new-run-webkit-tests output is different from (and worse than) original run-webkit-tests. The format is arbitrarily different, seems less conventiant, and has apparently broken links to wdiff output.
Created attachment 53584 [details] Original run-webkit-tests results
Created attachment 53585 [details] new-run-webkit tests results for same set of tests, same failures
We should remove the wdiff links now that we have pretty patch wired up.
(In reply to comment #3) > We should remove the wdiff links now that we have pretty patch wired up. So long as Chromium is OK with wdiff links being replaced by pretty diff links, that seems fine to me.
(In reply to comment #4) > (In reply to comment #3) > > We should remove the wdiff links now that we have pretty patch wired up. > > So long as Chromium is OK with wdiff links being replaced by pretty diff links, > that seems fine to me. I think wdiff is a little bit easier to read in the case of font only differences, but not so much that I think we should keep it. Instead, we should try to make pretty patch handle font only differences better. I'll upload some sample diffs to demonstrate.
Created attachment 53665 [details] pretty patch diff
Created attachment 53666 [details] wdiff or the same 2 files
CCing Adam Roben, the original PrettyPatch author.
That's really nice btw. I'm not sure we really want to kill wdiff. We shouldn't show the "wdiff" link on platforms which don't have wdiff installed however. :)
(In reply to comment #9) > That's really nice btw. I'm not sure we really want to kill wdiff. We > shouldn't show the "wdiff" link on platforms which don't have wdiff installed > however. :) Ideally we'd use wdiff it's available and pretty-patch otherwise, since wdiff output is a bit better. I don't think there are any cases where you actually want both.
(In reply to comment #10) > (In reply to comment #9) > > That's really nice btw. I'm not sure we really want to kill wdiff. We > > shouldn't show the "wdiff" link on platforms which don't have wdiff installed > > however. :) > > Ideally we'd use wdiff it's available and pretty-patch otherwise, since wdiff > output is a bit better. I don't think there are any cases where you actually > want both. We could change PrettyPatch to create wdiff-like output in the cases where it currently shows intra-line diffs. Maybe we'd want this to be a special mode for PrettyPatch, though, as there are some cases where I think that would be undesirable (e.g., when changing a comment in C++ code and re-wrapping the comment lines).
Ojan, I'm punting this one to you for now. Feel free to punt it back if you don't want it.
Created attachment 90253 [details] the new results.html format for new-run-webkit-tests
As of r84294, I think we can call this resolved. Please let me know (or file new bugs) if there are still concerns with the output format.