1331357 - Intermittent linux-qr debug | REFTEST TEST-UNEXPECTED-FAIL tests/reftest/tests/editor/reftests/caret_on_presshell_reinit-ref.html | assertion count 2 is more than expected 0 to 1 assertions

Status: RESOLVED FIXED

(Core :: Graphics: WebRender, defect, P3)

Description

[Kartikaya Gupta (email:[e-mail])]

See https://treeherder.mozilla.org/#/jobs?repo=graphics&fromchange=648d37cd4daa9a4bb3a6fc0c3ccefc96623d78a2&filter-searchStr=quantum%20ru8&selectedJob=68988061

I'm doing retriggers to track down if it's a regression. It only started showing up recently so I'm hopeful we'll isolate the change that introduced it.

Comment 1

[Kartikaya Gupta (email:[e-mail])]

Happening on R8 as well, not just Ru8.

Comment 2

[Kartikaya Gupta (email:[e-mail])]

Looks like it's actually been happening for a while, but maybe got more frequent recently.

Comment 3

[Intermittent Failures Robot]

14 failures in 690 pushes (0.02 failures/push) were associated with this bug in the last 7 days.  

Repository breakdown:
* graphics: 14

Platform breakdown:
* linux64-qr: 14

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1331357&startday=2017-01-16&endday=2017-01-22&tree=all

Comment 4

[Kartikaya Gupta (email:[e-mail])]

So the "assertion count 2" is because the same assertion is triggering twice. The test in question already has an asserts-if(0-1) to catch one instance of the assertion. I think it's reasonable to expand it to 0-2 to catch the two instances, although maybe just for webrender. The assertion is the one at [1].

[1] https://dxr.mozilla.org/mozilla-central/rev/0eef1d5a39366059677c6d7944cfe8a97265a011/layout/generic/nsTextFrame.cpp#2931

Comment 5

[Kartikaya Gupta (email:[e-mail])]

For reference the asserts-if(0-1) was added in bug 1218032, there is some discussion in bug 1218032 comment 29 and bug 1218032 comment 41. It looks like it was supposed to be investigated but never was.

Comment 6

[Kartikaya Gupta (email:[e-mail])]

Attachment: Bug 1331357 - Expand number of allowed assertions during a reftest, because in QR builds we sometimes get two.

Review commit: https://reviewboard.mozilla.org/r/112732/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/112732/

Comment 7

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Comment on attachment 8837662 [details]
Bug 1331357 - Expand number of allowed assertions during a reftest, because in QR builds we sometimes get two.

https://reviewboard.mozilla.org/r/112732/#review114478

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/05efc3638dd9
Expand number of allowed assertions during a reftest, because in QR builds we sometimes get two. r=masayuki

Comment 9

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/05efc3638dd9

Comment 10

[Intermittent Failures Robot]

13 failures in 833 pushes (0.016 failures/push) were associated with this bug in the last 7 days.  
Repository breakdown:
* mozilla-central: 5
* try: 4
* graphics: 4

Platform breakdown:
* linux64-qr: 13

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1331357&startday=2017-02-13&endday=2017-02-19&tree=all

Comment 11

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/projects/graphics/rev/4b47e1e71f37
Expand number of allowed assertions because sometimes we get 3 instead of 2. r=masayuki

Comment 12

[Kartikaya Gupta (email:[e-mail])]

Comment 11 landed with the wrong bug number. That should have gone on bug 1350004 instead. I backed it out and relanded with the right bug number. Backout is https://hg.mozilla.org/projects/graphics/rev/a61c5c0ac88bf104f0ccd1c92253f1138919c0f8.

Comment 13

[Kartikaya Gupta (email:[e-mail])]

https://hg.mozilla.org/mozilla-central/rev/4b47e1e71f37
and backout: https://hg.mozilla.org/mozilla-central/rev/a61c5c0ac88b