Bug 33280 – [RVCT] ACID3 test crash

• Status: RESOLVED
• Resolution: FIXED
• Component: Platform

[Laszlo Gombos]

To reproduce, load http://acid3.acidtests.org/ with QtLauncher or other launchers on a Symbian device.

Crash was reproduced with SVG turned off, but probably will happen if SVG turned on as well.

[Laszlo Gombos]

Created attachment 46013 [details]
RVCT workaround

[Simon Hausmann]

(In reply to comment #1)
> Created an attachment (id=46013) [details]
> RVCT workaround

This patch is now tracked in http://gitorious.org/+qtwebkit-developers/webkit/qtwebkit/commits/staging/4.6/webkit-bug-33280

[Simon Hausmann]

(In reply to comment #1)
> Created an attachment (id=46013) [details]
> RVCT workaround

Patch merged into qtwebkit-4.6 with b217cdfedc7b5e7581325bda718192247f03dd5d

[Simon Hausmann]

We haven't merged this into the 2.0 release branch yet. Laszlo, can you verify if the bug still exists, i.e. use week 14 build to try?

[Janne Koskinen]

(In reply to comment #4)
> We haven't merged this into the 2.0 release branch yet. Laszlo, can you verify
> if the bug still exists, i.e. use week 14 build to try?

This is RVCT 2.2 issue and I don't think it will go away. I have previously reproduced this with 3.1 and 5.0 and Petri came up with the same fix.

[Simon Hausmann]

(In reply to comment #5)
> (In reply to comment #4)
> > We haven't merged this into the 2.0 release branch yet. Laszlo, can you verify
> > if the bug still exists, i.e. use week 14 build to try?
> 
> This is RVCT 2.2 issue and I don't think it will go away. I have previously
> reproduced this with 3.1 and 5.0 and Petri came up with the same fix.

Surely RVCT doesn't change, but WebKit's code base changes rapidly :)

[Petri Ojala]

Tested with latest 4.7 (gitorious.org/qt/qt.git - 4.7 branch).

Crash exists, so actions needed. (no crash with 4.6)

[Laszlo Gombos]

Petri, thanks for picking this up. 

I take you mean that QtWebKit 4.6 works and WebKit 2.0 fails independently on the underlying Qt version (4.6 or 4.7). Can you confirm ?

[Petri Ojala]

Qt 4.6 + qtwebkit4.6 = OK
Qt 4.7 + webkit 2.0 = Crash

[Simon Hausmann]

(In reply to comment #9)
> Qt 4.6 + qtwebkit4.6 = OK
> Qt 4.7 + webkit 2.0 = Crash

Thank you, that means we should apply this patch to the release branch. I'm on it.

[Simon Hausmann]

Patch cherry-picked into the qtwebkit-2.0 branch with commit d91872130476bbc3ab17655693f4be1de7f9d471

[Simon Hausmann]

Patch cherry-picked into the qtwebkit-2.1 branch with commit 54ecd4d4473ddea4fe630726cab29ff06278ca2e

[Suresh Voruganti]

Simon, can you resolve the error as the patch is already picked for Qtwebkit 2.1 and 2.0

[Chang Shu]

This patch is never landed...

[Janne Koskinen]

Bump. Do we land this patch or not? I'll close this next week for "later" if no objections.

[Chang Shu]

somebody should try the rvct4 build without this patch.

[Laszlo Gombos]

Assigning to myself to test if this is still required for RVCT4.

[Siddharth Mathur]

Invalid with latest trunk + RVCT 4.0. Hence resolving.

[Ademar Reis]

(In reply to comment #18)
> Invalid with latest trunk + RVCT 4.0. Hence resolving.

I'm assuming you mean it's not needed anymore (and therefore it's FIXED), so removing the dependency from #32653.

[Siddharth Mathur]

(In reply to comment #19)
> (In reply to comment #18)
> > Invalid with latest trunk + RVCT 4.0. Hence resolving.
> 
> I'm assuming you mean it's not needed anymore (and therefore it's FIXED), so removing the dependency from #32653.

Ademar,
Changing WONTFIX->FIXED. Sorry for the confusion. 
BTW, Norbert verified the other day that we get 100% on ACID3 with trunk on Symbian.