User Agent: Mozilla/5.0 (Windows NT 6.3; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/39.0.2171.95 Safari/537.36 Steps to reproduce: Access this URL: http://alexgorbatchev.com/SyntaxHighlighter/manual/api/autoloader.html Actual results: Non e10s: FF stop working, no crash report e10s: Tab stop working, try to close tab and FF stop working, no crash report. I needed to kill FF. Expected results: On e10s, FF shows unresponsive script message, but stop working. Must show message and kill the script process.
There are 2 issues on that page with 2 different regression ranges: 1) FF not responding (this bug) 2) Examples not displayed (I filed bug 1111694) About 1) Regression range good=2014-05-16 bad=2014-05-17 http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=58c5a3427997&tochange=2893f60d5903 My guess is it's another issue with the new regexp engine: Bug 976446
Loic, what makes you think that bug 1111694 and this bug are two different things? IIUC, the examples would only be shown after the regexp has successfully run. Since it doesn't do that (this bug) no examples are shown (bug 1111694). Or am I missing something?
Yes, this regexp is probably uncorrectly-interpreted, so Firefox throws an execption and doesn't display the result (bug 1111694). But after the change of the regexp engine, the result is worse because FF fails to treat the regexp and becomes unresponding.
Brian, same question as in bug 1111694, but in the correct bug this time: a regexp contained in this page causes FF to become unresponsive without ever showing the slow script dialog. Sorry for the churn.
We're executing the long running regexp while trying to show the slow script dialog, which will cause a hang and is imo a design problem with Gecko. See bug 1012491 comment 13.