r110542: New fast/forms/label/labelable-elements.html fails

• Component: Forms
• Status: RESOLVED FIXED

[Ádám Kallai]

This tests fails on Qt and SnowLeopard from revision #110542.

--- /ramdisk/qt-linux-64-release/build/layout-test-results/fast/forms/label/labelable-elements-expected.txt 
+++ /ramdisk/qt-linux-64-release/build/layout-test-results/fast/forms/label/labelable-elements-actual.txt 
@@ -7,7 +7,7 @@
 Labelable: button [object HTMLButtonElement]
 PASS element.labels.length is 1
 
-Unsupported: datalist [object HTMLUnknownElement]
+Not labelable: datalist [object HTMLDataListElement]
 PASS element.labels is undefined.
 PASS element.labels is not null
 
@@ -45,7 +45,7 @@
 PASS element.labels is not null
 
 Labelable: output [object HTMLOutputElement]
-FAIL element.labels.length should be 1. Threw exception TypeError: Cannot read property 'length' of null
+FAIL element.labels.length should be 1. Threw exception TypeError: 'null' is not an object (evaluating 'element.labels.length')
 
 Labelable: progress [object HTMLProgressElement]
 PASS element.labels.length is 1


I will skip it until it is fixed.

[Ádám Kallai]

I skipped this test. 

http://trac.webkit.org/changeset/110558

[Jessie Berlin]

Is there some reason why the test should be skipped? We usually land expected failing results unless this is:

1. A feature that is not turned on for your port.
2. DRT/WKTR support for the test doesn't exist for your port.
3. It is crashing or causing crashes in other tests.

We should land mac-specific expected failing results for this test, since it is failing on Lion as well as SL:

http://build.webkit.org/results/SnowLeopard%20Intel%20Release%20(Tests)/r110576%20(38036)/fast/forms/label/labelable-elements-pretty-diff.html
http://build.webkit.org/results/SnowLeopard%20Intel%20Release%20(WebKit2%20Tests)/r110561%20(19017)/fast/forms/label/labelable-elements-pretty-diff.html
http://build.webkit.org/results/Lion%20Intel%20Debug%20(Tests)/r110578%20(4409)/fast/forms/label/labelable-elements-pretty-diff.html

[Jessie Berlin]

Might be related to / fixed by the fix for https://bugs.webkit.org/show_bug.cgi?id=80466

[Jessie Berlin]

Committed the expected failing results for Mac in http://trac.webkit.org/changeset/110614

[Kent Tamura]

http://trac.webkit.org/changeset/110947 made:

fast/forms/label/labelable-elements-expected.txt: Result for ENABLE_DATALIST platforms
platform/chromium/fast/forms/label/labelable-elements-expected.txt: Result for !ENABLE_DATALIST

Unskipped in platform/qt/Skipped and platform/gtk/Skipped.

[Csaba Osztrogonác]

(In reply to comment #2)
> Is there some reason why the test should be skipped? We usually land expected failing results unless this is:
> 
> 1. A feature that is not turned on for your port.
> 2. DRT/WKTR support for the test doesn't exist for your port.
> 3. It is crashing or causing crashes in other tests.

Historically we don't land expected failing results to platform/qt*
automatically. It is better to skip it to paint the bot green and
let the author check what happened.

It think it is a good way to have a plain text list of failing 
tests instead of zillion never closed bugzilla bug reports.