff --git a/winsup/cygwin/strfuncs.cc b/winsup/cygwin/strfuncs.cc
index 9d0e3ec08..0dac85e99 100644
--- a/winsup/cygwin/strfuncs.cc
+++ b/winsup/cygwin/strfuncs.cc
@@ -10,28 +10,44 @@ Cygwin license.  Please consult the file "CYGWIN_LICENSE" for
 details. */
 
 #include "winsup.h"
+#include <stdlib.h>
 #include <winbase.h>
 #include <winnls.h>
 #include <ntdll.h>
+#include "cygerrno.h"
+#include "security.h"
+#include "path.h"
+#include "fhandler.h"
+#include "dtable.h"
+#include "cygheap.h"
 
 codepage_type current_codepage = ansi_cp;
 
 UINT
 get_cp ()
 {
-  return current_codepage == ansi_cp ? GetACP() : GetOEMCP();
+  switch (current_codepage)
+    {
+      case oem_cp:
+	return GetOEMCP ();
+      case utf8_cp:
+	return CP_UTF8;
+      case ansi_cp:
+      default:
+	return GetACP ();
+    }
 }
 
 /* tlen is always treated as the maximum buffer size, including the '\0'
    character.  sys_wcstombs will always return a 0-terminated result, no
    matter what. */
 int __stdcall
-sys_wcstombs (char *tgt, int tlen, const WCHAR *src, int slen)
+sys_wcstombs (char *tgt, int tlen, const PWCHAR src, int slen)
 {
   int ret;
 
   ret = WideCharToMultiByte (get_cp (), 0, src, slen, tgt, tlen, NULL, NULL);
-  if (ret)
+  if (ret && tgt)
     {
       ret = (ret < tlen) ? ret : tlen - 1;
       tgt[ret] = '\0';
@@ -39,13 +55,60 @@ sys_wcstombs (char *tgt, int tlen, const WCHAR *src, int slen)
   return ret;
 }
 
+/* Allocate a buffer big enough for the string, always including the
+   terminating '\0'.  The buffer pointer is returned in *tgt_p, the return
+   value is the number of bytes written to the buffer, as usual.
+   The "type" argument determines where the resulting buffer is stored.
+   It's either one of the cygheap_types values, or it's "HEAP_NOTHEAP".
+   In the latter case the allocation uses simple calloc. */
+int __stdcall
+sys_wcstombs_alloc (char **tgt_p, int type, const PWCHAR src, int slen)
+{
+  int ret;
+
+  ret = WideCharToMultiByte (get_cp (), 0, src, slen, NULL, 0,NULL, NULL);
+  if (ret)
+    {
+      size_t tlen = (slen == -1 ? ret : ret + 1);
+
+      if (type == HEAP_NOTHEAP)
+        *tgt_p = (char *) calloc (tlen, sizeof (char));
+      else
+      	*tgt_p = (char *) ccalloc ((cygheap_types) type, tlen, sizeof (char));
+      if (!*tgt_p)
+        return 0;
+      ret = sys_wcstombs (*tgt_p, tlen, src, slen);
+    }
+  return ret;
+}
+
 int __stdcall
-sys_mbstowcs (WCHAR *tgt, const char *src, int len)
+sys_mbstowcs (PWCHAR tgt, const char *src, int len)
 {
   int res = MultiByteToWideChar (get_cp (), 0, src, -1, tgt, len);
   return res;
 }
 
+/* Same as sys_wcstombs_alloc, just backwards. */
+int __stdcall
+sys_mbstowcs_alloc (PWCHAR *tgt_p, int type, const char *src)
+{
+  int ret;
+
+  ret = MultiByteToWideChar (get_cp (), 0, src, -1, NULL, 0);
+  if (ret)
+    {
+      if (type == HEAP_NOTHEAP)
+        *tgt_p = (PWCHAR) calloc (ret, sizeof (WCHAR));
+      else
+      	*tgt_p = (PWCHAR) ccalloc ((cygheap_types) type, ret, sizeof (WCHAR));
+      if (!*tgt_p)
+        return 0;
+      ret = sys_mbstowcs (*tgt_p, src, ret);
+    }
+  return ret;
+}
+
 static WCHAR hex_wchars[] = L"0123456789abcdef";
 
 NTSTATUS NTAPI