September 2005 - Wine-Devel

Re: Interaction between shell32.ShellExecute, IContextMenu and the ShellLink object?

by Kevin Koltzau

On Friday 23 September 2005 4:10 am, Mike McCormack wrote: > Hi All, > The question is, through which interface (or set of interfaces) does > shell32 invoke the Shelllink object? > > From what I can determine, ShellExecute should use the .lnk extension > to look in the registry (HKCR\.lnk -> HKCR\lnkfile\CLSID) and get > CLSID_ShellLink. Using the CLSID, it calls CoCreateInstance, then > IShellExtInit->Initialize() which loads the .lnk file. > > Unfortunately, the next step is unclear to me. I think that > IContextMenu->InvokeCommand is used, but I'm not sure as to how. Pass the result of IShellLink::GetIDList to IShellFolder::GetUIObjectOf to get an IContextMenu Or if IShellLink::GetIDList fails, you could pass IShellLink::GetPath to IShellFolder::ParseDisplayName first Then call IContextMenu::InvokeCommand

Sept. 23, 2005

Re: Documentation volunteer(s) needed

by josephblack

In article <87oe6nlq4i.fsf(a)wine.dyndns.org>, julliard(a)winehq.org says... > Folks, > > As most of you probably know (at least those of you who managed to get > out of bed in time for my keynote ;-) we are supposed to release 0.9 > real soon now. We do have one remaining issue: the documentation > needs some major work. Not every bit of it is critical for the > release, but we need to fix at least the User's Guide to reflect the > recent configuration changes. Would anybody be interested in helping > with that? > www.wine-wiki.org has been compiling tips from the wine mailing lists. I had planned to use it to update the documentation, but keeping up with the lists has kept me busy. After some discussion, Jason (the owner) has confirmed he would like to raise an offer of wine-wiki.org and its contents to the Wine project. joseph black admin

Sept. 23, 2005

Interaction between shell32.ShellExecute, IContextMenu and the ShellLink object?

by Mike McCormack

Hi All, Currently the way that ShellExecute is implemented looks wrong. Resolving shortcuts is done by calling SHELL_ResolveShortCutW (line 273 of dlls/shell32/shlexec.c) instead of invoking ShellLink object and letting it resolve itself. The question is, through which interface (or set of interfaces) does shell32 invoke the Shelllink object? From what I can determine, ShellExecute should use the .lnk extension to look in the registry (HKCR\.lnk -> HKCR\lnkfile\CLSID) and get CLSID_ShellLink. Using the CLSID, it calls CoCreateInstance, then IShellExtInit->Initialize() which loads the .lnk file. Unfortunately, the next step is unclear to me. I think that IContextMenu->InvokeCommand is used, but I'm not sure as to how. If anybody (Michael Jung?) knows how this works, please let me know, as I'd like to make IShellLink work properly, complete MSI advertised shortcuts and clean up ShellExecute... thanks! Mike

Sept. 23, 2005

Re: WooHoo get a load of this :-)

by Marcus Meissner

> I'd say we should all wait till we see what's in the challenge registration > form, where they are actually supposed to reveal exactly what those $10k buys > them. > > One application doesn't mean much. It can be something huge. It can be > something that needs something that wine nowhere near implements. And so > on... > > If they're willing to pay someone $10k then either it's not such a trivial > job, or their VCs want some sort of demonstration and the SpecOpS people just > don't have enough expertise. > > Or it's all just blown out of proportion with $10k being way more than good > compensation for the work involved, but someone clueless enough having $10k > to toss around. > > Maybe it's just worth for some wine hackers to register and see what comes out > of it. Perhaps its some sort of hiring test. Well, we will see. Ciao, Marcus

Sept. 23, 2005

Re: autogen configure script

by Marcus Meissner

On Fri, Sep 23, 2005 at 02:31:55AM -0400, Steven Edwards wrote: > Hi, > A lot of people don't like the fact we diff configure and configure.ac > so here is a autogen.sh ripped from the Samba4 source. Can we get > something like this in cvs? Please keep this kind of abonimation from CVS. Just do not diff configure, Alexandre regenerates it. Ciao, Marcus

Sept. 23, 2005

Re: autogen configure script

by Mike McCormack

Steven Edwards wrote: > A lot of people don't like the fact we diff configure and configure.ac > so here is a autogen.sh ripped from the Samba4 source. Can we get > something like this in cvs? You don't need to diff configure, only configure.ac. Alexandre discards configure changes and regenerate configure with his version of autoconf, so there's no need to send patches changing it. I suspect that there may be a licensing issue if you add stuff coming from Samba (GPL), to Wine (LGPL). Mike

Sept. 23, 2005

QA'ing wine running elementary school software

by dank＠kegel.com

I'm still looking around to figure out what software it'd be worth QA'ing under Wine as Wine approaches 0.9 or 1.0. Here's a likely set of candidates in the educational software category, along with links to free demos and the appropriate appdb pages: "Kidspiration" Free trial: http://www.inspiration.com/freetrial Appdb: http://appdb.winehq.org/appview.php?appId=1091 Status: No info "Oregon Trail 5" Free trial: http://www.learningcompany.com/jump.jsp?itemID=147&itemType=PRODUCT&RS=1 Appdb: http://appdb.winehq.org/appview.php?versionId=2909 Status: won't run - can't switch screen resolution "Kid Pix Deluxe 4 for Schools" Free trial: http://rivapprod2.riverdeep.net/portal/page?_pageid=414,105278&_dad=portal&… Appdb: http://appdb.winehq.org/appview.php?appId=285 Status: no info - Dan p.s. To figure out what educational software is in use, I searched for approved software list elementary school A few hits included: http://www.riordanfoundation.org/What_We_Do/05%20CIC%20budget%20worksheet%2… http://edweb.tusd.k12.az.us/softwarelist.asp http://www.tvdsb.on.ca/currcomptech/elementary_software_approval_list_aug20… http://www.gmcs.org/images/technology/software.pdf http://dysart.schoolnet.com/outreach/dysart/departments/edtech/resource/app… http://www.indyschools.com/software/ That gave me a rough idea of what titles were popular. I then searched for $title approved software site:.us for likely values of $title, and counted hits. (Searching only .us sites filtered out a lot of noise, I think, since k-12 schools are much more likely than software vendors or reviews to have .us domains.) The results: 10,000 for "accelerated reader" approved software site:.us 9,640 for "print shop" approved software site:.us 502 for "kidspiration" approved software site:.us 420 for "kid pix" approved software site:.us 369 for "oregon trail" approved software site:.us 220 for "reader rabbit" approved software site:.us Taking them from the top: "Accellerated Reader" http://www.renlearn.com/ar/overview/requirements.htm says they have a web version that explicitly supports Linux clients! So maybe there's no reason to work on Accellerated Reader. "print shop" lives at http://www.broderbund.com/jump.jsp?itemID=834&mainPID=834&itemType=PRODUCT and sadly has no free demo. But $50 isn't too steep as these things go. "Kidspiration", free trial available at http://www.inspiration.com/freetrial/ "Kid Pix" - there are several versions mentioned (kid pix studio, kid pix deluxe, kid pix deluxe for schools). kid pix deluxe 4 for schools lives at http://rivapprod2.riverdeep.net/portal/page?_pageid=414,105278&_dad=portal&… and has a free demo, for educators at least. "Oregon Trail" - Oregon Trail 5 has a free demo, see http://www.learningcompany.com/jump.jsp?itemID=147&itemType=PRODUCT&RS=1 "Reader Rabbit" has so many versions I can't pick one to mention here.

Sept. 23, 2005

Re: listview crash fix

by Dimi Paun

Is this one working for you? Sorry, not much tested here other than a quick compile, it's past 2:30am already... -- Dimi Paun <dimi(a)lattica.com> Lattica, Inc.

Sept. 23, 2005

autogen configure script

by Steven Edwards

Hi, A lot of people don't like the fact we diff configure and configure.ac so here is a autogen.sh ripped from the Samba4 source. Can we get something like this in cvs? Thanks Steven

Sept. 23, 2005

Re: server: Try4: add case-sensitivity comparison flag to find_object & co

by Vitaliy Margolen

Sorry, wrong list x-| Thursday, September 22, 2005, 2:08:55 PM, Vitaliy Margolen wrote: > Use this one instead - it should apply to the corrected previous patch. > This is just a case-sensitivity patch to the server. It depends on previous > patch to pass OBJECT_ATTRIBUTES.Attributes to the server. > I removed case_sensitive flag from name space, since it's no use anymore. > Vitaliy Margolen > changelog: > server: Add case-sensitivity comparison flag to find_object & co

Sept. 22, 2005

Re: [wined3d] oprionally repack non-power2 textures [resend]

by Raphael

On Thursday 22 September 2005 22:06, Oliver Stieber wrote: > Hi, > This patch optionally repacks nonpower 2 textures to the expected span > length when a textures is locked. This is necessary for buggy games like > Warhammer 40k that don't work with the odd span sizes produce by default > nonpower 2 support. > > repacking can be quite CPU intensive, and warhammer 40k is the only game so > far that needs nonpower2 texture repacking so it's turned off by default > but can be turned on via the following registry key: > HKEY_CURRENT_USER/Software/Wine/Direct3D Nonpower2Mode [repack] > > Other options for nonpower2 textures will be added in the future, e.g. > OpenGL2 and texturerectangle. > > Oliver. You forget the patch :) Keep the good job Regards, Raphael

Sept. 22, 2005

Re: server: Try4: add case-sensitivity comparison flag to find_object & co

by Vitaliy Margolen

Use this one instead - it should apply to the corrected previous patch. This is just a case-sensitivity patch to the server. It depends on previous patch to pass OBJECT_ATTRIBUTES.Attributes to the server. I removed case_sensitive flag from name space, since it's no use anymore. Vitaliy Margolen changelog: server: Add case-sensitivity comparison flag to find_object & co

Sept. 22, 2005

Re: Problem in a installation on IBM zSeries s/390

by Vitaliy Margolen

Thursday, September 22, 2005, 12:55:18 PM, Ricardo Azzolini wrote: > I´m having intallation problems with WINE ( when interlocked.c is compiled > the following message appears: "error you must implement the interlocked* > functions for your CPU"). That means that this function is not implemented for your platform. It's only a first from lots more that are not implemented. > Have you ever seen this error before? Do you have any suggestions ? Yes, pull out your book on what ever CPU it's using and start coding. ;-) Vitaliy

Sept. 22, 2005

Problem in a installation on IBM zSeries s/390

by Ricardo Azzolini

IŽm having intallation problems with WINE ( when interlocked.c is compiled the following message appears: "error you must implement the interlocked* functions for your CPU"). I've been working with Linux SUSE Enterprise Server 9 for s/390 zSeries, and WINE-20050830. Have you ever seen this error before? Do you have any suggestions ? Thanks, Ricardo Azzolini

Sept. 22, 2005

Re: WooHoo get a load of this :-)

by Ivan Leo Puoti

Marcus Meissner wrote: >On Wed, Sep 21, 2005 at 06:39:06PM +0100, Ivan Leo Puoti wrote: > > >>Tom Wickline wrote: >> >> >>>4. System becomes the property of SpecOpS Laboratories. >>> >>> >>Oh yeah, someone writes a full windows API implementation in a month or so, >> >> > >You misread the mail... > >It says "One Application" and "get it running using WINE". > >I would consider myself nearly capable to do that myself. > >Ciao, Marcus > > > Well, that depends very much on the application, still I think those guys are smoking crack. Ivan.

Sept. 22, 2005

Re: Thanks for your quick reply...

by Vitaliy Margolen

Thursday, September 22, 2005, 11:23:43 AM, Esko Woudenberg wrote: > Would I be correct in guessing that: > "Relayexclude"="RtlEnterCriticalSection;RtlLeaveCriticalSection; > _EnterSysLevel;_LeaveSysLevel;_CheckNotSysLevel;NtCurrentTeb; > LdrAccessResource;RtlUpperChar;kernel32.97;kernel32.98;TlsGetValue" > are generally not important messages for debugging? I think wine beat them to death so you can be sure they do work ;-) Also they are the most commonly called functions that don't mean anything unless you are interested in that particular area. > (and also is +relay enough or should I use +all?) +relay is a good starting point to see what is going on. But having a source, you don't really need it. You should know where is your program fails what part of Wine (like ole, comctl, file ops, etc.) Then you can enable that trace and investigate that part. >> Ok, here it looks like it either really runs out of >> stack or just >> trying to access the last page of the stack. Wine >> has a problem in those >> areas. It doesn't grow stack and throws that >> exception when app tries to >> access last page of the stack, that works under >> windows. > If that is the case, is there a work-around? Sort of. I have a program that just clears whole stack after it's done. For this I just disabled guard pages that Wine uses to raise that exception. If your program really runs out of stack - i.e. ESP points to the last stack page then you need to find out why does it do that. It could be just a case of nested functions that never return because they not getting correct results from Wine. > Detecting Wine... >> Bad idea. You need to find out what is wrong with >> Wine and fix it. >> Or find the problem and let us fix it. > I understand your concerns but aim to follow parallel > paths... (Might end up prompting user "Wine detected > - Work around known problems? Y/N") Most of the time, you should be able to find the problem and fix it in Wine. The hardest part is to find it, not to fix it (unless that is something really complicated. But then you can't just remove huge chunk of your program anyway.) If you can find the problem, I'm sure people here will help you fix it, if you don't know how. What you can do in this case is to add a test to Wine's regression test suite. Vitaliy

Sept. 22, 2005

Re: Running Wine from source tree

by Detlef Riekenberg

Am Mittwoch, den 21.09.2005, 22:49 -0700 schrieb James Liggett: > I'm doing a lot of testing, Same for me. > and I was wondering if I can run Wine from > within its source tree without having to reinstall it after every build. I'm running wine out of source-tree within this directories: ~/wine.cvs ~/wine.cvs/bin ~/wine.cvs/docs ~/wine.cvs/src .. and added some alias in ~/.bashrc => ".bashrc_4_wine" I also have scripts for "wineprefixcreate" => "_wpc" and for compiling wine => "_makewine" Replace "wine.cvs" with "wine.6" (w6, w6s, w6b), "wine.7" (w7, w7s, w7b) and so on for running with monthly snapshots. Have fun. -- By By ... ... Detlef

Sept. 22, 2005

Thanks for your quick reply...

by Esko Woudenberg

Would I be correct in guessing that: "Relayexclude"="RtlEnterCriticalSection;RtlLeaveCriticalSection; _EnterSysLevel;_LeaveSysLevel;_CheckNotSysLevel;NtCurrentTeb; LdrAccessResource;RtlUpperChar;kernel32.97;kernel32.98;TlsGetValue" are generally not important messages for debugging? (and also is +relay enough or should I use +all?) > Ok, here it looks like it either really runs out of > stack or just > trying to access the last page of the stack. Wine > has a problem in those > areas. It doesn't grow stack and throws that > exception when app tries to > access last page of the stack, that works under > windows. If that is the case, is there a work-around? Detecting Wine... > Bad idea. You need to find out what is wrong with > Wine and fix it. > Or find the problem and let us fix it. I understand your concerns but aim to follow parallel paths... (Might end up prompting user "Wine detected - Work around known problems? Y/N") __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com

Sept. 22, 2005

Re: Newbie Wine question... Borland C++ Builder and Firebird via FIBplus

by Molle Bestefich

Esko Woudenberg wrote: > I found info on WINEDEBUG and am currently > running it set to +relay This results in 300+ > Megabyte debug file. (Ran system out of memory trying > to loot at it with Konq) I have then been using > "tail" to grab just the last few thousand lines of > it... It always ends with hundreds/thousands of calls > to ntdll.RtlEnterCriticalSection and > ntdll.RtsLeaveCriticalSection Wonderful tool, that WINEDEBUG! I'm very much a newbie too, and the WINEDEBUG stuff really helps me getting just a little bit of understanding about what is going on. Love it. In the manual here: http://www.winehq.com/docs/wine-user.html#CONFIG-FILE-HOW I've found that you can exclude those CriticalSection messages which you probably don't care about. That should give you a much smaller file and faster execution times, which is always nice when you're debugging. You need to put the following key/value somewhere in your registry - the manual and winehq pages are, ahem, not very explicit about where to put them, but supposedly somewhere under key named 'Debug'...: "RelayExclude" = "RtlEnterCriticalSection;RtlLeaveCriticalSection" > followed by > err:seh:setup_exception stack overflow ### bytes in > thread #### eip ###### esp ###### stack 0x##### - > 0x#### Found a bit in a wine debugging tutorial: http://wiki.winehq.org/Debugging_'Wild_Metal_Country' Which explains nicely that what you see is an infinite loop, I think somewhere in Wine's exception handling code, and that you should debug with +seh to see what's going on. > I am planning on using the HKLM\Software\Wine\Wine > existence to let the app know it is running on wine... Sounds hackish :-). > The app itself also has a "debug mode" that logs > events to a file while running - is there a way to > send these messages to the same log as WINEDEBUG? > That way I would be able to see what method or > function of the app resulted in the thousands of calls > into the WIN32 API... Quah? Mebbe. Easiest thing would probably be to make your app write to stderr instead of that file it's using? There's also a hack in the manual (see first link) that will send WINEDEBUG messages to a file, but then again so will 2> ...

Sept. 22, 2005

Re: Newbie Wine question... Borland C++ Builder and Firebird via FIBplus

by Vitaliy Margolen

Thursday, September 22, 2005, 10:12:52 AM, Esko Woudenberg wrote: > I keep getting "segmentation fault" when running the > app. I found info on WINEDEBUG and am currently > running it set to +relay This results in 300+ > Megabyte debug file. (Ran system out of memory trying you can add this to your user.reg file to limit size of it (It's two lines, split for readability): [Software\\Wine\\Debug] 1125240062 "Relayexclude"="RtlEnterCriticalSection;RtlLeaveCriticalSection; _EnterSysLevel;_LeaveSysLevel;_CheckNotSysLevel;NtCurrentTeb; LdrAccessResource;RtlUpperChar;kernel32.97;kernel32.98;TlsGetValue" > err:seh:setup_exception stack overflow ### bytes in > thread #### eip ###### esp ###### stack 0x##### - > 0x#### Ok, here it looks like it either really runs out of stack or just trying to access the last page of the stack. Wine has a problem in those areas. It doesn't grow stack and throws that exception when app tries to access last page of the stack, that works under windows. > I am planning on using the HKLM\Software\Wine\Wine > existence to let the app know it is running on wine... Bad idea. You need to find out what is wrong with Wine and fix it. Or find the problem and let us fix it. > The app itself also has a "debug mode" that logs > events to a file while running - is there a way to > send these messages to the same log as WINEDEBUG? Yes, just write to stderr. Vitaliy

Sept. 22, 2005

Re: WooHoo get a load of this :-)

by Kuba Ober

On Thursday 22 September 2005 11:37, Marcus Meissner wrote: > On Wed, Sep 21, 2005 at 06:39:06PM +0100, Ivan Leo Puoti wrote: > > Tom Wickline wrote: > > >4. System becomes the property of SpecOpS Laboratories. > > > > Oh yeah, someone writes a full windows API implementation in a month or > > so, > > You misread the mail... > > It says "One Application" and "get it running using WINE". > I would consider myself nearly capable to do that myself. I'd say we should all wait till we see what's in the challenge registration form, where they are actually supposed to reveal exactly what those $10k buys them. One application doesn't mean much. It can be something huge. It can be something that needs something that wine nowhere near implements. And so on... If they're willing to pay someone $10k then either it's not such a trivial job, or their VCs want some sort of demonstration and the SpecOpS people just don't have enough expertise. Or it's all just blown out of proportion with $10k being way more than good compensation for the work involved, but someone clueless enough having $10k to toss around. Maybe it's just worth for some wine hackers to register and see what comes out of it. The mere fact that they don't mention what application would it be sounds suspicious to me. Cheers, Kuba

Sept. 22, 2005

Re: IDE disk geometry

by Michael Ost

On Wed, 2005-09-21 at 00:57, Uwe Bonnes wrote: > Our IOCTL handling is a mess. Seems like it has only been needed and implemented for CDs, so far. > I had a patch that splitted out IOCTL code handling to separate, loadable > dlls, but the patch went by unnoticed. Look for Mail from around July 10 > "Uniform SYS/VXD Handling x/6" > In your case, when you did CreateFile("\\c:\"), in file.c it would be > noticed that a raw handle on the disk is needed and something like > "disk.sys" would be loaded, where you would implement the needed IOCTL. That sounds great. I wonder why it didn't go through...? But that's a bit too much heavy lifting of the wine goblet for me to undertake! - mo

Sept. 22, 2005

Newbie Wine question... Borland C++ Builder and Firebird via FIBplus

by Esko Woudenberg

I have a Borland C++ Application (with full source code) that I am trying to get to run under Wine 20050830. (I plan on switching to CVS builds once I see some progress) I keep getting "segmentation fault" when running the app. I found info on WINEDEBUG and am currently running it set to +relay This results in 300+ Megabyte debug file. (Ran system out of memory trying to loot at it with Konq) I have then been using "tail" to grab just the last few thousand lines of it... It always ends with hundreds/thousands of calls to ntdll.RtlEnterCriticalSection and ntdll.RtsLeaveCriticalSection followed by err:seh:setup_exception stack overflow ### bytes in thread #### eip ###### esp ###### stack 0x##### - 0x#### I am planning on using the HKLM\Software\Wine\Wine existence to let the app know it is running on wine... The app itself also has a "debug mode" that logs events to a file while running - is there a way to send these messages to the same log as WINEDEBUG? That way I would be able to see what method or function of the app resulted in the thousands of calls into the WIN32 API... I am trying to figure out how to: A> Turn the above errors into useful bug reports B> Be able to resolve limitations or errors in wine as I came across them... I do NOT speak for the U.S.Army Corps of Engineers. I am a contractor supporting one of their projects... The recent developments in Mass re Open Formats motivated me to try to get this app working under wine so that running Windows (only platform currently supported by this software) would not be a requirement in order to work on large USACE contracts... __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com

Sept. 22, 2005

Re: WooHoo get a load of this :-)

by Marcus Meissner

On Wed, Sep 21, 2005 at 06:39:06PM +0100, Ivan Leo Puoti wrote: > Tom Wickline wrote: > >4. System becomes the property of SpecOpS Laboratories. > > Oh yeah, someone writes a full windows API implementation in a month or so, You misread the mail... It says "One Application" and "get it running using WINE". I would consider myself nearly capable to do that myself. Ciao, Marcus

Sept. 22, 2005

Re: ntdll: Add few tests for object manager

by Vitaliy Margolen

Thursday, September 22, 2005, 3:54:06 AM, Alexandre Julliard wrote: >> Vitaliy Margolen <wine-patch(a)kievinfo.com> writes: >> All of them are to-do. Thanks to Robert Shearman for examples. ;-) >> Can't really test nothing more since most of the om functions are stubs. > Come on, you can test a lot more than that. Even if you only test > create and open functions, if you test the various case combinations, > with or without the CASE_INSENSITIVE flag, for mutexes, events, > semaphores, pipes, mappings, files, etc. you should be able to easily > come up with a few hundred tests. Well, yes you are right but these tests just enough to test case-sensitivity problem. All the rest of the objects well have the same problems as well. Things like pipes & files are a whole different story. They have their own name space that is different from the object manager name space. ATM wine doesn't have that, so it will be pointless testing something that does not exists at all. Like \\??\\PIPE is a symlink to \\Device\\NamedPipe which is a device. I don't know how windows behaves when using case-sensitive op and the name of the named pipe differs, but the path is correct. Likewise NtFsControlFile function does not take whole OBJECT_ATTRIBUTES nor FSCTL_PIPE_WAIT associated structure. So how does windows specify case-sensitivity? Like everything else new in wine this is just a starting point and can be extended as we go. I will probably have more tests as soon as I get name space re-implemented using directory object. There are number of interesting things to test ;-) Vitaliy

Sept. 22, 2005