No hide button in 0.9.60 ?

Spuuk · Post by **Spuuk** » Sat Apr 26, 2008 10:18 am

i am running flashfxp under wine now a while, recently i upgraded to kubuntu 8.04 and installed wine 0.9.60.

I discovered that then there is no hide button anymore in FlashFXP.

Onbly 2 Buttons to the left, the close and the maximize button.

I then re-installed 0.9.59 and with that i get again the minimse button back, unfortunaly this version has a bug not displaying the program in the taskbar.

I have then installed 0.9,58 version, with this one evrything works.

So is this a new feature of the 0.9.60 or have i done something wrong ?
Some other programs has the minimise button , just can't see it on flashfxp.

Zachary Goldberg · Post by **Zachary Goldberg** » Sat Apr 26, 2008 11:27 am

On Sat, Apr 26, 2008 at 11:18 AM, Spuuk <[email protected]> wrote:

i am running flashfxp under wine now a while, recently i upgraded to kubuntu 8.04 and installed wine 0.9.60.

I discovered that then there is no hide button anymore in FlashFXP.

Onbly 2 Buttons to the left, the close and the maximize button.

I then re-installed 0.9.59 and with that i get again the minimse button back, unfortunaly this version has a bug not displaying the program in the taskbar.

I have then installed 0.9,58 version, with this one evrything works.

So is this a new feature of the 0.9.60 or have i done something wrong ?
Some other programs has the minimise button , just can't see it on flashfxp.

Would it be possible for you to do a regression test to find out where
the button goes missing?

http://wiki.winehq.org/RegressionTesting

--
Zachary Goldberg
Computer Science & Engineering
Electrical Captain of Penn Electric Race Team
School of Engineering at the University of Pennsylvania

Spuuk · Post by **Spuuk** » Sat Apr 26, 2008 2:35 pm

Zachary Goldberg wrote:On Sat, Apr 26, 2008 at 11:18 AM, Spuuk <[email protected]> wrote:
i am running flashfxp under wine now a while, recently i upgraded to kubuntu 8.04 and installed wine 0.9.60.

I discovered that then there is no hide button anymore in FlashFXP.

Onbly 2 Buttons to the left, the close and the maximize button.

I then re-installed 0.9.59 and with that i get again the minimse button back, unfortunaly this version has a bug not displaying the program in the taskbar.

I have then installed 0.9,58 version, with this one evrything works.

So is this a new feature of the 0.9.60 or have i done something wrong ?
Some other programs has the minimise button , just can't see it on flashfxp.

Would it be possible for you to do a regression test to find out where
the button goes missing?

http://wiki.winehq.org/RegressionTesting

--
Zachary Goldberg
Computer Science & Engineering
Electrical Captain of Penn Electric Race Team
School of Engineering at the University of Pennsylvania

Well, i did this and got the following :
~/wine-git$ git bisect good
Bisecting: 271 revisions left to test after this
[732682b08cf3b64267045cf0ab051fb84af43aea] mshtml: Correctly handle NULL req in get_nscontainer_from_load_group.
~/wine-git$

Dont know what do due next

Zachary Goldberg · Post by **Zachary Goldberg** » Sat Apr 26, 2008 2:49 pm

On Sat, Apr 26, 2008 at 3:35 PM, Spuuk <[email protected]> wrote:

Zachary Goldberg wrote:
On Sat, Apr 26, 2008 at 11:18 AM, Spuuk <[email protected]> wrote:

i am running flashfxp under wine now a while, recently i upgraded to kubuntu 8.04 and installed wine 0.9.60.

I discovered that then there is no hide button anymore in FlashFXP.

Onbly 2 Buttons to the left, the close and the maximize button.

I then re-installed 0.9.59 and with that i get again the minimse button back, unfortunaly this version has a bug not displaying the program in the taskbar.

I have then installed 0.9,58 version, with this one evrything works.

So is this a new feature of the 0.9.60 or have i done something wrong ?
Some other programs has the minimise button , just can't see it on flashfxp.

Would it be possible for you to do a regression test to find out where
the button goes missing?

http://wiki.winehq.org/RegressionTesting

--
Zachary Goldberg
Computer Science & Engineering
Electrical Captain of Penn Electric Race Team
School of Engineering at the University of Pennsylvania

Well, i did this and got the following :
~/wine-git$ git bisect good
Bisecting: 271 revisions left to test after this
[732682b08cf3b64267045cf0ab051fb84af43aea] mshtml: Correctly handle NULL req in get_nscontainer_from_load_group.
~/wine-git$

Dont know what do due next

You now need to compile wine, then test the program. If it works then
do another "git bisect good", if it doesn't work then do "git bisect
bad". Then again compile, test, etc. Until there it says something
other than "x revisions left". It will tell you the exact revision
where there is a problem. You should only have to do this process at
most 5 to 10 times. (Maybe 15 or 20 minutes? The first compile will
take a while but subsequent ones will be very fast)

--
Zachary Goldberg
Computer Science & Engineering
Electrical Captain of Penn Electric Race Team
School of Engineering at the University of Pennsylvania

Spuuk · Post by **Spuuk** » Sun Apr 27, 2008 4:24 am

I compiled it 4 times, until i came to the "bad" status , and got this line :

Bisecting: 17 revisions left to test after this
[c5ebc4acb0af3da2799e58fab6c7f73a3184c51b] kernel32: Fix a test that fails in win2k and up.
./wine "C:\Programme\FlashFXP\FlashFXP.exe"
GST:~/wine-git$ git bisect bad

Bisecting: 8 revisions left to test after this
[fc29e334f6dcaf566bc17c5c5514929a395686f8] start.exe: Add /Unix switch for native file managers.
./wine "C:\Programme\FlashFXP\FlashFXP.exe"
GST:~/wine-git$ git bisect bad

Bisecting: 3 revisions left to test after this
[dd66542011c3a4376638bf4a9efb05fdfa346b85] urlmon: Add stub for CoInternetSetFeatureEnabled.
./wine "C:\Programme\FlashFXP\FlashFXP.exe"
GST:~/wine-git$ git bisect good

Bisecting: 1 revisions left to test after this
[d9e98ea3d0cfe3889afaa848e6dc7173993ff925] d3dx9: Add d3dx9tex.h.
./wine "C:\Programme\FlashFXP\FlashFXP.exe"
GST:~/wine-git$ git bisect good

0 revisions left to test after this
[313a85f1630d2fc2f69815a6cf43e42e0c4829f8] Spelling fixes.

None of this gives me reason whats broken between 0.9.59 anf 0.9.60 .

Edit: i did again a git bisect good and got this :

fc29e334f6dcaf566bc17c5c5514929a395686f8 is first bad commit
commit fc29e334f6dcaf566bc17c5c5514929a395686f8
Author: Vincent Povirk <[email protected]>
Date: Sat Apr 12 20:30:42 2008 -0400

start.exe: Add /Unix switch for native file managers.

:040000 040000 5733b96c3e0b65c9a02620c8b070c36cadf82698 016c4b86a6518cd781a0b1ceffa5e0ea12cc851b M programs

L. Rahyen · Post by **L. Rahyen** » Sun Apr 27, 2008 4:45 am

On Sunday April 27 2008 09:24:54 Spuuk wrote:

I compiled it 4 times, until i came to the "bad" status , and got this line
:

Bisecting: 17 revisions left to test after this
[c5ebc4acb0af3da2799e58fab6c7f73a3184c51b] kernel32: Fix a test that fails
in win2k and up. ./wine "C:\Programme\FlashFXP\FlashFXP.exe"
GST:~/wine-git$ git bisect bad

Bisecting: 8 revisions left to test after this
[fc29e334f6dcaf566bc17c5c5514929a395686f8] start.exe: Add /Unix switch for
native file managers. ./wine "C:\Programme\FlashFXP\FlashFXP.exe"
GST:~/wine-git$ git bisect bad

Bisecting: 3 revisions left to test after this
[dd66542011c3a4376638bf4a9efb05fdfa346b85] urlmon: Add stub for
CoInternetSetFeatureEnabled. ./wine "C:\Programme\FlashFXP\FlashFXP.exe"
GST:~/wine-git$ git bisect good

Bisecting: 1 revisions left to test after this
[d9e98ea3d0cfe3889afaa848e6dc7173993ff925] d3dx9: Add d3dx9tex.h.
./wine "C:\Programme\FlashFXP\FlashFXP.exe"
GST:~/wine-git$ git bisect good

0 revisions left to test after this
[313a85f1630d2fc2f69815a6cf43e42e0c4829f8] Spelling fixes.

None of this gives me reason whats broken between 0.9.59 anf 0.9.60 .

You need to do one more step (git bisect bad/good) to get the patch that
broke your application.

Spuuk · Post by **Spuuk** » Sun Apr 27, 2008 5:06 am

this one ?

spuuk@GST:~/wine-git$ git bisect good
fc29e334f6dcaf566bc17c5c5514929a395686f8 is first bad commit
commit fc29e334f6dcaf566bc17c5c5514929a395686f8
Author: Vincent Povirk <[email protected]>
Date: Sat Apr 12 20:30:42 2008 -0400

start.exe: Add /Unix switch for native file managers.

:040000 040000 5733b96c3e0b65c9a02620c8b070c36cadf82698 016c4b86a6518cd781a0b1ceffa5e0ea12cc851b M programs

Zachary Goldberg · Post by **Zachary Goldberg** » Sun Apr 27, 2008 10:19 am

On Sun, Apr 27, 2008 at 6:06 AM, Spuuk <[email protected]> wrote:

this one ?

spuuk@GST:~/wine-git$ git bisect good
fc29e334f6dcaf566bc17c5c5514929a395686f8 is first bad commit
commit fc29e334f6dcaf566bc17c5c5514929a395686f8
Author: Vincent Povirk <[email protected]>
Date: Sat Apr 12 20:30:42 2008 -0400

start.exe: Add /Unix switch for native file managers.

:040000 040000 5733b96c3e0b65c9a02620c8b070c36cadf82698 016c4b86a6518cd781a0b1ceffa5e0ea12cc851b M programs

Perfect; now if you could please go to http://bugs.winehq.org/ and
file a new bug report including the information you learned in the
regression test, and what you posted in the first comment that would
be the best way to have your bug be fixed!

With that information we know that it was a patch by Vincent Povirk
which seems to have caused the regression your seeing. So please also
add him to the "CC" list when your making the bugzilla report.
Hopefully then it'll be fixed in a short time.

--
Zachary Goldberg
Computer Science & Engineering
Electrical Captain of Penn Electric Race Team
School of Engineering at the University of Pennsylvania

Vincent Povirk · Post by **Vincent Povirk** » Sun Apr 27, 2008 10:27 am

On Sun, Apr 27, 2008 at 11:19 AM, Zachary Goldberg <[email protected]> wrote:

Perfect; now if you could please go to http://bugs.winehq.org/ and
file a new bug report including the information you learned in the
regression test, and what you posted in the first comment that would
be the best way to have your bug be fixed!

With that information we know that it was a patch by Vincent Povirk
which seems to have caused the regression your seeing. So please also
add him to the "CC" list when your making the bugzilla report.
Hopefully then it'll be fixed in a short time.

--

Zachary Goldberg
Computer Science & Engineering
Electrical Captain of Penn Electric Race Team
School of Engineering at the University of Pennsylvania

Yep, that is the next step. I really don't think my patch could have
caused this, but I'll find out when I can look at this problem in more
detail.

--
Vincent Povirk

Spuuk · Post by **Spuuk** » Sun Apr 27, 2008 1:33 pm

Yep, that is the next step. I really don't think my patch could have
caused this, but I'll find out when I can look at this problem in more
detail.

--
Vincent Povirk

Well sounds intressting for me, none of the results displayed by this software makes sense for me.

Some more intressting is, that when i enable "Desktop Effects" (Standard or Extra) and restart the x-server, i get again the 3th (minimise) button.

If i disable Desktop Effects the minimise button is gone with 0.9.60, and only be visible in 0.9.59 and before.

The FlashFXP Version i use is 2.1 build 924.

austin987 · Post by **austin987** » Sun Apr 27, 2008 6:22 pm

On Sun, Apr 27, 2008 at 5:06 AM, Spuuk <[email protected]> wrote:

this one ?

spuuk@GST:~/wine-git$ git bisect good
fc29e334f6dcaf566bc17c5c5514929a395686f8 is first bad commit
commit fc29e334f6dcaf566bc17c5c5514929a395686f8
Author: Vincent Povirk <[email protected]>
Date: Sat Apr 12 20:30:42 2008 -0400

start.exe: Add /Unix switch for native file managers.

:040000 040000 5733b96c3e0b65c9a02620c8b070c36cadf82698 016c4b86a6518cd781a0b1ceffa5e0ea12cc851b M programs

Doesn't sound like it should be the cause. Try:
$ git revert fc29e334f6dcaf566bc17c5c5514929a395686f8

recompile and see if it works then.

vitamin · Post by **vitamin** » Sun Apr 27, 2008 6:38 pm

austin987 wrote:On Sun, Apr 27, 2008 at 5:06 AM, Spuuk <[email protected]> wrote:
this one ?

spuuk@GST:~/wine-git$ git bisect good
fc29e334f6dcaf566bc17c5c5514929a395686f8 is first bad commit
commit fc29e334f6dcaf566bc17c5c5514929a395686f8
Author: Vincent Povirk <[email protected]>
Date: Sat Apr 12 20:30:42 2008 -0400

start.exe: Add /Unix switch for native file managers.

:040000 040000 5733b96c3e0b65c9a02620c8b070c36cadf82698 016c4b86a6518cd781a0b1ceffa5e0ea12cc851b M programs

Doesn't sound like it should be the cause. Try:
$ git revert fc29e334f6dcaf566bc17c5c5514929a395686f8

recompile and see if it works then.

PLEASE DON'T!!!
The best way to do this is

Code: Select all

git show fc29e334f6dcaf566bc17c5c5514929a395686f8 | patch -p1 -R

austin987 · Post by **austin987** » Sun Apr 27, 2008 6:41 pm

On Sun, Apr 27, 2008 at 6:38 PM, vitamin <[email protected]> wrote:

austin987 wrote:
On Sun, Apr 27, 2008 at 5:06 AM, Spuuk <[email protected]> wrote:

this one ?

spuuk@GST:~/wine-git$ git bisect good
fc29e334f6dcaf566bc17c5c5514929a395686f8 is first bad commit
commit fc29e334f6dcaf566bc17c5c5514929a395686f8
Author: Vincent Povirk <[email protected]>
Date: Sat Apr 12 20:30:42 2008 -0400

start.exe: Add /Unix switch for native file managers.

:040000 040000 5733b96c3e0b65c9a02620c8b070c36cadf82698 016c4b86a6518cd781a0b1ceffa5e0ea12cc851b M programs

Doesn't sound like it should be the cause. Try:
$ git revert fc29e334f6dcaf566bc17c5c5514929a395686f8

recompile and see if it works then.
PLEASE DON'T!!!
The best way to do this is

Code:
git show fc29e334f6dcaf566bc17c5c5514929a395686f8 | patch -p1 -R

What's wrong with git revert?

James McKenzie · Post by **James McKenzie** » Sun Apr 27, 2008 7:50 pm

Austin English wrote:

On Sun, Apr 27, 2008 at 6:38 PM, vitamin <[email protected]> wrote:

austin987 wrote:
On Sun, Apr 27, 2008 at 5:06 AM, Spuuk <[email protected]> wrote:
Doesn't sound like it should be the cause. Try:
$ git revert fc29e334f6dcaf566bc17c5c5514929a395686f8

recompile and see if it works then.
PLEASE DON'T!!!
The best way to do this is

Code:
git show fc29e334f6dcaf566bc17c5c5514929a395686f8 | patch -p1 -R

What's wrong with git revert?

Vitamin and Austin:

It appears that you both are attempting to do the same thing.
git-revert with the commit number should reverse the patch as does the
git-show | patch -p1 -R command. So, like Austin asked, is there
something wrong with git-revert? If so, this should be reported to the
git team so they can fix it.

James McKenzie

vitamin · Post by **vitamin** » Sun Apr 27, 2008 8:56 pm

James McKenzie wrote:Austin English wrote:
On Sun, Apr 27, 2008 at 6:38 PM, vitamin <[email protected]> wrote:

austin987 wrote: PLEASE DON'T!!!
The best way to do this is

Code:
git show fc29e334f6dcaf566bc17c5c5514929a395686f8 | patch -p1 -R

What's wrong with git revert?

Vitamin and Austin:

It appears that you both are attempting to do the same thing.
git-revert with the commit number should reverse the patch as does the
git-show | patch -p1 -R command. So, like Austin asked, is there
something wrong with git-revert? If so, this should be reported to the
git team so they can fix it.

James McKenzie

They do both alter source code you are correct. However, that's the only thing my method does. What Austin suggests alters index.

What "git revert" does is it creates and commits a reverse patch on top of HEAD. This something for developers to deal with. And they know where to look and what to do.

For users it will create extra set of problems. As they can't do 'git pull' anymore - their "master" is no longer the same as "origin". And they explicitly have to remove that commit with "git reset --hard". And if you forget to do this and continue on the standard path 'git fetch; git rebase origin' you will end up with problems down the round. You still have that revert patch in your tree!

Zachary Goldberg · Post by **Zachary Goldberg** » Sun Apr 27, 2008 9:10 pm

On Sun, Apr 27, 2008 at 9:56 PM, vitamin <[email protected]> wrote:

James McKenzie wrote:
Austin English wrote:

On Sun, Apr 27, 2008 at 6:38 PM, vitamin <[email protected]> wrote:

What's wrong with git revert?

Vitamin and Austin:

It appears that you both are attempting to do the same thing.
git-revert with the commit number should reverse the patch as does the
git-show | patch -p1 -R command. So, like Austin asked, is there
something wrong with git-revert? If so, this should be reported to the
git team so they can fix it.

James McKenzie
They do both alter source code you are correct. However, that's the only thing my method does. What Austin suggests alters index.

What "git revert" does is it creates and commits a reverse patch on top of HEAD. This something for developers to deal with. And they know where to look and what to do.

For users it will create extra set of problems. As they can't do 'git pull' anymore - their "master" is no longer the same as "origin". And they explicitly have to remove that commit with "git reset --hard". And if you forget to do this and continue on the standard path 'git fetch; git rebase origin' you will end up with problems down the round. You still have that revert patch in your tree!

That was extraordinarily informative and useful. Can this information
be put in the wiki?

--
Zachary Goldberg
Computer Science & Engineering
Electrical Captain of Penn Electric Race Team
School of Engineering at the University of Pennsylvania

James McKenzie · Post by **James McKenzie** » Sun Apr 27, 2008 10:13 pm

vitamin wrote:

[text on git-revert and git-show | patch -p1 -R removed]

They do both alter source code you are correct. However, that's the only thing my method does. What Austin suggests alters index.

What "git revert" does is it creates and commits a reverse patch on top of HEAD. This something for developers to deal with. And they know where to look and what to do.

Thank you for the explanation. However, will the user have to grab a
new git or can this simply be undone with the 'git-show | patch -p1'
sequence? This should be explained as well in a Wiki article.

For users it will create extra set of problems. As they can't do 'git pull' anymore - their "master" is no longer the same as "origin". And they explicitly have to remove that commit with "git reset --hard". And if you forget to do this and continue on the standard path 'git fetch; git rebase origin' you will end up with problems down the round. You still have that revert patch in your tree!

This would seriously break the git. I usually do a git and then branch
into another so that I can work with the problem as defined in issue
6254. This eliminates not being able to pull the next update. However,
this does lead to a serious use of hard drive space.

Lastly, thank you for the explanation of why git-revert is bad. This
should be added to the git page on the Wine Wiki IMHO.

James McKenzie

vitamin · Post by **vitamin** » Sun Apr 27, 2008 10:50 pm

James McKenzie wrote:vitamin wrote:
[text on git-revert and git-show | patch -p1 -R removed]

They do both alter source code you are correct. However, that's the only thing my method does. What Austin suggests alters index.

What "git revert" does is it creates and commits a reverse patch on top of HEAD. This something for developers to deal with. And they know where to look and what to do.

Thank you for the explanation. However, will the user have to grab a
new git or can this simply be undone with the 'git-show | patch -p1'
sequence?

That command modifies source only not the git index. Git is different from cvs in the way that all the information is being kept internally. When you modifying source - your are not altering git repository itself.

To revert source changes you can do number of things:
1. git show <patch-id> | patch -p1 - exact reverse of the original command
2. git checkout -f - reset all the source changes you might have made
3. git reset --hard - used this way is exactly the same as #2

James McKenzie wrote:This should be explained as well in a Wiki article.
For users it will create extra set of problems. As they can't do 'git pull' anymore - their "master" is no longer the same as "origin". And they explicitly have to remove that commit with "git reset --hard". And if you forget to do this and continue on the standard path 'git fetch; git rebase origin' you will end up with problems down the round. You still have that revert patch in your tree!

This would seriously break the git. I usually do a git and then branch
into another so that I can work with the problem as defined in issue
6254. This eliminates not being able to pull the next update. However,
this does lead to a serious use of hard drive space.

Creating branches in git is almost free. The only cost is one file 41 bytes long (not counting additional patches specific for that branch).

However in your case you should really commit that patch into your tree. Then doing 'git fetch && git rebase origin' when updating tree. What that operation does is:
- Downloads new patches from upstream (winehq.org)
- Removes your patches and "forwards" your repository to where upstream is
- Re-applies your extra patches on top.

This is a standard procedure Wine developers go through until their patches are accepted upstream.