Re: Automated pkgsrc commit testing

[Taylor R Campbell <campbell+netbsd-tech-pkg%mumble.net@localhost>]

> Date: Mon, 5 Aug 2024 13:36:59 +0100
> From: Jonathan Perkin <jperkin%mnx.io@localhost>
> 
> * On 2024-08-05 at 13:31 BST, Taylor R Campbell wrote:
> 
> >> pbulk-resolve: No match found for dependency py39-scipy-[0-9]* of package py39-qtgraph-qt5-0.12.0nb1
> >> [...]
> >
> >Aha!  Now that is not reported to pkgsrc-bulk@ as far as I know.
> >
> >I used to receive messages with these reports, but the last one I
> >received appears to have been on 2023-08-11, just under a year ago.
> >What happened to them?
> 
> People complained I was being grumpy by sending them, so I stopped 
> sending them publicly.
> 
> It appears I can't win either way.  I should just stop playing.

That's surprising to me, because (a) they essentially always indicate
a real technical problem and (b), unless things are continuously
broken for a long time, they are low-volume.  So unlike the other
daily/weekly packages cron job mail, I always found these reports
useful even if the bugs weren't my doing.

I don't know who complained about these reports, but it strikes me as
much worse to hear vague comments like `severely broken with no bulk
builds running at all due to untested py-scipy changes' that I have no
idea how to substantiate (and led me to spend half an hour chasing a
wild goose down pkgsrc-bulk@) than to hear useful specific actionable
automated alerts.

Was there specific criticism about the reports that we could find some
way to improve the experience for whoever it was who objected?