Hi
Ah! It never occurred to me that "Reply" meant "Commit" / "Push" /
"Merge" or even "Review". I saw only "Cherry Pick" / " Cancel" which did
not match my requirements. What a truly awful UI.
What if I don't want to "Merge" since it leads to double history entries
that on normal projects are difficult to disentangle? Ok, SimRel commits
are highly orthogonal. But Push gives a nice clean history and forces a
Rebase that just occasionally Merge can fail to replicate. IMHO if a
Merge is necessary the Gerrit Review should iterate.
Regards
Ed Willink
On 13/11/2018 05:46, Ed Merks wrote:
Ed,
Normally you would use the Gerrit review link to finish the
processing. E.g., for the last commit I used this link:
https://git.eclipse.org/r/#/c/132049/
After the initial commit, I waited for the build to finish so that CI
Bot (one of the automatic reviewers) adds a +1. That takes about 5
minutes. Then I used the Reply... (or the Review +2 button, which
will be there after the successful build) to make it possible to
"Submit" the changes to master, i.e., the Submit button will be there
once all the reviewers (CI Bot and you) have voted the changes up to
the necessary level.
So the chain of events looks like this in the review:
After submitting, when I do a pull on the repo, my changes are pulled
and the repo is up-to-date (no longer one commit behind master).
I believe a non-dilegent user could remove CI Bot from the review to
submit their changes even when those did not pass the aggregation
build but it appears to me that there really is no good reason to
allow direct push to master as a way to completely bypass the CI Bot
review.
Regards,
Ed
On 12.11.2018 22:04, Ed Willink wrote:
Hi
Sorry I must have missed something. My no doubt flawed recollection
was that we were assured (again) that direct push would continue to
be allowed.
I certainly use it every time since I see no other way to do it.
I push to Gerrit, check for build success, then Push to master.
Given that the bulk of failures are surely due to those updating
magic 'latest' contributions without any commit, why change the rules
for more diligent users?
Regards
Ed Willink
On 12/11/2018 17:34, Frederic Gurr wrote:
Hi,
As discussed before, direct push to master for the SimRel aggregation
build repository
(https://git.eclipse.org/r/simrel/org.eclipse.simrel.build) has been
disabled. So, all future commits have to go through a Gerrit review and
should get a +1 from the CI server before being merged.
The reasoning behind this is, that commits directly pushed to master
caused build failures and delays in the past.
Please let me know, if you have any questions or concerns.
Regards,
Fred
---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus
_______________________________________________
cross-project-issues-dev mailing list
cross-project-issues-dev@eclipse.org
To change your delivery options, retrieve your password, or
unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/cross-project-issues-dev
_______________________________________________
cross-project-issues-dev mailing list
cross-project-issues-dev@eclipse.org
To change your delivery options, retrieve your password, or unsubscribe from
this list, visit
https://www.eclipse.org/mailman/listinfo/cross-project-issues-dev
_______________________________________________
cross-project-issues-dev mailing list
cross-project-issues-dev@eclipse.org
To change your delivery options, retrieve your password, or unsubscribe from
this list, visit
https://www.eclipse.org/mailman/listinfo/cross-project-issues-dev