February 2007 - Wine-Devel - list.winehq.org

Re: Commit 37591409b28c2000e70bd0d3c654a3a7559a4a26 breaks accentuation

by Dmitry Timoshkov

"Pedro Araujo Chaves Jr." <inckie(a)gmail.com> wrote: > After some regression testing, I found out that commit > 37591409b28c2000e70bd0d3c654a3a7559a4a26 by Dmitry Timoshkov breaks > accentuation - at least for Brazilian Portuguese in Lotus Notes R5. > > I did some testing and found out that only TrueType fonts are affected > - that is, the fonts Notes lists as 'default foo' without the double-T > icon are all accented properly. To make matters worse, the affected > fonts don't behave the same way. > > For example, the string > > áéíóú ÁÉÍÓÚ àèìòù ÀÈÌÒÙ ãõ ÃÕ âêîôû ÂÊÎÔÛ äëïöü ÄËÏÖÜ > > renders as > > įéķóś ĮÉĶÓŚ ąčģņł ĄČĢŅŁ ćõ ĆÕ āźīōū ĀŹĪŌŪ äėļöü ÄĖĻÖÜ > > with Trebuchet MS. You can see they're pretty different - well, most > of you, at least. Please open a bug report regarding this problem, with all the appropriate info: an aplication that shows the problem, the fonts used, etc. If you could add a test to the existing tests in the above mentioned commit that shows the problem that would be great. -- Dmitry.

Feb. 27, 2007

Commit 37591409b28c2000e70bd0d3c654a3a7559a4a26 breaks accentuation

by Pedro Araujo Chaves Jr.

Hi all, After some regression testing, I found out that commit 37591409b28c2000e70bd0d3c654a3a7559a4a26 by Dmitry Timoshkov breaks accentuation - at least for Brazilian Portuguese in Lotus Notes R5. I did some testing and found out that only TrueType fonts are affected - that is, the fonts Notes lists as 'default foo' without the double-T icon are all accented properly. To make matters worse, the affected fonts don't behave the same way. For example, the string áéíóú ÁÉÍÓÚ àèìòù ÀÈÌÒÙ ãõ ÃÕ âêîôû ÂÊÎÔÛ äëïöü ÄËÏÖÜ renders as įéķóś ĮÉĶÓŚ ąčģņł ĄČĢŅŁ ćõ ĆÕ āźīōū ĀŹĪŌŪ äėļöü ÄĖĻÖÜ with Trebuchet MS. You can see they're pretty different - well, most of you, at least. However, that's only a display-and-print issue, as I can copy the text and paste it elsewhere, and the accents are in their right places (and in the right letters, for that matter) - as one would expect. Regarding Microsoft fonts, by the way, Calibri, Candara, Consolas, Constantia, Corbel, Segoe Print, and Segoe Script (couldn't test Cambria) all behave exactly like Trebuchet MS, whereas Segoe UI behaves just like Times New Roman, which shows problems only with graves and tildes. However the behavior is kinda random among the many other fonts (from many unrelated sources, of course) I tested, so I (still) didn't do an exhaustive check for a pattern. Would anybody have any word on this? Cheers, Pedro.

Feb. 27, 2007

Re: [3/12 WineD3D/DDraw: Forward DDSCL_MULTITHREADED to WineD3D

by Stefan Dösinger

> What exactly is this supposed to fix? We already pass the behaviour > flags through to IWineD3DImpl_CreateDevice(), so shouldn't wined3d > just handle it there? In DirectDraw the multithreading flag is set after creating the device, so I need a seperate method to set it in wined3d, because I don't know it at initialization time. This method has to initialize the thread safety things in wined3d. Now d3d8 and d3d9 call SetMultithreaded too because it is clearer to have SetMultithreaded called by ddraw, d3d8 and d3d9 instead of ddraw and wined3d. On the other hand, there are now 2 places where the multithreaded flag is passed to wined3d now, CreateDevice and SetMultithreaded. I could call SetMultithreaded from wined3d::CreateDevice, and not call it from d3d8/9, or I could remove the WINED3DCREATE_MULTITHREADED flag from the public wined3d_interface.h header, sinaling that CreateDevice ignores that flag? Or anything I missed?

Feb. 27, 2007

Re: dplay: directplay should initialize session Guid patch +conformance tests

by H. Verbeet

On 27/02/07, Dmitry Timoshkov <dmitry(a)codeweavers.com> wrote: > > +BOOL FAR PASCAL EnumConnectionsCallback(LPCGUID lpguidSP, LPVOID lpConnection, > > + DWORD dwConnectionSize, LPCDPNAME lpName, DWORD dwFlags, > > + LPVOID lpContext) > > +{ > > This doesn't look like a proper type of callb ack for IDirectPlayX_EnumConnections. > I wondered about that as well, but it does appear to be what the DXSDK headers specify.

Feb. 27, 2007

Re: dplay: directplay should initialize session Guid patch +conformance tests

by Dmitry Timoshkov

"Alessandro Pignotti" <alex.pigna(a)inventati.org> wrote: > +BOOL FAR PASCAL EnumConnectionsCallback(LPCGUID lpguidSP, LPVOID lpConnection, > + DWORD dwConnectionSize, LPCDPNAME lpName, DWORD dwFlags, > + LPVOID lpContext) > +{ This doesn't look like a proper type of callb ack for IDirectPlayX_EnumConnections. > + bzero(&sessionDesc,sizeof(DPSESSIONDESC2)); > + bzero(&zeroGuid,16); Use memset instead of non-portable bzero. > + LPDIRECTPLAY4 pDP = NULL; This initialization is redundant. If you also could use 4 spaces indents instead of tabs, and add an empty line after data declarations that would be nice as well. And a personal opinion: all that lines without a single space delimiting variables, functions, equal signs are absolutely not readable. -- Dmitry.

Feb. 27, 2007

Re: ntdll: resend 1/4 NtCreateMailslotFile tests

by Alexandre Julliard

Jeff Latimer <lats(a)yless4u.com.au> writes: > diff --git a/dlls/ntdll/tests/Makefile.in b/dlls/ntdll/tests/Makefile.in > index cccc4c5..4ceccfd 100644 > --- a/dlls/ntdll/tests/Makefile.in > +++ b/dlls/ntdll/tests/Makefile.in > @@ -14,6 +14,7 @@ CTESTS = \ > generated.c \ > info.c \ > large_int.c \ > + mailslot.c \ Naming it file.c would be better, since that's the name of the corresponding source file, and that makes it more general so it can later contain other tests too. -- Alexandre Julliard julliard(a)winehq.org

Feb. 27, 2007

Re: [4/6] cmd.exe : Prompt during rd /s command

by Alexandre Julliard

"Ann & Jason Edmeades" <us(a)edmeades.me.uk> writes: > + /* Load the translated 'Are you sure' message */ > + LoadString (hinst, WCMD_YESORN, buffer, sizeof(buffer)); > + > + /* Loop waiting on a Y or N */ > + while (answer[0] != 'Y' && answer[0] != 'N') { Making the message translatable is good, but the Y/N answers need to be translatable too... -- Alexandre Julliard julliard(a)winehq.org

Feb. 27, 2007

Re: [1/12] - [12/12] WineD3D (multithreading)

by Stefan Dösinger

Am Dienstag 27 Februar 2007 09:54 schrieb Mirek: Looks like we have a rouge multithreaded game here - which is pretty strange since it worked before. > Here is output with only patches 1,2,3,4,5,6,7,10,11,12 (without patch 8 > and 9): > fixme:d3d:ActivateContext Context creation for a new thread not > implemented yet Can you send me a +d3d9,tid trace from that?

Feb. 27, 2007

Re: dplay: directplay should initialize session Guid patch + conformance tests

by Kai Blin

On Saturday 24 February 2007 22:10, Alessandro Pignotti wrote: > Thanks for the suggestions, this version of the patch uses IsEqualGUID and > eliminates C++ style comments and insert a proper copyright notice in the > added file You might want to use #include "wine/test.h" instead of #include <wine/test.h>, all the other tests do that. Also, you should split out the test from the START_TEST(dplayx) function into a function called e.g. test_session_guid and do all the testing there. Also, please use IsEqualGUID() in EnumConnectionsCallback, too. Sorry for not replying earlier, but last week was kind of a mess, and I'm slowly catching up. Cheers, Kai -- Kai Blin, <kai Dot blin At gmail Dot com> WorldForge developer http://www.worldforge.org/ Wine developer http://wiki.winehq.org/KaiBlin/ -- Will code for cotton.

Feb. 27, 2007

Re: GDI surfaces, GDI heap and wined3d's VideoMemorySize

by H. Verbeet

On 27/02/07, Chris Robinson <chris.kcat(a)gmail.com> wrote: > I thought it was added to the registry because Wine can't figure out the video > memory amount itself from OpenGL or X, and that allows the user to make it > report the proper amount (or at least, more proper). Before it was hard-coded > to 64MB. > Pretty much.

Feb. 27, 2007