http://bugs.winehq.org/show_bug.cgi?id=32772
Bug #: 32772 Summary: Built-in iexplore crashes on exit with Flash player plug-in installed Product: Wine-gecko Version: unspecified Platform: x86 URL: http://download.macromedia.com/pub/flashplayer/install ers/archive/fp_11.5.502.146_archive.zip OS/Version: Linux Status: NEW Keywords: download, source Severity: normal Priority: P2 Component: wine-gecko-unknown AssignedTo: jacek@codeweavers.com ReportedBy: dmitry@baikal.ru Classification: Unclassified
Since winetricks has not been updated to support recent Flash player releases it needs to be installed manually (flash 11.5.502.146 is the latest one):
rm -rf ~/.wine wget http://download.macromedia.com/pub/flashplayer/installers/archive/fp_11.5.50... wine fp_11.5.502.146_archive/11_5_r502_146/flashplayer11_5r502_146_winax.exe wine iexplore http://www.adobe.com/software/flash/about
About Flash page correctly reports "You have version 11,5,502,146 installed", but Flash demo actually doesn't show (although comments in the bug 15092 claim that at one point Flash worked fine in built-in iexplore and youtube).
Press Alt+F4, and iexplore crashes. This is last lines of the console output, pointing to the crash in mozalloc:
Fault in cycle collector: traversed refs exceed refcount (ptr: 03964f80) ###!!! ABORT: cycle collector fault: file /build/wine-mozilla-1.9/xpcom/base/nsCycleCollector.cpp, line 1111 ###!!! ABORT: cycle collector fault: file /build/wine-mozilla-1.9/xpcom/base/nsCycleCollector.cpp, line 1111 wine: Unhandled page fault on write access to 0x00000000 at address 0x61e474b3 (thread 0029), starting debugger... Unhandled exception: page fault on write access to 0x00000000 in 32-bit code (0x61e474b3). Register dump: CS:0023 SS:002b DS:002b ES:002b FS:0063 GS:006b EIP:61e474b3 ESP:010de2c0 EBP:6af20810 EFLAGS:00010246( R- -- I Z- -P- ) EAX:0000000a EBX:00000003 ECX:010de280 EDX:00000000 ESI:6af1fbe0 EDI:00000000 Stack dump: 0x010de2c0: 0000000a 7d929320 6af20810 008d4d51 0x010de2d0: 7bcaaff4 00000060 00000050 69db67f0 0x010de2e0: 010de314 6af1fc26 00000003 6af1fbe0 0x010de2f0: 010de85c 00000000 6af20810 69c65213 0x010de300: 7d929320 6af1fc26 010de314 00000457 0x010de310: 00000002 21232323 41202121 54524f42 Backtrace: =>0 0x61e474b3 in mozalloc (+0x74b3) (0x6af20810) 0x61e474b3: movl $0x7b,0x00000000
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #1 from Dmitry Timoshkov dmitry@baikal.ru 2013-01-21 03:23:09 CST --- For the reference here is an official archive of various Flash versions for testing and/or winetricks usage: http://helpx.adobe.com/flash-player/kb/archived-flash-player-versions.html
http://bugs.winehq.org/show_bug.cgi?id=32772
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |austinenglish@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=32772
Dmitry Timoshkov dmitry@baikal.ru changed:
What |Removed |Added ---------------------------------------------------------------------------- See Also| |http://bugs.winehq.org/show | |_bug.cgi?id=32946
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #2 from Dmitry Timoshkov dmitry@baikal.ru 2013-02-10 20:48:32 CST --- It's pretty disappointing to see no activity on this bug. One would assume that Flash player is pretty common and widely used plugin to support, and crashes are serious enough on there own to fix at least them (not mentioning Flash basic functionality which is reported separately as bug 32946).
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #3 from Jacek Caban jacek@codeweavers.com 2013-02-11 04:26:18 CST ---
It's pretty disappointing to see no activity on this bug.
You're a developer, you should know better that sometimes it takes a while before one can work on a given bug. That said, you also didn't bother to run a regression test and your assumptions are based on a comment from unrelated bug, which seems to be about NPAPI version of Flash, but you seem to be testing ActiveX version. Still, that worked at some point too.
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #4 from Dmitry Timoshkov dmitry@baikal.ru 2013-02-11 05:02:22 CST --- (In reply to comment #3)
You're a developer, you should know better that sometimes it takes a while before one can work on a given bug. That said, you also didn't bother to run a regression test and your assumptions are based on a comment from unrelated bug, which seems to be about NPAPI version of Flash, but you seem to be testing ActiveX version. Still, that worked at some point too.
This bug is about a crash in gecko, and I didn't claim that there was no crash before, so this bug is not about a regression.
Speaking about Flash plugin functionality, how do you know that bug 15092 is about NPAPI version of flash? Does built-in iexplore support such version of Flash plugin or supported at some point?
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #5 from Jacek Caban jacek@codeweavers.com 2013-02-11 05:14:17 CST --- (In reply to comment #4)
This bug is about a crash in gecko, and I didn't claim that there was no crash before, so this bug is not about a regression.
Well, it still seems like a regression. Anyway, if you don't think it's a regression, then your complains about lack of activity on the bug is even more disappointing.
Speaking about Flash plugin functionality, how do you know that bug 15092 is about NPAPI version of flash?
That bug predates our support for AcviveX in MSHTML.
Does built-in iexplore support such version of Flash plugin or supported at some point?
Yes, it should still work (although it won't integrate well with scripts on the site, so you may expect some sites to not work in this configuration).
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #6 from Dmitry Timoshkov dmitry@baikal.ru 2013-02-11 05:27:59 CST --- (In reply to comment #5)
This bug is about a crash in gecko, and I didn't claim that there was no crash before, so this bug is not about a regression.
Well, it still seems like a regression. Anyway, if you don't think it's a regression, then your complains about lack of activity on the bug is even more disappointing.
By "disappointing" I meant that apparently at some point Flash plugin worked, comment http://bugs.winehq.org/show_bug.cgi?id=15092#c21 already contains a lot of results of investigation/attempts on Flash functionality in built-in iexplore, and it's sad that there was no progress since then, and that one may expect that Flash is very widely used, and supposed to be at least actively worked on. Of course, I understand priorities and such, but posting a comment explaining that and updating status doesn't take too much time.
Speaking about Flash plugin functionality, how do you know that bug 15092 is about NPAPI version of flash?
That bug predates our support for AcviveX in MSHTML.
Does built-in iexplore support such version of Flash plugin or supported at some point?
Yes, it should still work (although it won't integrate well with scripts on the site, so you may expect some sites to not work in this configuration).
Did you test it with recent versions of gecko+iexplore? Do you have some suggestions/instructions how to install and use NPAPI version of Flash with built-in iexplore? (So that I don't waste time doing it wrong if you know how to do/test it properly).
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #7 from Jacek Caban jacek@codeweavers.com 2013-02-11 07:35:59 CST --- (In reply to comment #6)
Did you test it with recent versions of gecko+iexplore?
No, unfortunatelly I haven't tested it for a while.
Do you have some suggestions/instructions how to install and use NPAPI version of Flash with built-in iexplore? (So that I don't waste time doing it wrong if you know how to do/test it properly).
AFAIR Flash has an installer similar to its ActiveX version for NPAPI. Simply installing it should be enough.
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #8 from Austin English austinenglish@gmail.com 2013-02-11 20:30:53 CST --- 'winetricks flash11' should work. I'm not sure why Dan made it separate, presumably since it doesn't work yet.
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #9 from Dmitry Timoshkov dmitry@baikal.ru 2013-02-11 21:34:27 CST --- (In reply to comment #8)
'winetricks flash11' should work. I'm not sure why Dan made it separate, presumably since it doesn't work yet.
It doesn't work (and apparently it tries to download an ActiveX Flash):
$winetricks flash11 Executing w_do_call flash11 Executing load_flash11 Executing mkdir -p /home/dmitry/.cache/winetricks/flash11 Downloading http://fpdownload.macromedia.com/get/flashplayer/pdc/11.1.102.55/install_fla... to /home/dmitry/.cache/winetricks/flash11 --2013-02-12 12:28:01-- http://fpdownload.macromedia.com/get/flashplayer/pdc/11.1.102.55/install_fla... Resolving fpdownload.macromedia.com... 184.85.178.70 Connecting to fpdownload.macromedia.com|184.85.178.70|:80... connected. HTTP request sent, awaiting response... 404 Not Found 2013-02-12 12:28:02 ERROR 404: Not Found.
------------------------------------------------------ Downloading http://fpdownload.macromedia.com/get/flashplayer/pdc/11.1.102.55/install_fla... failed ------------------------------------------------------
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #10 from Dmitry Timoshkov dmitry@baikal.ru 2013-02-11 21:55:09 CST --- (In reply to comment #7)
Do you have some suggestions/instructions how to install and use NPAPI version of Flash with built-in iexplore? (So that I don't waste time doing it wrong if you know how to do/test it properly).
AFAIR Flash has an installer similar to its ActiveX version for NPAPI. Simply installing it should be enough.
Installing a not ActiveX version of Flash plugin works, and executing 'wine iexplore http://www.adobe.com/software/flash/about' reports correct Flash version. But either exiting iexplore or pressing Refresh button lead to the same crash in gecko. Also actual flash animation is not displayed.
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #11 from Austin English austinenglish@gmail.com 2013-02-12 12:37:31 CST --- (In reply to comment #9)
(In reply to comment #8)
'winetricks flash11' should work. I'm not sure why Dan made it separate, presumably since it doesn't work yet.
It doesn't work (and apparently it tries to download an ActiveX Flash):
$winetricks flash11 Executing w_do_call flash11 Executing load_flash11 Executing mkdir -p /home/dmitry/.cache/winetricks/flash11 Downloading http://fpdownload.macromedia.com/get/flashplayer/pdc/11.1.102.55/install_fla...
Your version of winetricks is out of date, it should be getting 11.5.502.135 (https://code.google.com/p/winetricks/source/detail?r=930)
http://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #12 from Dmitry Timoshkov dmitry@baikal.ru 2013-02-12 20:26:16 CST --- (In reply to comment #11)
http://fpdownload.macromedia.com/get/flashplayer/pdc/11.1.102.55/install_fla...
Your version of winetricks is out of date, it should be getting 11.5.502.135 (https://code.google.com/p/winetricks/source/detail?r=930)
http://wiki.winehq.org/winetricks points to http://winetricks.org/winetricks and that one wants to download 11.1.102.55. You may want to update the thing you are distributing.
https://bugs.winehq.org/show_bug.cgi?id=32772
Joel Holdsworth joel@airwebreathe.org.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |joel@airwebreathe.org.uk
--- Comment #13 from Joel Holdsworth joel@airwebreathe.org.uk --- I've seen this bug on some pages, and I don't have flash installed. I can't give links though, because these are integrated into my product. The product does however use an ActiveX control in it.
https://bugs.winehq.org/show_bug.cgi?id=32772
joaopa jeremielapuree@yahoo.fr changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jeremielapuree@yahoo.fr
--- Comment #14 from joaopa jeremielapuree@yahoo.fr --- Created attachment 68781 --> https://bugs.winehq.org/attachment.cgi?id=68781 screenshot showing flashplayer is installed
Looks like the bug is fixed in wine--RC1 Can an administrator close this bug as FIXED?
Original reported version of Flash can be downloaded here: https://web.archive.org/web/20190403182418if_/http://fpdownload.macromedia.c...
https://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #15 from joaopa jeremielapuree@yahoo.fr --- Created attachment 68782 --> https://bugs.winehq.org/attachment.cgi?id=68782 Screenshot showing animation is playing
https://bugs.winehq.org/show_bug.cgi?id=32772
Anastasius Focht focht@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|Built-in iexplore crashes |Built-in iexplore crashes |on exit with Flash player |on exit with Flash player |plug-in installed |11 ActiveX plug-in | |installed URL|http://download.macromedia. |https://web.archive.org/web |com/pub/flashplayer/install |/20190403182418if_/http://f |ers/archive/fp_11.5.502.146 |pdownload.macromedia.com/pu |_archive.zip |b/flashplayer/installers/ar | |chive/fp_11.5.502.146_archi | |ve.zip Status|NEW |RESOLVED Resolution|--- |FIXED CC| |focht@gmx.net
--- Comment #16 from Anastasius Focht focht@gmx.net --- Hello folks,
confirming. The crash was fixed a while ago.
List of Wine-Gecko versions vs. Wine releases:
https://gist.github.com/rmi1974/61ba2bbb629fd23ebb8201e2b882e2b6
--- snip --- $ git log --oneline | \ sed -nre 's/(\w+).*wine gecko.* ([0-9]+([.][0-9]+)+).*/\1 \2/ipg' | \ xargs -l bash -c 'echo -n "Wine-Gecko $1 | " ; \ git -c "versionsort.suffix=-" tag --sort=version:refname --contains $0 | \ head -n1'
Wine-Gecko 2.47.2 | wine-6.0-rc1 Wine-Gecko 2.47.1 | wine-5.0-rc1 Wine-Gecko 2.47 | wine-1.9.13 Wine-Gecko 2.44 | wine-1.9.3 Wine-Gecko 2.40 | wine-1.7.50 Wine-Gecko 2.36 | wine-1.7.38 Wine-Gecko 2.34 | wine-1.7.31 Wine-Gecko 2.24 | wine-1.7.3 Wine-Gecko 2.21 | wine-1.5.31 Wine-Gecko 1.9 | wine-1.5.22 Wine-Gecko 1.8 | wine-1.5.15 Wine-Gecko 1.7 | wine-1.5.10 Wine-Gecko 1.6 | wine-1.5.7 Wine-Gecko 1.5 | wine-1.5.0 Wine-Gecko 1.4 | wine-1.3.33 Wine-Gecko 1.3 | wine-1.3.27 Wine-Gecko 1.2.0 | wine-1.3.16 Wine-Gecko 1.1.0 | wine-1.3.2 Wine-Gecko 1.0.0 | wine-1.1.27 Wine-Gecko 0.9.1 | wine-1.1.15 Wine-Gecko 0.9.0 | wine-1.1.12 Wine-Gecko 0.0.1 | wine-0.9.58 Wine-Gecko 0.1.0 | wine-0.9.39 --- snip ---
Just checking the last ones:
--- snip --- # testing Wine-Gecko 2.47.1 | wine-5.0-rc1
$ export PATH=/home/focht/projects/wine/mainline-install-5.0-rc1-x86_64/bin:$PATH
$ wine --version wine-5.0-rc1
$ rm -rf ~/.wine
$ wine fp_11.5.502.146_archive/11_5_r502_146/flashplayer11_5r502_146_winax.exe /install
# shows page properly, doesn't show animation (bug 32946), alt+f4 ok $ wine iexplore.exe http://helpx.adobe.com/flash-player.html
# shows page properly $ wine iexplore http://www.adobe.com/software/flash/about
# shows page properly $ wine iexplore http://helpx.adobe.com/flash-player/kb/archived-flash-player-versions.html
--- snip ---
--- snip --- # testing Wine-Gecko 2.47 | wine-1.9.13
$ export PATH=/home/focht/projects/wine/mainline-install-1.9.13-x86_64/bin:$PATH
$ wine --version wine-1.9.13
$ rm -rf ~/.wine
$ wine fp_11.5.502.146_archive/11_5_r502_146/flashplayer11_5r502_146_winax.exe /install
# shows empty page, when being resized crashes $ wine iexplore.exe http://helpx.adobe.com/flash-player.html
# shows empty page $ wine iexplore http://www.adobe.com/software/flash/about
# shows empty page $ wine iexplore http://helpx.adobe.com/flash-player/kb/archived-flash-player-versions.html --- snip ---
--- snip --- # testing Wine-Gecko 2.40 | wine-1.7.50
$ export PATH=/home/focht/projects/wine/mainline-install-1.7.50-x86_64/bin:$PATH
$ wine --version wine-1.7.50
$ rm -rf ~/.wine
$ wine fp_11.5.502.146_archive/11_5_r502_146/flashplayer11_5r502_146_winax.exe /install
# shows empty page $ wine iexplore.exe http://helpx.adobe.com/flash-player.html
# shows empty page, alt+f4 crashed once on exit (different problem, unrelated fp11) $ wine iexplore http://www.adobe.com/software/flash/about
# shows empty page $ wine iexplore http://helpx.adobe.com/flash-player/kb/archived-flash-player-versions.html --- snip ---
So it was fixed with Wine-Gecko 2.47.1 release, part of Wine 5.0-rc1 release.
There is no Wine-Gecko product version by design (reserved for Wine releases) hence one needs to do some work to keep track of Wine-Gecko -> Wine release integration mappings.
In this case it was most likely Wine 1.5-ish release (guessed from the report date).
--- snip --- 2013-02-01 09:10:43 +0100 (tag: wine-1.5.23) Release 1.5.23. 2013-01-18 21:44:53 +0100 (tag: wine-1.5.22) Release 1.5.22. 2013-01-04 21:45:01 +0100 (tag: wine-1.5.21) Release 1.5.21. --- snip ---
====
Not specifically targeting this bug but a more general observation for 'Wine-Gecko' product bug reports...
https://bugs.winehq.org/buglist.cgi?bug_status=UNCONFIRMED&bug_status=NE...
I'm ok with bugs related to packaging issues and support for other platforms etc. but various other bugs are simply misplaced and buried. I never look at these because they are automatically filtered out by product (I have certain default filters).
If bugs are reported with 'Wine-Gecko' product I propose to always create a pair of bugs: one against the component 'Wine-Gecko' (= internal tracking) and another one for the component integration (upgrade) in Wine release -> product 'Wine'.
The one against Wine product is the important one. It appears in official release notes = visible to end users.
$ sha1sum fp_11.5.502.146_archive.zip db5fd38c4f4335e24e53b76e41526bd69f42ec3fb fp_11.5.502.146_archive.zip
$ du -sh fp_11.5.502.146_archive.zip 147M fp_11.5.502.146_archive.zip
$ wine --version wine-5.0-rc1
Regards
https://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #17 from Jacek Caban jacek@codeweavers.com --- (In reply to Anastasius Focht from comment #16)
I'm ok with bugs related to packaging issues and support for other platforms etc. but various other bugs are simply misplaced and buried. I never look at these because they are automatically filtered out by product (I have certain default filters).
I reviewed those bugs and almost all of them were already fixed or belong to mshtml component. Separated product made more sense in the past, when Wine Gecko development was very active. The package is in maintenance mode now and I'd be in favour of retiring bugzilla product and (ab)using mshtml for Gecko bugs. The only bug left in this component now is bug 43938 and having it in mshtml component wouldn't be wrong. Distinguishing between wine-gecko and mshtml has always been challenging for reporters (for understandable reasons), at this point I don't think that the complication is justified.
https://bugs.winehq.org/show_bug.cgi?id=32772
Jacek Caban jacek@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Assignee|jacek@codeweavers.com |wine-bugs@winehq.org Component|wine-gecko-unknown |mshtml Product|Wine-gecko |Wine
--- Comment #18 from Jacek Caban jacek@codeweavers.com --- BTW, I usually moved bugs to Wine component after fixing it so that they ride the usual announce+closure ride.
https://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #19 from Dmitry Timoshkov dmitry@baikal.ru --- (In reply to Jacek Caban from comment #17)
(In reply to Anastasius Focht from comment #16)
I'm ok with bugs related to packaging issues and support for other platforms etc. but various other bugs are simply misplaced and buried. I never look at these because they are automatically filtered out by product (I have certain default filters).
I reviewed those bugs and almost all of them were already fixed or belong to mshtml component. Separated product made more sense in the past, when Wine Gecko development was very active. The package is in maintenance mode now and I'd be in favour of retiring bugzilla product and (ab)using mshtml for Gecko bugs. The only bug left in this component now is bug 43938 and having it in mshtml component wouldn't be wrong. Distinguishing between wine-gecko and mshtml has always been challenging for reporters (for understandable reasons), at this point I don't think that the complication is justified.
Bug 49448 is clearly a bug in Gecko (despite its component is set to mshtml).
https://bugs.winehq.org/show_bug.cgi?id=32772
--- Comment #20 from Anastasius Focht focht@gmx.net --- Hello folks,
--- quote --- Distinguishing between wine-gecko and mshtml has always been challenging for reporters (for understandable reasons), at this point I don't think that the complication is justified. --- quote ---
thanks for trying to make things easier for bug reporters and triagers.
--- quote --- Bug 49448 is clearly a bug in Gecko (despite its component is set to mshtml). --- quote ---
Most people can't tell the difference. Not sure if it's really worth to distinguish, given the small number of Gecko bugs.
Maybe the 'mshtml' component description could be enhanced in WineHQ bugzilla:
https://bugs.winehq.org/editcomponents.cgi?action=edit&product=Wine&...
--- snip --- HTML, CSS parser and renderer --- snip ---
to reflect that "Wine-Gecko" is covered as well.
Alternatively, new component could be created under 'Wine' product - but again, is it worth? Same argument: too few bugs, too much overhead to correct bug reporter input.
Regards
https://bugs.winehq.org/show_bug.cgi?id=32772
Jacek Caban jacek@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jacek@codeweavers.com
--- Comment #21 from Jacek Caban jacek@codeweavers.com --- (In reply to Dmitry Timoshkov from comment #19)
Bug 49448 is clearly a bug in Gecko (despite its component is set to mshtml).
Oh, right, this would be another one with current definition. But if we changed the definition as Anastasius suggested, it would fit mshtml. One way to look at Wine Gecko is that it's an internal (but significant) implementation detail of mshtml and as such bugs belong there. The fact that the code that needs to be changed is in Gecko may be expressed in a summary.
In fact, it has always been partially this way, the separated product was meant as a tool to properly track status. For bugs that require Gecko changes, we can resolve them in gecko product as soon as it they are fixed in Wine Gecko git and later move resolved bug to mshtml component when Wine is upgraded to use the fixed version of Gecko. By retiring separated product we lose the ability to have 'fixed in Wine Gecko git, but not yet in Wine' state. The volume of such bugs is low enough that I don't mind.
And to be fair, it's usually more than two bugs that belong to gecko product under current definitions. The fact that there are no more bugs is partially because I recently released Gecko update that fixed a few of them. There is still quite a bit of work that would be nice to have done on Gecko side. It's just that majority of remaining mshtml work is in Wine part of mshtml using the existing Gecko (that is no longer updated every few months with tons of new features and problems merged from upstream).
https://bugs.winehq.org/show_bug.cgi?id=32772
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #22 from Alexandre Julliard julliard@winehq.org --- Closing bugs fixed in 6.0-rc3.
https://bugs.winehq.org/show_bug.cgi?id=32772
Anastasius Focht focht@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Fixed by SHA1| |5f0b5d350566a46f0f999e4cff7 | |ad9e280fcfa05 Version|unspecified |1.5.22
-
wine-bugs＠winehq.org -
WineHQ Bugzilla