[zeromq-dev] Builds failing
Arnaud Loonstra
arnaud at sphaero.org
Sat Feb 6 13:38:51 CET 2016
I agree with the reasoning of Pieter but a bit of common sense would be
in place I guess (and thanks about the city, I live there ;) ).
We welcome all patches but when a PR is merged within seconds why not
wait for a minute to let Travis finish. This is especially the case with
concurrent merges and the inter dependencies of projects. Setting up
Travis for your own branch doesn't prevent these situations if people
don't do that.
I think this also prevents running in circles. From what I get from the
failing czmq build is that it is the same problem we had at the
hackathon. This was fixed. Why did it reappear? The problem we have with
the failing czmq test.py was the same problem we had before which was
also fixed. How come it has returned? I've been more busy chasing
failing builds than adding new functionalities and it gets harder when I
have to dig in the history to find when the build was last successful.
So my 2 cents would be to let yourself be informed before clicking the
merge button, its just a matter between seconds and minutes anyway.
Rg,
Arnaud
On 2016-02-05 19:06, Constantin Rack wrote:
> See previous discussion here:
>
> http://lists.zeromq.org/pipermail/zeromq-dev/2015-November/029814.html
> [5]
>
> In summary: we usually do not wait for CI to finish before merge.
>
> If you want submit only passing pull requests, you could enable
> Travis-CI
> for your fork repository and let it check there before submitting the
> PR.
>
> -Constantin
>
> Am 05.02.2016 um 18:47 schrieb Osiris Pedroso:
>
>> I second this request.
>> I would rather have a clean build that new functionality being added
>> that may break the build it even further.
>>
>> Best regards,
>> Osiris
>>
>> On Fri, Feb 5, 2016 at 10:33 AM Arnaud Loonstra <arnaud at sphaero.org
>> [3]> wrote:
>>
>>> Hi all,
>>>
>>> I was trying to get the zyre CI builds to run successfully again.
>>> It
>>> failed I think because of a memory error in czmq.
>>>
>>> I was wondering what policy do we use for pull requests while the
>>> CI
>>> test hasn't finished yet. I personally like to refrain from
>>> merging
>>> until all tests have finished. Any other perspectives?
>>>
>>> Rg,
>>>
>>> Arnaud
>>> _______________________________________________
>>> zeromq-dev mailing list
>>> zeromq-dev at lists.zeromq.org [1]
>>> http://lists.zeromq.org/mailman/listinfo/zeromq-dev [2]
>> _______________________________________________
>> zeromq-dev mailing list
>> zeromq-dev at lists.zeromq.org [4]
>> http://lists.zeromq.org/mailman/listinfo/zeromq-dev
>
>
>
> Links:
> ------
> [1] mailto:zeromq-dev at lists.zeromq.org
> [2] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
> [3] mailto:arnaud at sphaero.org
> [4] mailto:zeromq-dev at lists.zeromq.org
> [5]
> http://lists.zeromq.org/pipermail/zeromq-dev/2015-November/029814.html
More information about the zeromq-dev
mailing list