Iulia Cristescu, QA [:JuliaC] (away, please needinfo? cornel.ionce@softvision.ro) Reporter
	Description • 6 years ago

This bug was filed from the Socorro interface and is
report bp-75c8bc5a-5fd5-4c79-b16d-9c7100180419.
=============================================================

Top 10 frames of crashing thread:

0 XUL CrashReporter::TerminateHandler toolkit/crashreporter/nsExceptionHandler.cpp:1434
1 libc++abi.dylib std::__terminate 
2 libc++abi.dylib __cxa_throw 
3 libobjc.A.dylib objc_begin_catch 
4 CoreFoundation -[NSMutableArray replaceObjectsInRange:withObjectsFromSet:] 
5 AppKit AppKit@0x19eca0 
6 AppKit AppKit@0x19eb5d 
7 AppKit AppKit@0x1624d7 
8 AppKit AppKit@0x19eb02 
9 AppKit AppKit@0x202ed9 

=============================================================
[Note]:
- The issue is reproducible in calls between macOS 10.13.2 (Firefox Beta) and macOS 10.12.6 (Firefox Beta)

[Affected versions]:
- 60.0b13 build1 (20180416175245)

[Affected platforms]:
- macOS 10.13.2

[Steps to reproduce]:
1. Launch Firefox
2. Create and join a https://talky.io/ chat room from the station (A)
3. Join the same room from the station (B)
4. Initiate the screen sharing on station (A) 

[Expected result]:
- The selected screen is successfully shared from (A) to (B)

[Actual result]:
- The screen sharing process is not started, the browser crashes instead on station (A)

[Regression range]:
- The issue is not reproducible on latest Nightly 61.0a1 (2018-04-19))
- We will provide more related details as soon as possible

	Nils Ohlmeier [:drno]
	Comment 1 • 6 years ago

Since this appears to be fixed in 61, this might be a duplicate of what you fixed in 61 Dan? Which bug number was that?

	Dan Minor [:dminor]
	Comment 2 • 6 years ago

My fixes were in Bug 1432793 and Bug 1453740 but those have both been uplifted and verified as fixed in beta13. The crash signature was different too, so at first glance this looks unrelated.

	Nils Ohlmeier [:drno]
	Comment 3 • 6 years ago

Lulia if your time allows it it would be a great help if you could could run an inverse regression rage search to figure out what change in 61 has fixed this issue, so we could look at uplifting that into 60.

	Nils Ohlmeier [:drno]
	Comment 4 • 6 years ago

Forgot to set NI

	Timea Zsoldos [:zstimi/tzsoldos], Desktop QA
	Comment 5 • 6 years ago

 This issue is triggered on calls between certain OS and build combinations, screen sharing from macOS 10.13 to macOS 10.12: 
  - calls between macOS 10.13 (Firefox Beta) and macOS 10.12 (Firefox Beta) 
  - calls between macOS 10.13 (Firefox Nightly) and macOS 10.12 (Firefox Nightly)
  - calls between macOS 10.13 (Firefox Release) and macOS 10.12 (Firefox Release) 
  - calls between macOS 10.13 (Firefox Nightly) and macOS 10.12 (Firefox Release)
  - calls between macOS 10.13 (Firefox Release) and macOS 10.12 (Firefox Beta) 
  - calls between macOS 10.13 (Firefox Nightly) and macOS 10.12 (Firefox Beta) 
  - calls between macOS 10.13 (Firefox Release 57.0.2) and macOS 10.12 (Firefox Release 57.0.2) 

If you start a screenshare from macOS 10.12 to macOS 10.13 when select a window to share was displayed a warning message: "Only share Firefox with sites you trust. Sharing can alow deceptive sites to browse as you and steal your private data. Learn more".

This warning message was displayed when starts a screenshare from macOS 10.13 to mac 10.12 but quickly Firefox crashed. 

[Regression range]:
Last good:275588f4d852d7dc183a9dcc70a311413dc7a063
First bad:6d4b65abc362bd5f7b6406d3b98fa538a1f492eb

Pushlog:https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=275588f4d852d7dc183a9dcc70a311413dc7a063&tochange=6d4b65abc362bd5f7b6406d3b98fa538a1f492eb

	Julien Cristau [:jcristau]
	Comment 6 • 6 years ago

That fix range in comment 5 doesn't look right, it only touched a(n unrelated) test.

	Marcia Knous [:marcia]
	Comment 7 • 6 years ago

Timea, can you double check the regression range per Comment 6? Thanks.

	Timea Zsoldos [:zstimi/tzsoldos], Desktop QA
	Comment 8 • 6 years ago

 I checked again this issue in order to find the right regression range, but I couldn't reproduce the crash anymore. I double checked every build combination from comment 5, but no crash occurred.
I also can confirm that neither Firefox 60.0b16 is affected.

	Marcia Knous [:marcia]
	Comment 9 • 6 years ago

(In reply to Timea Zsoldos from comment #8)
>  I checked again this issue in order to find the right regression range, but
> I couldn't reproduce the crash anymore. I double checked every build
> combination from comment 5, but no crash occurred.
> I also can confirm that neither Firefox 60.0b16 is affected.

It looks from crash-stats that 2 people did hit the crash in 60 b16.

	Tetsuharu OHZEKI [:tetsuharu] (UTC+9)
	Comment 11 • 6 years ago

I faced to this crash if I input some characters via a keyboard. Then, I also opened Google Hangout in a background tab (The foreground tab is different page).

	Henrik Skupin [:whimboo][⌚️UTC+1]
	Comment 12 • 6 years ago

Please note that more recent crashes could also be related to bug 1270217 which caused similar crashes eg. for printing like bug 1467582.

	Stephen A Pohl [:spohl] Assignee
	Comment 13 • 6 years ago

Bug 1290972 landed, which appears to fix this.