Bug 153938 – Eclipse JVM crashes in libjavascriptcoregtk JSC::JSCell::getPrimitiveNumber

• Component: WebKitGTK
• Status: RESOLVED WONTFIX

Created attachment 270789 [details]
JVM crash log

This crash happened while using Eclipse Mars.1 which had:
- a markdown editor tab showing the Preview view (web-like); the markdown has no javascript
- a web browser tab open on:
http://www.brpreiss.com/books/opus8/html/page96.html#SECTION004300000000000000000

Eclipse was running with:
java version "1.8.0_51"
Java(TM) SE Runtime Environment (build 1.8.0_51-b16)
Java HotSpot(TM) 64-Bit Server VM (build 25.51-b03, mixed mode)
from jdk-8u51-linux-x64 downloaded as a tarball from the web.

The log says:
Stack: [0x00007f4bab57b000,0x00007f4bab67c000],  sp=0x00007f4bab676cb8,  free space=1007k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
C  [libjavascriptcoregtk-3.0.so.0+0x4c2ec0]  JSC::JSCell::getPrimitiveNumber(JSC::ExecState*, double&, JSC::JSValue&) const+0x0
C  0x00007f4a3271ee61
Java frames: (J=compiled Java code, j=interpreted, Vv=VM code)
J 14833  org.eclipse.swt.internal.gtk.OS._gtk_widget_destroy(J)V (0 bytes) @ 0x00007f4b97aa4681 [0x00007f4b97aa4640+0x41]
J 42420 C2 org.eclipse.swt.widgets.Widget.destroyWidget()V (30 bytes) @ 0x00007f4b9c36671c [0x00007f4b9c3665a0+0x17c]
J 27969 C2 org.eclipse.swt.widgets.Widget.release(Z)V (90 bytes) @ 0x00007f4b97bb5d1c [0x00007f4b97bb5c40+0xdc]
J 41828 C1 org.eclipse.swt.widgets.Control.release(Z)V (115 bytes) @ 0x00007f4b977dd834 [0x00007f4b977dd1c0+0x674]
J 36561 C2 org.eclipse.swt.widgets.Widget.dispose()V (27 bytes) @ 0x00007f4b9aa99d88 [0x00007f4b9aa99d20+0x68]
j  org.eclipse.e4.ui.workbench.renderers.swt.SWTPartRenderer.disposeWidget(Lorg/eclipse/e4/ui/model/application/ui/MUIElement;)V+40
[...]

Core dumps were disabled, so I don't have any.

If this is insufficient to investigate, feel free to close as unreproducible.  This is the first time I encounter this crash.

Thanks!

Best regards
Fabrice

[Michael Catanzaro]

WebKit1 bug

[Michael Catanzaro]

Sorry, I should be more clear when closing new reports.

You are using an old version of WebKitGTK+ that is no longer supported.

I discuss this a bit at the bottom of: https://blogs.gnome.org/mcatanzaro/2016/02/01/on-webkit-security-updates/

[Carlos Garcia Campos]

We don't work on wk1 anymore, but anybody is free to provide a patch for any issue that applies in the 2.4 branch, and I'll happily apply it and make a release.

[Michael Catanzaro]

(In reply to comment #3)
> We don't work on wk1 anymore, but anybody is free to provide a patch for any
> issue that applies in the 2.4 branch, and I'll happily apply it and make a
> release.

Indeed, the WONTFIX state indicates that nobody is going to look into this upstream, but patches are still accepted. I will start replying with this to future WK1 bugs, so you don't have to keep correcting me. :)

[Carlos Garcia Campos]

(In reply to comment #4)
> (In reply to comment #3)
> > We don't work on wk1 anymore, but anybody is free to provide a patch for any
> > issue that applies in the 2.4 branch, and I'll happily apply it and make a
> > release.
> 
> Indeed, the WONTFIX state indicates that nobody is going to look into this
> upstream, but patches are still accepted. I will start replying with this to
> future WK1 bugs, so you don't have to keep correcting me. :)

How do you know nobody is going to look into this? How is anybody going to look into this if we close the bug as wontfix? While the 2.4 branch is alive, anybody is free to work on it, the fact that I'm not going to work on it doesn't mean that the bug will not be fixed, because everybody is free to propose patches. People are not going to contribute a patch to a bug closed as wontfix.

[Michael Catanzaro]

Definitely time to close 2.4 bugs.