Bug 71665 - [GTK] platform/gtk/accessibility/name-from-label.html fails
Summary: [GTK] platform/gtk/accessibility/name-from-label.html fails
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebKitGTK (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Mario Sanchez Prada
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-11-07 04:39 PST by Philippe Normand
Modified: 2011-11-22 00:30 PST (History)
3 users (show)

See Also:


Attachments
Patch (2.64 KB, patch)
2011-11-21 06:15 PST, Mario Sanchez Prada
pnormand: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Philippe Normand 2011-11-07 04:39:27 PST
Currently skipped in GTK as anonymous flaky but failure looks consistent here:

--- /home/phil/gst/jhbuild/build/WebKit/WebKitBuild/Release/layout-test-results/platform/gtk/accessibility/name-from-label-expected.txt
+++ /home/phil/gst/jhbuild/build/WebKit/WebKitBuild/Release/layout-test-results/platform/gtk/accessibility/name-from-label-actual.txt
@@ -4,8 +4,8 @@
 On success, you will see a series of "PASS" messages, followed by "TEST COMPLETE".
 
 
-PASS child.role is 'AXRole: entry'
-PASS child.title is 'AXTitle: Full Name:'
+FAIL child.role should be AXRole: entry. Was AXRole: paragraph.
+FAIL child.title should be AXTitle: Full Name:. Was AXTitle: .
 PASS successfullyParsed is true
 
 TEST COMPLETE
Comment 1 Mario Sanchez Prada 2011-11-21 06:15:31 PST
Created attachment 116073 [details]
Patch
Comment 2 Mario Sanchez Prada 2011-11-21 06:19:51 PST
(In reply to comment #1)
> Created an attachment (id=116073) [details]
> Patch

It seems this has been a problem even before November the 7th (tried with a commit from Nov 1st, and was failing as well there). I have been trying to find out the exact point where it start failing but couldn't find it. 

However, as per visual inspection in the code, I now wonder how it could be working fine before, since its obvious that the form object in the middle was not being taking into account when it should, so either form objects were explicitly not being exposed before (which is not the case I think) or there was a bug somewhere else (e.g. determining when the webArea is focusable, determining if a form is visible or not...) that was fixed at some point recently, unleashing the problem with this wrongly implemented layout test (that I wrote btw, so assuming my fault on that :().

My guessing is that the reason why it started failing recently was the later one (a bug somewhere else that got fixed, unveiling the problem with this test), so that's why I propose now this patch, which is basically a rewrite of the layout test that properly considers the a11y hierarchy.
Comment 3 Philippe Normand 2011-11-21 06:23:00 PST
It was skipped *before* I opened this bug, which I did when I moved the flaky tests in tests_expectations.txt
Comment 4 Philippe Normand 2011-11-21 07:15:53 PST
Comment on attachment 116073 [details]
Patch

Looks sane to me!
Comment 5 Mario Sanchez Prada 2011-11-22 00:30:57 PST
Committed r100979: <http://trac.webkit.org/changeset/100979>