February 2009 - Wine-Devel - list.winehq.org

WINEDEBUG/Test suite

by Austin English

Howdy, Got a bit bored over the weekend and updated my build script to build wine daily while I'm at work, then automatically run the test suite under different (currently 6) configurations. Currently, it tests: no winedebug (ae-ub-810) Wine running under a virtual desktop (ae-ub-810-virtdesktop) # this one was broken in today's test...err, not run under a virtual desktop, fixed for tomorrow's. Needed a sleep 5s to get the virtual desktop setting saved before running the test suite. WINEDEBUG=+heap (ae-ub-810-heap) WINEDEBUG=+relay (ae-ub-810-relay) WINEDEBUG=+seh (ae-ub-810-seh) WINEDEBUG=+all (ae-ub-810-all) It's got some interesting results: http://test.winehq.org/data/3fa05ae6a7c08f9534badfeffbe02db9e00ab47a/#group… Currently, several tests timeout because of it (not worried about those...besides, I just ordered parts for a quad core on newegg, so those may be gone soon). The more interesting thing is that some test fail with memory corruption, etc. Most notably, crypt32:chain, crypt32:encode, crypt32:sip, d3d9:visual, ddraw:visual, oleaut32:tmarshal, riched20:editor, riched32:editor, and rpcrt4:cstub. There are several crashes as well (netapi32:access, jscript:run, msi:msi) all with a similar error. I filed http://bugs.winehq.org/show_bug.cgi?id=17415 for that. I filed a few other bugs for some of these issues, but not all of them...hopefully some of them will be fixed before I get the chance to :-p. I'd appreciate if everyone took a couple minutes to look at any areas of the code they're familiar with and see if they can fix these failures. I plan to test around with various optimizations, e.g., -O0, etc., since I recall at WineConf 2008 there were some failures only showing up under those configurations. If that shows anything consistent, I'll add them. -- -Austin

Feb. 17, 2009

Re: advapi32: GetNamedSecurityInfoExA Stub

by Ricardo Filipe

Ah, right you are, thanks. > > I guess that should have read 'most obviously incorrect'. > > I cannot find any reference to GetNamedSecurityInfoEx with or without > the A or W on http://www.msdn.microsoft.com. I have looked several > times over the last six months or so. Is there another MSDN somewhere ? > > GetSecurityInfoExW and (there is no stub for) GetSecurityInfoExA have > prototypes in include/aclapi.h. Would that be an appropriate .h to add > prototypes for GetNamedSecurityInfoExA and GetNamedSecurityInfoExW to ? > > If you would like I could add a stub for GetSecurityInfoExA too. One > patch or two ? i think you can use aclapi.h for it's prototypes, GetNamedSecurityInfoA/W are already there so... it's weird that these are not documented in the SDK i think... but it's not the first. you should just do stubs of what is needed, so if there is no imeadiate need to stub GetSecurityInfoExA... but you can do it if you feel like :P send one patch for each AW pair, stub, spec and prototype in each patch.

Feb. 16, 2009

big patch to implement SetupPromptForDisk (please comment)

by Ricardo Filipe

hello guys. i have took the endeavour of making my first dialog patch, managed to do it this weekend, to implement SetupPromptForDisk, which i stubbed but had some problems on the stub. gave me some big headaches to make this work x) so here is the patch, i've made the two versions of the dialog, but before submiting i'd like your comments on it. what is wrong in it? in which file should it be? i left it in stubs.c but i think misc.c might be a better place. is it testable? i once asked this on IRC and got a "no" for response, is it not testable? should i put some comments? code documentation (or lack there of) is a flaw of mine... i'm hoping for your replies so that i can submit as soon as possible. i've made some tests for "manual" testing and i have tested in "A vampire Story", the game that needed the stub, and it all worked fine. yhey! it lacks LOTS of details though, i just implemented the juice of it and some simple details. regards.

Feb. 16, 2009

Re: advapi32: GetNamedSecurityInfoExA Stub

by Paul Bryan Roberts

Dmitry Timoshkov wrote: > "Paul Bryan Roberts" <pbronline-wine(a)yahoo.co.uk> wrote: > >> I guess this stub falls in to the 'not obviously correct' category. >> >> I could do with some feedback on what might be an acceptable patch. > > It appears that GetNamedSecurityInfoExA signature doesn't match > the MSDN/PSDK one. Also it's a common practice to simultaneously > add both A and W versions, and add prototypes to an appropriate > .h file. > Ah, right you are, thanks. I guess that should have read 'most obviously incorrect'. I cannot find any reference to GetNamedSecurityInfoEx with or without the A or W on http://www.msdn.microsoft.com. I have looked several times over the last six months or so. Is there another MSDN somewhere ? GetSecurityInfoExW and (there is no stub for) GetSecurityInfoExA have prototypes in include/aclapi.h. Would that be an appropriate .h to add prototypes for GetNamedSecurityInfoExA and GetNamedSecurityInfoExW to ? If you would like I could add a stub for GetSecurityInfoExA too. One patch or two ?

Feb. 16, 2009

search path redux - if office 2007 always uses a private riched20, why does wine interpose its own global one?

by Dan Kegel

See http://bugs.winehq.org/show_bug.cgi?id=14980 Essentially, to run Office 2007, you have to set an override for riched20. Since Office installs a new, spiffier version of riched20 in its own private directory, and expects to find it there, isn't it a bug that we don't let it have it? Sounds like it's time to write a conformance test... sure would be nice if that override wasn't needed.

Feb. 16, 2009

Re: urlmon: Correctly test to see if PutProperty_CLASSIDPROP is called

by Alistair Leslie-Hughes

Hi Pual, "Paul Vriens" <paul.vriens.wine(a)gmail.com> wrote in message news:4999D04B.3010304(a)gmail.com... > This patch introduced some test failures on Win98 up to W2K. Could you > have a look? > > Looks like we need some broken or optional logic. > Sure, I'll have a look tonight. Best Regards Alistair Leslie-Hughes

Feb. 16, 2009

Re: office 2007 student trial downloaded, activated, and started ok!

by Dan Kegel

On Sun, Feb 15, 2009 at 9:56 PM, Dan Kegel <dank(a)kegel.com> wrote: > Golly, Wine must be getting good. The student trial (though not > the pro trial) of office 2007 from http://office.microsoft.com > seems to install, activate, and start properly with no hacks > at all. > > Can't edit the default document though, it says "selection locked". > Guess I'll try the overrides listed in the appdb. Yep, just overriding riched20 in winecfg fixed that! OMG it plays my presentation properly!! Wine ascendant, indeed!

Feb. 16, 2009

Re: richedit: Removed unused ME_Document structure.

by Dylan Smith

On Mon, Feb 16, 2009 at 3:19 PM, James McKenzie <jjmckenzie51(a)earthlink.net>wrote: > Would this be needed for EM_FORMATRANGE support or can this safely be > removed. Don't want to delete and then add back code. > This wouldn't be needed for EM_FORMATRANGE. If it were to duplicate the text to cache the formatting it could also use ME_TextBuffer, however the previous patch sent for EM_FORMATRANGE didn't even use that, it just duplicated a the text and stored the pointer to the linked list for the text as local variables and destroyed them before returning from ME_FormatRange. The easier way to handle this message would probably be to rewrap the text, for the specified target device, and then rewrap it again to restore the text before returning from ME_FormatRange. This would then allow ME_SETTARGETDEVICE to be implemented properly, and wouldn't be as bad as duplicating all the text and destoying it while processing each EM_FORMATRANGE message. The wrapping code needs to be looked at first though, because currently it is slow since it joins and splits runs each time it rewraps the text.

Feb. 16, 2009

Re: urlmon: Correctly test to see if PutProperty_CLASSIDPROP is called

by Paul Vriens

Alistair Leslie-Hughes wrote: > Hi, > > Changelog: > urlmon: Correctly test to see if PutProperty_CLASSIDPROP is called > > Best Regards > Alistair Leslie-Hughes > > > > > > > ------------------------------------------------------------------------ > > Hi Alistair, This patch introduced some test failures on Win98 up to W2K. Could you have a look? Looks like we need some broken or optional logic. -- Cheers, Paul.

Feb. 16, 2009

Re: Debian Etch, Lenny and Squeeze packages

by Ben Klein

2009/2/17 James McKenzie <jjmckenzie51(a)earthlink.net>: > Ben Klein wrote: >> Debian Lenny has gone stable. YAY! I've already got chroots and my >> build scripts set up for Debian Squeeze, which is the new testing, so >> I'll be able to produce packages of Wine for Squeeze without trouble, >> but there are some issues I'd like comment on. >> >> 1) I will not supply packages for Etch/amd64 (this used to be a TODO >> for me). No negotiation on this, but I am willing to assist people to >> compile/package it themselves if there's a good reason for not >> upgrading to Lenny. >> 2) I'll still build Etch/i386 packages (because it's damn easy for me >> to do so), but how long should this go on for? Ideally, we should >> encourage people to upgrade to Lenny. >> > Is there a must have feature in Lenny? Then I would give people 3 > months to upgrade. I'm not sure I understand what you mean here, but Lenny has the traits of any new Debian release (a whole lot of new), including libgnutls that works wine Wine, as well as making it easy to build amd64 packages. Etch/amd64 was missing vital lib32 packages. > Is there a platform that Lenny does NOT support that > is in major use (witness the same thing with MacOSX 10.3 to 10.4 > upgrade. Some platforms, about five percent of all Mac users at the > time, were left behind.) If this is true, the time of support may need > to be longer. Essentially, Wine only works on ia32. The only ia32 architectures that Debian supports are i386, amd64 and ia64. Although it should be possible to support ia64 (since the binaries are ia32 anyway, it should just be a matter of changing the architecture in the control file), it's not practical due to 1) Limitations of ia64 architecture in running ia32 applications 2) Inability for me to ensure the packages will install (I have no ia64 CPU) 3) Unpopularity of ia64 compared to amd64 So we don't need to worry about this. And no, it doesn't look like Debian has dropped any architectures. >> 3) Do we want to set up a Squeeze repository with 1.1.15 in it, or >> should we wait until the next release to set up a Squeeze repository? >> If the former, I will build packages specifically on/for Squeeze - >> best to do it this way for maximum compatibility. >> > The builds for Etch/Lenny/Squeeze need to be separate unless Lenny > builds will run on Squeeze. Yes, I know. But should we wait until the next release of Wine to set up the Squeeze repository? One of the Lenny or Sid packages is pretty much guaranteed to install on Squeeze. >> Ideally, I'd like a standardised policy on points #2 and #3, so that >> we know what's going on for the next release of Debian (I know, >> long-term planning). Scott, I'd value your opinion too, even though >> you only do Ubuntu packages now. >> > Yes, the policy needs to be in writing. Debian updates take a long time > and the next release takes a long time. However there is a good reason > for this. Once we make a final decision here, it can go on the Wiki or the download page.

Feb. 16, 2009