Bug 56450

Summary:

Need to do an experiment to make sure that reftests don't cause any issues on all buildbots

Product:

WebKit

Reporter:

Hayato Ito <hayato>

Component:

Tools / Tests

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

dpranke, hamaji, ojan

Priority:

Version:

528+ (Nightly build)

Hardware:

OS:

OS X 10.5

Bug Depends on:

Bug Blocks:

36065

Attachments:

Description	Flags
the-first-reftest	none
move-to-harness	ojan: review+

Hayato Ito

Reported 2011-03-15 23:18:23 PDT

I'll submit the first reftest after all necessary changes are landed.

Attachments
the-first-reftest (1.62 KB, patch) 2011-03-15 23:20 PDT, Hayato Ito	no flags	Details Formatted Diff Diff
move-to-harness (1.65 KB, patch) 2011-03-20 21:04 PDT, Hayato Ito	ojan: review+	Details Formatted Diff Diff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.

Hayato Ito

Comment 1 2011-03-15 23:20:14 PDT

Created attachment 85913 [details] the-first-reftest

Hayato Ito

Comment 2 2011-03-15 23:21:15 PDT

I've checked it as 'r?', but I won't submit this patch soon until all necessary changes are landed.

Eric Seidel (no email)

Comment 3 2011-03-17 04:29:57 PDT

Comment on attachment 85913 [details] the-first-reftest What identifies this as a reftest? expected.html I assume?

Hayato Ito

Comment 4 2011-03-17 04:39:39 PDT

Yeah, we use a naming convention to identify. Please see a master bug to know how and why: https://bugs.webkit.org/show_bug.cgi?id=36065 And the Wiki page I started to write is here: http://trac.webkit.org/wiki/Writing%20Reftests (In reply to comment #3) > (From update of attachment 85913 [details]) > What identifies this as a reftest? expected.html I assume?

Eric Seidel (no email)

Comment 5 2011-03-18 14:34:17 PDT

Comment on attachment 85913 [details] the-first-reftest Seems this should go into the harness directory (which is about testing DRT/RWT) since this test doesn't really add much to our CSS testing.

Hayato Ito

Comment 6 2011-03-18 19:17:02 PDT

This is an experiment reftest, so the test itself doesn't matter. If you think this test doesn't add much to our css testing, I am happy to revert it after we confirm all buildbots are happy with this reftest. We decided to use LayoutTest directory to where '-expected.html' is added as other existing LayoutTests do. So '-expected.html' must go into LayoutTests directory to do an experiment. (In reply to comment #5) > (From update of attachment 85913 [details]) > Seems this should go into the harness directory (which is about testing DRT/RWT) since this test doesn't really add much to our CSS testing.

Ojan Vafai

Comment 7 2011-03-20 15:24:29 PDT

I think Eric was just saying that the test should go in fast/harness instead of fast/css.

Hayato Ito

Comment 8 2011-03-20 21:04:14 PDT

Created attachment 86293 [details] move-to-harness

Hayato Ito

Comment 9 2011-03-20 21:08:29 PDT

Hi Ojan, thank you for clarifying that. Although I am not 100% sure what the purpose of the 'fast/harness' directory is, I've moved the test to 'fast/harness' directory. Could you review it? (In reply to comment #7) > I think Eric was just saying that the test should go in fast/harness instead of fast/css.

Hayato Ito

Comment 10 2011-03-21 22:00:03 PDT

Committed r81644: <http://trac.webkit.org/changeset/81644>

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