Jump to content

Bugs and problems, discussion


Slice
 Share

5 posts in this topic

Recommended Posts

Here I want to report about problems and solutions if found.

Hope someone sometimes will help me somehow.

 

at this moment

$ git describe head

wine-1.7.44-183-g61028e0

 

but there is no our own numeration. How to do this?

 

I implemented my patches and step by step impelent patches from stage because something is not working.

Namely

3DMark05 was working with Sherry-102 but not working now. "Unknown exception".

Step by step I make sources will be the same...

Link to comment
Share on other sites

Resolved!

$ git describe head

wine-1.7.44-185-g52a0a5e

 

Now I want to show master-class. A simple lesson.

1. I created new blank WineskinWrapper with vanilla Wine-1.7.44 and installed the application. One trick: uninstall embedded Futuremark System Info and install version 4.2.8 - only working version.

And I see exactly the same stop!

conclusion: The problem is in pure Wine as in Stage it is working. I have to searc unimplied patches. 

 

2. Sure it is not DirectX patches. I may change them different way.

 

3. Test with several other Wine engines 1.7.41, 1.7.39, 1.7.17, 1.7.2 and I have random success!

Compare logs and see next messages about files and they related to ntdll.

Conclusion: search patches related to Files.

 

4. Look wine-stage list patches about ntdll-File. Apply one by one with dependences and test each result. Success!

 

The changes committed and for now I will stop looking into stage project. I am satisfied with current Sherry.

I will do better graphics.

Link to comment
Share on other sites

Next useful patch I accepted is implement CorValidateImage by Michael Muller.

Without it I can't start OpenGL Extension Viewer (for windows). Now I can

Screen Shot 2015-06-15 at 14.06.43.png

Link to comment
Share on other sites

At the same computer

Screen Shot 2015-06-17 at 12.52.37.png

Why?

 

I just found that original Wine works faster. But I remember that old Sherry works faster then Wine.

So making several tests I found there is a regression in new Sherry. OK, will search.

 

Git Bisecting

 

This is very useful feature!

Propose we know old revision that works good and know new revision that works bad.

Do the follow

$ git bisect start
$ git bisect good 1dcc531ec1af9a5d9aa0fad2101009f098184e3e
$ git bisect bad head

Then it said

Bisecting: 8 revisions left to test after this (roughly 3 steps)

and set some revision to start.

Compile it and test. See result and inform git about this revision

$ git bisect good

or 

$ git bisect bad

this is my log (partially)

Sergeys-iMac:Wine-Sherry slice$ git bisect bad ba69ae20762e3e2148049d4edc76a51168c27da7
Bisecting: 8 revisions left to test after this (roughly 3 steps)
[649366c536a70767b3f193ac8fb3ca3e4753d49b] Merge source code for Wineskin
Sergeys-iMac:Wine-Sherry slice$ git bisect good
Bisecting: 4 revisions left to test after this (roughly 2 steps)
[387895e1d7aeeec5fac3408a31a63644b93c33c8] move .gitignore from the Wineskin dir to the root project
Sergeys-iMac:Wine-Sherry slice$ git bisect bad
Bisecting: 1 revision left to test after this (roughly 1 step)
[0a296b4eca2635988024b770076ffbbbeedee1cf] accept patch by Linus Unnebäck from bug34166
Sergeys-iMac:Wine-Sherry slice$ git bisect bad
Bisecting: 0 revisions left to test after this (roughly 0 steps)
[f7176db439974152afa8a731918e4e1f87c561e0] accept patch fromstage kernel32-named-pipe
Sergeys-iMac:Wine-Sherry slice$ git bisect good
0a296b4eca2635988024b770076ffbbbeedee1cf is the first bad commit
commit 0a296b4eca2635988024b770076ffbbbeedee1cf
Author: Sergey Isakov <sergey.slice@gmail.com>
Date:   Thu Jun 11 16:13:12 2015 +0300

    accept patch by Linus Unnebäck from bug34166

:040000 040000 f3e3ca7908edd4797115ecf4b3087dc1ccd70edb 1121aaf0310179b56d497f0123683faec4bae891 M	dlls

Looking into this patch I understand that early it works for me but with updated Wine it became not good.

OK, I reverted it and got expected good result!

 

Finish bisecting

git stash
git bisect reset

Apply wrong patch to revert it and commit as new revision.

Link to comment
Share on other sites

I mostly implemented all patches known for me.

1. Latest wine-hq 1.7.45-213 but partially

2. Latest wine-stage but only chosen patches

3. All my investigations

4. Some patches from Bugzilla (6955, 23807, 35397, others ) not accepted neither to wine-hq nor to wine-stage. These are good patches but commiters didn't confirm stupid Wine Rules. 

Now we can move forward.

 

This is latest engine

WS9WineXSherry1.7.45-219g.tar.7z.zip

Link to comment
Share on other sites

 Share

×
×
  • Create New...