> Can we point to a commit before Yetus-994 instead of master and get away with
> this?
I'm fine with that approach not to breaking more.
Filed https://issues.apache.org/jira/browse/HADOOP-17297
Thanks,
Akira
On Mon, Oct 5, 2020 at 10:18 AM Akira Ajisaka wrote:
>
> Hi Mastake and Ayush,
>
>
Hi Mastake and Ayush,
After YETUS-994, Yetus updates the commit status instead of writing a comment.
Now I think Yetus still fail to update the commit status, however, now
Jenkins can successfully update the commit status instead and the
result is shown in the link.
Example:
Hi Akira,
Thanx for working on this.
Is there anything blocking here? I think it isn’t working still.
Can we point to a commit before Yetus-994 instead of master and get away with
this? or if there is no solution handy we may go ahead with approach #4 and
then figure out how to proceed.
-Ayush
Hi Akira,
Thanks for working on this.
It looks like you are trying token on PR #2348.
https://github.com/apache/hadoop/pull/2348
It it does not work, I think option #4 is reasonable.
I prefer using fixed/static version of toolchain in order to
avoid surprising the developers.
We can upgrade
I've got a GitHub App token from the infra team and set this to the
hadoop-multibranch job.
The commit statuses in GitHub pull requests will be updated by Jenkins
and you'll need to check these statuses.
-Akira
On Tue, Sep 29, 2020 at 2:55 AM Akira Ajisaka wrote:
>
> Apache Hadoop developers,
>
Apache Hadoop developers,
After YETUS-994, the jenkins job updates the commit status instead of
writing a comment to the pull request. It requires an OAuth token with
write access to "repo:status" but now there is no such token for
Apache Hadoop. I asked the infra team to create a token:
