[Bug 9583] New: CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

Summary: CompareStringW gives incorrect result for some wide
                    strings
           Product: Wine
           Version: CVS/GIT
          Platform: Other
        OS/Version: other
            Status: UNCONFIRMED
          Severity: enhancement
          Priority: P2
         Component: wine-kernel
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

It seems like CompareStringW does not work correctly, when using wide/unicode
strings. I first discovered this when I saw that listbox controls sorted
swedish characters (say, LATIN SMALL LETTER A WITH RING ABOVE) side by side
with the base character (LATIN SMALL LETTER A). An example is
http://www.cendio.se/~astrand/wine/40-listbox-sort/list.exe.

The cause seems to be that CompareStringW gives the wrong result. I've created
the small test example
http://www.cendio.se/~astrand/wine/40-listbox-sort/comparestring.exe, with
source available in the same directory. On Windows, the result is that the
string "äpple" is greater than "orange", which is correct. With the latest
Wine CVS version, the result is that "äpple" is lesser than "orange", which is
wrong.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

Dmitry Timoshkov  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|CVS/GIT                     |0.9.44.

--- Comment #1 from Dmitry Timoshkov   2007-09-04 10:24:40 ---
What is your locale under Windows? Could you run your test with Windows locale
set to English and see if that changes the result?

(Note: the tests above are not Wine related)

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #2 from Peter Åstrand   2007-09-05 02:21:11 ---
My locale is 0x41d. If I change the locale in Windows to English (USA) (0x409),
it  changes the result, the string is now compared less instead. Here's a log:

H:>comparestring
locale is 0x41d
string1 is greater

H:>comparestring
locale is 0x409
string1 is less

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #3 from Dmitry Timoshkov   2007-09-05 02:31:05 ---
This means that Windows uses different sort weight tables depending on
locale, and that's not really a unicode sort (which is locale agnostic).

Does calling lstrcmpW show the same result?

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #4 from Dmitry Timoshkov   2007-09-05 02:52:38 ---
Also it would be interesting to see what sort keys LCMapStringW returns
in both locales and where they differ.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #5 from Dmitry Timoshkov   2007-09-05 03:04:31 ---
Another point that may make the testing easier: most likely just specifying
an appropriate locale (0x41d or 0x409) in the 1st argument of CompareStringW
and LCMapStringW is enough to avoid system locale changes.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #6 from Peter Åstrand   2007-09-05 05:50:52 ---
> Does calling lstrcmpW show the same result?

Yes. With locale 0x41d, under Windows, both gives "greater", while under Wine,
both gives "less".
> Also it would be interesting to see what sort keys LCMapStringW returns
> in both locales and where they differ.

I'm not familiar with LCMapStringW. What kind of flags should I call it with?
> Another point that may make the testing easier: most likely just specifying
> an appropriate locale (0x41d or 0x409) in the 1st argument of CompareStringW
> and LCMapStringW is enough to avoid system locale changes.

Since the test program prints out the locale at startup, there's no risk of
mistakes. Using GetUserDefaultLCID() is convenient, since it allows us to test
different locale just by changing the locale in the control panel, instead of
having to recompile.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #7 from Dmitry Timoshkov   2007-09-05 05:57:42 ---
> I'm not familiar with LCMapStringW. What kind of flags should I call it with?

LCMAP_SORTKEY
> > Another point that may make the testing easier: most likely just specifying
> > an appropriate locale (0x41d or 0x409) in the 1st argument of CompareStringW
> > and LCMapStringW is enough to avoid system locale changes.
>
> Since the test program prints out the locale at startup, there's no risk of
> mistakes. Using GetUserDefaultLCID() is convenient, since it allows us to test
> different locale just by changing the locale in the control panel, instead of
> having to recompile.

Passing explicit locale to CompareStringW and LCMapStringW makes sure that
you are testing real and not implicit behaviour. Also having a common testing
function that takes a locale id as a parameter helps to avoid recompilations.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #8 from Peter Åstrand   2007-09-05 07:37:10 ---
> LCMAP_SORTKEY

Ok, I've tried calling LCMapStringW now, in both locales, but I don't really
know how to interpret the result:

locale is 0x41d
CompareStringW: string1 is greater
lstrcmpW: string1 is greater
LCMapStringW result: 15
0xe 0xaf 0xe 0x7e 0xe 0x7e 0xe 0x48 0xe 0x21 0x1 0x1 0x1 0x1 0x0

locale is 0x409
CompareStringW: string1 is less
lstrcmpW: string1 is less
LCMapStringW result: 16
0xe 0x2 0xe 0x7e 0xe 0x7e 0xe 0x48 0xe 0x21 0x1 0x13 0x1 0x1 0x1 0x0

An updated source file is available on
http://www.cendio.se/~astrand/wine/40-listbox-sort/.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #9 from Dmitry Timoshkov   2007-09-05 23:12:11 ---
> Ok, I've tried calling LCMapStringW now, in both locales, but I don't really
> know how to interpret the result:

In order to make the interpretation of the result easier it would help to
call LCMapStringW for each character, and print original unicode code point
together with resulting sort key.
> An updated source file is available on
> http://www.cendio.se/~astrand/wine/40-listbox-sort/.

Looks like comparestring.cpp provided there isn't really the latest source.
Also please attach the tests here in the bug.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #10 from Peter Åstrand   2007-09-06 03:36:24 ---
> In order to make the interpretation of the result easier it would help to
> call LCMapStringW for each character, and print original unicode code point
> together with resulting sort key.

Perhaps it would be easier if you modify the test program yourself, the way you
want it?
> Looks like comparestring.cpp provided there isn't really the latest source.

Sorry about that, this should be fixed now.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #11 from Dmitry Timoshkov   2007-09-06 03:43:38 ---
I have no time at the moment to work on this, so if you feel like motivated
enough I'd prefer just provide additional help if needed.

First we need to understand what is the difference in behaviour, and how
locale (and which locale) affects it. Once investigated we will need to
start thinking about possible solutions.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #12 from Peter Åstrand   2007-09-06 07:13:22 ---
Ok, I've extend the test program now, to call LCMapStringW for each character.
Updated source and binary in the usual location. I've also tried running it
both under Wine and Windows, bith both US and Swedish locale. Here's the
result:

Windows
-------

locale is 0x409
CompareStringW: string1 is less
lstrcmpW: string1 is less
LCMapStringW result: 16
0xe 0x2 0xe 0x7e 0xe 0x7e 0xe 0x48 0xe 0x21 0x1 0x13 0x1 0x1 0x1 0x0

char: 0xe4, sortkeys: 0xe 0x2 0x1 0x13 0x1 0x1 0x1 0x0
char: 0x70, sortkeys: 0xe 0x7e 0x1 0x1 0x1 0x1 0x0
char: 0x70, sortkeys: 0xe 0x7e 0x1 0x1 0x1 0x1 0x0
char: 0x6c, sortkeys: 0xe 0x48 0x1 0x1 0x1 0x1 0x0
char: 0x65, sortkeys: 0xe 0x21 0x1 0x1 0x1 0x1 0x0
char: 0x0, sortkeys: 0x1 0x1 0x1 0x1 0x0

locale is 0x41d
CompareStringW: string1 is greater
lstrcmpW: string1 is greater
LCMapStringW result: 15
0xe 0xaf 0xe 0x7e 0xe 0x7e 0xe 0x48 0xe 0x21 0x1 0x1 0x1 0x1 0x0

char: 0xe4, sortkeys: 0xe 0xaf 0x1 0x1 0x1 0x1 0x0
char: 0x70, sortkeys: 0xe 0x7e 0x1 0x1 0x1 0x1 0x0
char: 0x70, sortkeys: 0xe 0x7e 0x1 0x1 0x1 0x1 0x0
char: 0x6c, sortkeys: 0xe 0x48 0x1 0x1 0x1 0x1 0x0
char: 0x65, sortkeys: 0xe 0x21 0x1 0x1 0x1 0x1 0x0
char: 0x0, sortkeys: 0x1 0x1 0x1 0x1 0x0

Wine
----

locale is 0x409
CompareStringW: string1 is less
lstrcmpW: string1 is less
LCMapStringW result: 29
0xa 0x15 0xb 0x67 0xb 0x67 0xb 0x3 0xa 0x65 0x1 0x2 0x2 0x2 0x2 0x2 0x1 0x2 0x2
0x2 0x2 0x2 0x1 0xe4 0x70 0x70 0x6c 0x65 0x1

char: 0xe4, sortkeys: 0xa 0x15 0x1 0x2 0x1 0x2 0x1 0xe4 0x1
char: 0x70, sortkeys: 0xb 0x67 0x1 0x2 0x1 0x2 0x1 0x70 0x1
char: 0x70, sortkeys: 0xb 0x67 0x1 0x2 0x1 0x2 0x1 0x70 0x1
char: 0x6c, sortkeys: 0xb 0x3 0x1 0x2 0x1 0x2 0x1 0x6c 0x1
char: 0x65, sortkeys: 0xa 0x65 0x1 0x2 0x1 0x2 0x1 0x65 0x1
char: 0x0, sortkeys: 0x1 0x1 0x1 0x1

locale is 0x41d
CompareStringW: string1 is less
lstrcmpW: string1 is less
LCMapStringW result: 29
0xa 0x15 0xb 0x67 0xb 0x67 0xb 0x3 0xa 0x65 0x1 0x2 0x2 0x2 0x2 0x2 0x1 0x2 0x2
0x2 0x2 0x2 0x1 0xe4 0x70 0x70 0x6c 0x65 0x1

char: 0xe4, sortkeys: 0xa 0x15 0x1 0x2 0x1 0x2 0x1 0xe4 0x1
char: 0x70, sortkeys: 0xb 0x67 0x1 0x2 0x1 0x2 0x1 0x70 0x1
char: 0x70, sortkeys: 0xb 0x67 0x1 0x2 0x1 0x2 0x1 0x70 0x1
char: 0x6c, sortkeys: 0xb 0x3 0x1 0x2 0x1 0x2 0x1 0x6c 0x1
char: 0x65, sortkeys: 0xa 0x65 0x1 0x2 0x1 0x2 0x1 0x65 0x1
char: 0x0, sortkeys: 0x1 0x1 0x1 0x1

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #13 from Austin English   2008-06-12 11:16:23 ---
Is this still an issue in current (1.0-rc4 or newer) wine?

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #14 from Peter Åstrand   2008-06-13 01:32:36 ---
Yes. Tested with the latest CVS version (2007-06-12).

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #15 from Austin English   2008-12-24 10:06:13 ---
Created an attachment (id=18173)
 --> (http://bugs.winehq.org/attachment.cgi?id=18173)
output under wine in git

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #16 from Austin English   2008-12-24 10:08:37 ---
Created an attachment (id=18174)
 --> (http://bugs.winehq.org/attachment.cgi?id=18174)
output under windows 2000

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
           Keywords|                            |download, source

--- Comment #17 from Austin English   2008-12-24 10:09:24 ---
http://www.cendio.com/~astrand/wine/40-listbox-sort/comparestring.cpp

Still present in git. I used the comparestring.exe.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |testcase

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                URL|                            |http://www.cendio.se/~astra
                   |                            |nd/wine/40-listbox-sort/com
                   |                            |parestring.exe

--- Comment #18 from Austin English   2010-06-13 04:49:18 ---
Still present in 1.2-rc3.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

http://bugs.winehq.org/show_bug.cgi?id=9583

Julian Rüger  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                URL|http://www.cendio.se/~astra |http://www.cendio.com/~astr
                   |nd/wine/40-listbox-sort/com |and/wine/40-listbox-sort/co
                   |parestring.exe              |mparestring.exe

--- Comment #19 from Austin English  ---
austin@aw25 ~ $ wine comparestring.exe 
locale is 0x409
CompareStringW: string1 is less
lstrcmpW: string1 is less
LCMapStringW result: 30
0xa 0x15 0xb 0x67 0xb 0x67 0xb 0x3 0xa 0x65 0x1 0x2 0x2 0x2 0x2 0x2 0x1 0x2 0x2
0x2 0x2 0x2 0x1 0xe4 0x70 0x70 0x6c 0x65 0x1 0x0

char: 0xe4, sortkeys: 0xa 0x15 0x1 0x2 0x1 0x2 0x1 0xe4 0x1 0x0 
char: 0x70, sortkeys: 0xb 0x67 0x1 0x2 0x1 0x2 0x1 0x70 0x1 0x0 
char: 0x70, sortkeys: 0xb 0x67 0x1 0x2 0x1 0x2 0x1 0x70 0x1 0x0 
char: 0x6c, sortkeys: 0xb 0x3 0x1 0x2 0x1 0x2 0x1 0x6c 0x1 0x0 
char: 0x65, sortkeys: 0xa 0x65 0x1 0x2 0x1 0x2 0x1 0x65 0x1 0x0 
char: 0x0, sortkeys: 0x1 0x1 0x1 0x1 0x0 
austin@aw25 ~ $ wine --version
wine-1.7.12

still present.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

Virgo Pärna  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #20 from Virgo Pärna  ---
Problem affects also CompareStringA. 
For testing I made Delphi 5 executable with:
CompareStringA(1061, 0, PChar('AA'), 2, PCHar('ÄÄ'), 2)
comparision. And it returns 1 in "Windows 7", but 2 in wine 1.4.1.
It seems, that wine does that comparsion Accent Insensitive way, which is
incorrect. 
And I suspect, that base issue is same with Ansi and Wide version. 
And this is major issue, because any sorting or character comparision using
those functions will work incorrectly.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #21 from Virgo Pärna  ---
Problem seems to be, that call to CompareStringEx passes NULL as Locale Name.
Which actually should be LOCALE_NAME_USER_DEFAULT, so in the case, when
GetUserDefaultLCID returns same value, it should work correctly. So it should
use call to LCIDToLocaleName to get proper locale name for passing to
CompareStringEx

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #22 from Virgo Pärna  ---
I'm not a C programmer, but as  far as can see:
Any call to CompareString or CompareStringEx ends up calling
wine_compare_string. And at that moment any locale information passed together
with CompareString or CompareStringEx call (as lcid or locale parameter) is
lost. 
wine_comparer_string ends up calling compare_unicode_weights. And
CompareString/CompareStringEx end up reporting strings equal only if that call
reports  them equal (returns 0).
And compare_unicode_weights uses collation_table. And that part of the code is
little bit too complex to me. But it appears that collation_table is compiled
in and is incorrect for this use.
Someone, that is better at C, can correct me, if I misunderstood something.
CompareStringEx is in locale.c
wine_compare_string and compare_unicode_weights are in sortkey.c 
collation_table is in collation.c

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #23 from Virgo Pärna  ---
This is starting to get quite complex.
If I understand correctly, then compare_unicode_weights is supposed to give
same result for A and Ä. But then compare_diacritic_weights should show the
difference. 
Just for testing I extracted relavent parts from wine and I made a test program
to see, what happens. And it seems that those lines:
        ce1 = collation_table[collation_table[*str1 >> 8] + (*str1 & 0xff)];
        ce2 = collation_table[collation_table[*str2 >> 8] + (*str2 & 0xff)];
give the same result for A and Ä.
And that part is exactly same in compare_unicode_weights and
compare_diacritic_weights and compare_case_weights. So that would always give
same results AFAIU.
And I really don't understand, what part of the
http://www.unicode.org/reports/tr10/allkeys.txt collation.c contains.
> From the allkeys.txt

0041  ; [.0A15.0020.0008.0041] # LATIN CAPITAL LETTER A
00C1  ; [.0A15.0020.0008.0041][.0000.0032.0002.0301] # LATIN CAPITAL LETTER A
WITH ACUTE; QQCM

How does this information translate to collation_table? My perl is
unfortunately even worse than my C. But as far as I understand the array is
generated in make_unicode script.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #24 from Virgo Pärna  ---
> 00C1  ; [.0A15.0020.0008.0041][.0000.0032.0002.0301] # LATIN CAPITAL LETTER
> A WITH ACUTE; QQCM

Öfcause Ä is actually 
00C4  ; [.0A15.0020.0008.0041][.0000.0047.0002.0308] # LATIN CAPITAL LETTER A
WITH DIAERESIS; QQCM

Anyway, 
collation_table gives 0x0A150151 for both A and Ä.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

Luke Bratch  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #25 from Virgo Pärna  ---
Created attachment 49623
  --> https://bugs.winehq.org/attachment.cgi?id=49623
Testcase patch

Additional testcases for locale.c showing the problem. Also testbot job 9085.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

Nikolay Sivov  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #49623|application/mbox            |text/plain
          mime type|                            |

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

Ken Sharp  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|enhancement                 |minor

[Bug 9583] CompareStringW gives incorrect result for some wide strings

https://bugs.winehq.org/show_bug.cgi?id=9583

[e-mail] changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 9583] CompareStringW gives incorrect result for some wide strings

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=9583

Fabian Maurer  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #26 from Fabian Maurer  ---
I'm currently working on a proper solution for this. Just need to get it
upstreamed, which will probably take some time still.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=9583

[e-mail] changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 9583] CompareStringW gives incorrect result for some wide strings

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #27 from Fabian Maurer  ---
Essentially the same reason as bug 10767 - different collation tables. This
should be fixed in staging, can you retest please?
Also, should we merge those bugs?

[Bug 9583] CompareStringW gives incorrect result for some wide strings

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #28 from Virgo Pärna  ---
Tested with custom made Delphi5 executable.

Windows: 
CompareStringA(1024, SORT_STRINGSORT, PChar('VAA'), 3, PChar('VÄÄ'), 3) returns
1
CompareStringA(0, SORT_STRINGSORT, PChar('VAA'), 3, PChar('VÄÄ'), 3) returns 1

wine-4.0 (Debian 4.0-2):
CompareStringA(1024, SORT_STRINGSORT, PChar('VAA'), 3, PChar('VÄÄ'), 3) returns
2
CompareStringA(0, SORT_STRINGSORT, PChar('VAA'), 3, PChar('VÄÄ'), 3) returns 2

wine-6.0-101-g6c8029d8d0a
CompareStringA(1024, SORT_STRINGSORT, PChar('VAA'), 3, PChar('VÄÄ'), 3) returns
1
CompareStringA(0, SORT_STRINGSORT, PChar('VAA'), 3, PChar('VÄÄ'), 3) returns 1

So it seems to be much better. Bear and wheelbarrow are no longer same word...

[Bug 9583] CompareStringW gives incorrect result for some wide strings

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=9583

VZ  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 9583] CompareStringW gives incorrect result for some wide strings

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=9583

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
      Fixed by SHA1|                            |d8c973ad95ba5e8a9a51df0dd9b
                   |                            |e587950179ec3
             Status|NEW                         |RESOLVED

--- Comment #29 from Alexandre Julliard  ---
This should be fixed by d8c973ad95ba5e8a9a51df0dd9be587950179ec3.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=9583

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #30 from Alexandre Julliard  ---
Closing bugs fixed in 7.10.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #31 from VZ  ---
Sorry, but this doesn't seem to be really fixed for me with Wine 7.0.

E.g. CompareStringEx("de-DE", NORM_IGNORECASE, "ß", "ss") yields 0 under native
MSW but -1 with Wine. And CompareStringEx("sv", 0, "ä", "ae") gives -1 instead
of the expected 1 for Swedish sort order.

[Bug 9583] CompareStringW gives incorrect result for some wide strings

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=9583

--- Comment #32 from VZ  ---
(In reply to VZ from comment #31)
> Sorry, but this doesn't seem to be really fixed for me with Wine 7.0.

Damn, sorry for the noise, I've somehow misread 7.10 in the previous comment as
7.0 and didn't realize that my version of Wine didn't have the fix yet. Of
course, I've noticed it immediately after posting the comment above. Please
disregard it and sorry again.