Version: (using KDE 4.2.3) OS: Linux Installed from: RedHat RPMs 1) go to: http://service.futuremark.com/peacekeeper/browserStatistics.action where Firefox, Safari, Chrome, IE, Opera. Ignore warning that Konq is unsupported, and run benchmark. Boom. 2) Attempt to log into bugs.kde.org after restarting Konq, find broken image symbols on every page, be forced to log in every step of the way, despite accepting cookies, having JS enabled for kde.org, etc. 3) Look at bug search 'feature', contemplate manually checking every URL that shows 'crash', laugh at implicit assumption of infinite available time. 4) Get very frustrated with Version 4.2.3 (KDE 4.2.3) on Fedora 10 test, and never-ending history of JavaScript shortcomings on *any* platform. 5) Go back to mostly using a poorly-integrated browser with a worse UI: Firefox. At least it doesn't fall over, and bookmark editing mostly works.
Try to not fix different issues in the same report. (policy here is one report per issue) -Did you got a backtrace of the crash in the mentioned site? (I'm going to try to reproduce) - The broken images on bugs.kde.org is already reported. - I created a new crash reporting tool that will search similar crashes for you (and you will not need to login/browse bugs.kde.org). It will be available on KDE4.3 Thanks
Here using: Qt: 4.5.1 (qt-copy 958974) KDE: 4.2.86 (KDE 4.2.86 (KDE 4.3 >= 20090514)) kdelibs svn rev. 968443 / kdebase svn rev. 968445 on ArchLinux i686 - Kernel 2.6.29.3 I couldn't reproduce the crash on the mentioned benchmarks site. (and I got 743 Points)
I got 1242 points and no crash. kde4-konqueror-4.2.3-80.6 (using KDE 4.2.3) All packages from openSUSE 11.1 x86_64
Interestingly, 1242 points was also what I got with Firefox 3.6a1pre (aka Minefield) built from the developer sources aquired last tuesday. In case you can't believe Konq keeping up with a fresh Firefox: http://nerdvar.com/bilder/skjerm/javascript_konq_vs_firefox.png
@Greg Metcalfe: are you there?
Yes Andres, I'm here. Just wasn't sure what else I could contribute. I've just repeated the test. It runs smoothly for a few seconds, up to 385 points, then froze with 5 minutes still showing on screen--that never decremented. Eventually Konqueror crashed, taking this bug entry screen with it :( Repeat, and this time run the test in another Konqueror window instead of another tab (still via clicking on the URL above. This time it ran for only a few seconds, and both Konqueror windows. Including this bug screen again, of course :( OK, what about a completely separate instance of Konqueror? Same thing. Ran only a few seconds, and crashed. kdebase-4.2.3-1.fc10.i386 qt3-3.3.8b-17.fc10.i386 and qt-4.5.1-10.fc10.i386 kernel 2.6.29.3-60.fc10.i686 Just for bubbles and grins, I then ran it with Firefox 3.0.8, installed via firefox-3.0.10-1.fc10.i386. It would run the test without cookies. OK, allow service.futuremark.com. Still no joy, so tell Firefox to accept cookies. It ran to completion for a score of 322. Went back to figure out what third-party !@#$ cookies it might have planted on me, and didn't see anything out of the ordinary--just 2 jssession cookies. Another new Konqueror window, and uncheck the bit about accepting cookies only from orginating server, and run the test again. This time it completed, and reported: Konqueror(v4.2) Scored: 334 Points So it seems down to a cookie issue. I double-checked in Konq, and I am very certain no third-party cookie was dropped on me. The list is trivial to check as Konq is my 'paranoid' browser--I only accept cookies from a couple of bugzillas, etc. I received 4 from futuremark.com: name JSESSIONID path /peacekeeper expires end of session name JSESSIONID path /ecom expires end of session name JSESSIONID path /ecom expires end of session name __utmc path / expires end of session So the only difference I can see is that I had to uncheck originating server. Looking at the cookie screens, it's hard for a user to see what might have been the problem. Was there another IP address at futuremark.com shipping cookies, and that caused the problem? There's no way to tell from Konqueror, as it reports the domain, not the server. *If* that's the case, it would mean an error in the semantics of what is displayed in the cookies setup screen and the management screen.
Thanks for the further testing. A backtrace would also be useful: http://techbase.kde.org/Development/Tutorials/Debugging/How_to_create_useful_crash_reports
Unfortunately, Fedora packages don't give me symbols. Tried it, just in case they had changed that, but no joy. I afraid I can't help on that.
Dario, I was wrong about that last. http://fedoraproject.org/wiki/StackTraces is old (they talk about Fedora Core 6, which hit EOL in November 07) but I'll try to give it a spin within the next day or two.
I could run the benchmark sucessfully using: Konqueror Version 4.2.92 (KDE 4.2.92 (KDE 4.3 >= 20090617)) Using KDE 4.2.92 (KDE 4.2.92 (KDE 4.3 >= 20090617)) - svn r984201 qt-copy r978427
Futuremark has rewrite the complete Benchmark and this is now available under: http://peacekeeper.futuremark.com/ So crash isn't reproduceable and should be closed. Yet existing problems with peacekeeper is reported in #301901
(In reply to comment #11) > Futuremark has rewrite the complete Benchmark and this is now available > under: http://peacekeeper.futuremark.com/ > > So crash isn't reproduceable and should be closed. > > Yet existing problems with peacekeeper is reported in #301901 ...closing.