DryrunTest fails on every platform other than mac
Windows can be fixed. But linux can't. We should re-write the test, or change how dryrun works to not assume Chromium in the linux case. For now I'm going to disable the test on non-mac to make the bots green.
Created attachment 55461 [details] Patch
Comment on attachment 55461 [details] Patch LGTM, r=me.
Committed r58999: <http://trac.webkit.org/changeset/58999>
Comment on attachment 55461 [details] Patch webkit-patch land --no-close should have obsoleted this patch. Something must have regressed in webkit-patch.
Created attachment 55623 [details] Patch
I re-wrote the test and fixed what was breaking on chromium windows. If we want to test more ports we can add them now.
Comment on attachment 55623 [details] Patch LGTM. Why does the mac port need to be run dryrun on mac only?
Grr. Actually, this patch is bad, because it'll still try to run the chromium tests in a non-chromium checkout and fail.
Created attachment 55637 [details] Patch
(In reply to comment #8) > (From update of attachment 55623 [details]) > LGTM. Why does the mac port need to be run dryrun on mac only? If you run it on windows, the mac port tries to pull in server_process which pulls in fcntl, which doesn't exist.
It seems like there should be a better way to get coverage of the platforms other than the one you're running on, but I'm not sure what it is. Also, for platforms with multiple implementations (e.g., mac, chromium-mac) we need a way to figure out which is which.
Comment on attachment 55637 [details] Patch We should add a comment in the code as to why mac is run on mac-only. Your explanation seemed to suggest it should be run on everything except for "win32". Also: + if sys.platform not in 'mac': Why change the quotes? For better or worse our python guidelines in the wiki suggest " over ': http://trac.webkit.org/wiki/PythonGuidelines
Created attachment 55640 [details] Patch
(In reply to comment #13) > (From update of attachment 55637 [details]) > We should add a comment in the code as to why mac is run on mac-only. Your explanation seemed to suggest it should be run on everything except for "win32". > > Also: > > + if sys.platform not in 'mac': > > Why change the quotes? For better or worse our python guidelines in the wiki suggest " over ': http://trac.webkit.org/wiki/PythonGuidelines That was not intentional. Fixed.
Comment on attachment 55640 [details] Patch LGTM. Seems we still could make the check be != "win32" no?
Comment on attachment 55640 [details] Patch Oh, and I should have fixed: sys.platform not in "mac" that shoudl read "sys.platform != "mac"" not in happens to work but is not as intended.
(In reply to comment #17) > (From update of attachment 55640 [details]) > Oh, and I should have fixed: > sys.platform not in "mac" > > that shoudl read "sys.platform != "mac"" > > not in happens to work but is not as intended. I was wondering about that. As to the != 'win32', I dunno; I haven't tested this on non-chromium linux. I think it'll work.
Committed r59182: <http://trac.webkit.org/changeset/59182>
http://trac.webkit.org/changeset/59182 might have broken GTK Linux 32-bit Debug The following changes are on the blame list: http://trac.webkit.org/changeset/59182 http://trac.webkit.org/changeset/59183