[Popcon-developers] Bug#709892: Experimental popcon package
- Subject: [Popcon-developers] Bug#709892: Experimental popcon package
- From: email@example.com (Bill Allombert)
- Date: Mon, 8 Jun 2015 00:20:17 +0200
- Message-id: <[🔎] 20150607222017.GA5722@yellowpig>
- In-reply-to: <20140729212548.GA16412@yellowpig>
- References: <20140725141653.GC21723@yellowpig> <20140729191304.GA7206@doriath.local> <20140729212548.GA16412@yellowpig>
On Tue, Jul 29, 2014 at 11:25:49PM +0200, Bill Allombert wrote:
> On Tue, Jul 29, 2014 at 09:13:04PM +0200, Ernest Adrogu? wrote:
> > Hi,
> > 2014-07-25, 16:16 (+0200); Bill Allombert escriu:
> > > I have made a preliminary package available at:
> > > <http://people.debian.org/~ballombe/popcon/debs/popularity-contest_1.62~pre1_all.deb>
> > >
> > > I would be extremly glad if you could experiment with it especially if you are
> > > affected by bug #709892. This kind of change is hard to test, and we have little time
> > > before the freeze.
> > I installed your experimental package yesterday. So far, it works.
> > However, because the non-random submission is attempted before the other
> > one, I still get a failure to submit report:
> > # grep popularity /var/log/syslog
> > Jul 29 07:36:08 doriath popularity-contest: unable to submit report to http://popcon.debian.org/cgi-bin/popcon.cgi.
> > Jul 29 15:49:01 doriath /USR/SBIN/CRON: (root) CMD ( test -x /etc/cron.daily/popularity-contest && /etc/cron.daily/popularity-contest)
> > The 15:49 attempt (from cron.d) did succeed.
> Thanks for confirming that.
> > If I understand correctly this failure will only occur once, as 7 days from
> > Jul 29 07:36:08 will be less than a week since the latest submission and the
> > script will do nothing, right?
> Unfortunately the delay is actually of 6.5 days currently.
> I agree that loging a failure every week might be annoying.
> Not sure how to avoid that.
I will try to move forward with this.
The remaining issues are:
1) the cron.d script might run after the cron.daily has already failed.
2) there is no way to detect that the SMTP submission succeeded.
Fixing 2) does not seem possible.
However fixing 1) should be easy: set a slightly longer delay for
cron.daily than for cron.d.
To mitigate 2) we can decide that only failure from cron.d trigger the
I have made a new test package:
Please test, and sorry for the delay. It is difficult to experiment with popcon.
Bill. <ballombe at debian.org>
Imagine a large red swirl here.