firefox should kill non-responding instance and start up

[Artem Baguinski]

Binary package hint: mozilla-firefox

Once firefox crashed for one reason or another one can't start new instance anymore because there is still a firefox-bin process running, which causes firefox to suggest "close all instances or restart the system".

I can "killall firefox-bin" and open firefox after, but this is not the same as "close all instances", and suggesting to "restart the system" seems too disruptive to the user workflow, especially considering that it isn't necessary.

to help users who don't know about "killall", or try to avoid using command line and to save keystrokes for those who can use it alright but get annoyed by the error message, firefox could do something along the lines: if it detected that other instance is running but isn't responding - kill that other instance and continue starting up. it shouldn't be just "killall" because it could be running on a multiuser setup, it should only kill firefox processes by the same user.

[John Dong (jdong)]

I also agree it's a usability issue that Firefox sometimes tells the user there's a hanging firefox but doesn't offer a workaround for it.

As a note, MIT's variant of Firefox deployed on the clusters will offer on that dialog to kill the offending Firefox process and/or remove stale locks. It'd be nice if we can do the same.

[Duncan Lithgow (duncan-lithgow)]

I have been frustrated by this bug for a long time. I don't remember ever seeing it on windows for some reason. It sounds like either of the suggestions above are workable.

I'm surprised this is only set to importance 'wishlist' it's a serious issue which stops users from starting ff. If we accept that suggesting to reboot is a solution then a lot of other apps should also be let off the hook for freezing! Surely that's not what we want.

[Artem Baguinski (femistofel)]

For the record: this dialog only happens to me when using alternative window managers or xfce; metacity, when application doesn't responde to close request, offers to kill the it and does actually kill firefox "to death" - starting another instace directly after works.

[Alexander Sack (asac)]

can this still be reproduced reliably?

 status incomplete

 - Alexander

[Artem Baguinski (femistofel)]

On Wed, May 14, 2008 at 12:59 PM, Alexander Sack <email address hidden> wrote:
> can this still be reproduced reliably?
>
> status incomplete
>
> - Alexander

i hit it again after just having upgraded to hardy. not sure what was
causing crashes though and i can't recreate them at the moment.

--
cheers,
artm

http://lab.v2.nl/

[Alexander Sack (asac)]

On Thu, May 15, 2008 at 06:44:42PM -0000, Artem Baguinski wrote:
> On Wed, May 14, 2008 at 12:59 PM, Alexander Sack <email address hidden> wrote:
> > can this still be reproduced reliably?
> >
> > status incomplete
> >
> > - Alexander
>
> i hit it again after just having upgraded to hardy. not sure what was
> causing crashes though and i can't recreate them at the moment.
>

Firefox needs to be restarted after upgrading. until you do that,
everything can happen

 status invalid

 - Alexander