Killing Flash Player Process causes SM Crash

[Zerebralus]

Just noticed this issue with Version 2.32 (betas), also get some crash reports (sent).

Maybe this issue causes plugin container failure (not sure at all)

[therube]

Last few crash report URLs (about:crashes) ?

[Zerebralus]

The last three:

Date 05/01/2015 00:30 https://crash-stats.mozilla.com/report/index/681f4726-c1db-4448-bf85-61abf2150104
Date 05/01-2015 00:25 https://crash-stats.mozilla.com/report/index/1929ba6c-6657-47f0-bfba-fed312150104
Date 04/01/2015 16:12 https://crash-stats.mozilla.com/report/index/e0eadba5-9073-410e-980a-6165d2150104

atm Seamonkey 2.32 Beta 3

[Zerebralus]

same again with the final Version 2.32

Just ended my browser game but the flashplayer plugin still uses about 400 MB. then (in the taskmanager) killing the flashplayer process... also crashes the browser.

Then changed "dom.ipc.plugins.enabled" in the browser config from true to false and then tried again to kill the flashplayer process. Only the flashplayer in the browser (game) window normally crashed. But not the browser as before...

so what's wrong?

[patrickjdempsey]

Why force-crash Flash at all? If you leave it alone it should close out when it's done doing it's thing in the background. It usually takes a minute or so... may take longer for heavy use like that.

Also, when you disable dom.ipc.plugins.enabled you shouldn't even see plugin-container as Flash will not be in a separate process so you can't force-crash Flash out. Note that running in that mode is not recommended because if Flash crashes it will *definitely* take out SeaMonkey.

[Zerebralus]

the curious thing is... although dom.ipc.plugins.enabled is set to false, the plugin container is still running in the taskmanager

[patrickjdempsey]

Plugin-container will open on restart as part of a crash recovery. So if you force-crash the browser, next time it opens plugin-container will also open. I've seen cases where that may cause SeaMonkey to immediately crash out, but that involved running out of page-file space. Ending the crash-recovery plugin-container process has absolutely nothing to do with closing Flash. If you close and open SeaMonkey normally, you should not see plugin-container with dom.ipc.plugins.enabled false. (But of course you won't have crash-protection)

[therube]

dom.ipc.plugins.enabled is set to false

If that is the case, then why would you not expect killing Flash to cause SeaMonkey to crash?

The is precisely the purpose of out-of-process plugins, to prevent a Plugin crash from taking down the browser.

[Zerebralus]

First Seamonkey 2.31 tried again without problems when the flash kill process. Then online update on Seamonkey 2:32 and the same tried. This time with no problems. Only the window with the Flash Game crashed (as expected)

So I guess that a file in the profile was probably a mistake because I had started with an almost new profile.

[patrickjdempsey]

Right. Once you start having crashes, or are force-crashing the browser you greatly increase the chance of corrupted files which can lead to further instability. This is one reason why it's really not recommended to force-crash programs.