1. Install Spider https://bclary.com/projects/spider/spider-0.1.0.5-sm+tb+fx+an+fn.xpi 2. firefox -spider -start -quit -url 'http://reservations.airarabia.com/agents/private/showPageTimedOut' -hook https://bclary.com/projects/spider/test-crash-on-load.js ###!!! ASSERTION: Why is there a document channel?: '!chan', file /home/worker/workspace/build/src/toolkit/components/windowwatcher/nsWindowWatcher.cpp, line 2181 Assertion failure: aNextTabParentId == 0, at /home/worker/workspace/build/src/xpfe/appshell/nsXULWindow.cpp:2016 happens on shutdown. http://reservations.airarabia.com/agents/private/showPageTimedOut and 15 other urls reproduce this in Bughunter. The common behavior between the urls is the opening of a new window. It *requires* Spider *and* the use of the -hook in order to reproduce so this may be invalid. If so, please mark it so. Ehsan: You introduced the assertion in bug 1356922.

(In reply to Bob Clary [:bc:] from comment #0) > Created attachment 8872134 [details] > log with assertion stack > > 1. Install Spider > https://bclary.com/projects/spider/spider-0.1.0.5-sm+tb+fx+an+fn.xpi This link is 404ing for me... Also if this extension is an old-school unsigned extension installing it and then running something from the command line isn't possible any longer since unsigned extensions can now only be installed manually through about:debugging through a pretty painful process which would preclude me from running the step 2 of your instructions. I'm mentioning it because I ran into a similar issue with a fuzzing extension recently to minimize turn around time. :-) Please let me know how to proceed. Thanks!

(In reply to :Ehsan Akhgari (needinfo please, extremely long backlog) from comment #1) > (In reply to Bob Clary [:bc:] from comment #0) > > Created attachment 8872134 [details] > > log with assertion stack > > > > 1. Install Spider > > https://bclary.com/projects/spider/spider-0.1.0.5-sm+tb+fx+an+fn.xpi > > This link is 404ing for me... > Sorry. Fixed. > Also if this extension is an old-school unsigned extension installing it and > then running something from the command line isn't possible any longer since > unsigned extensions can now only be installed manually through > about:debugging through a pretty painful process which would preclude me > from running the step 2 of your instructions. I'm mentioning it because I > ran into a similar issue with a fuzzing extension recently to minimize turn > around time. :-) > It is old school but is signed. > Please let me know how to proceed. Thanks! The link should work and since it is signed you should be able to download and install it. It does run from the command line. I'll have to update it soon to deal with the deprecation of extensions, but for the moment it still works.

Sorry, I updated the web site but forgot to update the link here. https://bclary.com/projects/spider/spider-0.1.0.5-an+fn+fx+sm+tb.xpi

I didn't manage to reproduce this myself but I spoke to Bob on IRC. One thing we got to was the primary=true attribute missing from https://hg.mozilla.org/automation/sisyphus/file/tip/spider/chrome/content/spider/spider.xul#l201 which could have caused this to fire. Bob said he will try adding it to see what will happen.

Still get the assertion with this patch to Spider.

I tried with the assertion disabled and it seemed to be about the same. Both with and without the assertion I got several firefox: fccache.c:548: FcCacheFini: Assertion `fcCacheChains[i] == NULL' failed. though with the assertion disabled, I got the fccache assert more since I didn't die due to this fatal assertion. Also, with Spider updated to use primary="true" I lose the scrollbars on the xul:browser which I am loath to do.

(In reply to Bob Clary [:bc:] from comment #6) > I tried with the assertion disabled and it seemed to be about the same. Both > with and without the assertion I got several > > firefox: fccache.c:548: FcCacheFini: Assertion `fcCacheChains[i] == NULL' > failed. > > though with the assertion disabled, I got the fccache assert more since I > didn't die due to this fatal assertion. Huh, I don't know what that is, I don't think that comes from our code. Does that go away if you back out bug 1356922 locally? If no, I think removing the assertion is the right thing to do. Can you please post a diff of the changes that you have applied? > Also, with Spider updated to use primary="true" I lose the scrollbars on the > xul:browser which I am loath to do. It seems that didn't help anyway.

This is what I did. The fccache thing has existed for a while I think and is not related.

Fwiw, I hit this assertion when running the cross_fuzz fuzzer. http://lcamtuf.coredump.cx/cross_fuzz/