[PATCH 1/2] comctl32/tests: Fix a test failure on Win10 21H1 with Hebrew locale.

[Zhiyi Zhang]

Signed-off-by: Zhiyi Zhang <zzhang(a)codeweavers.com>
---
 dlls/comctl32/tests/propsheet.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/dlls/comctl32/tests/propsheet.c b/dlls/comctl32/tests/propsheet.c
index 9acdea85c66..04a540b07e6 100644
--- a/dlls/comctl32/tests/propsheet.c
+++ b/dlls/comctl32/tests/propsheet.c
@@ -78,8 +78,8 @@ static int CALLBACK sheet_callback(HWND hwnd, UINT msg, LPARAM lparam)
         size = SizeofResource(module, hrsrc);
         ok(size != 0, "Failed to get size of propsheet dialog resource\n");
         buffer_size = HeapSize(GetProcessHeap(), 0, (void *)lparam);
-        ok(buffer_size == 2 * size, "Unexpected template buffer size %u, resource size %u\n",
-                buffer_size, size);
+        ok(buffer_size == 2 * size || buffer_size == 2 * size + 8 /* Win10 21H1 Hebrew (Israel) */,
+           "Unexpected template buffer size %u, resource size %u\n", buffer_size, size);
         break;
       }
     case PSCB_INITIALIZED:
-- 
2.32.0

[Marvin]

Hi,

While running your changed tests, I think I found new failures.
Being a bot and all I'm not very good at pattern recognition, so I might be
wrong, but could you please double-check?

Full results can be found at:
https://testbot.winehq.org/JobDetails.pl?Key=102697

Your paranoid android.


=== w1064_tsign (64 bit report) ===

comctl32:
propsheet.c:821: Test failed: property sheet with custom window proc: the msg 0x001c was expected, but got msg 0x0047 instead

[Francois Gouget]

I see that this patch has not been applied. Maybe because:

On Thu, 25 Nov 2021, Marvin wrote:
[...]
> === w1064_tsign (64 bit report) ===
>
> comctl32:
> propsheet.c:821: Test failed: property sheet with custom window proc: the msg 0x001c was expected, but got msg 0x0047 instead

This is an unrelated preexisting but random failure.
So I think it should not prevent the patch from being applied.

The patch itself fixes a systematic failure in the Hebrew locale:
https://bugs.winehq.org//show_bug.cgi?id=52195

That said, maybe the reason the patch was not applied is because we 
don't know why Windows returns an extra 8 bytes specifically in the 
Hebrew locale? Or maybe we want to consider this behavior broken?

-- 
Francois Gouget <fgouget(a)codeweavers.com>

[Zhiyi Zhang]

On 12/9/21 22:58, Francois Gouget wrote:
> I see that this patch has not been applied. Maybe because:
>
> On Thu, 25 Nov 2021, Marvin wrote:
> [...]
> > === w1064_tsign (64 bit report) ===
> >
> > comctl32:
> > propsheet.c:821: Test failed: property sheet with custom window proc: the msg 0x001c was expected, but got msg 0x0047 instead
> > This is an unrelated preexisting but random failure.
> > So I think it should not prevent the patch from being applied.

The random message failure was fixed by 3c789e2, which was the 2/2 patch of this series.
> The patch itself fixes a systematic failure in the Hebrew locale:
> https://bugs.winehq.org//show_bug.cgi?id=52195
>
> That said, maybe the reason the patch was not applied is because we 
> don't know why Windows returns an extra 8 bytes specifically in the 
> Hebrew locale? Or maybe we want to consider this behavior broken?

[Francois Gouget]

On Thu, 9 Dec 2021, Zhiyi Zhang wrote:
[...]
> > > comctl32:
> > > propsheet.c:821: Test failed: property sheet with custom window proc: the msg 0x001c was expected, but got msg 0x0047 instead
> > > This is an unrelated preexisting but random failure.
> > > So I think it should not prevent the patch from being applied.
>
> The random message failure was fixed by 3c789e2, which was the 2/2 
> patch of this series.

Yes. That helped a lot. There was still a failure happening regularly on 
the cw-gtx560 machine in the Windows 10 1709 and 1809 cases (got 0x0046 
nistead of 0x0210) but I figured out it was because Windows was asking 
at boot about whether the machine should be discoverable on the network. 
That sidebar dialog caused the message ordering issues. So I dismissed 
the dialog and redid the snapshots and these failures are gone.

The same failure happened once after the patch on fgtb-w10pro64-rx550 
but I'll assume this was caused by some previous test crashing / leaving 
a dialog up. It has not happened again so I think comctl32:propsheet 
will be all clean once the Hebrew situation is resolved.

Hence my interest in getting this patch or some variant committed :-)


-- 
Francois Gouget <fgouget(a)codeweavers.com>