We would like to collect failures with one set of EWS run as much as possible. Once one of test puts merge-blocked label, then the other EWS tasks in the same set of EWS stops, which enforces us to upload a patch repeatedly to collect failures on each platform, which is not efficient and regression from the current form of EWS.
I'm now getting it in https://github.com/WebKit/WebKit/pull/420 too, which EWS tests are blocked by mips compilation failure.
I think merge-blocked label should not affect on EWS test runs since we didn't have such a thing in bugzilla. We can stop EWS if a new change is pushed in PR (as we did in bugzilla), but if it is not, I don't think we should stop EWS testings.
<rdar://problem/92752163>
Apologies for the forward dupe, this has been fixed. *** This bug has been marked as a duplicate of bug 240169 ***