Manual Good and Bad Builds:
Good Build:55.0.2845.0
Bad Build: 55.0.2846.0

MAMUAL CHANGELOG:
https://chromium.googlesource.com/chromium/src/+log/55.0.2845.0..55.0.2846.0?pretty=fuller&n=10000

Confirm at MacOS chrome Version: 55.0.2846.0

Able to reproduce this on the latest canary(55.0.2846.0) on Mac OS 10.11.6 and Linux Ubuntu 14.04. Mac/Linux canary is unusable and crashes every min or so. Works fine on Windows-10 with the same chrome version.

Stack trace of bf484cd100000000 from Mac: 

Thread 18 CRASHED [EXC_BAD_INSTRUCTION / EXC_I386_INVOP @ 0x000000010d5c26d6 ] MAGIC SIGNATURE THREAD
0x000000010d5c26d6	(Google Chrome Framework -thread_watcher.cc:389 )	ThreadWatcher::StartWatching(ThreadWatcher::WatchingParams const&)
0x000000010d5c3909	(Google Chrome Framework -thread_watcher.cc:585 )	ThreadWatcherList::InitializeAndStartWatching(unsigned int, std::__1::map<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, ThreadWatcherList::CrashDataThresholds, std::__1::less<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::allocator<std::__1::pair<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const, ThreadWatcherList::CrashDataThresholds> > > const&)
0x000000010d92f9fa	(Google Chrome Framework -callback.h:388 )	base::debug::TaskAnnotator::RunTask(char const*, base::PendingTask const&)
0x000000010d95127b	(Google Chrome Framework -message_loop.cc:488 )	base::MessageLoop::RunTask(base::PendingTask const&)
0x000000010d9515bb	(Google Chrome Framework -message_loop.cc:497 )	base::MessageLoop::DeferOrRunPendingTask(base::PendingTask)
0x000000010d951b3c	(Google Chrome Framework -message_loop.cc:660 )	base::MessageLoop::DoDelayedWork(base::TimeTicks*)
0x000000010d953988	(Google Chrome Framework -message_pump_mac.mm:334 )	base::MessagePumpCFRunLoopBase::RunWork()
0x000000010d947729	(Google Chrome Framework + 0x018ec729 )	base::mac::CallWithEHFrame(void () block_pointer)
0x000000010d953373	(Google Chrome Framework -message_pump_mac.mm:306 )	base::MessagePumpCFRunLoopBase::RunWorkSource(void*)
0x00007fff92551880	(CoreFoundation + 0x000aa880 )	__CFRUNLOOP_IS_CALLING_OUT_TO_A_SOURCE0_PERFORM_FUNCTION__
0x00007fff92530fbb	(CoreFoundation + 0x00089fbb )	__CFRunLoopDoSources0
0x00007fff925304de	(CoreFoundation + 0x000894de )	__CFRunLoopRun
0x00007fff9252fed7	(CoreFoundation + 0x00088ed7 )	CFRunLoopRunSpecific
0x000000010d953d4e	(Google Chrome Framework -message_pump_mac.mm:554 )	base::MessagePumpCFRunLoop::DoRun(base::MessagePump::Delegate*)
0x000000010d9537c3	(Google Chrome Framework -message_pump_mac.mm:238 )	base::MessagePumpCFRunLoopBase::Run(base::MessagePump::Delegate*)
0x000000010d96dbb0	(Google Chrome Framework -run_loop.cc:35 )	base::RunLoop::Run()
0x000000010d996e3e	(Google Chrome Framework -thread.cc:301 )	base::Thread::ThreadMain()
0x000000010d992f76	(Google Chrome Framework -platform_thread_posix.cc:71 )	base::(anonymous namespace)::ThreadFunc(void*)
0x00007fff974f599c	(libsystem_pthread.dylib + 0x0000399c )	_pthread_body
0x00007fff974f5919	(libsystem_pthread.dylib + 0x00003919 )	_pthread_start
0x00007fff974f3350	(libsystem_pthread.dylib + 0x00001350 )	thread_start
0x000000010d992f1f	(Google Chrome Framework + 0x01937f1f )	

Based on the above changelog from C#1, suspecting: https://codereview.chromium.org/2296793003 for 'thread_watcher.cc' related change.

jlevasseur@: Could you please take a look at this crash.

Thank you!

#1 browser crash on Windows(32 bit) and Mac canary based on the available crash data in the crash server.

Note:
=====
Crash data for Windows(64 bit) is not showing as of now.

Canary upgraded to 55.0.2846.0 on Windows 10 and this is crashing every 2 mins.

jlevasseur@: Can we get an update on this issue get the suspected CL reverted if that is related.

 

Issue 643122 has been merged into this issue.

Issue 643159 has been merged into this issue.

Looks like https://codereview.chromium.org/2295983002/ was reverted and should fix this crash.

Can we build and push a new canary?

+Desktop PMO (go/chromeschedule doesn't yet have a specific owner for M55 desktop), can we reroll Canary when the revert (https://codereview.chromium.org/2299453005/) lands?

(monorail doesn't highlight added CC's? FYI, #11 was specifically aimed at govind/bustamante)

Greetings, same here, 4 times out of 5 Canary crashes before it can check whether there is a new version.

jkummerow@, can you please merge the revert (https://codereview.chromium.org/2299453005/) to canary branch 2846 once it lands. So we can trigger new canary from same branch.

Geetha/Di, please trigger new canary from branch 2846 once above revert is merged. Thank you.

jkummerow@ is in EMEA, I'll handle the merge.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/ec1da671f97f53a3fa24505c80593229264941fd

commit ec1da671f97f53a3fa24505c80593229264941fd
Author: jkummerow <jkummerow@chromium.org>
Date: Thu Sep 01 15:17:48 2016

Revert of Fix memory leak in ThreadWatcher. (patchset #3 id:40001 of https://codereview.chromium.org/2295983002/ )

Reason for revert:
Current Canary is unusable due to browser process crashes in line 88 (Windows) and 389 (Mac) in thread_watcher.cc, see  crbug.com/643109 .

Original issue's description:
> Fix memory leak in ThreadWatcher.
>
> BUG=none
> TEST=unit tests
>
> Committed: https://crrev.com/5ca77e86703646cb84be35ef827068ac6951be1e
> Cr-Commit-Position: refs/heads/master@{#415729}

TBR=isherman@chromium.org,rtenneti@chromium.org,wzhong@chromium.org,alokp@chromium.org,jlevasseur@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTRY=true
BUG= 643109 

Review-Url: https://codereview.chromium.org/2299453005
Cr-Commit-Position: refs/heads/master@{#415955}

[modify] https://crrev.com/ec1da671f97f53a3fa24505c80593229264941fd/chrome/browser/metrics/thread_watcher.cc
[modify] https://crrev.com/ec1da671f97f53a3fa24505c80593229264941fd/chrome/browser/metrics/thread_watcher.h
[modify] https://crrev.com/ec1da671f97f53a3fa24505c80593229264941fd/chrome/browser/metrics/thread_watcher_unittest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1917cde9d763b20e310b04da3c8d336b2fafbc5b

commit 1917cde9d763b20e310b04da3c8d336b2fafbc5b
Author: Gabriel Charette <gab@chromium.org>
Date: Thu Sep 01 15:23:20 2016

[Merge to 55.0.2846.2] Revert of Fix memory leak in ThreadWatcher. (patchset #3 id:40001 of https://codereview.chromium.org/2295983002/ )

Reason for revert:
Current Canary is unusable due to browser process crashes in line 88 (Windows) and 389 (Mac) in thread_watcher.cc, see  crbug.com/643109 .

Original issue's description:
> Fix memory leak in ThreadWatcher.
>
> BUG=none
> TEST=unit tests
>
> Committed: https://crrev.com/5ca77e86703646cb84be35ef827068ac6951be1e
> Cr-Commit-Position: refs/heads/master@{#415729}

TBR=isherman@chromium.org,rtenneti@chromium.org,wzhong@chromium.org,alokp@chromium.org,jlevasseur@chromium.org
NOPRESUBMIT=true
NOTRY=true
BUG= 643109 

Review-Url: https://codereview.chromium.org/2299453005
Cr-Commit-Position: refs/heads/master@{#415955}
(cherry picked from commit ec1da671f97f53a3fa24505c80593229264941fd)

Review URL: https://codereview.chromium.org/2299163003 .

Cr-Commit-Position: refs/branch-heads/2846@{#4}
Cr-Branched-From: c5c7ef8980b50bc69d99046118ec7ff767c27fa2-refs/heads/master@{#415833}

[modify] https://crrev.com/1917cde9d763b20e310b04da3c8d336b2fafbc5b/chrome/browser/metrics/thread_watcher.cc
[modify] https://crrev.com/1917cde9d763b20e310b04da3c8d336b2fafbc5b/chrome/browser/metrics/thread_watcher.h
[modify] https://crrev.com/1917cde9d763b20e310b04da3c8d336b2fafbc5b/chrome/browser/metrics/thread_watcher_unittest.cc

Merge complete @pucchakayala for Canary re-roll.

Oops

Thank you gab@. I'm triggering new Canary now from branch 2846. 

Adding the gCon hotlist - this is a top issue reported by users on m55 as well: https://screenshot.googleplex.com/bBwveboTfzZ

Issue 643218 has been merged into this issue.

Issue 643216 has been merged into this issue.

Issue 643210 has been merged into this issue.

 Issue 643209  has been merged into this issue.

Issue 643206 has been merged into this issue.

Issue 643203 has been merged into this issue.

Issue 643163 has been merged into this issue.

Users experienced this crash on the following builds:

Mac Canary 55.0.2846.0 -  8642.09 CPM, 3099 reports, 2303 clients (signature ThreadWatcher::StartWatching)

If this update was incorrect, please add "Fracas-Wrong" label to prevent future updates.

- Go/Fracas

Issue 643261 has been merged into this issue.

Issue 643273 has been merged into this issue.

 Issue 643271  has been merged into this issue.

In case it's useful: on the macOS Sierra beta this Canary printed a lot of warning lines to the Terminal (when launched from there) like:

objc[1414]: __weak variable at 0x7f8531eb5020 holds 0x2121212121212121 instead of 0x7f852eff20f0. This is probably incorrect use of objc_storeWeak() and objc_loadWeak(). Break on objc_weak_error to debug.

Issue 643277 has been merged into this issue.

 Issue 643301  has been merged into this issue.

Issue 643343 has been merged into this issue.

 Issue 643344  has been merged into this issue.

 Issue 643308  has been merged into this issue.

Any ETA on when the fix will rollout to Mac?

And Windows as well. Really waiting to use canary. Is there any other way a normal user can get the fixed build before updates are rolled out ? 

Is there a new build available (for Windows) yet? 

Issue 643373 has been merged into this issue.

I've gotten 55.0.2846.4, and it seems to be much more stable now

New canary 55.0.2846.4 was triggered this morning and its available now as per https://omahaproxy.appspot.com/

Thank you so much! Just updated and now it seems fine on my pc. 

Issue 643403 has been merged into this issue.

Issue 643319 has been merged into this issue.

 Issue 643709  has been merged into this issue.

 Issue 643708  has been merged into this issue.

Issue 643614 has been merged into this issue.

Issue 643755 has been merged into this issue.

Issue 643754 has been merged into this issue.

Issue 644255 has been merged into this issue.

Issue 644331 has been merged into this issue.

Issue 644330 has been merged into this issue.

 Issue 644289  has been merged into this issue.

Spike levels dropped in feedback reports as well.
https://screenshot.googleplex.com/W0XyzTQjivc

Issue 643244 has been merged into this issue.

Issue 643170 has been merged into this issue.