Bug 134123

Summary:

AX: Remove tight platform expectation coupling form roles-exposed layout test.

Product:

WebKit

Reporter:

Samuel White <samuel_white>

Component:

Accessibility

Assignee:

Samuel White <samuel_white>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

aboxhall, apinheiro, buildbot, bunhere, cdumez, cfleizach, commit-queue, dmazzoni, gyuyoung.kim, jcraig, jdiggs, mario, rniwa, sergio, webkit-bug-importer

Priority:

Keywords:

InRadar

Version:

528+ (Nightly build)

Hardware:

Mac (Intel)

OS:

OS X 10.9

Attachments:

Description	Flags
Patch.	cfleizach: review+, buildbot: commit-queue-
Archive of layout-test-results from webkit-ews-10 for mac-mountainlion-wk2	none
Patch.	none

Samuel White

Reported 2014-06-20 14:30:05 PDT

We have some platform specific expectations baked into the roles-exposed.html test that limit our ability to make changes downstream. These type of expectations should be validated via the roles-exposed-expected.txt text diffing (rather than explicit .js checks in the test itself) so we can make platform changes downstream.

Attachments
Patch. (1.80 KB, patch) 2014-06-20 16:12 PDT, Samuel White	cfleizach: review+ buildbot: commit-queue-	Details Formatted Diff Diff
Archive of layout-test-results from webkit-ews-10 for mac-mountainlion-wk2 (531.60 KB, application/zip) 2014-06-20 18:50 PDT, Build Bot	no flags	Details
Patch. (107.88 KB, patch) 2014-06-24 15:56 PDT, Samuel White	no flags	Details Formatted Diff Diff
Show Obsolete (2) View All Add attachment proposed patch, testcase, etc.

Radar WebKit Bug Importer

Comment 1 2014-06-20 14:30:16 PDT

<rdar://problem/17401006>

Samuel White

Comment 2 2014-06-20 16:12:55 PDT

Created attachment 233477 [details] Patch.

James Craig

Comment 3 2014-06-20 18:12:39 PDT

You might as well pull out the role conditional check and te role description expectations too, no? It doesn’t make sense to have a PASS notice on the role if the subrole or role description are there but not being used. Just print out the actual results for each.

Build Bot

Comment 4 2014-06-20 18:50:37 PDT

Comment on attachment 233477 [details] Patch. Attachment 233477 [details] did not pass mac-wk2-ews (mac-wk2): Output: http://webkit-queues.appspot.com/results/6302044176515072 New failing tests: media/W3C/video/networkState/networkState_during_loadstart.html

Build Bot

Comment 5 2014-06-20 18:50:40 PDT

Created attachment 233496 [details] Archive of layout-test-results from webkit-ews-10 for mac-mountainlion-wk2 The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews. Bot: webkit-ews-10 Port: mac-mountainlion-wk2 Platform: Mac OS X 10.8.5

Samuel White

Comment 6 2014-06-24 15:56:41 PDT

Created attachment 233756 [details] Patch. Removing the remaining platform specific information from the test itself and the PASS/FAIL output. Platform agnostic tests such as this one should rely on the text diffs to determine PASS/FAIL.

Samuel White

Comment 7 2014-06-24 16:01:41 PDT

(In reply to comment #3) > You might as well pull out the role conditional check and te role description expectations too, no? It doesn’t make sense to have a PASS notice on the role if the subrole or role description are there but not being used. Just print out the actual results for each. Agreed. Done.

chris fleizach

Comment 8 2014-06-24 16:59:57 PDT

Comment on attachment 233756 [details] Patch. gtk will also need to be updated, so please monitor the gtk bot in order to land that result

WebKit Commit Bot

Comment 9 2014-06-24 17:56:42 PDT

Comment on attachment 233756 [details] Patch. Clearing flags on attachment: 233756 Committed r170412: <http://trac.webkit.org/changeset/170412>

WebKit Commit Bot

Comment 10 2014-06-24 17:56:48 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.