-1 to forcing the use of refs/for/x
The validation results are unreliable as the aggregation build is not safely reproducible. A failed validation result is just as likely to be due to someone changing their already-contributed repository.
Thanks,
- Konstantin
On 01/07/2016 03:34 PM, David M Williams wrote:
It is already 'required' that contributors "go though Gerrit" ... but it is allowed that the 'review/validation' can be skipped, if "refs/heads/master" used instead of "refs/for/master".
Right, by enforcing Gerrit, I was meaning "enforcing review".
But I myself would not like to see it *required* to go through "refs/for/master".
Why so? Have you tried using it?
I think most already do go through "refs/for/master" and the few times they do not
It's about half-half: http://git.eclipse.org/c/simrel/org.eclipse.simrel.build.git/log/ . All changes that don't have a "refs/changes/..." tag where pushed directly to master, without review and preliminary validation.
I would assume they have a good reason for it.
I would assume it's more that they need to be educated/encouraged/forced to use Gerrit.
Can you point out (or monitor for) cases where people go directly to "refs/heads/master" and it causes problems?
First, there are all failing builds that pre-dates Gerrit usage ;)
I have troubles to identify where to get a history of the SimRel builds and to associate it with the actual commits that were involved. The CI jobs on http://hudson.eclipse.org/simrel do not show easy to consume data. Is there somewhere else I can look at to first get a list of recent-ish failure of simrel for Neon?
--
Mickael Istria
Eclipse developer at JBoss, by Red Hat
My blog - My Tweets