-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

Hello,

on 05/31/04 21:11, Tim Hentenaar wrote:

A friend of mine who has a recent release of wine uses Photoshop 7 just fine. It was he and I who took the screenshot that's in the app database (we were using 2003-12-12 at the time). He just updated his wine install via gentoo's emerge about a week and a half ago.

this sounds me strange, as in current CVS from command cvs -d $CVSROOT update -D "2004-06-01 CST" the XP extensions which cause the regression are still present.

BTW, in the bug #2017 [1] there're some reports by Gentoo users about the same problem... Could you try to find which version does your friend use?

Thx, bye, Gismo / Luca

[1] http://bugs.winehq.org/show_bug.cgi?id=2017

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

Hello,

on 06/01/04 00:04, Mike Hearn wrote:

For now try this patch, does it work?

yes, it works on CVS-20040130-165721 and CVS-20040601.

Actually, the problem is specific to this declaration: @ stdcall QueryActCtxW(long ptr ptr long ptr long ptr) If I leave it, I got the 'success' window, while if I comment out it, Photoshop 7.0 starts with no problem. I tested it on CVS-20040601. Please, could someone other confirm this?

thanks -mike

You're welcome, I'm not a Photoshop user (thanks GIMP), but a collegue of mine needs it...

Index: kernel32.spec

RCS file: /home/wine/wine/dlls/kernel/kernel32.spec,v retrieving revision 1.133 diff -u -r1.133 kernel32.spec --- kernel32.spec 23 Apr 2004 02:46:18 -0000 1.133 +++ kernel32.spec 31 May 2004 22:02:35 -0000

BTW, which is the correct command to apply this patch? Because I got an HUNK failed (and applied it by hand): ===== luca@gismo:~/cvs/wine-cvs-20040601$ cd dlls/kernel/ luca@gismo:~/cvs/wine-cvs-20040601/dlls/kernel$ patch -p0 < /home/luca/wine-photoshop7-mike-1.diff patching file kernel32.spec Hunk #1 FAILED at 1078. 1 out of 1 hunk FAILED -- saving rejects to file kernel32.spec.rej luca@gismo:~/cvs/wine-cvs-20040601/dlls/kernel$ =====

Thx, bye, Gismo / Luca

PS obviously, I'm here for more tests :-)

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

Hello,

on 06/01/04 20:10, Mike Hearn wrote:

OK, could you submit a patch to comment that line out? These functions are just stubs anyway. When they are more complete it's trivial to renable them.

patch posted on bug #2017 [1]. Direct link to the patch [2].

Thx, bye, Gismo / Luca

[1] http://bugs.winehq.org/show_bug.cgi?id=2017 [2] http://bugs.winehq.org/attachment.cgi?id=598&action=view

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

Hello,

on 06/02/04 13:35, Mike Hearn wrote:

Normally you would post the patch to wine-patches for review. Could you do that please? We don't use bugzilla for patch tracking.

sure, done [1]. Sorry for my confusion.

Thx, bye, Gismo / Luca

[1] http://www.winehq.org/hypermail/wine-patches/2004/06/0020.html

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

Hello,

on 06/01/04 00:34, Troy Rollo wrote:

On Tue, 1 Jun 2004 01:53, Luca Capello wrote:

...

as reported on bug #2017 [1], Adobe Photoshop 7.0 stops working after wine-20031212 (in the majority of cases).

*In what way* does it stop working? It might be that you're using a

feature

that Wine coders using Photoshop 7.0 don't use.

I thought it was superfluous to report again the problem, as it's perfect reported in the bug #2017 [1] and in the Wine App DB (Photoshop 7.0, [2]). Anyway, here the description: ===== luca@gismo:~/cvs/wine-cvs-20040130-165721$ ./wine /home/luca/.wine/fake_windows/Program\ Files/Adobe/Photoshop\ 7.0/Photoshop.exe fixme:keyboard:X11DRV_KEYBOARD_DetectLayout Your keyboard layout was not found! Using closest match instead (United States keyboard layout (phantom key version)) for scancode mapping. Please define your layout in dlls/x11drv/keyboard.c and submit them to us for inclusion into future Wine releases. See the Wine User Guide, chapter "Keyboard" for more information. fixme:win32:SetCriticalSectionSpinCount critsection=0x40c338e0: spincount=20 not supported fixme:win32:SetCriticalSectionSpinCount critsection=0x40c33fa4: spincount=500 not supported fixme:win32:SetCriticalSectionSpinCount critsection=0x40c33aa4: spincount=500 not supported fixme:win32:SetCriticalSectionSpinCount critsection=0x40c33ca4: spincount=500 not supported fixme:win32:SetCriticalSectionSpinCount critsection=0x40c34028: spincount=500 not supported fixme:win32:SetCriticalSectionSpinCount critsection=0x40c3414c: spincount=500 not supported fixme:win32:SetCriticalSectionSpinCount critsection=0x40c341c8: spincount=50 not supported fixme:actctx:QueryActCtxW stub! fixme:actctx:QueryActCtxW stub! fixme:actctx:QueryActCtxW stub! fixme:actctx:QueryActCtxW stub! fixme:actctx:QueryActCtxW stub! fixme:actctx:QueryActCtxW stub! [a small window appears with "Success" and the 'OK' button, clicking on the button closes the window and exits wine with value 0] luca@gismo:~/cvs/wine-cvs-20040130-165721$ =====

If needed, I can provide the full 'WINEDEBUG=+warn,all ./wine' output.

Thx, bye, Gismo / Luca

[1] http://bugs.winehq.org/show_bug.cgi?id=2017 [2] http://appdb.winehq.org/appview.php?appId=17&versionId=1336

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

Hello,

on 06/01/04 16:55, Rein Klazes wrote:

I anything, a WINEDEBUG=+relay trace would be more useful.

sure, the problem is that it's quite big... ===== luca@gismo:~/cvs/wine$ script ~/winedebug+relay_wine-cvs-20040601_photoshop7.0 <cut> WINEDEBUG=+relay ./wine /home/luca/.wine/fake_windows/Program\ Files/Adobe/Photoshop\ 7.0/Photoshop.exe <cut>

luca@gismo:~/cvs/wine$ ll ~/winedebug+relay_wine-cvs-20040601_photoshop7.0 - -rw-r--r-- 1 luca luca 46757031 Jun 1 17:47 /home/luca/winedebug+relay_wine-cvs-20040601_photoshop7.0

luca@gismo:~/cvs/wine$ gzip ~/winedebug+relay_wine-cvs-20040601_photoshop7.0 luca@gismo:~/cvs/wine$ ll ~/winedebug+relay_wine-cvs-20040601_photoshop7.0.gz - -rw-r--r-- 1 luca luca 1169175 Jun 1 17:47 /home/luca/winedebug+relay_wine-cvs-20040601_photoshop7.0.gz

luca@gismo:~/cvs/wine$ bzip2 ~/winedebug+relay_wine-cvs-20040601_photoshop7.0 luca@gismo:~/cvs/wine$ ll ~/winedebug+relay_wine-cvs-20040601_photoshop7.0.bz2 - -rw-r--r-- 1 luca luca 415184 Jun 1 17:47 /home/luca/winedebug+relay_wine-cvs-20040601_photoshop7.0.bz2 =====

I can easily upload the file on my web, if needed, but before I'd like to know if the command I used is the correct one.

Thx, bye, Gismo / Luca

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

Hello,

on 06/01/04 18:28, Rein Klazes wrote:

That is one possibility. You can also attach it in bugs.winehq.org to the bug report that you mentioned before.

I'm so stupid I didn't think at it. Debug attached on bug #2017 [1].

Thx, bye, Gismo / Luca

[1] http://bugs.winehq.org/show_bug.cgi?id=2017

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

Hello,

on 06/02/04 10:55, Rein Klazes wrote:

Here are the details of the QueryActCtxW call:

|0009:Ret advapi32.RegCloseKey() retval=00000000 ret=00be869f |0009:Call

kernel32.QueryActCtxW(80000010,0116283c,00000000,00000001,406aebc0,00000008,00000000) ret=004bf17f

|fixme:actctx:QueryActCtxW stub! |0009:Ret kernel32.QueryActCtxW() retval=00000000 ret=004bf17f |0009:Call kernel32.GetLastError() ret=006068aa |0009:Ret kernel32.GetLastError() retval=00000000 ret=006068aa

As expected the program inspects the last error value for details why the calls fails. Setting the last error value is missing form the QueryActCtxW stub, so the program may well be confused (especially when it receives ERROR_SUCCESS). So if you to try something you may try whether inserting a line like:

SetLastError(ERROR_NOT_SUPPORTED);

will make the program do the right thing (forgetting about using these functions). There are several more error codes that you may try:

http://msdn.microsoft.com/library/default.asp?url=/library/en-us/debug/base/...

with a number of specific activation context error codes in the 12000-14999 region.

simple problem: I'm not a good programmer (some BASH, C, Perl and php knowledge), so where should I insert the above line?

If not, you can always follow Mike's advice.

Already done.

The real problem: which is the best solution for this problem? Commenting out the declaration following Mike's advice (as it seems that that it isn't used) or trying to find the correct error following your advice? I've no problem at making test to find the error code...

Another fact that I forgot to tell you is that on the latest CrossOffice 3.0 demo, Photoshop 7.0 works like a charm with no problem at all.

Thx, bye, Gismo / Luca

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

Hello,

on 06/02/04 14:19, Rein Klazes wrote:

It is a small function, hard to do it wrong there. Just put it before the line with return .

I'm so stupid, I didn't think that I should see in the other files modified by the patch. Thank you for the explanation.

After you did that and saved your changes, do a make followed by make install in the directory dll's/kernel. That saves a lot of time by not having to recurse twice through the tree.

I haven't installed the wine-cvs, so I always made - - make distclean - - cvs -d... [to update] - - ./configure && make depend && make - - ./wine [...]/Photoshop.exe

Now for these tests I saw that after a change, just 'make' it's ok.

Mike's solution seems certain to work. Disadvantage is that it may break another program (that /needs/ the presence of this stub).

As I supposed...

Fixing it the way I described is a far better solution, but I have no idea if it even has a chance to succeed.

If I were you, I would try a few values and if unsuccessful go for Mike's solution (but I am not you, neither have I easy access to the program in question).

Ok, as I told you, I've no problem to make tests... I mean, now that we've a temporary solution, I can try different error values in background of my normal work. Just to see if I fall on the 'right' one. I tested only the error values the description of which contains 'activation context'. I reported the results in these form:

error number, error description yes(work)/no(doesn't), text included in the window that appears

===== ERROR_NOT_SUPPORTED no, Request not supported

13016, The application attempted to activate a disabled activation context. no, Cannot load Photoshop.dll: Unkwown error [copying 'Photoshop.dll' in 'fake_windows/windows/system/' didn't solve it]

13017, The requested lookup key was not found in any active activation context. no, Cannot load Photoshop.dll: Unkwown error

14000, The requested section was not present in the activation context. no, Cannot load Photoshop.dll: Unkwown error

14001, This application has failed to start because the application configuration is incorrect. Reinstalling the application may fix this problem. no, Cannot load Photoshop.dll: Unkwown error

14002, The application binding data format is invalid. no, Cannot load Photoshop.dll: Unkwown error

14006, The application attempted to activate a disabled activation context. no, Cannot load Photoshop.dll: Unkwown error

14007, The requested lookup key was not found in any active activation context. no, Cannot load Photoshop.dll: Unkwown error

14008, A component version required by the application conflicts with another component version already active. no, Cannot load Photoshop.dll: Unkwown error

14009, The type requested activation context section does not match the query API used. no, Cannot load Photoshop.dll: Unkwown error

14010, Lack of system resources has required isolated activation to be disabled for the current thread of execution. no, Cannot load Photoshop.dll: Unkwown error

14011, An attempt to set the process default activation context failed because the process default activation context was already set. no, Cannot load Photoshop.dll: Unkwown error

14082, The activation context activation stack for the running thread of execution is corrupt. no, Cannot load Photoshop.dll: Unkwown error

14084, The activation context being deactivated is not the most recently activated one. no, Cannot load Photoshop.dll: Unkwown error

14085, The activation context being deactivated is not active for the current thread of execution. no, Cannot load Photoshop.dll: Unkwown error

14086, The activation context being deactivated has already been deactivated. no, Cannot load Photoshop.dll: Unkwown error

14088, A kernel mode component is releasing a reference on an activation context. no, Cannot load Photoshop.dll: Unkwown error

14089, The activation context of system default assembly could not be generated. no, Cannot load Photoshop.dll: Unkwown error =====

So, it seems that there's only Mike's solution, if someone has no other ideas...

Thx, bye, Gismo / Luca

Luca Capello [email protected] writes:

So, it seems that there's only Mike's solution, if someone has no other ideas...

If simply returning an error doesn't work, you need to improve the stub to provide enough functionality for the app to be happy.

On Fri, 04 Jun 2004 15:34:41 +0200, you wrote:

...

If simply returning an error doesn't work, you need to improve the stub to provide enough functionality for the app to be happy.

well, the problem is that I'm not a programmer and I don't know in which way I should improve the stub. So I think for the moment I'm going to use 'wine' commenting the QueryActCtxW call out.

OK, I managed to get a copy of Photoshop and fixed the problem here. It is a Dutch version, please verify that it works with what you have.

I was not so far from a fix after all, just needed another error code.

Changelog: dlls/kernel : actctx.c Set last error value to ERROR_CALL_NOT_IMPLEMENTED in QueryActCtxW() to make Photoshop 7.0 work again.

Rein.

On Wed, 02 Jun 2004 14:19:07 +0200, Rein Klazes wrote:

Mike's solution seems certain to work. Disadvantage is that it may break another program (that /needs/ the presence of this stub).

That's unlikely at this point. The activation context APIs are XP/Server2003 only, very very few apps (if any) will have a hard dependency on them.

I agree we should fill out these APIs, but this could be a bit tricky. They are rather poorly explained on MSDN and appear to exist to support .NET, which we don't bother with right now. For instance the whole manifest/assembly concept is AFAICT useful mostly for apps which can use the .NET runtime ...

thanks -mike