Re: MAKE_JOBS_SAFE= yes for ham/gnuradio-doxygen

[Roland Illig <roland.illig%gmx.de@localhost>]

Am 05.05.2019 um 21:19 schrieb maya%netbsd.org@localhost:
> I would personally prefer not to have this be a pkglint error.
>
> If I write:
> - Doesn't work: contains no additional information
> - Didn't work in version: obtainable via cvs annotate
> - Doesn't work for <details of wrong dependency>: too much work IMO,
>   like 80% of the way into fixing the parallel jobs issue.
>
> IMO, we shouldn't spend extra effort on solving parallel jobs issues
> because we have a valid quick fix for them, unless someone is
> specifically interested in doing this.

My current draft looks like this:
https://github.com/rillig/pkglint/blob/14963f4c13/mklinechecker.go#L445

In cases where MAKE_JOBS_SAFE is set to "no", I'd like to know:

- What are the symptoms?
- Any specific error messages?
- Has it been reported upstream?

That way, when someone wants to clean up the MAKE_JOBS_SAFE variables
some years later, they have a quick way to see whether something has
changed.

> Maybe if it contained the error it would be something, but I don't think
> this information is that useful. You will need to re-test for parallel
> jobs issues after fixing one issue, because there may be more.

Of course there may be more, but having the information about the one
error that has definitely occured makes it easy to decide if it is worth
retrying now.