Filed by: rmaries [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer.html#?job_id=313997971&repo=autoland
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/UuO_QiCtRRSuMq7izDfVAg/runs/0/artifacts/public/logs/live_backing.log
Reftest URL: https://hg.mozilla.org/mozilla-central/raw-file/tip/layout/tools/reftest/reftest-analyzer.xhtml#logurl=https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/UuO_QiCtRRSuMq7izDfVAg/runs/0/artifacts/public/logs/live_backing.log&only_show_unexpected=1

Using a bug separate from bug 827183.

I wonder if there's a pernosco recording of this.

Ugh, it won't because these are disabled on Linux... Will push some try runs with this enabled to see if I catch it.

Stack:

#13 0x000007a85baa987e in NS_DebugBreak(uint32_t, char const*, char const*, char const*, int32_t)
    (aSeverity=1, aStr=0x7a857a8f9ae "User did not call nsIContentViewer::Destroy", aExpr=0x7a857522f7c "!mPresShell && !mPresContext", aFile=0x7a8578a2000 "/home/emilio/src/moz/gecko-2/layout/base/nsDocumentViewer.cpp", aLine=574) at /home/emilio/src/moz/gecko-2/xpcom/base/nsDebugImpl.cpp:366
#14 0x000007a85eed57cb in nsDocumentViewer::~nsDocumentViewer() (this=0x9aa12589900) at /home/emilio/src/moz/gecko-2/layout/base/nsDocumentViewer.cpp:573
#15 0x000007a85eed6162 in nsDocumentViewer::~nsDocumentViewer() (this=0x9aa12589900) at /home/emilio/src/moz/gecko-2/layout/base/nsDocumentViewer.cpp:559
#16 0x000007a85eed5522 in nsDocumentViewer::Release() (this=0x9aa12589900) at /home/emilio/src/moz/gecko-2/layout/base/nsDocumentViewer.cpp:547
#17 0x000007a85f207cec in nsPagePrintTimer::~nsPagePrintTimer() (this=0x6febd280) at /home/emilio/src/moz/gecko-2/layout/printing/nsPagePrintTimer.cpp:17
#18 0x000007a85bb87119 in mozilla::Runnable::Release() (this=0x6febd280) at /home/emilio/src/moz/gecko-2/xpcom/threads/nsThreadUtils.cpp:68
#19 0x000007a85bb8808c in nsTimerImpl::Callback::clear() (this=0x7ffe6e6f2f10) at /home/emilio/src/moz/gecko-2/xpcom/threads/nsTimerImpl.h:107
#20 nsTimerImpl::Callback::~Callback() (this=0x7ffe6e6f2f10) at /home/emilio/src/moz/gecko-2/xpcom/threads/nsTimerImpl.h:104
#21 nsTimerImpl::Fire(int) (this=0x6fe9a190, aGeneration=1) at /home/emilio/src/moz/gecko-2/xpcom/threads/nsTimerImpl.cpp:597
#22 0x000007a85bb87d34 in nsTimerEvent::Run() (this=0x6fed5140) at /home/emilio/src/moz/gecko-2/xpcom/threads/TimerThread.cpp:251

It is the print document, set here, but the docshell gets destroyed before the print finishes:

#0  nsCOMPtr<nsIContentViewer>::assign_assuming_AddRef(nsIContentViewer*) (this=0x17046a731aa0, aNewPtr=0x0) at /home/emilio/src/moz/gecko-2/obj-debug-no-sccache/dist/include/nsCOMPtr.h:413
#1  nsCOMPtr<nsIContentViewer>::operator=(decltype(nullptr)) (this=0x17046a731aa0) at /home/emilio/src/moz/gecko-2/obj-debug-no-sccache/dist/include/nsCOMPtr.h:684
#2  nsDocShell::Destroy() (this=0x17046a731800) at /home/emilio/src/moz/gecko-2/docshell/base/nsDocShell.cpp:4223
#3  0x000007a85fe9f50a in nsWebBrowser::SetDocShell(nsDocShell*) (this=0x6fe8b4c0, aDocShell=0x0) at /home/emilio/src/moz/gecko-2/toolkit/components/browser/nsWebBrowser.cpp:1131
#4  0x000007a85fe9ec41 in nsWebBrowser::InternalDestroy() (this=0x6fe8b4c0) at /home/emilio/src/moz/gecko-2/toolkit/components/browser/nsWebBrowser.cpp:174
#5  0x000007a85fea1a8b in non-virtual thunk to nsWebBrowser::Destroy() () at /home/emilio/src/moz/gecko-2/toolkit/components/browser/nsWebBrowser.cpp:855

We do call nsDocumentViewer::Destroy from there, but we hit this condition and thus keep stuff alive.

That "blocked destroy count" gets decremented here, right before dropping the content viewer entirely:

#0  0x000007a85eedf5db in non-virtual thunk to nsDocumentViewer::DecrementDestroyBlockedCount() () at /home/emilio/src/moz/gecko-2/layout/base/nsDocumentViewer.cpp:3727
#1  0x000007a85f207c8e in nsPagePrintTimer::~nsPagePrintTimer() (this=0x6febd280) at /home/emilio/src/moz/gecko-2/layout/printing/nsPagePrintTimer.cpp:19
#2  0x000007a85f207cec in nsPagePrintTimer::~nsPagePrintTimer() (this=0x6febd280) at /home/emilio/src/moz/gecko-2/layout/printing/nsPagePrintTimer.cpp:17
#3  0x000007a85bb87119 in mozilla::Runnable::Release() (this=0x6febd280) at /home/emilio/src/moz/gecko-2/xpcom/threads/nsThreadUtils.cpp:68
#4  0x000007a85bb8808c in nsTimerImpl::Callback::clear() (this=0x7ffe6e6f2f10) at /home/emilio/src/moz/gecko-2/xpcom/threads/nsTimerImpl.h:107
#5  nsTimerImpl::Callback::~Callback() (this=0x7ffe6e6f2f10) at /home/emilio/src/moz/gecko-2/xpcom/threads/nsTimerImpl.h:104
#6  nsTimerImpl::Fire(int) (this=0x6fe9a190, aGeneration=1) at /home/emilio/src/moz/gecko-2/xpcom/threads/nsTimerImpl.cpp:597
#7  0x000007a85bb87d34 in nsTimerEvent::Run() (this=0x6fed5140) at /home/emilio/src/moz/gecko-2/xpcom/threads/TimerThread.cpp:251
#8  0x000007a85bb7994d in mozilla::SchedulerGroup::Runnable::Run() (this=0x6fea84c0) at /home/emilio/src/moz/gecko-2/xpcom/threads/SchedulerGroup.cpp:146
#9  0x000007a85bb80255 in mozilla::RunnableTask::Run() (this=0x6febd900) at /home/emilio/src/moz/gecko-2/xpcom/threads/TaskController.cpp:242

By the time we close the window, we're already done, nsPagePrintTimer::mDone is true.

Do it as soon as we're done printing, rather than waiting until the
print timer dies.

Set release status flags based on info from the regressing bug 1636728

Backed out for perma failures.

Push with failures: https://treeherder.mozilla.org/#/jobs?repo=autoland&selectedTaskRun=WmF9Fi1PRgSB86EVtowM1Q.0&resultStatus=testfailed%2Cbusted%2Cexception&revision=dd7fb8f29da26ebcd74febb515588efe97e964fc&failure_classification_id=2

Logs:
https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=314278143&repo=autoland&lineNumber=39433
https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=314278756&repo=autoland&lineNumber=60520

Backout: https://hg.mozilla.org/integration/autoland/rev/cd9583400fea833d307c7b95a97498a82561999c

https://hg.mozilla.org/mozilla-central/rev/e6ef511d20c9

Comment on attachment 9172635 [details]
Bug 1661137 - Decrement the document viewer's destroy blocker earlier. r=bobowen

Beta/Release Uplift Approval Request

• User impact if declined: Printing project
• Is this code covered by automated tests?: No
• Has the fix been verified in Nightly?: Yes
• Needs manual test from QE?: No
• If yes, steps to reproduce:
• List of other uplifts needed: None
• Risk to taking this patch: Medium
• Why is the change risky/not risky? (and alternatives if risky):
• String changes made/needed:

Comment on attachment 9172635 [details]
Bug 1661137 - Decrement the document viewer's destroy blocker earlier. r=bobowen

Approved for 81.0b6.

https://hg.mozilla.org/releases/mozilla-beta/rev/6081b620e74f