RESOLVED FIXED 57414
[Qt] segfault on javascript alert() 
https://bugs.webkit.org/show_bug.cgi?id=57414
Summary [Qt] segfault on javascript alert()
Thijs
Reported 2011-03-29 19:02:16 PDT
Full bug report (including nice formatting + stack trace) here: http://developer.qt.nokia.com/forums/viewthread/4760/ Summary: very simple demo application crashes after clicking the button on a javascript alert(). I'm using Qt Creator 2.1.0 (based on Qt 4.7.1 / 32 bit, Windows XP)). It's QtWebKit version is supposedly based on SVN version 56441, so I'm not sure which version to pick exactly.
Attachments
Add attachment proposed patch, testcase, etc.
Benjamin Poulain
Comment 1 2011-03-30 04:47:07 PDT
Confirming the P1 since it is a crash. You are adding a blocking call in the frame loader, this is really nasty way of doing JavaScript, but that should not crash.
Andreas Kling
Comment 2 2011-03-30 12:16:54 PDT
Is this reproducible with WebKit trunk?
Thijs
Comment 3 2011-03-30 18:48:22 PDT
(In reply to comment #1) > You are adding a blocking call in the frame loader, > this is really nasty way of doing JavaScript, but that should not crash. Yes, it's mostly for testing, but I can imagine this situation will also happen in our product. (In reply to comment #2) > Is this reproducible with WebKit trunk? Yesterday and today I'm trying to build QtWebKit 2.1 and/or WebKit from trunk. I'm not a Windows developer, so learning how to setup my environment & build everything from source is taking some time. If/when I get them compiled, I will try again.
Benjamin Poulain
Comment 4 2011-03-31 04:19:37 PDT
> (In reply to comment #2) > > Is this reproducible with WebKit trunk? > Yesterday and today I'm trying to build QtWebKit 2.1 and/or WebKit from trunk. I'm not a Windows developer, so learning how to setup my environment & build everything from source is taking some time. If/when I get them compiled, I will try again. In case you have not seen it yet, here is the doc to build WebKit trunk on Windows: http://trac.webkit.org/wiki/BuildingQtOnWindows Jocelyn, would you have a recent build to test this?
Jocelyn Turcotte
Comment 5 2011-03-31 08:14:49 PDT
(In reply to comment #4) > > (In reply to comment #2) > > > Is this reproducible with WebKit trunk? > > Yesterday and today I'm trying to build QtWebKit 2.1 and/or WebKit from trunk. I'm not a Windows developer, so learning how to setup my environment & build everything from source is taking some time. If/when I get them compiled, I will try again. > > In case you have not seen it yet, here is the doc to build WebKit trunk on Windows: http://trac.webkit.org/wiki/BuildingQtOnWindows > > > Jocelyn, would you have a recent build to test this? Loading the html with qttestbrowser on trunk doesn't crash.
Benjamin Poulain
Comment 6 2011-03-31 08:20:46 PDT
> Loading the html with qttestbrowser on trunk doesn't crash. Reducing the priority since this is already fixed in trunk.
Alexis Menard (darktears)
Comment 7 2011-03-31 09:20:56 PDT
(In reply to comment #5) > (In reply to comment #4) > > > (In reply to comment #2) > > > > Is this reproducible with WebKit trunk? > > > Yesterday and today I'm trying to build QtWebKit 2.1 and/or WebKit from trunk. I'm not a Windows developer, so learning how to setup my environment & build everything from source is taking some time. If/when I get them compiled, I will try again. > > > > In case you have not seen it yet, here is the doc to build WebKit trunk on Windows: http://trac.webkit.org/wiki/BuildingQtOnWindows > > > > > > Jocelyn, would you have a recent build to test this? > > Loading the html with qttestbrowser on trunk doesn't crash. Well it has to be on a server no? I believe loading local file triggers a different path and the report says it was on a server.
Benjamin Poulain
Comment 8 2011-03-31 09:35:23 PDT
(In reply to comment #7) > Well it has to be on a server no? I believe loading local file triggers a different path and the report says it was on a server. Good point, here is the test case: http://chaos.troll.no/~poulain/57414/ (trunk does not crash for me on Linux)
Jocelyn Turcotte
Comment 9 2011-03-31 09:37:05 PDT
(In reply to comment #8) > (In reply to comment #7) > > Well it has to be on a server no? I believe loading local file triggers a different path and the report says it was on a server. > > Good point, here is the test case: http://chaos.troll.no/~poulain/57414/ (trunk does not crash for me on Linux) Doesn't crash on Windows neither.
Alexis Menard (darktears)
Comment 10 2011-03-31 09:50:29 PDT
Ademar what you want to do from that bug?
Ademar Reis
Comment 11 2011-04-11 11:36:50 PDT
(In reply to comment #10) > Ademar what you want to do from that bug? I'm not sure yet. :P I'm inclined to close it as FIXED (since it's fixed on trunk), but since it's a crash with a proper testcase, I'll keep it open until I finish my current list of pending issues, then I'll revisit it.
Ademar Reis
Comment 12 2011-04-12 06:03:26 PDT
Closing the bug as fixed and adding it to the qtwebkit-2.0 maintainance bug to keep track of it.
Thijs
Comment 13 2011-04-13 00:53:18 PDT
I just wanted to let you know that after upgrading to QtWebKit branch 2.1 (I finally got it to compile on Linux!) this was also solved for me. Good news it's also solved in the 2.0 trunk already!
Ademar Reis
Comment 14 2011-04-13 05:36:04 PDT
(In reply to comment #13) > I just wanted to let you know that after upgrading to QtWebKit branch 2.1 (I finally got it to compile on Linux!) this was also solved for me. Good news it's also solved in the 2.0 trunk already! Unfortunately it has not been fixed in 2.0. It's marked as a blocker, so it's a candidate for a future maintainance release, but the fix is not in the branch yet.
wolfy
Comment 15 2012-08-01 02:11:41 PDT
Could anybody reopen this? This bug appeared again in QT 4.8.1 and 4.8.2.
Note You need to log in before you can comment on or make changes to this bug.