449533 - <input-text-size-1.html> regressed on MacOSX 10.4 SeaMonkey tinderbox

Status: VERIFIED FIXED

(Core :: Graphics, defect)

Description

[Serge Gautherie (:sgautherie)]

Started to fail between

<http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1217992881.1217997477.24789.gz&fulltext=1>
"MacOSX 10.4 comm-central dep unit test on 2008/08/05 20:21:21"
rev:ea479bed5229
m-c:e6555c9b6357

and

<http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1217997270.1218001811.3744.gz&fulltext=1>
"MacOSX 10.4 comm-central dep unit test on 2008/08/05 21:34:30"
rev:ea479bed5229
m-c:2e0d459691a5

Comment 1

[Serge Gautherie (:sgautherie)]

Attachment: Error Log

Comment 2

[Robert Kaiser]

I don't see failures other than bug 447999 and what bug 444161 just has caused on current cycles - is this still happening?

Comment 3

[Serge Gautherie (:sgautherie)]

(In reply to comment #2)
> I don't see failures other than bug 447999 and what bug 444161 just has caused
> on current cycles - is this still happening?

You commented while I was looking into the fix timeframe ;->

***

Stopped failing between

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1220691025.1220696641.3532.gz
MacOSX 10.4 comm-central dep unit test on 2008/09/06 01:50:25
m-c:f7414f6cb95b

and

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1220696379.1220699848.12423.gz
MacOSX 10.4 comm-central dep unit test on 2008/09/06 03:19:39
m-c:14ce7619e9c1


"added 3 changesets with 52 changes to 49 files":
Bug 243519. Rework root element frame construction and the CanvasFrame so that [...]
Bug 381349 - Session Restore: [... Firefox specific]
Bug 243519. Fix the crashtest so that it doesn't bomb out with a JS error without completing the actual test. [...]

***

And now, roc backed out his two patches, before I submitted this comment ;-<

Comment 4

[Serge Gautherie (:sgautherie)]

(In reply to comment #3)
> And now, roc backed out his two patches, before I submitted this comment ;-<

As expected, it reappeared on
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1220705289.1220710432.30199.gz
MacOSX 10.4 comm-central dep unit test on 2008/09/06 05:48:09

Comment 5

[Serge Gautherie (:sgautherie)]

Bug 243519 2nd landing did not fix this bug
...
Actually, it didn't fix it in comment 3 either:
in fact, this test (and others) simply did not run during that timeframe :-/

Comment 6

[Serge Gautherie (:sgautherie)]

I think this test (too) should be disabled on MacOSX 10.4, as a workaround.

Comment 7

[Serge Gautherie (:sgautherie)]

(In reply to comment #0)
> "MacOSX 10.4 comm-central dep unit test on 2008/08/05 20:21:21"
> m-c:e6555c9b6357
> "MacOSX 10.4 comm-central dep unit test on 2008/08/05 21:34:30"
> m-c:2e0d459691a5

http://hg.mozilla.org/mozilla-central/pushloghtml?startdate=2008-08-05+18%3A58%3A37&enddate=2008-08-05+21%3A34%3A29
[
John Daggett — Bug 437356. Set up unified font entry object. r+sr=stuart
]

***

Fwiw, this test already had issues with fonts:
http://hg.mozilla.org/mozilla-central/log/eb49f27bbd6e/layout/reftests/forms/input-text-size-1.html

CC'ing kartl too, as he knows this test well ;->

Comment 8

[Karl Tomlinson (:karlt)]

Looks like the monospace font is not being detected as being monospace.
Does SeaMonkey have different default monospace font family or size to Firefox?

Comment 9

[Boris Zbarsky [:bzbarsky]]

It almost certainly doesn't, but the OS setup on the test machines might be different...

Comment 10

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

(In reply to comment #8)
> Looks like the monospace font is not being detected as being monospace.
> Does SeaMonkey have different default monospace font family or size to Firefox?

Any chance the code in bug 428458 (see bug 428458 comment 44) was removed/bypassed/broken by whatever checkin(s) you've determined has caused this?  John added an explicit tag to Courier and Monaco on 10.4 because for whatever reason they're not properly tagged as monospaced fonts on 10.4.

My understanding is that the MoCo boxes on 1.9.1 are all 10.5, so they wouldn't see this bug....

Comment 11

[John Daggett (:jtd)]

(In reply to comment #10)
> Any chance the code in bug 428458 (see bug 428458 comment 44) was
> removed/bypassed/broken by whatever checkin(s) you've determined has caused
> this?  

This looks like exactly what happened, I moved around the IsFixedPitch method when I created the unified gfxFontEntry object and didn't adjust the code in gfxQuartzFontCache::SetFixedPitch.  I'll post a patch for this in a little bit.

Comment 12

[John Daggett (:jtd)]

Attachment: patch, v.1, set the mFixedPitch flag within SetFixedPitch [Checkin: Comment 13]

set the mFixedPitch flag within SetFixedPitch

With this change reftest now passes on 10.4

Comment 13

[Serge Gautherie (:sgautherie)]

Comment on attachment 338265 [details] [diff] [review]
patch, v.1, set the mFixedPitch flag within SetFixedPitch
[Checkin: Comment 13]

http://hg.mozilla.org/mozilla-central/rev/39169ef45e62

Comment 14

[Serge Gautherie (:sgautherie)]

V.Fixed between
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1221199678.1221204492.6248.gz
MacOSX 10.4 comm-central dep unit test on 2008/09/11 23:07:58
and
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1221204882.1221209751.19074.gz
MacOSX 10.4 comm-central dep unit test on 2008/09/12 00:34:42