From: Piotr Caban <piotr@codeweavers.com>
Subject: [PATCH 2/2 v2] msvcrt: Implement opening Unicode files with no BOM in _wsopen_dispatch.
Message-Id: <25263c53-3bc0-0043-a81b-511add2ecada@codeweavers.com>
Date: Thu, 26 Nov 2020 15:26:35 +0100


Spotted by Alistair Leslie-Hughes.

Signed-off-by: Piotr Caban <piotr@codeweavers.com>
---
  dlls/msvcrt/file.c       |  8 ++-----
  dlls/msvcrt/tests/file.c | 49 ++++++++++++++++++++++++++++++++++++++++
  2 files changed, 51 insertions(+), 6 deletions(-)

diff --git a/dlls/msvcrt/file.c b/dlls/msvcrt/file.c
index b3e0b4abeaa..c913f86df89 100644
--- a/dlls/msvcrt/file.c
+++ b/dlls/msvcrt/file.c
@@ -2174,17 +2174,15 @@ static int check_bom(HANDLE h, int oflags, BOOL seek)
     char bom[sizeof(utf8_bom)];
     DWORD r;
 
-    oflags &= ~(MSVCRT__O_WTEXT|MSVCRT__O_U16TEXT|MSVCRT__O_U8TEXT);
-
     if (!ReadFile(h, bom, sizeof(utf8_bom), &r, NULL))
         return oflags;
 
     if (r==sizeof(utf8_bom) && !memcmp(bom, utf8_bom, sizeof(utf8_bom))) {
-        oflags |= MSVCRT__O_U8TEXT;
+        oflags = (oflags & ~(MSVCRT__O_WTEXT | MSVCRT__O_U16TEXT)) | MSVCRT__O_U8TEXT;
     }else if (r>=sizeof(utf16_bom) && !memcmp(bom, utf16_bom, sizeof(utf16_bom))) {
         if (seek && r>2)
             SetFilePointer(h, 2, NULL, FILE_BEGIN);
-        oflags |= MSVCRT__O_U16TEXT;
+        oflags = (oflags & ~(MSVCRT__O_WTEXT | MSVCRT__O_U8TEXT)) | MSVCRT__O_U16TEXT;
     }else if (seek) {
         SetFilePointer(h, 0, NULL, FILE_BEGIN);
     }
@@ -2284,8 +2282,6 @@ int CDECL MSVCRT__wsopen_dispatch( const MSVCRT_wchar_t* path, int oflags, int s
           oflags = check_bom(hand, oflags, FALSE);
           CloseHandle(hand);
       }
-      else
-          oflags &= ~(MSVCRT__O_WTEXT|MSVCRT__O_U16TEXT|MSVCRT__O_U8TEXT);
   }
 
   hand = CreateFileW(path, access, sharing, &sa, creation, attrib, 0);
diff --git a/dlls/msvcrt/tests/file.c b/dlls/msvcrt/tests/file.c
index fc6dacd71ff..a413d674164 100644
--- a/dlls/msvcrt/tests/file.c
+++ b/dlls/msvcrt/tests/file.c
@@ -971,6 +971,38 @@ static void test_fgetwc_unicode(void)
     ok(ch == WEOF, "got %04hx, expected WEOF (unicode)\n", ch);
     fclose(tempfh);
 
+    tempfh = fopen(tempfile, "r,ccs=utf-8");
+    ok(tempfh != NULL, "can't open tempfile\n");
+    for (i = 1; i < ARRAY_SIZE(wchar_text); i++)
+    {
+        ch = fgetwc(tempfh);
+        ok(ch == wchar_text[i],
+                "got %04hx, expected %04x (unicode[%d])\n", ch, wchar_text[i], i-1);
+    }
+    ch = fgetwc(tempfh);
+    ok(ch == WEOF, "got %04hx, expected WEOF (unicode)\n", ch);
+    fclose(tempfh);
+
+    tempfh = fopen(tempfile, "a,ccs=utf-16le");
+    ok(tempfh != NULL, "can't open tempfile\n");
+    ch = fputwc('a', tempfh);
+    ok(ch == 'a', "fputwc returned %x\n", ch);
+    fclose(tempfh);
+
+    tempfh = fopen(tempfile, "a+,ccs=utf-8");
+    ok(tempfh != NULL, "can't open tempfile\n");
+    for (i = 1; i < ARRAY_SIZE(wchar_text); i++)
+    {
+        ch = fgetwc(tempfh);
+        ok(ch == wchar_text[i],
+                "got %04hx, expected %04x (unicode[%d])\n", ch, wchar_text[i], i-1);
+    }
+    ch = fgetwc(tempfh);
+    ok(ch == 'a', "got %04x, expected 'a'\n", ch);
+    ch = fgetwc(tempfh);
+    ok(ch == WEOF, "got %04hx, expected WEOF (unicode)\n", ch);
+    fclose(tempfh);
+
     tempfh = fopen(tempfile, "wb");
     ok(tempfh != NULL, "can't open tempfile\n");
     ret = WideCharToMultiByte(CP_UTF8, 0, wchar_text, ARRAY_SIZE(wchar_text),
@@ -990,6 +1022,23 @@ static void test_fgetwc_unicode(void)
     ch = fgetwc(tempfh);
     ok(ch == WEOF, "got %04hx, expected WEOF (utf8)\n", ch);
     fclose(tempfh);
+
+    tempfh = fopen(tempfile, "wb");
+    ok(tempfh != NULL, "can't open tempfile\n");
+    fwrite(wchar_text+1, 1, sizeof(wchar_text)-1, tempfh);
+    fclose(tempfh);
+
+    tempfh = fopen(tempfile, "rt,ccs=utf-16le");
+    ok(tempfh != NULL, "can't open tempfile\n");
+    for (i = 1; i < ARRAY_SIZE(wchar_text); i++)
+    {
+        ch = fgetwc(tempfh);
+        ok(ch == wchar_text[i],
+                "got %04hx, expected %04x (unicode[%d])\n", ch, wchar_text[i], i-1);
+    }
+    ch = fgetwc(tempfh);
+    ok(ch == WEOF, "got %04hx, expected WEOF (unicode)\n", ch);
+    fclose(tempfh);
     unlink(temppath);
 }