873974 - Test failure 'Disk cache has no entries - '1' should equal '0' ' in /testPrivateBrowsing/testAboutCache.js

Status: RESOLVED FIXED

(Mozilla QA Graveyard :: Mozmill Tests, defect)

Description

[Daniela Petrovici]

This error appeared in 19th of May on Windows 7 x86 with Nightly FR build:
http://mozmill-ci.blargon7.com/#/functional/report/14f8bc4e22e61353662cded4c2d371c0

Comment 1

[Andrei Eftimie]

Also failed on XP, Nightly:
http://mozmill-ci.blargon7.com/#/functional/report/91511d5536e962e0de7f57ed1e0c144c
(updated malformed dashboard URL)

Comment 2

[Andreea Matei [:AndreeaMatei]]

It's reproducing for me on Windows 8 as well.

Comment 3

[mario garbi]

Happened again today on Win8 x86 with Firefox 24.0a2 fr
http://mozmill-ci.blargon7.com/#/functional/report/a1b02004612785c13cf7c6bf1e6766e2

Comment 4

[Andreea Matei [:AndreeaMatei]]

Attachment: pbCache.patch

Failed again today on Win 7 with Nightly - it locale:
http://mozmill-daily.blargon7.com/#/functional/report/180cf2548ef2865af3ae441d61729b88

I reproduced it (4 times in 50 single test runs) on my Win 8 vm, while trying to get the other failure from this test (Memory cache has no entries after PB mode - 'undefined' should equal '0). 
What I found is that the initialized disk entries count, after clearing the cache in setup, is returning 1 instead of 0. 

So I moved closeAllTabs() before clearing the cache. I think that closeAllTabs it's closing each tab and opens about:newtab where sometimes is a loading timing issue and it gets incremented into the disk entries. 

I ran the test for 200 times, no fail. Also with aurora.

Comment 5

[Henrik Skupin [:whimboo][⌚️UTC+1]]

Comment on attachment 779229 [details] [diff] [review]
pbCache.patch

Review of attachment 779229 [details] [diff] [review]:
-----------------------------------------------------------------

Looks good. But please add a comment why we clear the cache at the end. So that we don't flip this order again later. r+ with the comment added.

Comment 6

[Andreea Matei [:AndreeaMatei]]

Added the comment and landed:
http://hg.mozilla.org/qa/mozmill-tests/rev/8f7f992d58fb (default)

It is possible that bug 888924 was fixed by this as well, but I'll run some more tests today to make sure. That one was reproducing more rarely.

I'll also check backporting.

Comment 7

[Henrik Skupin [:whimboo][⌚️UTC+1]]

(In reply to Andreea Matei [:AndreeaMatei] from comment #6)
> Added the comment and landed:
> http://hg.mozilla.org/qa/mozmill-tests/rev/8f7f992d58fb (default)

Please make sure to update the attachment on this bug too. Otherwise it will differ from what has been landed.

Comment 8

[Andreea Matei [:AndreeaMatei]]

Attachment: patch v2

Comment 9

[Andreea Matei [:AndreeaMatei]]

Attachment: [beta, release] pbCache.patch

It's not applying on beta and release due to aModule which is missing there and also the "use strict" line.

Comment 10

[Andreea Matei [:AndreeaMatei]]

Attachment: [esr17] pbCache.patch

We need an update for esr17 branch as it has an extra check, with these lines:
// Make sure we are not in PB mode and don't show a prompt
pb.enabled = false;
pb.showPrompt = false;

Transplanted the main patch on Aurora:
http://hg.mozilla.org/qa/mozmill-tests/rev/2e5287d875fc (aurora)

Comment 11

[Henrik Skupin [:whimboo][⌚️UTC+1]]

Comment on attachment 779660 [details] [diff] [review]
[beta, release] pbCache.patch

Review of attachment 779660 [details] [diff] [review]:
-----------------------------------------------------------------

Splinter review has problems to show this patch. But it looks fine.

Comment 12

[Andreea Matei [:AndreeaMatei]]

Thanks, landed as:
http://hg.mozilla.org/qa/mozmill-tests/rev/fe47e939f71c (beta)
http://hg.mozilla.org/qa/mozmill-tests/rev/a524d2f206f9 (release)
http://hg.mozilla.org/qa/mozmill-tests/rev/2ffa35db1bb3 (esr17)

I managed to still reproduce the other failure so I'll move to bug 888924 for the investigation there.