Closed Bug 850457 Opened 11 years ago Closed 11 years ago

reftest 759036-1.html fails on Win8 - differing pixels: 77608

Categories

(Core :: Layout, defect)

x86_64
Windows 8
defect
Not set
normal

Tracking

()

RESOLVED FIXED
mozilla22

People

(Reporter: jimm, Assigned: jimm)

References

Details

Attachments

(1 file, 2 obsolete files)

This is a pretty bad failure. We should investigate. Might be related to the color shifting problem in bug 848936.

759036-1.html | image comparison (==), max difference: 1, number of differing pixels: 77608
Another color problem here. The ref image uses 0x7fFF7f, the resulting test image renders 0x80ff80.
Attached patch patch (obsolete) — Splinter Review
This work around was discussed in bug 848936.
Assignee: nobody → jmathies
Attachment #724901 - Flags: review?(netzen)
Comment on attachment 724901 [details] [diff] [review]
patch

Review of attachment 724901 [details] [diff] [review]:
-----------------------------------------------------------------

Hey sorry Jim I don't think I can review this, maybe roc?
Attachment #724901 - Flags: review?(netzen)
Comment on attachment 724901 [details] [diff] [review]
patch

No problem. Roc, one more skip-if test..
Attachment #724901 - Flags: review?(roc)
Why are we skipping instead of using fuzz?
(In reply to Robert O'Callahan (:roc) (Mozilla Corporation) from comment #5)
> Why are we skipping instead of using fuzz?

With a difference of 77,608 pixels it just seemed prudent. Easy to switch to fuzzy though if you like.
Attached patch fuzzy patch (obsolete) — Splinter Review
Attachment #724901 - Attachment is obsolete: true
Attachment #724901 - Flags: review?(roc)
Comment on attachment 725357 [details] [diff] [review]
fuzzy patch

roc is into his weekend, so requesting review from matt. updated per what I inferred from roc's comment - change from skip to fuzzy.
Attachment #725357 - Flags: review?(mbrubeck)
Comment on attachment 725357 [details] [diff] [review]
fuzzy patch

Isn't 77608 pixels basically the whole image?  I think I prefer your original skip-if patch.  Maybe fails-if(d2d) would be best, since at least then we would be force to update the annotation when the test is fixed.

Heading out on PTO today, so I'm afraid you'll need to swap out reviewers again...
yeah, fails may well be better than fuzzy
Attachment #725357 - Flags: review?(mbrubeck)
Attached patch patchSplinter Review
now using fails-if(d2d).
Attachment #725357 - Attachment is obsolete: true
Attachment #725872 - Flags: review?(mbrubeck)
Note this fails-if applies to all of our windows tests. Win7 and xp will migrate to the new test slave hardware after win8 work is complete.
Attachment #725872 - Flags: review?(mbrubeck) → review+
https://hg.mozilla.org/mozilla-central/rev/56aa0efd1144
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla22
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: