|Summary:||Improve layout test result display|
|Component:||Tools / Tests||Assignee:||Nobody <firstname.lastname@example.org>|
|Severity:||Enhancement||CC:||email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org|
|OS:||Mac OS X 10.4|
The HTML output of run-webkit-tests, used to examine test failures, can be improved in several ways. I am attaching a very crude mockup of what the display of a single test failure might look like.
Created an attachment (id=12005) [details] Mockup (very rough draft)
Some notes on the mockup: the slider is intended to control an SVG filter applied to the difference image to enhance small differences (basically a linear transformation). In fact, I think it might be possible to not save the difference image as a file at all and generate it on the fly using SVG. Also, since the width is 800px, there's plenty of room for a sidebar with the list of all failing tests.
Oh, and while I don't find it particularly useful, the "alternate between actual and expected images" functionality could be preserved, activated say by clicking a small icon between the actual and expected tabs. It would also be nice to have the text diff colorized.
(In reply to comment #1) > Created an attachment (id=12005)  [details] > Mockup (very rough draft) None of the expected, actual or diff images show up when I open the .webarchive, although I've confirmed that they're all in file. Not sure why, though. If I convert the .webarchive to a "web page, complete" format, it seems to work fine.
Created an attachment (id=12007) [details] Mockup Recreated Converted Attachment 12005 [details] to a "web page, complete" format using a Perl script I've been working on, then converted it back to a .webarchive. Works for me even after the "original" files have been removed.
(In reply to comment #4) > (In reply to comment #1) > > Created an attachment (id=12005)  [details] > > Mockup (very rough draft) > > None of the expected, actual or diff images show up when I open the > .webarchive, although I've confirmed that they're all in file. Not sure why, > though. If I convert the .webarchive to a "web page, complete" format, it > seems to work fine. This is probably a bug when reloading the .webarchive, although it would be useful to have the original sources to make it reproducible.