Bug 29985 – Make LayoutTests/editing/selection/hit-test-anonymous.html resistant to GDI/CG differences

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Julie Parent]

This test measures the offsetLeft of an element, adds 200, and then clicks in that position, and expects the click to happen at a specific character (specifically, for the selection offset to be at a certain character). However, the character occurs at a different x position when using GDI or CG (off by 8 pixels), so this test fails for platforms using GDI.

Possible solutions:
1. Add in a fudge factor, documented with an explanation, for the selection offset, aka, it can be either character at offset 24 or 25.
2. Make the click more specific about where it occurs. Since we can't easily measure where a character is, we could wrap it in an inline tag, measure the location of the inline tag, and specifically click there, rather than just adding 200. This shouldn't effect the quality of the test.
3. GDI platforms just expect failure here.

Thoughts?  Other solutions?  I lean towards #2.

cc'ing Darin since he authored the test in https://bugs.webkit.org/show_bug.cgi?id=24726.

[Darin Adler]

We should just make the font size larger and tweak the numeric offset and change the expected result to match. We're just getting unlucky here that the slight differences in fonts are giving us different results. With a larger font I think we'd be fine.

[Julie Parent]

Created attachment 40484 [details]
Updates the test to use a larger font size and updates the offsets accordingly.

[Julie Parent]

Created attachment 40485 [details]
Fixes whitespace changes.

[WebKit Commit Bot]

Comment on attachment 40485 [details]
Fixes whitespace changes.

Rejecting patch 40485 from commit-queue.

[e-mail] does not have committer permissions according to http://trac.webkit.org/browser/trunk/WebKitTools/Scripts/modules/committers.py.

[WebKit Commit Bot]

Comment on attachment 40485 [details]
Fixes whitespace changes.

Rejecting patch 40485 from commit-queue.

[e-mail] does not have committer permissions according to http://trac.webkit.org/browser/trunk/WebKitTools/Scripts/modules/committers.py.

[Julie Parent]

Sorry for all the commit queue noise here.  I just realized that my bugs.webkit.org account is with my @google address and my committer account is with @chromium, so I can't set my own commit queue bit.  I asked Ojan to set it for me since his accounts match, but I think the commit-queue is using a stale copy of the committers list, since I had just updated it ~30 mins ago to include Ojan and myself.

Will try again once queue syncs that file.

[Eric Seidel (no email)]

The commit-queue is not smart enough to update the committers list before checking the list.  I should make it do so.  Right now it updates it when landing a patch though, so you have to wait for anohter patch to land via the queue after you add yourself to committers.py.  I'll file a bug and fix it. :( 

The email address in committers.py should be your *bugzilla* email, not your committer email.  committers.py attempts to document this, but doesn't do a very good job.

[Eric Seidel (no email)]

Actually, I'm not sure my previous assessment was correct.  It's possible the queue will cache the committers list even across updates...  Filed bug 30084.

[Eric Seidel (no email)]

Comment on attachment 40485 [details]
Fixes whitespace changes.

Julie is busy as Chromium Sheriff today, so she asked that I set this for her.

[WebKit Commit Bot]

Comment on attachment 40485 [details]
Fixes whitespace changes.

Rejecting patch 40485 from commit-queue.

Failed to run "['WebKitTools/Scripts/run-webkit-tests', '--no-launch-safari', '--quiet', '--exit-after-n-failures=1']" exit_code: 1
Running build-dumprendertree
Running tests from /Users/eseidel/Projects/CommitQueue/LayoutTests
Testing 11378 test cases.
fast/dom/prototype-inheritance.html -> failed

Exiting early after 1 failures. 5487 tests run.
90.74s total testing time

5486 test cases (99%) succeeded
1 test case (<1%) had incorrect layout
1 test case (<1%) had stderr output

[Eric Seidel (no email)]

Comment on attachment 40485 [details]
Fixes whitespace changes.

Hyatt broke the tree, and the bots had not caught up yet.

--- /tmp/layout-test-results/fast/dom/prototype-inheritance-expected.txt        2009-10-05 14:23:37.000000000 -0700
+++ /tmp/layout-test-results/fast/dom/prototype-inheritance-actual.txt  2009-10-05 14:23:37.000000000 -0700
@@ -7,6 +7,8 @@
 PASS inner.Attr.constructor.isInner is true
 PASS inner.Audio.isInner is true
 PASS inner.Audio.constructor.isInner is true
+PASS inner.BeforeLoadEvent.isInner is true
+PASS inner.BeforeLoadEvent.constructor.isInner is true
 PASS inner.CDATASection.isInner is true
 PASS inner.CDATASection.constructor.isInner is true
 PASS inner.CSSCharsetRule.isInner is true

[WebKit Commit Bot]

Comment on attachment 40485 [details]
Fixes whitespace changes.

Rejecting patch 40485 from commit-queue.

Failed to run "['WebKitTools/Scripts/run-webkit-tests', '--no-launch-safari', '--quiet', '--exit-after-n-failures=1']" exit_code: 1
Running build-dumprendertree
Running tests from /Users/eseidel/Projects/CommitQueue/LayoutTests
Testing 11378 test cases.
fast/dom/prototype-inheritance.html -> failed

Exiting early after 1 failures. 5487 tests run.
101.12s total testing time

5486 test cases (99%) succeeded
1 test case (<1%) had incorrect layout
1 test case (<1%) had stderr output

[Eric Seidel (no email)]

Comment on attachment 40485 [details]
Fixes whitespace changes.

The buildbots were behind.  bugzilla-tool checked them, saw they were green, but they shouldn't have been.  This was not your fault.

[Eric Seidel (no email)]

False rejection was caused by bug 30098.

[WebKit Commit Bot]

Comment on attachment 40485 [details]
Fixes whitespace changes.

Clearing flags on attachment: 40485

Committed r49127: <http://trac.webkit.org/changeset/49127>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.