On 6/13/19 6:59 PM, Adam Williamson wrote:
On Thu, 2019-06-13 at 13:25 +0300, Panu Matilainen wrote:
> On 6/13/19 12:54 PM, Miroslav Suchý wrote:
>> Dne 13. 06. 19 v 11:43 Peter Lemenkov napsal(a):
>>> Hello All!
>>> I've noticed that I cannot build Elixir in Rawhide anymore. It got
>>> stuck at tests and all I've got is a cryptic (at least to me) message:
>>>
>>>
>>> + RPM_EC=0
>>> BUILDSTDERR: ++ jobs -p
>>> + exit 0
>>>
>>>
>>> See this link for full build log:
>>>
>>> *
https://kojipkgs.fedoraproject.org//work/tasks/8021/35518021/build.log
>>> *
https://koji.fedoraproject.org/koji/taskinfo?taskID=35517975
>>>
>>> For comparison here is how successful build log for F-30 looks like
>>> (the same package)
>>>
>>> *
https://kojipkgs.fedoraproject.org//work/tasks/8987/35518987/build.log
>>> *
https://koji.fedoraproject.org/koji/taskinfo?taskID=35518984
>>>
>>> Are there any differences between Koji settings for Rawhide and F-30
>>> which we should know about? Selinux, resource constraints etc?
>>>
>>
>> This is wrong (not sure if the culprit)
>>
>> %endif %{__with_rebar3}
>>
>> I would rewrite it to:
>>
>> %endif # __with_rebar3
>
> Actually both are wrong, and rpm >= 4.15 will complain (unlike old
> versions). Rpm only supports comments at beginning of line, and this
> only ever worked by accident.
Oh dear, that's unfortunate. Because of the lack of indentation,
parsing a nested set of %ifs in spec files has always been difficult,
and adding an end-of-line comment after the %endif to help is
definitely a pattern I've seen in multiple spec files.
I know. While the new warning catched several severely broken usages
that actually have side-effects, outlawing innocent comments feels a bit
like throwing the baby out with the bathwater in the name of consistency.
It is just a warning though, so there's no need to go mass-removing
those comments. We might want to revisit that issue, one way or the other.
- Panu -