-------- Forwarded Message -------- Subject: Patch from Bug 1621274 has landed. Date: Wed, 30 Sep 2020 08:02:02 +0900 From: ISHIKAWA,chiaki <ishikawa@yk.rim.or.jp> To: dev-apps-thunderbird@lists.mozilla.org https://bugzilla.mozilla.org/show_bug.cgi?id=1621274 This eliminates the major headache in looking at other people's mochitest log for debug build on try-comm-central. That is, there are so many hits of an assertion that is related to incorrect calculation for XUL box  usage. The stacktraces from the assertion  bloats the log size pretty much, and because the stacktraces contain so many function names that they often get in the way of search something else. If the search string matches function names in the stacktraces, it is really a bother. I have no idea why others are not disturbed these stackdumps in the log. Makes me wonder if I am the only person in the world who checks the content of the log from time to time to make sure my locally created patches are sane enough. Just checking the unreliable PASS/FAIL outcome of the test is insufficient for the following reasons. Mozilla try-comm-central test framework DOES NOT report or abort tests that have serious javascript errors, etc. I can't trust a test framework which reports test success even when the executed tests throw uncaught errors, for example. We really need to check the log output more carefully especially after new features are introduced. We need a log police or something like that IMHO. Chiaki _______________________________________________ dev-apps-thunderbird mailing list dev-apps-thunderbird@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-apps-thunderbird

Thanks for giving it a gentle push across the finishing line!  -Magnus On 2020-09-30 02:25, ISHIKAWA,chiaki wrote: > > > > -------- Forwarded Message -------- > Subject:     Patch from Bug 1621274 has landed. > Date:     Wed, 30 Sep 2020 08:02:02 +0900 > From:     ISHIKAWA,chiaki <ishikawa@yk.rim.or.jp> > To:     dev-apps-thunderbird@lists.mozilla.org > > > > https://bugzilla.mozilla.org/show_bug.cgi?id=1621274 > > This eliminates the major headache in looking at other people's > mochitest log for debug build on try-comm-central. > > That is, there are so many hits of an assertion that is related to > incorrect calculation for XUL box  usage. > The stacktraces from the assertion  bloats the log size pretty much, > and because the stacktraces contain so many function names that they > often get in the way of search something else. If the search string > matches function names in the stacktraces, it is really a bother. > > I have no idea why others are not disturbed these stackdumps in the log. > Makes me wonder if I am the only person in the world who checks the > content of the log from time to time to make sure my locally created > patches are sane enough. > > Just checking the unreliable PASS/FAIL outcome of the test is > insufficient for the following reasons. > > Mozilla try-comm-central test framework DOES NOT report or abort tests > that have serious javascript errors, etc. > I can't trust a test framework which reports test success even when > the executed tests throw uncaught errors, for example. > > We really need to check the log output more carefully especially after > new features are introduced. > We need a log police or something like that IMHO. > > Chiaki > > > _______________________________________________ > dev-apps-thunderbird mailing list > dev-apps-thunderbird@lists.mozilla.org > https://lists.mozilla.org/listinfo/dev-apps-thunderbird > > _______________________________________________ > Maildev mailing list > Maildev@lists.thunderbird.net > http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net >

On 2020/09/30 15:01, Magnus Melin wrote: > Thanks for giving it a gentle push across the finishing line! > >  -Magnus > You are welcome. I think there was one line missing in the pushed patch, and the corrected one was pushed again. Stay tuned. Chiaki > On 2020-09-30 02:25, ISHIKAWA,chiaki wrote: >> >> >> >> -------- Forwarded Message -------- >> Subject:     Patch from Bug 1621274 has landed. >> Date:     Wed, 30 Sep 2020 08:02:02 +0900 >> From:     ISHIKAWA,chiaki <ishikawa@yk.rim.or.jp> >> To:     dev-apps-thunderbird@lists.mozilla.org >> >> >> >> https://bugzilla.mozilla.org/show_bug.cgi?id=1621274 >> >> This eliminates the major headache in looking at other people's >> mochitest log for debug build on try-comm-central. >> >> That is, there are so many hits of an assertion that is related to >> incorrect calculation for XUL box  usage. >> The stacktraces from the assertion  bloats the log size pretty much, >> and because the stacktraces contain so many function names that they >> often get in the way of search something else. If the search string >> matches function names in the stacktraces, it is really a bother. >> >> I have no idea why others are not disturbed these stackdumps in the log. >> Makes me wonder if I am the only person in the world who checks the >> content of the log from time to time to make sure my locally created >> patches are sane enough. >> >> Just checking the unreliable PASS/FAIL outcome of the test is >> insufficient for the following reasons. >> >> Mozilla try-comm-central test framework DOES NOT report or abort >> tests that have serious javascript errors, etc. >> I can't trust a test framework which reports test success even when >> the executed tests throw uncaught errors, for example. >> >> We really need to check the log output more carefully especially >> after new features are introduced. >> We need a log police or something like that IMHO. >> >> Chiaki >> >> >> _______________________________________________ >> dev-apps-thunderbird mailing list >> dev-apps-thunderbird@lists.mozilla.org >> https://lists.mozilla.org/listinfo/dev-apps-thunderbird >> >> _______________________________________________ >> Maildev mailing list >> Maildev@lists.thunderbird.net >> http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net >> > > _______________________________________________ > Maildev mailing list > Maildev@lists.thunderbird.net > http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net >

Now that the obnoxious assert() is gone, a few others have become dominant. We should look into them. For example, one is [15592, Main Thread] ###!!! ASSERTION: How did we end up with a negative min-size?: 'aNewMinSize >= 0', file /builds/worker/checkouts/gecko/layout/generic/nsFlexContainerFrame.cpp, line 590 and the other is ###!!! ASSERTION: Overwriting an existing document channel!: '(loadFlags & nsIChannel::LOAD_REPLACE) || !(mDocumentRequest.get())', file /builds/worker/checkouts/gecko/uriloader/base/nsDocLoader.cpp, line 453 the former is rather new. The latter has been there for some time. They can be seen in, say, https://firefoxci.taskcluster-artifacts.net/CqVWrP30Sl6pSARUdVJE0w/0/public/logs/live_backing.log of https://treeherder.mozilla.org/#/jobs?repo=try-comm-central&selectedTaskRun=JoE9TsVLSG2FoqWjkxkOQg.0&revision=3623448b2133267fee4c3b90bdac59762599c874 TIA Chiaki On 2020/10/01 4:56, ISHIKAWA,chiaki wrote: > On 2020/09/30 15:01, Magnus Melin wrote: >> Thanks for giving it a gentle push across the finishing line! >> >>  -Magnus >> > You are welcome. > I think there was one line missing in the pushed patch, and the > corrected one was pushed again. > > Stay tuned. > > Chiaki > > >> On 2020-09-30 02:25, ISHIKAWA,chiaki wrote: >>> >>> >>> >>> -------- Forwarded Message -------- >>> Subject:     Patch from Bug 1621274 has landed. >>> Date:     Wed, 30 Sep 2020 08:02:02 +0900 >>> From:     ISHIKAWA,chiaki <ishikawa@yk.rim.or.jp> >>> To:     dev-apps-thunderbird@lists.mozilla.org >>> >>> >>> >>> https://bugzilla.mozilla.org/show_bug.cgi?id=1621274 >>> >>> This eliminates the major headache in looking at other people's >>> mochitest log for debug build on try-comm-central. >>> >>> That is, there are so many hits of an assertion that is related to >>> incorrect calculation for XUL box  usage. >>> The stacktraces from the assertion  bloats the log size pretty much, >>> and because the stacktraces contain so many function names that they >>> often get in the way of search something else. If the search string >>> matches function names in the stacktraces, it is really a bother. >>> >>> I have no idea why others are not disturbed these stackdumps in the >>> log. >>> Makes me wonder if I am the only person in the world who checks the >>> content of the log from time to time to make sure my locally created >>> patches are sane enough. >>> >>> Just checking the unreliable PASS/FAIL outcome of the test is >>> insufficient for the following reasons. >>> >>> Mozilla try-comm-central test framework DOES NOT report or abort >>> tests that have serious javascript errors, etc. >>> I can't trust a test framework which reports test success even when >>> the executed tests throw uncaught errors, for example. >>> >>> We really need to check the log output more carefully especially >>> after new features are introduced. >>> We need a log police or something like that IMHO. >>> >>> Chiaki >>> >>> >>> _______________________________________________ >>> dev-apps-thunderbird mailing list >>> dev-apps-thunderbird@lists.mozilla.org >>> https://lists.mozilla.org/listinfo/dev-apps-thunderbird >>> >>> _______________________________________________ >>> Maildev mailing list >>> Maildev@lists.thunderbird.net >>> http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net >>> >> >> _______________________________________________ >> Maildev mailing list >> Maildev@lists.thunderbird.net >> http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net >> > > > > _______________________________________________ > Maildev mailing list > Maildev@lists.thunderbird.net > http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net >