Bug 95579

Summary:

[chromium] Need to rebaseline chromium tests after hidpi changes

Product:

WebKit

Reporter:

alexst

Component:

Tools / Tests

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

alexst, eric.carlson, feature-media-reviews, jamesr, webkit.review.bot

Priority:

Version:

528+ (Nightly build)

Hardware:

Unspecified

OS:

Unspecified

Attachments:

Description	Flags
Patch	none
Patch	none

alexst

Reported 2012-08-31 08:26:54 PDT

chromium DRT is now outputting 1600x1200 images for hidpi tests, need to rebaseline for all chromium platforms.

Attachments
Patch (1.33 MB, patch) 2012-08-31 08:43 PDT, alexst	no flags	Details Formatted Diff Diff
Patch (1.28 MB, patch) 2012-09-06 08:41 PDT, alexst	no flags	Details Formatted Diff Diff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.

alexst

Comment 1 2012-08-31 08:43:55 PDT

Created attachment 161708 [details] Patch

James Robinson

Comment 2 2012-08-31 14:18:06 PDT

Comment on attachment 161708 [details] Patch Are you sure that's the correct hidpi broken image icon? It looks pretty small and doesn't match the outline well

alexst

Comment 3 2012-09-01 16:57:36 PDT

(In reply to comment #2) > (From update of attachment 161708 [details]) > Are you sure that's the correct hidpi broken image icon? It looks pretty small and doesn't match the outline well You are correct, the broken image is small. I just checked, it does differ from the mac version as well, which I think is right. It's a separate bug unrelated to this CL, though. Asset selection seems broken. I don't believe it to be a regression introduced by this change (or the related one I am rebaselining for), all this is doing is adding the missing 3/4 of the hdpi image. I think the correct thing to do here is file the asset selection bug and rebaseline again once that's fixed. Do you agree?

alexst

Comment 4 2012-09-04 08:27:17 PDT

James Robinson

Comment 5 2012-09-04 09:01:39 PDT

OK, then can we rebaseline all the other tests (which do seem to be passing) and just leave a bug for the hidpi ones?

alexst

Comment 6 2012-09-04 09:47:53 PDT

(In reply to comment #5) > OK, then can we rebaseline all the other tests (which do seem to be passing) and just leave a bug for the hidpi ones? James, could you please clarify? I am not sure I understand your comment.

James Robinson

Comment 7 2012-09-04 10:26:59 PDT

Don't rebaseline the broken image icons in this patch since the new baselines are wrong. Instead, mark (or leave the marked) as expected failures with a link to the correct bug.

alexst

Comment 8 2012-09-04 10:45:30 PDT

(In reply to comment #7) > Don't rebaseline the broken image icons in this patch since the new baselines are wrong. Instead, mark (or leave the marked) as expected failures with a link to the correct bug. I could do that, but I think it will be a bit confusing for whoever will work on the hidpi broken image. The new baseline is no more wrong than the one it replaces as far as the broken image is concerned. If I leave them marked as broken, the test will fail because the resolutions mismatch, not because the old test had the correct image and the new one didn't. If anything, I think it would make more sense to rebaseline them with Mac versions of the 1600x1200 images and then mark as broken with a new bug.

alexst

Comment 9 2012-09-06 08:41:33 PDT

Created attachment 162514 [details] Patch

WebKit Review Bot

Comment 10 2012-09-06 12:30:10 PDT

Comment on attachment 162514 [details] Patch Clearing flags on attachment: 162514 Committed r127773: <http://trac.webkit.org/changeset/127773>

WebKit Review Bot

Comment 11 2012-09-06 12:30:13 PDT

All reviewed patches have been landed. Closing bug.

Note You need to log in before you can comment on or make changes to this bug.