[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Request for Code: MIA database (python)



Hi,

I've been procrastinating coding up some extra features in the mia
database management code[1] that would make my work somewhat easier, and
thus would make me more productive in this task. Therefore, instead of
procrastinating longer, and also since I'm not a very good python
hacker, I hereby ask for volonteers to come up with some patch(es) that
(partially or fully) implement my wishes outlined below.

Currently, mia-check will search through the database for active
items[2], and list those that have had no activity for a configurable
number of days (21 by default). However, there are two issues that make
this less than perfect:

1) It is maintainer-dependent what the delay is I want to impose on
   myself to look at this maintainer again. For example, if I add a
   'hint' regarding MIA'ness to a maintainer, I want no delay, and otoh,
   if I decide with the maintainer to see again in 3 months how it's
   going, I don't want him to show up in the list for the next 3 months.
   Implementation wise, I was thinking about some extra header to MIA
   mails or something, or an extra code addded to it (like '{+90d}' or
   something)
2) (Less important if (1) is implemented) I want to be able to locally
   procrastinate on a certain maintainer: the delay has passed, but I'm
   yet unsure what to do next. The probably means it's a borderline
   case, and it can wait. I could use the feature at (1) for this, but
   that'd involve writing to the database, and sending of mails etc,
   while I want a lower-level and easier 'postpone' command. Thus, this
   should be using some locally configuarable timestamp cache or
   something.


Another issue is that of identity -- many maintainers use multiple
different email addresses to maintain packages. A map of those email
addresses (esp. which of them belong to the same person) would be useful
throughout all of QA. Part of it can be automatically generated from
uid's on GPG keys, as the PTS somewhat tries to do, but part of it
cannot. echelon also uses some kind of mapping, I don't know which, but
I noticed the activity-from matches all kinds of alternative email
addresses of maintainers too (which is nice). This map and associated
scripts should be a separate part in QA, but once it exists, MIA-DB
should make use of it, for example, by making it possible for it to show
all packages maintained by that maintainer, and to link queries and
pings to eachother in some way.

If you want to help on code of either of those issues, please do know
that: Martin Michlmayr is the author and maintainer of the mia tracking
scripts, and as such has a say of what code is acceptable and what not
(although if _I_ like the code, I'd use it regardless).  Also, I think
it's better to first discuss your implementation plans before actually
implementing it, to prevent work for nothing.

Thanks,
--Jeroen

[1] http://cvs.debian.org/mia/?cvsroot=qa, the actual database is
    readable for all DD in /org/qa.debian.org/mia/db on merkel
[2] files, dossiers, on maintainers that are maybe MIA, and are not
    completely dealt with to my satisfaction yet

-- 
Jeroen van Wolffelaar
Jeroen@wolffelaar.nl (also for Jabber & MSN; ICQ: 33944357)
http://Jeroen.A-Eskwadraat.nl



Reply to: