I got two crashes in objc_msgSend within an hour of each other today. I wasn't at my computer for the first one, but the second happened when I tried to close a tab I had just opened using Cmd-W. http://crash-stats.mozilla.com/report/index/6b937267-5273-4467-bfed-7e1552091006?p=1 http://crash-stats.mozilla.com/report/index/e8010c00-d963-4319-8bd5-52fe72091006?p=1 just filing this to have this on record; this is apparently a topcrash on trunk.

same here, happened during normal browsing but don't have clear steps to reproduce so far. http://crash-stats.mozilla.com/report/index/bp-68ac1c00-afb8-4959-b450-a2a402091006

Looks like the first reported crash was on 9/29.

(In reply to comment #2) > Looks like the first reported crash was on 9/29. Don't you mean 9-23? http://crash-stats.mozilla.com/report/list?product=Firefox&version=Firefox%3A3.7a1pre&platform=mac&query_search=signature&query_type=startswith&query=objc_msgSend%20|%20-[NSWindow%20sendEvent%3A]&date=&range_value=4&range_unit=weeks&do_query=1&signature=objc_msgSend%20|%20-[NSWindow%20sendEvent%3A] Or are you separating all of those crashes out by "Address" (or some other field) and only considering certain ones to be "this crash"?

I'm getting these too... one on 10/2: bp-fc190fa5-13d0-413b-8789-e53b92091002 three yesterday: bp-cafe45a1-47be-4000-ae05-eb9602091007 bp-4957f4b4-75d4-4138-becf-52a122091007 bp-a6c89231-c790-40e8-844f-15bdc2091007

If this started on the 23rd then bug 515003 falls into the regression range. I can back it out so that we can see if it stops the crashes.

It seems pretty likely that bug 515003 is the problem. If you don't have a fix soon it should probably be backed out. You might also want to try adding some debug output to mozilla-central for a day or two if you can't reproduce and the stack traces aren't helpful enough.

(In reply to comment #6) > It seems pretty likely that bug 515003 is the problem. If you don't have a fix > soon it should probably be backed out. You might also want to try adding some > debug output to mozilla-central for a day or two if you can't reproduce and the > stack traces aren't helpful enough. Any chance bug 515003 could be backed out? I've hit this crash a fair number of times (recently including 8 times on Sunday and 4 times on Monday). bp-c402146a-7a17-4581-8244-762c92091011 (Sun Oct 11) bp-9f6034c1-10e7-48f8-aef0-8661b2091011 (Sun Oct 11) bp-3d6bb1fb-ce65-4d77-8596-20fda2091011 (Sun Oct 11) bp-222b71ae-cbe0-4730-b616-e7e212091011 (Sun Oct 11) bp-37e444dc-1c06-4233-bf39-f267b2091011 (Sun Oct 11) bp-4cc049bf-58a5-4356-ad57-9c9b22091011 (Sun Oct 11) bp-1ac01c3d-4aaa-44fb-b486-600132091011 (Sun Oct 11) bp-abe720b9-3992-401f-aead-dd6fd2091011 (Sun Oct 11) bp-2bed8d1e-9c03-43b1-8907-561bd2091012 (Mon Oct 12) bp-b7b88abc-2b0c-47e0-9c24-104102091012 (Mon Oct 12) bp-e1674e40-9990-457a-b543-480332091012 (Mon Oct 12) bp-228a176a-c838-425e-baf7-4acd82091012 (Mon Oct 12) (Around lunchtime on Monday I then switched back to using the 20090922/3.7a1pre build and I haven't seen this bug since then.)

I've just pushed bug 520044 which might fix this. Let's wait and see.

(In reply to comment #8) > I've just pushed bug 520044 which might fix this. Let's wait and see. FWIW, I've been running 20091021/3.7a1pre (b0b361366219) all day and I've only had one crash (bp-7297da03-3eda-4cf6-8ae3-3e1a92091021) and that one crash wasn't even related to this bug (it was a Flash crash). From an @objc_msgSend perspective, I think things are looking pretty good so far.

Crash stats also doesn't have any crashes with that build yet.

Seems fixed.