Not logged in
Log in now
Create an account
Subscribe to LWN
LWN.net Weekly Edition for May 23, 2013
An "enum" for Python 3
An unexpected perf feature
LWN.net Weekly Edition for May 16, 2013
A look at the PyPy 2.0 release
Firefox 3.0.6 released
Posted Feb 4, 2009 15:14 UTC (Wed) by k8to (subscriber, #15413)
Posted Feb 4, 2009 15:53 UTC (Wed) by MisterIO (guest, #36192)
Posted Feb 4, 2009 16:10 UTC (Wed) by jengelh (subscriber, #33263)
Posted Feb 4, 2009 16:36 UTC (Wed) by Los__D (guest, #15263)
Posted Feb 4, 2009 23:41 UTC (Wed) by thoffman (subscriber, #3063)
Anyway, the random things might be the fault of the page more than Firefox.
Posted Feb 4, 2009 23:45 UTC (Wed) by kpfleming (subscriber, #23250)
Posted Feb 5, 2009 7:46 UTC (Thu) by jengelh (subscriber, #33263)
Posted Feb 5, 2009 7:50 UTC (Thu) by Los__D (guest, #15263)
Posted Feb 5, 2009 8:02 UTC (Thu) by dlang (✭ supporter ✭, #313)
Posted Feb 5, 2009 9:13 UTC (Thu) by dgm (subscriber, #49227)
Posted Feb 4, 2009 23:58 UTC (Wed) by MisterIO (guest, #36192)
Crash crash crash
Posted Feb 4, 2009 20:06 UTC (Wed) by ncm (subscriber, #165)
I'm resuming 42 windows and hundreds of tabs, so there's no telling which page reload it's tripping over, if indeed it's some particular page. Maybe I'll file a proper bug report with my files.
Posted Feb 4, 2009 23:57 UTC (Wed) by ncm (subscriber, #165)
RESOLVED INVALID This is an issue with a debian build of firefox. You should report it to them,
then get it kicked back upstream if it ends up being a firefox thing. Or you
could download the official Mozilla build, create a new profile, and try to
reproduce with no addons.
Posted Feb 5, 2009 3:26 UTC (Thu) by bignose (subscriber, #40)
I'm fairly sympathetic to complaints about Mozilla Corp., but not in this case. The response you got was polite, appropriate, and gave good options for what needs to happen next. (I'm assuming that the build in which you experienced the bug *is*, as they said, a Debian build.)
If you don't want to go to the bother of obtaining and running the official Mozilla build (I certainly wouldn't), then submit the bug report to Debian. Sorting out whether it's a bug in Debian's build or the upstream is squarely within the role of the Debian iceweasel package maintainer.
Posted Feb 5, 2009 5:02 UTC (Thu) by k8to (subscriber, #15413)
Posted Feb 5, 2009 5:19 UTC (Thu) by ncm (subscriber, #165)
Posted Feb 5, 2009 7:51 UTC (Thu) by dlang (✭ supporter ✭, #313)
Posted Feb 5, 2009 17:14 UTC (Thu) by nye (guest, #51576)
Posted Feb 5, 2009 21:11 UTC (Thu) by dlang (✭ supporter ✭, #313)
Posted Feb 6, 2009 3:28 UTC (Fri) by k8to (subscriber, #15413)
Posted Feb 6, 2009 3:46 UTC (Fri) by dlang (✭ supporter ✭, #313)
they choose not to do that (which they have a perfect right to do), but then to complain that upstream is not supporting the debian changes is unreasonable.
and exactly what bylaws would they have violated by shipping the upstream version?
so three choices
1. ship the upstream and get support from upstream developers with bug reports (but possibly violating debian bylaws)
2. get patches approved by upstream and ship the result, getting support from upstream developers with bug reports (painful coordination needed, there is a possibility that upstream will not sign off on a patch that debian feels is important)
3. fork the project, remove the mozilla trademarks, and do all the support themselves (as is normal with any other forked project)
they choose to do #3
Posted Feb 6, 2009 11:16 UTC (Fri) by TRS-80 (subscriber, #1804)
Posted Feb 6, 2009 18:43 UTC (Fri) by dlang (✭ supporter ✭, #313)
they have said that there are too many important changes in the new versions to backport them all.
as such, claims that they can't ship the upstream becouse of support end of life doesn't sound reasonable.
Posted Feb 8, 2009 10:39 UTC (Sun) by cortana (subscriber, #24596)
Social Contract -> DFSG -> 3. Derived Works:
"The license must allow modifications and derived works, and must allow them to be distributed under the same terms as the license of the original software."
Firefox's license does not permit its artwork to be modified. The official builds also include various other proprietary components, such as the crash reporting mechanism.
Posted Feb 6, 2009 6:10 UTC (Fri) by njs (guest, #40338)
Or put another way, if there were websites popping up all the time that were selling people PostgreSQL for $30 (marked down from $50! what a deal! pay no attention to the added spyware!), I bet the Postgres folks would start worrying about the legal minutiae of trademark enforcement real quick. For example.
(But wouldn't it be awesome to live in the world where they had that problem?)
(I'm sure that in the long run there are better approaches to this and by the time PostgreSQL gets there we'll know what they are, but in the mean time attacking our most successful projects *because* they are so successful that they run into new challenges is not helping us get there. And I say all this as someone who tends to think that Debian also made the right decision.)
Posted Feb 9, 2009 11:20 UTC (Mon) by nye (guest, #51576)
Their only other option was to ship Firefox in non-free, unless they felt like rewriting the social contract to allow for restrictions on what users could do with the software in main - not exactly likely to go down well either; it's not like they were just being obstructive for the sake of it.
Posted Feb 9, 2009 11:31 UTC (Mon) by dlang (✭ supporter ✭, #313)
Posted Feb 5, 2009 8:07 UTC (Thu) by graydon (subscriber, #5009)
The guy who resolved your bug is a volunteer doing triage. He's doing triage because mozilla has a limited supply of people on paid staff to fix the hundreds of thousands of bugs generated by the hundreds of millions of users using the product. We absolutely value bug reports, but we also have to have -- by necessity -- some sort of system of sorting, categorizing, and (sometimes) turning away bugs as unlikely-fixable, based on a cursory inspection.
Goodness knows we'd love to fix all the bugs reachable from our own code. Note the 15,000+ crashdumps we've received due to that dastardly bug holding down the #2 slot in our topcrash list. That's the flash player; not much we can do there. Your situation is back inside the realm of plausible debug-ability, but it's still outside our normal support ballpark, and we're not magicians: starting from a build-we-made and a fresh profile (and uh, not 100+ tabs) is a pretty good first stab at diagnosing. If you consider that an offensive response, I don't know what else to say. We can only focus on so many things at a time, and sorting like this is a necessary part of deciding what to focus on.
Put another way: when you go to the hospital, do you take the judgment of the triage nurse there personally? Even if they mark you down as "non-critical, take a tylenol and go home, come back if it gets worse"? Sometimes that's the best they can do with limited resources.
Posted Feb 5, 2009 8:04 UTC (Thu) by TRS-80 (subscriber, #1804)
Posted Feb 6, 2009 2:24 UTC (Fri) by rsidd (subscriber, #2582)
When my firefox session crashes with more than about 10 tabs open, I opt to start a new session rather than restore the old one.
Posted Feb 6, 2009 3:48 UTC (Fri) by dlang (✭ supporter ✭, #313)
there shouldn't be any problem with doing so. the fact that there is indicates some serious bugs inside the codebase.
Firefox stability with many tabs
Posted Feb 6, 2009 9:49 UTC (Fri) by Cato (subscriber, #7643)
Once I converted the open tabs into links in a web page (there should really be better tools for this, it involved some Perl hacking), and cut back to 50 tabs or less, it was magically much more stable. You can also use Tree Style Tabs extension which makes it easy to bookmark a whole set of tabs.
I have found this bug recurs with just 90+ tabs, but nowhere near as often. Clearly there are bugs in this area, but it really helps to cut down the tabs dramatically.
Generally Firefox does seem less stable on Linux than Windows, although with Windows when using many Firefox tabs, resource usage by Firefox seems to also make the whole OS less usable (e.g. windows fail to appear), resulting in the need for a Windows reboot.
Posted Feb 6, 2009 11:24 UTC (Fri) by TRS-80 (subscriber, #1804)
There's a bookmark all tabs command in Firefox, no need for an extension. I guess I could bookmark half my tabs, but the bookmark manager is so useless using tabs works better. Now there's a bug for Mozilla - make bookmarks work for power users, then they wouldn't have these bug reports from people with hundreds of tabs open.
Copyright © 2009, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds