Bug 120057

Created attachment 209171 [details] Patch

Comment on attachment 209171 [details] Patch Why not a reference test?

(In reply to comment #2) > (From update of attachment 209171 [details]) > Why not a reference test? Sounds good, I'll start working on it.

Created attachment 209990 [details] Patch

Comment on attachment 209990 [details] Patch Attachment 209990 [details] did not pass mac-ews (mac): Output: http://webkit-queues.appspot.com/results/1623620 New failing tests: css3/compositing/background-blend-mode-gif-color.html

Created attachment 209996 [details] Archive of layout-test-results from webkit-ews-08 for mac-mountainlion The attached test failures were seen while running run-webkit-tests on the mac-ews. Bot: webkit-ews-08 Port: mac-mountainlion Platform: Mac OS X 10.8.4

Comment on attachment 209990 [details] Patch Attachment 209990 [details] did not pass mac-wk2-ews (mac-wk2): Output: http://webkit-queues.appspot.com/results/1591981 New failing tests: css3/compositing/background-blend-mode-gif-color.html

Created attachment 210022 [details] Archive of layout-test-results from webkit-ews-13 for mac-mountainlion-wk2 The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews. Bot: webkit-ews-13 Port: mac-mountainlion-wk2 Platform: Mac OS X 10.8.4

Created attachment 210074 [details] Patch

Comment on attachment 210074 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=210074&action=review r=me > LayoutTests/ChangeLog:4 > + Test background blend modes with gif background layers. > + https://bugs.webkit.org/show_bug.cgi?id=120057 Why does it need to be an animated gif? The ChangeLog should be more precise about that. Also, shouldn't animated gif's influence blending as well (when the content of the gif changes)? Is it possible to write tests for that?

Created attachment 210107 [details] Patch

(In reply to comment #10) > (From update of attachment 210074 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=210074&action=review > > r=me > > > LayoutTests/ChangeLog:4 > > + Test background blend modes with gif background layers. > > + https://bugs.webkit.org/show_bug.cgi?id=120057 > > Why does it need to be an animated gif? The ChangeLog should be more precise about that. Also, shouldn't animated gif's influence blending as well (when the content of the gif changes)? Is it possible to write tests for that? I've added another test which validates that blend mode update when the content of the gif changes.

Comment on attachment 210107 [details] Patch Attachment 210107 [details] did not pass mac-ews (mac): Output: http://webkit-queues.appspot.com/results/1633107 New failing tests: css3/compositing/background-blend-mode-gif-color-2.html css3/compositing/background-blend-mode-gif-color.html

Created attachment 210114 [details] Archive of layout-test-results from webkit-ews-07 for mac-mountainlion The attached test failures were seen while running run-webkit-tests on the mac-ews. Bot: webkit-ews-07 Port: mac-mountainlion Platform: Mac OS X 10.8.4

Comment on attachment 210107 [details] Patch Attachment 210107 [details] did not pass mac-wk2-ews (mac-wk2): Output: http://webkit-queues.appspot.com/results/1603054 New failing tests: css3/compositing/background-blend-mode-gif-color-2.html css3/compositing/background-blend-mode-gif-color.html

Created attachment 210145 [details] Archive of layout-test-results from webkit-ews-12 for mac-mountainlion-wk2 The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews. Bot: webkit-ews-12 Port: mac-mountainlion-wk2 Platform: Mac OS X 10.8.4

Created attachment 210278 [details] Patch

Comment on attachment 210278 [details] Patch Attachment 210278 [details] did not pass mac-wk2-ews (mac-wk2): Output: http://webkit-queues.appspot.com/results/1609958 New failing tests: css3/compositing/background-blend-mode-gif-color-2.html css3/compositing/background-blend-mode-gif-color.html

Created attachment 210284 [details] Archive of layout-test-results from webkit-ews-09 for mac-mountainlion-wk2 The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews. Bot: webkit-ews-09 Port: mac-mountainlion-wk2 Platform: Mac OS X 10.8.4

Comment on attachment 210278 [details] Patch Attachment 210278 [details] did not pass mac-ews (mac): Output: http://webkit-queues.appspot.com/results/1688050 New failing tests: css3/compositing/background-blend-mode-gif-color-2.html css3/compositing/background-blend-mode-gif-color.html

Created attachment 210298 [details] Archive of layout-test-results from webkit-ews-08 for mac-mountainlion The attached test failures were seen while running run-webkit-tests on the mac-ews. Bot: webkit-ews-08 Port: mac-mountainlion Platform: Mac OS X 10.8.4

Comment on attachment 210278 [details] Patch Attachment 210278 [details] did not pass mac-ews (mac): Output: http://webkit-queues.appspot.com/results/1688065 New failing tests: css3/compositing/background-blend-mode-gif-color-2.html css3/compositing/background-blend-mode-gif-color.html

Created attachment 210300 [details] Archive of layout-test-results from webkit-ews-02 for mac-mountainlion The attached test failures were seen while running run-webkit-tests on the mac-ews. Bot: webkit-ews-02 Port: mac-mountainlion Platform: Mac OS X 10.8.4

Created attachment 210328 [details] Patch

There were inconsistencies between the local results and the ones from the bots, in other words, my tests would only pass locally. Converting them back to pixel tests.

Comment on attachment 210328 [details] Patch Can we do these as reference tests instead of pixel tests?

(In reply to comment #26) > (From update of attachment 210328 [details]) > Can we do these as reference tests instead of pixel tests? The test scenario implies blending an animated gif with a background color, and comparing it with another div filled with the expected resulted color. I've tried doing this, however, there were inconsistencies between the reference html that passed on my local machine and the bot-machine results. What I suspect is that blending operations output slightly different results across platforms (+/- a few units per color component), making ref tests unusable in this scenario, as I've seen with the previous two versions of this patch. This is the reason I've decided to rework this patch and use pixel tests instead.

Comment on attachment 210328 [details] Patch Attachment 210328 [details] did not pass mac-ews (mac): Output: http://webkit-queues.appspot.com/results/1692190 New failing tests: http/tests/appcache/resource-redirect.html http/tests/appcache/abort-cache-ondownloading-resource-404.html http/tests/appcache/abort-cache-onprogress.html http/tests/appcache/abort-cache-ondownloading.html http/tests/appcache/resource-redirect-2.html http/tests/appcache/interrupted-update.html http/tests/inspector/appcache/appcache-manifest-with-non-existing-file.html http/tests/appcache/404-resource.html

Created attachment 210339 [details] Archive of layout-test-results from webkit-ews-04 for mac-mountainlion The attached test failures were seen while running run-webkit-tests on the mac-ews. Bot: webkit-ews-04 Port: mac-mountainlion Platform: Mac OS X 10.8.4

Created attachment 210456 [details] Patch

Comment on attachment 210456 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=210456&action=review > LayoutTests/ChangeLog:9 > + * css3/compositing/background-blend-mode-gif-color-2.html: Added. > + * css3/compositing/background-blend-mode-gif-color.html: Added. Can we make these reference tests instead of pixel tests?

Comment on attachment 210456 [details] Patch Clearing flags on attachment: 210456 Committed r155097: <http://trac.webkit.org/changeset/155097>

All reviewed patches have been landed. Closing bug.