WebKit Bugzilla
New
Browse
Search+
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
[patch]
Patch
bug-69444-20120215200017.patch (text/plain), 4.21 KB, created by
Dirk Pranke
on 2012-02-15 20:00:18 PST
(
hide
)
Description:
Patch
Filename:
MIME Type:
Creator:
Dirk Pranke
Created:
2012-02-15 20:00:18 PST
Size:
4.21 KB
patch
obsolete
>Subversion Revision: 107867 >diff --git a/Tools/ChangeLog b/Tools/ChangeLog >index f85eee7afe9ed179e3427388513c75fc0d1ea691..a0308abf322c2dca734db7d7f663268c78ef27a6 100644 >--- a/Tools/ChangeLog >+++ b/Tools/ChangeLog >@@ -1,5 +1,25 @@ > 2012-02-15 Dirk Pranke <dpranke@chromium.org> > >+ The new run-webkit-tests needs to dump out pixel hash failures even if the pixel test passes. >+ https://bugs.webkit.org/show_bug.cgi?id=69444 >+ >+ Reviewed by NOBODY (OOPS!). >+ >+ Currently if a test's image checksum doesn't match the checksum >+ in the baseline, but the images themselves match according to >+ ImageDiff, new-run-webkit-tests ignores the problem. This is >+ probably bad, but it's not yet clear what the right thing to do >+ is. This patch will log a warning to stderr, at least (but the >+ test will still pass). >+ >+ * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: >+ (SingleTestRunner._compare_image): >+ (SingleTestRunner._run_reftest): >+ * Scripts/webkitpy/layout_tests/port/test.py: >+ (unit_test_list): >+ >+2012-02-15 Dirk Pranke <dpranke@chromium.org> >+ > webkitpy: add a worker_args concept to start_worker() > https://bugs.webkit.org/show_bug.cgi?id=78572 > >diff --git a/Tools/Scripts/webkitpy/layout_tests/controllers/single_test_runner.py b/Tools/Scripts/webkitpy/layout_tests/controllers/single_test_runner.py >index 89ae5bdcaea945e8cd2f7e93af766b4ec27940d3..c903cb3494a20e425ddbb03b9f0585c0953df69a 100644 >--- a/Tools/Scripts/webkitpy/layout_tests/controllers/single_test_runner.py >+++ b/Tools/Scripts/webkitpy/layout_tests/controllers/single_test_runner.py >@@ -264,6 +264,10 @@ class SingleTestRunner: > driver_output.image_diff = diff_result[0] > if driver_output.image_diff: > failures.append(test_failures.FailureImageHashMismatch(diff_result[1])) >+ else: >+ # See https://bugs.webkit.org/show_bug.cgi?id=69444 for why this isn't a full failure. >+ _log.warning('%s had mismatched checksums but matching images (exp %s, got %s)' % >+ (self._test_name, expected_driver_output.image_hash, driver_output.image_hash)) > return failures > > def _run_reftest(self): >@@ -272,8 +276,8 @@ class SingleTestRunner: > reference_output = None > test_result = None > >- # A reftest can have multiple match references and multiple mismatch references; >- # the test fails if any mismatch matches and all of the matches don't match. >+ # A reftest can have multiple match references and multiple mismatch references; >+ # the test fails if any mismatch matches and all of the matches don't match. > # To minimize the number of references we have to check, we run all of the mismatches first, > # then the matches, and short-circuit out as soon as we can. > # Note that sorting by the expectation sorts "!=" before "==" so this is easy to do. >diff --git a/Tools/Scripts/webkitpy/layout_tests/port/test.py b/Tools/Scripts/webkitpy/layout_tests/port/test.py >index b3aecfd2bd368286296bc439427c742387ee6d0f..3ddac1801b45831d3942e89bcf7ade960343565b 100644 >--- a/Tools/Scripts/webkitpy/layout_tests/port/test.py >+++ b/Tools/Scripts/webkitpy/layout_tests/port/test.py >@@ -100,8 +100,6 @@ class TestList(object): > > def unit_test_list(): > tests = TestList() >- tests.add('failures/expected/checksum.html', >- actual_checksum='checksum_fail-checksum') > tests.add('failures/expected/crash.html', crash=True) > tests.add('failures/expected/exception.html', exception=True) > tests.add('failures/expected/timeout.html', timeout=True) >@@ -171,6 +169,10 @@ layer at (0,0) size 800x34 > expected_checksum=None, > expected_image='tEXtchecksum\x00checksum_in_image-checksum') > >+ # Note that here the checksums don't match but the images do, so this test passes. >+ # See https://bugs.webkit.org/show_bug.cgi?id=69444 . >+ tests.add('passes/checksum.html', actual_checksum='checksum_fail-checksum') >+ > # Text output files contain "\r\n" on Windows. This may be > # helpfully filtered to "\r\r\n" by our Python/Cygwin tooling. > tests.add('passes/text.html',
You cannot view the attachment while viewing its details because your browser does not support IFRAMEs.
View the attachment on a separate page
.
View Attachment As Diff
View Attachment As Raw
Actions:
View
|
Formatted Diff
|
Diff
Attachments on
bug 69444
:
127303
|
127305
|
127306
|
127506