[PATCH 0/2] MR5866: kernel32, dwrite: Tweak some tests to reduce test report sizes.

[PATCH 1/2] kernel32/tests: Break debugger loop on unexpected result.

[Rémi Bernon]

From: Rémi Bernon

Avoid infinite loop that sometimes happen on macOS and make the test
spew hundred of thousand of "Test failed" lines.
---
 dlls/kernel32/tests/debugger.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/dlls/kernel32/tests/debugger.c b/dlls/kernel32/tests/debugger.c
index 301714342e4..924a998bd3d 100644
--- a/dlls/kernel32/tests/debugger.c
+++ b/dlls/kernel32/tests/debugger.c
@@ -2190,8 +2190,10 @@ static void test_debugger(const char *argv0)
     {
         next_event(&ctx, WAIT_EVENT_TIMEOUT);
         ok (ctx.ev.dwDebugEventCode != EXCEPTION_DEBUG_EVENT, "got exception\n");
+        if (ctx.ev.dwDebugEventCode != EXCEPTION_DEBUG_EVENT) break;
     }
     while (ctx.ev.dwDebugEventCode != EXIT_PROCESS_DEBUG_EVENT);
+    if (ctx.ev.dwDebugEventCode != EXIT_PROCESS_DEBUG_EVENT) TerminateProcess(pi.hProcess, 0);

ret = CloseHandle(event);
     ok(ret, "CloseHandle failed, last error %ld.\n", GetLastError());

[PATCH 2/2] dwrite/tests: Ignore macOS specific "flip" sbix format.

[Rémi Bernon]

From: Rémi Bernon

---
 dlls/dwrite/tests/font.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/dlls/dwrite/tests/font.c b/dlls/dwrite/tests/font.c
index e2b5b9163f1..4bd193af4cb 100644
--- a/dlls/dwrite/tests/font.c
+++ b/dlls/dwrite/tests/font.c
@@ -9297,6 +9297,9 @@ static DWORD get_sbix_formats(IDWriteFontFace4 *fontface)
             case MS_TIFF_TAG:
                 ret |= DWRITE_GLYPH_IMAGE_FORMATS_TIFF;
                 break;
+            case DWRITE_MAKE_OPENTYPE_TAG('f','l','i','p'):
+                /* ignore macOS-specific tag */
+                break;
             default:
                 ok(0, "unexpected format, %#lx\n", GET_BE_DWORD(format));
             }

[Nikolay Sivov (＠nsivov)]

Looks like some new addition that has no directwrite equivalent. I could find that fonttools added support for it earlier this year, so it must be useful for something.

[PATCH 0/2] MR5866: kernel32, dwrite: Tweak some tests to reduce test report sizes. - approved

[Nikolay Sivov (＠nsivov)]

This merge request was approved by Nikolay Sivov.

[Rémi Bernon (＠rbernon)]

Fwiw I don't think the test failures are related. There's some `gdi32:brush` Linux failures which are likely not, and the kernel32 tests fail because I have an accent in my name that ends up in the GIT environment variables and which the kernel32 environ tests don't like, and which break the Gitlab runner text output (see https://gitlab.winehq.org/wine/wine/-/merge_requests/4362#note_53581).

[eric pouech (＠epo)]

eric pouech (@epo) commented about dlls/kernel32/tests/debugger.c:
>  {
>      next_event(&ctx, WAIT_EVENT_TIMEOUT);
>      ok (ctx.ev.dwDebugEventCode != EXCEPTION_DEBUG_EVENT, "got exception\n");
> -
>    if (ctx.ev.dwDebugEventCode != EXCEPTION_DEBUG_EVENT) break;
> }
>    while (ctx.ev.dwDebugEventCode != EXIT_PROCESS_DEBUG_EVENT);
> - if (ctx.ev.dwDebugEventCode != EXIT_PROCESS_DEBUG_EVENT) TerminateProcess(pi.hProcess, 0);

If the thousand of "Test failed" lines come from the ok() inside the above loop, I wonder if the test to break out of the loop isn't inverted?
Otherwise you just skip the whole teardown, making the whole loop not useful.

[Rémi Bernon (＠rbernon)]

On Tue Jun 18 11:16:55 2024 +0000, eric pouech wrote:
> If the thousand of "Test failed" lines come from the ok() inside the
> above loop, I wonder if the test to break out of the loop isn't inverted?
> Otherwise you just skip the whole teardown, making the whole loop not useful.

Oh yeah indeed... thanks