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

Re: ddraw: Fixed in part regression bug #13277
by Jens Albretsen 23 May '08

23 May '08

On Friday 23 May 2008 06:35:22 you wrote: > Jens Albretsen wrote: > > On Thursday 22 May 2008 23:56:42 you wrote: > >> I don't have a compiler on my windows box, nor do I have cross > >> development tools on my linux box, but if you get me an exe I can test > >> on win xp (nvidia 7200 GT) or 2k (intel 865, or a variety of other > >> boxes). > >> > >> -Austin > > > > I have the same problem, I have absolutely no M$ stuff at home. > > I'm pretty sure what I did is correct, since like I said the game works > > on windows and now it also does in wine. Before wine was crashing so it > > can only be better. And next time someone runs the test on the M$ > > platforms we'll see if the test fails or what? Anyway without the patch > > we just get a crash. > > That's the part your have to prove with tests and test results. Just the > fact that it fixes one program doesn't mean it's correct and won't break > all other programs. > > This is been the worst part about Wine and d3d in particular. That small > "fixes" fix one game/problem yet break everything else. > > Vitaliy. In this case it was the function that did'nt check the arguments correctly, so it would break nothing only get more stable, but anyway I added a test, and It's been verified to work on windows XP. I can understand if it was like the other part I submitted where I was changing the public ref count and failing the tests. /Jens

1 0

Re: ddraw: Fixed in part regression bug #13277
by Jens Albretsen 23 May '08

23 May '08

On Thursday 22 May 2008 23:56:42 you wrote: > I don't have a compiler on my windows box, nor do I have cross > development tools on my linux box, but if you get me an exe I can test > on win xp (nvidia 7200 GT) or 2k (intel 865, or a variety of other > boxes). > > -Austin I have the same problem, I have absolutely no M$ stuff at home. I'm pretty sure what I did is correct, since like I said the game works on windows and now it also does in wine. Before wine was crashing so it can only be better. And next time someone runs the test on the M$ platforms we'll see if the test fails or what? Anyway without the patch we just get a crash. > On 5/22/08, Jens Albretsen <jens(a)albretsen.dk> wrote: > > On Thursday 22 May 2008 23:39:33 you wrote: > >> On 5/22/08, Jens Albretsen <jens(a)albretsen.dk> wrote: > >> > I have no way of testing what is does native? Only got windows at > >> > work. > >> > >> If you need something tested on windows, or on a specific > >> version/locale you don't have access to, send a test app to wine-users > >> or wine-devel asking for testers. Odds are *someone* has the > >> configuration needed and time to do it. > > > > Anyone who want to apply the patches I did and do a "make test" of > > dlls/ddraw > > on windows?

2 1

Re: Cyrillic glyphs and bitmap fonts for Tahoma Bold
by Dmitry Timoshkov 23 May '08

23 May '08

"Igor Tarasov" <tarasov.igor(a)gmail.com> wrote: > diff --git a/fonts/tahomabd.sfd b/fonts/tahomabd.sfd > index b9fdfa0..e42678a 100644 > --- a/fonts/tahomabd.sfd > +++ b/fonts/tahomabd.sfd > @@ -1,4 +1,4 @@ > -SplineFontDB: 1.0 > +SplineFontDB: 2.0 As I pointed out many times this is a fatal change, please leave it at 1.0. -- Dmitry.

1 0

Re: Alternate shell?
by Susan Cragin 22 May '08

22 May '08

http://shellcity.net/cgi-bin/DB_Search/db_search.cgi?setup_file=topicpage.s… Don't forget this list. -----Original Message----- >From: Dan Kegel <dank(a)kegel.com> >Sent: May 22, 2008 6:16 PM >To: Wine Develop <wine-devel(a)winehq.org> >Subject: Alternate shell? > >For a maximal dogfood experience, I was looking around >for a way to use a replacement Windows shell with >Wine as my desktop environment instead of gnome >or kde. > >So far, all the ones I've found are either broken, hard to use, >or just plain weird. > >I've been working through the list at >http://www.dmoz.org/Computers/Software/Operating_Systems/Microsoft_Windows/… > >http://www.astonshell.com/aston/ - couldn't figure out how to run it? > >http://www.emergedesktop.org/ - kind of works, but very minimal, no start menu > >http://www.labyrinth.net.au/~mosses/rob/liteshell.html - works, but >very minimal, no start menu > >http://purels.org - no help, just exits, saying "ha ha, you didn't >configure me". Obviously not meant for me :-) > >http://www.lighttek.com/talisman.htm - pretty, but no Programs menu?! >(Maybe that's there but broken, wine bug?) > >http://www.sky.franken.de/explorer/download.html - ROS explorer - >prebuilt binary >there is 3 years out of date, doesn't work too well > >Anyone have a favorite I should try? >- Dan > >

1 0

Re: ddraw: Fixed in part regression bug #13277
by Alexander Dorofeyev 22 May '08

22 May '08

Jens Albretsen wrote: > Changelog: > > Fixed Failure when texture was loading onto itself. > Fixes in part d3d7 regression bug #13277. > Still need to find a way to make d3d7 stay alive after the game releases the > interface. > Did you check what native does when asked to Load texture from itself? Is it D3D_OK or maybe something like INVALIDCALL? There already should be a few tests for IDirect3DTexture_Load in ddraw/tests/d3d.c so you can add a few lines there that test this.

3 3

Re: [rsaenh] don't throw error on salt with more that 128bit
by Juan Lang 22 May '08

22 May '08

- /* salt length can't be greater than 128 bits = 16 bytes */ - if (blob->cbData > 16) - { - SetLastError(ERROR_INVALID_PARAMETER); - return FALSE; - } memcpy(pCryptKey->abKeyValue + pCryptKey->dwKeyLen, blob->pbData, blob->cbData); You can't just remove the check, an app that passes in a bad blob will overwrite memory. If that's not allowed in Windows, it shouldn't be in Wine, either. You really need to write a test case that shows what the correct limit for the salt length is, or show that Windows just crashes if you pass in too much salt data. That'll prevent future regressions from cropping up, too. --Juan

1 0

Re: ddraw: Fixed last part of regression bug #13277
by Stefan Dösinger 22 May '08

22 May '08

Am Donnerstag, 22. Mai 2008 16:17:26 schrieb Jens Albretsen: > Changelog: > d3d1-7 surfaces was not ref counting ddraw so if there was surface and > ddraw was released, wine would kill everything. Now the Return to > Mysterious Island works and does a clean exit afterwards. I think the patch is wrong. There's a test that shows that this is not happening. If it doen't break any existing tests, please add a test to demonstrate that the change is correct

3 4

Re: ole32/tests: Add a trailing '\n' to ok() calls.
by Michael Karcher 22 May '08

22 May '08

Am Donnerstag, den 22.05.2008, 10:37 +0200 schrieb Francois Gouget: > - ok(len == 39, "len: %d (expected 39)",len); > + ok(len == 39, "len: %d (expected 39)\n", len); [and so on] You are right. This was missing from my patch that got applied. I sent an updated version one hour later[1] (got through moderation 6 hours later or so), which does - Include the missing "\n" characters - Makes this function indented like the others in ole32/tests/compobj.c - Test another case that wine gets wrong (with todo_wine) In direct sucession[2], I sent a patch which corrects the problem in wine and removes the todo_wine from the testcase again. I did it that way to seperate the two patches, still preventing the testsuite to show an additional failure. Was something wrong with the way I handled the first not-that-good patch (apart from sending it with the missing new lines at all, of course)? Regards, Michael Karcher [1]: <1211272286.9165.17.camel(a)hermes2.karcher.local> [2]: <1211268226.9165.13.camel(a)hermes2.karcher.local>

1 0

Re: kernel32:Fix a test that fails for all platforms
by Alexandre Julliard 22 May '08

22 May '08

"James Hawkins" <jhawkins(a)codeweavers.com> writes: > If you don't like this version, let me know. The other option is to > convert any forwards slashes in the result string to backslashes. This should most likely be fixed in winetest. -- Alexandre Julliard julliard(a)winehq.org

1 0

stub implementation of GdipImageGetFrameDimensionsCount
by Petr Dlouhý 22 May '08

22 May '08

Hello, I want to ask, why my patch (http://www.winehq.org/pipermail/wine-patches/2008-May/054973.html) didn't get committed? -- Petr Dlouhý

1 0