June 2010 - Wine-Devel - list.winehq.org

Re: user32: Recurse into WS_EX_CONTROLPARENT controls in GetDlgItem

by Dmitry Timoshkov

André Hentschel <nerv(a)dawncrow.de> wrote: > This fixes bug 22862 which blocks bug 12804 (1.2) > We already do a similar recursion in DEFDLG_FindDefButton and GetNextDlgGroupItem > > I didnt found any other changes in the tests... i expected some more "Test succeeded inside todo block" You need to add explicit tests for GetDlgItem() inside of user32/tests. -- Dmitry.

June 30, 2010

Documentation - Reference to MSDN?

by Max TenEyck Woodbury

I've been reading the Wine code and noticed that some of the external interfaces are practically undocumented. I did a web search on some of the names and found descriptions in MSDN. I realize that copying the information from MSDN directly into the code is a poor idea (like copyright violation) so I have a couple of questions: 1) Would including the URL of the MSDN article be useful/a good idea? 2) Would enumerating coded values and flags be allowed? 3) Where should data structures be documented?

June 30, 2010

Re: add russian translation for winerrs messages [try2]

by Paul Vriens

On 06/30/2010 03:43 PM, Dmitry Timoshkov wrote: > Dmitry Timoshkov<dmitry(a)codeweavers.com> wrote: > >> Vitaly Lipatov<lav(a)etersoft.ru> wrote: >> >>> +LanguageNames=(RUS=0x19:winerr) >>> +CodePages=(FRA=65001:0) >> >> 'FRA'? This doesn't look correct. > > Also it doesn't make much sense to send a half completed translation. > I could agree but if you look at the Dutch kernel32 resource it's even worse I'm afraid. This at least gives us a starting point. (Try3 was sent with at least the 'FRA' fixed). -- Cheers, Paul.

June 30, 2010

Re: add russian translation for winerrs messages [try2]

by Dmitry Timoshkov

Dmitry Timoshkov <dmitry(a)codeweavers.com> wrote: > Vitaly Lipatov <lav(a)etersoft.ru> wrote: > > > +LanguageNames=(RUS=0x19:winerr) > > +CodePages=(FRA=65001:0) > > 'FRA'? This doesn't look correct. Also it doesn't make much sense to send a half completed translation. -- Dmitry.

June 30, 2010

Re: add russian translation for winerrs messages [try2]

by Dmitry Timoshkov

Vitaly Lipatov <lav(a)etersoft.ru> wrote: > +LanguageNames=(RUS=0x19:winerr) > +CodePages=(FRA=65001:0) 'FRA'? This doesn't look correct. -- Dmitry.

June 30, 2010

Re: user32/tests: Use if(0) instead of a comment to make sure code still compiles

by testbot＠testbot.winehq.org

Hi, While running your changed tests on Windows, 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 http://testbot.winehq.org/JobDetails.pl?Key=3023 Your paranoid android. === W7PRO (32 bit msg) === msg.c:9516: Test failed: wrong qstatus 00110011 msg.c:9523: Test failed: wrong qstatus 00190008 msg.c:9529: Test failed: wrong qstatus 00390020 msg.c:9538: Test failed: wrong qstatus 00790040 msg.c:9559: Test failed: wrong qstatus 00390000 msg.c:9567: Test failed: wrong qstatus 00790040 msg.c:9578: Test failed: wrong qstatus 00390000 msg.c:9589: Test failed: wrong qstatus 00310000 msg.c:9594: Test failed: PeekMessageA should have returned FALSE instead of msg 0113 === W7PROX64 (64 bit msg) === Timeout

June 29, 2010

Re: wininet: fix the handling of insufficient buffers in InternetGetCookie*()

by testbot＠testbot.winehq.org

Hi, While running your changed tests on Windows, 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 http://testbot.winehq.org/JobDetails.pl?Key=3020 Your paranoid android. === W98SE (32 bit internet) === internet.c:440: Test failed: A=B present internet.c:442: Test failed: E=F present internet.c:444: Test failed: I=J present internet.c:445: Test failed: K=L present internet.c:452: Test failed: A=B present internet.c:454: Test failed: E=F present internet.c:456: Test failed: I=J present internet.c:457: Test failed: K=L present internet.c:1158: Test failed: Retrieved proxy server should've been L"proxy.example", was: L"\0001\a000\2684X\0fb4h\0200" internet.c:1164: Test failed: Expected regtype to be REG_DWORD, was: 3 internet.c:1192: Test failed: Retrieved flags should've been PROXY_TYPE_DIRECT, was: 5767180 internet.c:1195: Test failed: Expected regtype to be REG_DWORD, was: 3 internet.c:1222: Test failed: Retrieved flags should've been PROXY_TYPE_PROXY | PROXY_TYPE_DIRECT, was: 5767180 internet.c:1225: Test failed: Expected regtype to be REG_DWORD, was: 3 === W2KPROSP4 (32 bit internet) === internet.c:440: Test failed: A=B present internet.c:442: Test failed: E=F present internet.c:444: Test failed: I=J present internet.c:445: Test failed: K=L present internet.c:452: Test failed: A=B present internet.c:454: Test failed: E=F present internet.c:456: Test failed: I=J present internet.c:457: Test failed: K=L present === WXPPROSP3 (32 bit internet) === internet.c:517: Test failed: bad last error {error = 122} internet.c:545: Test failed: the function failed internet.c:546: Test failed: bad last error {error = 122} === WVISTAADM (32 bit internet) === internet.c:517: Test failed: bad last error {error = 122} internet.c:545: Test failed: the function failed internet.c:546: Test failed: bad last error {error = 122} === W2K8SE (32 bit internet) === internet.c:517: Test failed: bad last error {error = 122} internet.c:545: Test failed: the function failed internet.c:546: Test failed: bad last error {error = 122} === W7PRO (32 bit internet) === internet.c:474: Test failed: bad last error {error = 5023} internet.c:502: Test failed: bad last error {error = 5023} internet.c:517: Test failed: bad last error {error = 5023} internet.c:545: Test failed: the function failed internet.c:546: Test failed: bad last error {error = 122} === W7PROX64 (32 bit internet) === internet.c:474: Test failed: bad last error {error = 5023} internet.c:502: Test failed: bad last error {error = 5023} internet.c:517: Test failed: bad last error {error = 5023} internet.c:545: Test failed: the function failed internet.c:546: Test failed: bad last error {error = 122} === W7PROX64 (64 bit internet) === internet.c:474: Test failed: bad last error {error = 5023} internet.c:502: Test failed: bad last error {error = 5023} internet.c:517: Test failed: bad last error {error = 5023} internet.c:545: Test failed: the function failed internet.c:546: Test failed: bad last error {error = 122}

June 29, 2010

Re: Using string resources in conformance tests.

by Michael Mc Donnell

Thanks for taking the time to write a comment Juan. On Tue, Jun 29, 2010 at 5:55 PM, Juan Lang <juan.lang(a)gmail.com> wrote: > Hi Michael, > >> I use the LoadString function to get the "New folder" string from >> shell32.dll that is used for creating a new folder. The problem is >> that the location where it is stored in shell32.dll varies between the >> different versions of Windows. I can only make the test work on one >> version at a time. Even the capitalization varies between the >> different versions of Windows, i.e. "New Folder" vs. "New folder". > > It appears as though you shouldn't be doing this in a conformance > test. It varies enough that applications can't depend on the resource > ID or string value being the same in different versions of Windows. > Our conformance tests try to test for invariants across Windows > versions, i.e. behavior that applications may depend on or expect. Yeah that sounds reasonable. I guess it means there will be very few Windows applications that depend on the folder getting the correct "New folder" name and we can therefore ignore this issue. > You might check the contents of the containing folder before and after > the command is executed, and make sure that at least one new folder > exists after the command is executed. Super, my test already does this. I'll cut out the resource handling part and clean the patch up so it will be ready for when the code freeze is lifted. Michael Mc Donnell

June 29, 2010

Re: [2/2] urlmon: implement HlinkSimpleNavigateToMoniker

by Jacek Caban

On 6/28/10 9:17 PM, André Hentschel wrote: > --- > dlls/urlmon/umon.c | 8 ++++++-- > 1 files changed, 6 insertions(+), 2 deletions(-) > > diff --git a/dlls/urlmon/umon.c b/dlls/urlmon/umon.c > index e122e07..12b1b0c 100644 > --- a/dlls/urlmon/umon.c > +++ b/dlls/urlmon/umon.c > @@ -734,8 +734,12 @@ HRESULT WINAPI HlinkSimpleNavigateToMoniker(IMoniker *pmkTarget, > LPCWSTR szLocation, LPCWSTR szTargetFrameName, IUnknown *pUnk, > IBindCtx *pbc, IBindStatusCallback *pbsc, DWORD grfHLNF, DWORD dwReserved) > { > - FIXME("stub\n"); > - return E_NOTIMPL; > + URLMoniker *This = MONIKER_THIS(iface); You can't do this. There is no guaranty that the moniker is URL moniker. Jacek

June 29, 2010

Re: Using string resources in conformance tests.

by Juan Lang

Hi Michael, > I use the LoadString function to get the "New folder" string from > shell32.dll that is used for creating a new folder. The problem is > that the location where it is stored in shell32.dll varies between the > different versions of Windows. I can only make the test work on one > version at a time. Even the capitalization varies between the > different versions of Windows, i.e. "New Folder" vs. "New folder". It appears as though you shouldn't be doing this in a conformance test. It varies enough that applications can't depend on the resource ID or string value being the same in different versions of Windows. Our conformance tests try to test for invariants across Windows versions, i.e. behavior that applications may depend on or expect. You might check the contents of the containing folder before and after the command is executed, and make sure that at least one new folder exists after the command is executed. --Juan

June 29, 2010