June 2008 - Wine-devel - list.winehq.org

Regression in winebrowser found and identified
by Pavel Troller 07 Jun '08

07 Jun '08

Hi! Today we've found that winebrowser gets invalid URL consisting of many question marks instead of correct URL, in at least two cases: 1) Attempt to buy a game in Steam (yes, we are regularly doing that) 2) Trying to buy a program called CPT-Master from 3am systems, free demo which shows the problem when "Buy now" button is pressed in its welcome screen, can be downloaded at http://www.3amsystems.com/products/CPT-Master.exe The following commit has been identified as making the regression: e547cf7043e36fecfe5e23d1d6f93a9c5ba89c46 is first bad commit commit e547cf7043e36fecfe5e23d1d6f93a9c5ba89c46 Author: Hans Leidekker <hans(a)it.vu.nl> Date: Sun Apr 6 14:44:31 2008 +0200 winebrowser: Convert to Unicode. :040000 040000 4f3a034b18392677fc0dfd1ed6df7fefe843a10e 0ab8603a13e836a0b91dd2f146df8da59f7e79e4 M programs It really looks related to the problem which I'm reporting :-). BTW, my system is NOT running in Unicode. Maybe is it related ? I hope it is not yet mandatory. I'm using iso-8859-2 encoding, which works much better in my normal working environment. Please do your best to fix it before 1.0, I feel it really bad and I think that many programs may be affected by it. Should I submit a bug too ? With regards, Pavel Troller

2 2

Re: kernel: updated French translation
by Laurent Vromman 07 Jun '08

07 Jun '08

Hi, There is a mistake here : "L'acquéreur de le sémaphore est mort" The right translation is : "L'acquéreur du sémaphore est mort" The translation of "broken pipe" is not good too, but I can't find anything better on the internet. You should maybe try something like "Tube indisponible (broken pipe)", if you think it is better. Laurent Jonathan Ernst wrote: > ------------------------------------------------------------------------ > > >

2 1

French translation only partial?
by Frans Kool 07 Jun '08

07 Jun '08

Hi, I just had a look at the French translation of the dlls/kernel32/nls/winerr_fra.mc file. I noticed only the first 54 messages have been translated, whereas the other 1876 are still the English originals. Was this on purpose? Or was this a partial translation which needs to be finalized (the remaining 97.2%)? I'm all for translating Wine in as many languages as possible, but I doubt this is the way to get there. Is there some way to detect if the 'translation' of a message is identical to the original (English) one and have it show up as 'partially translated' in the translation overview [1]? As a side note, is there any review on translations or is the policy "any translation is better than no translation"? When Wine 1.0 is going to be released, how can we be sure that the application is actually completely translated. For example, if I really wanted to harm translations, I could make a copy of all English resource files, just change the 'LANGUAGE LANG_ENGLISH, SUBLANG_DEFAULT' to whatever language I would like to 'translate' and submit them. As far as I know, there is no way of detecting this. Anyone got ideas? Frans. [1] http://pf128.krakow.sdi.tpnet.pl/wine-transl/index.php

2 1

Re: [PATCH] GetUrlCacheEntryInfo*: Handle null dest pointer when length is given
by Rob Shearman 07 Jun '08

07 Jun '08

2008/6/6 Jon Griffiths <jon_p_griffiths(a)yahoo.com>: > - if (dwRequiredSize > *lpdwBufferSize) > - { > - *lpdwBufferSize = dwRequiredSize; > + ret = dwRequiredSize > *lpdwBufferSize? FALSE : TRUE; > + *lpdwBufferSize = dwRequiredSize; > + > + if (!ret) > SetLastError(ERROR_INSUFFICIENT_BUFFER); > - return FALSE; > + else if (!lpCacheEntryInfo) > + { > + SetLastError(ERROR_INVALID_PARAMETER); > + ret = FALSE; > } > - *lpdwBufferSize = dwRequiredSize; > - return TRUE; > + return ret; > } Internal functions should return an error code instead of calling SetLastError. This avoids situations like this: > @@ -1531,6 +1539,9 @@ BOOL WINAPI GetUrlCacheEntryInfoA( > FALSE /* ANSI */)) > { > URLCacheContainer_UnlockIndex(pContainer, pHeader); > + /* This case is inconsistent w.r.t RetrieveUrlCacheEntryFile */ > + if (!lpCacheEntryInfo && GetLastError() == ERROR_INVALID_PARAMETER) > + SetLastError(ERROR_INSUFFICIENT_BUFFER); > return FALSE; > } > TRACE("Local File Name: %s\n", debugstr_a(lpCacheEntryInfo->lpszLocalFileName)); -- Rob Shearman

2 2

Re: mshtml: Added FIXME about QUERY_USES_HISTORYFOLDER in about protocol.
by Paul Vriens 07 Jun '08

07 Jun '08

Jacek Caban wrote: > --- > dlls/mshtml/protocol.c | 2 ++ > 1 files changed, 2 insertions(+), 0 deletions(-) > > > ------------------------------------------------------------------------ > > Hi Jacek, return E_NOTIMPL; + case QUERY_USES_HISTORYFOLDER: + FIXME("Unsupported option QUERY_USES_HISTORYFOLDER\n"); default: return E_FAIL; } shouldn't there be a "return E_NOTIMPL" after the FIXME as we do with the others? -- Cheers, Paul.

3 2

RE: d3d8: Fix the null condition in ApplyStateBlock
by Stefan Dösinger 06 Jun '08

06 Jun '08

What would be interesting to know is which effect this has on the states, but not in this patch I'd say, and not before the 1.0 release. In the long run it might be an idea to port the d3d9:stateblock test to d3d8 and see how d3d8 behaves

1 0

Re: urlmon: Fix a few failing tests in win2k3
by Jacek Caban 06 Jun '08

06 Jun '08

James Hawkins wrote: > Hi, > > Changelog: > * Fix a few failing tests in win2k3. You've fixed tests on IE7, not only win2k3. It would be better to change Wine to behave like IE7. Also your patch lefts some variables in wrong state. I will send patches to fix it. Jacek

2 4

Re: [urlmon/tests 2/2] Fix test on IE7 (Resend)
by Jacek Caban 06 Jun '08

06 Jun '08

Hi Paul, Paul Vriens wrote: > Hi, > > This (together with number 1 of course) fixes the tests on IE7 boxes. > > Not sure if a patch for urlmon/protocol.c is needed as well. Let me > know and > I'll sent one. > > Changelog > Fix test on IE7 While fixing such tests it's good to add FIXME in code so info about it won't be lost. It's not documented anywhere which protocol handles which options. I will send a patch. Jacek

1 0

Re: Remove unused variable in dlls/mlang/tests/mlang.c
by James Hawkins 06 Jun '08

06 Jun '08

2008/6/5 Gerald Pfeifer <gerald(a)pfeifer.com>: > > ChangeLog: > Remove unused variable in test_EnumScripts(). > > Index: dlls/mlang/tests/mlang.c > =================================================================== > RCS file: /home/wine/wine/dlls/mlang/tests/mlang.c,v > retrieving revision 1.24 > diff -u -3 -p -r1.24 mlang.c > --- dlls/mlang/tests/mlang.c 4 Jun 2008 11:37:37 -0000 1.24 > +++ dlls/mlang/tests/mlang.c 5 Jun 2008 15:17:51 -0000 > @@ -594,7 +594,6 @@ static void test_EnumScripts(IMultiLangu > > for (i = 0; pGetCPInfoExA && i < n; i++) > { > - CPINFOEXA cpinfoex; > #ifdef DUMP_SCRIPT_INFO > trace("SCRIPTINFO #%u:\n" > "ScriptId %08x\n" > > > http://winehq.org/pipermail/wine-cvs/2008-June/044013.html -- James Hawkins

2 1

Compile wine on Red Hat 9
by Wen Huang 06 Jun '08

06 Jun '08

Hi everybody, I was trying to compile a winehq on Red Hat 9 and I downloaded the source 0.9.60 from sourceforge. When I tried to compile by ./configure make depend make during the make step, it gave an error ..... gcc -c -I. -I. -I../../include -I../../include -I/usr/include/ libxml2 -I/usr/include/libxml2 -D__WINESRC__ -DCOM_NO_WINDOWS_H - D_REENTRANT -fPIC -Wall -pipe -fno-strict-aliasing -Wdeclaration-after- statement -Wwrite-strings -Wpointer-arith -g -O2 -o cdata.o cdata.c cdata.c: In function `domcdata_appendData': cdata.c:601: void value not ignored as it ought to be make[2]: *** [cdata.o] Error 1 make[2]: Leaving directory `/home/tus/wine-0.9.60/dlls/msxml3' make[1]: *** [msxml3] Error 2 make[1]: Leaving directory `/home/tus/wine-0.9.60/dlls' make: *** [dlls] Error 2 Can anybody kindly offer some help? By the way, is there rpm package for wine I can use on Red Hat 9, I tried to use the script on Wine Red Hat wiki to generate a rpm, but it failed at the same step... Thanks, Wen

2 3