Brackets crashes when backspacing to delete the css shape while Live Preview is still loading the page.

[RaymondLim]

I was trying to reproduce #7437 by typing shape-inside: circle(); and launching Live Preview. For this specific crash, I did Live preview with shape-inside: circle(); already in the code. Before the browser finish loading I went back to Brackets and hit backspace a couple times to delete from the end of circle();. Then I see black screen inside the Brackets window.

@JeffryBooher See you can map to source from this crash log to get some useful info.

Note that I was testing in the version before this pull request #7424 landed in master.

[RaymondLim]

Not sure what we can do with the crash log, but I can't reproduce it any more since it depends on the timing of loading and deleting in Brackets.

[dangoor]

I've been trying to reproduce this, but so far I've only been able to crash Chrome (34.0.1847.116 beta)

[dangoor]

Ahh, what I'm seeing is #7422

[pthiess]

@redmunds couldn't reproduce, seems like an edge case

[oslego]

@RaymondLim is this not a case on #7422 ?

[RaymondLim]

@oslego No, this is a crash in Brackets, not in the browser. We have the crash log to investigate. Otherwise, this is utr and should be closed.

[oslego]

Ok, thanks. What does UTR stand for?

[RaymondLim]

Unable To Reproduce.

[zhorvath]

FYI, since Shapes Level 1 contains only shape-outside, shape-inside has been removed from WebKit&Blink. Shape-inside support is only available until the following nightly builds: WebKit r166290 (2014-03-26); Chromium 260092 (2014-03-28)

[oslego]

@zhorvath thanks for the details! Acknowledged.

The shapes editor will turn off silently when it does't find support for shape-inside in the browser - it is using feature detectior. The editor will work again once shape-inside resurfaces.

We'll need to post a blog update on Brackets when the shape-inside-less Chrome hits stable so people don't get confused. We're already communicating to people that shape-inside is out of Blink temporarily.

[pthiess]

@njx @RaymondLim I'm inclined to even close this for now since there is work ongoing with the Shapes editor and I'm not sure where cause and effect is to be presumed. The crash log indicates one of the CEF helper libraries is involved (org.cef.cefclient.helper). What's your take?

[njx]

Closing - our current version of CEF is old and it's actually hard to find symbols for it :) Given that this was a one-time crash that we can't repro, and we're upgrading to a much newer CEF anyway, let's close this and then just see if it comes back later, rather than taking the time to symbolize it now.