Don't call create-patch twice during post-diff. It's slow.
Created attachment 45421 [details] Patch
Created attachment 45422 [details] Patch
style-queue ran check-webkit-style on attachment 45422 [details] without any errors.
Comment on attachment 45421 [details] Patch r=me
Comment on attachment 45422 [details] Patch again - r=me
http://trac.webkit.org/changeset/52523
Comment on attachment 45422 [details] Patch I wonder if we shouldn't cache this off of scm instead? Not sure. Manually storing/retrieving it from state seems odd.
We really need some sort of test for these commands to validate that they don't call scm too much...
There's a related patch for land-diff that tries caching in SCM. It's unclear which is better. I think it depends on what contract you think SCM is enforcing.
Created attachment 45520 [details] Patch
style-queue ran check-webkit-style on attachment 45520 [details] without any errors.
Comment on attachment 45520 [details] Patch I think you might have posted this patch to the wrong bug...
Created attachment 45536 [details] Patch
Comment on attachment 45536 [details] Patch My fault, I still had some testing changes in my tree.