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

More consistent organisation of Debian meta-data


Currently, all kind of metadata (like bug summaries in the form of
PTS/DDPO bugs.txt, lintian checks, etc) are scattered all over the
place, and not easily reusable by others.

What about this:
for every site that has an interface to some metadata (qa.d.o, bugs.d.o,
etc), make sure that in the http subdirectory 'data', the raw data is
present. So, if I go to http://qa.debian.org/data, I can find there the
bugs.txt for DDPO, the extract, the raw depcheck data, etc.

Second, I think it'd also be nice for debugging and seeing where the
cron job times are not tweaked well enough if there is some kind of
trace in these data files. By trace, I think something like a 'TRACE'
field in it, where for the bugs.db file for DDPO you might find for

spohr.d.o:/org/bugs.d.o/db-h (timestamp for last bts-update run) \
	[rsync] merkel.d.o:...db-h (timestamp for last sporh->merkel update) \
	[aba's bts2ldap] merkel.d.o:~aba/bts2ldap/index (timestamp for \
		last bts2ldap run) \
	[merkel.d.o:~qa/ddpo/update-bugs] ~qa/bugs.txt (timestamp) \
	[db_load] ~qa/bugs.db (timestamp)

This way, you can really show on pages when the last update was, by
taking the first (source) timestamp, rather than what typically happens
the last one, which doesn't mean anything if one step earlier something
is broken/outdated.

And of course, you can then very easily verify whether stuff is


Jeroen van Wolffelaar
Jeroen@wolffelaar.nl (also for Jabber & MSN; ICQ: 33944357)

Reply to: