$ curl 'https://gerrit-review.googlesource.com/changes/?q=I18b391f5c858b8608745435d465a63596b3eccac&n=25&O=3' -H 'accept: application/json' -H 'cache-control: no-cache' --compressed | tail -1 | jq '.'
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 355 0 355 0 0 1661 0 --:--:-- --:--:-- --:--:-- 1666
[
{
"id": "gerrit~master~I18b391f5c858b8608745435d465a63596b3eccac",
"project": "gerrit",
"branch": "master",
"hashtags": [],
"change_id": "I18b391f5c858b8608745435d465a63596b3eccac",
"subject": "Add RepoOnlyOps into BatchUpdate",
"status": "NEW",
"created": "2016-05-31 21:40:59.000000000",
"updated": "2016-06-17 01:13:30.000000000",
"submit_type": "MERGE_IF_NECESSARY",
"mergeable": false,
"submittable": false,
"insertions": 155,
"deletions": 1,
"_number": 78186,
"owner": {
"_account_id": 1026258
},
"labels": {
"Verified": {},
"Code-Review": {}
},
"revisions": {}
}
]
How is that possible? It never happened before :-O
Can this be a side-effect of the NoteDB roll-out?
P.S. This condition caused the Gerrit CI to blow-up as he did not what to build. Will push a change on the gerrit-ci-scripts to discard the change and display a warning in the meantime.
Luca.
--
--
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en
---
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
I see, this is a kind of tricky thing as this change has no current patch-sets and it *is correct* .... but it seems that this situation breaks the Gerrit UX as well.I honestly never pushed a draft patch-set to a published change and neither I thought that I could actually then "unpublish" partially a change in this way.If you open https://gerrit-review.googlesource.com/#/c/78186/ you get a nice 500 :-OThe REST API that are failing with a 500 are:Can this be related to the same condition?Is this really a fully supported use-case?
On 17 Jun 2016, at 15:47, Edwin Kempin <eke...@google.com> wrote:On Fri, Jun 17, 2016 at 4:35 PM, lucamilanesio <luca.mi...@gmail.com> wrote:I see, this is a kind of tricky thing as this change has no current patch-sets and it *is correct* .... but it seems that this situation breaks the Gerrit UX as well.I honestly never pushed a draft patch-set to a published change and neither I thought that I could actually then "unpublish" partially a change in this way.If you open https://gerrit-review.googlesource.com/#/c/78186/ you get a nice 500 :-OThe REST API that are failing with a 500 are:Can this be related to the same condition?Is this really a fully supported use-case?Draft patch sets on published changes are pretty broken. It's better to use change edits in this case.
I see, this is a kind of tricky thing as this change has no current patch-sets and it *is correct* .... but it seems that this situation breaks the Gerrit UX as well.I honestly never pushed a draft patch-set to a published change and neither I thought that I could actually then "unpublish" partially a change in this way.If you open https://gerrit-review.googlesource.com/#/c/78186/ you get a nice 500 :-OThe REST API that are failing with a 500 are:Can this be related to the same condition?
Is this really a fully supported use-case?
+czhen thanks for showing us he interesting use cases :)
Yeah the draft patch sets are pretty broken