snapshot.d.o fuse filesystem
Hi,
(Continuing here the discussion from IRC)
I'm interested in using snapshot.d.o to do some data-mining on the
evolution of development practices in Debian. I'd like to look at
various things over the 2005-2011 timeframe:
- VCSes for package maintenance (using Vcs-* fields)
- patch systems and new source formats
- classic debhelper vs cdbs vs dh7
- Homepage field
- generalization of team maintenance (either by real teams
(@lists.(alioth)?.debian.org) or co-maintenance).
snapshot.d.o is a fantastic tool for that. I'd like to proceed by
looking at the Debian archive at regular intervals (once every one or
two months).
What I'm missing is a way to programatically:
- retrieve the Sources file for unstable at a given date
- retrieve specific (source,version) couples from the farm (so that I
can uncompress them, and look at them)
I understand that I could use the pgsql DB directly, but I'd like to
look at other easier ways first, if there are some. The fuse filesystem
looked exactly like what I need.
> <weasel> the fuse filesystem has problems with binnmus
> <weasel> (one source,version builds multiple versions of a single binpackagename)
> <weasel> if that gets fixed we could probably try mounting it again
I won't need to deal with binNMUs, since I'm only interested in the source
packages. So would the FUSE filesystem be usable for my task?
If not, are you planning to fix it?
If not, how do you suggest I try to fix it myself?
- Lucas
Reply to: