Re: Why is package X not in testing yet?
On Mon, Mar 08, 2004 at 10:21:56PM +0100, Richard B. Kreckel wrote:
> On Sat, 6 Mar 2004, Colin Watson wrote:
> > Richard, do you read debian-release? If not, it would be helpful if you
> > had a Mail-Followup-To: header or similar saying so.
>
> In other free software ecosystems it is customary to CC the poster when
> replying. Even if he/she is known to be the list administrator.
Not the case here:
http://www.debian.org/MailingLists/#codeofconduct
I'll try to remember in your case, though, as I'd rather not have the
argument here (it's somewhat long-running).
> On Tue, 2 Mar 2004, Colin Watson wrote:
> > manual attention sometimes. The reason this doesn't show up in
> > update_output is slightly unfortunate: for performance reasons, the
> > testing script only shows problems on the first architecture in
> > alphabetical order, and qalculate was only in testing for i386 and ia64,
> > unlike ginac which was in testing for alpha.
>
> It's not so easy not to comment on an optimization which discards valuable
> information. In fact, the crucial piece of information, in this case.
> This is not the first time people have found the output irritating beyond
> hope because of missing clues. Sigh.
It was necessary when ftp-master was auric; at one point (although
pre-reimplementation-in-Python, I think) it was known for the testing
run to take more than a day, which wasn't too useful. On newraff that's
not a problem, although you'd have to get aj to comment on the memory
implications of discarding this optimization.
I do agree that testing i386 first would be helpful, considering the
magic hack that gets applied to Architecture: all packages.
Anyway, if you have problems understanding what's going on in the
future, please feel free to ask debian-release.
--
Colin Watson [cjwatson@flatline.org.uk]
Reply to: