Bug 104849 – Web Inspector: Don't throw exceptions when we don't have a callstack

• Component: Web Inspector (Deprecated)
• Status: RESOLVED FIXED

[johnjbarton]

If WebInspector.debuggerModel.debuggerPausedDetails() gives details.reason === 'other', then we try to execute
  details.callFrames[0].createLiveLocation(didGetUILocation.bind(this));
but if callFrames.length === 0 we fail.

I suspect that the 'other' is related to liveEdit.

[johnjbarton]

Created attachment 179132 [details]
Patch

[Pavel Feldman]

Comment on attachment 179132 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=179132&action=review

> Source/WebCore/inspector/front-end/ScriptsPanel.js:307
> +            if (details.callFrames.length) 

Out of curiosity: have you seen this happening? We expect at least one call frame at all times. Could be related to live editing?

> Source/WebCore/inspector/front-end/ScriptsPanel.js:310
> +                this.sidebarPanes.callstack.setStatus(WebInspector.UIString("Paused, reason unknown."));

I don't think we need this line since we will anyways see the problem due to no call frames.

[johnjbarton]

(In reply to comment #2)
> (From update of attachment 179132 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=179132&action=review
> 
> > Source/WebCore/inspector/front-end/ScriptsPanel.js:307
> > +            if (details.callFrames.length) 
> 
> Out of curiosity: have you seen this happening? 

Yes, I found this by having break-on-exceptions set on devtools-on-devtools.

> We expect at least one call frame at all times. Could be related to live editing?

I'm pretty sure the case I hit was live edit.


> 
> > Source/WebCore/inspector/front-end/ScriptsPanel.js:310
> > +                this.sidebarPanes.callstack.setStatus(WebInspector.UIString("Paused, reason unknown."));
> 
> I don't think we need this line since we will anyways see the problem due to no call frames.

But as you say above the no-call-frames case is unexpected. So this message is really intended for us, so we can be alert for the no-call-frame case and learn about it.

However, I cannot find the output from setStatus() in the UI anyway.

[johnjbarton]

Created attachment 179784 [details]
Patch

[johnjbarton]

Created attachment 180372 [details]
Patch

[johnjbarton]

(In reply to comment #2)
> (From update of attachment 179132 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=179132&action=review
> 
> > Source/WebCore/inspector/front-end/ScriptsPanel.js:307
> > +            if (details.callFrames.length) 
> 
> Out of curiosity: have you seen this happening? We expect at least one call frame at all times. Could be related to live editing?

I hit this without doing live editing today.  In this case I had devtools running on devtools so this one may be related to the known incorrect behavior of js debugging in this set up.

[johnjbarton]

Can we push this to commit-queue?

[WebKit Review Bot]

Comment on attachment 180372 [details]
Patch

Clearing flags on attachment: 180372

Committed r139285: <http://trac.webkit.org/changeset/139285>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.