Gentoo Archives: gentoo-science

From: Auke Booij <auke@××××××.com>
To: gentoo-soc@l.g.o, gentoo-science@l.g.o
Subject: [gentoo-science] G-CRAN weekly report #8
Date: Mon, 02 Aug 2010 12:22:37
Okay, I haven't sent a weekly report in a while now, because of some absence.

Last report, I promised I'd work on external dependencies and g-common
(remember, g-common is the tool that will handle non-ebuild
repositories in a transparent way). Well, what can I say, I did.
External dependencies were implemented using a big list of
hand-written regexes (ah yes, it's a bit of a hack, but it'll work for
most CRAN packages. bioconductor doesn't give us external dependency
data at all, unfortunately), and g-common is in a usable state. As far
as I'm concerned, what's left is:
-enabling g-common to output real ebuilds with proper formatting
instead of writing symlinks. this seems to be a feature that may be
used later by some overlay teams. this has no priority, but I think
it'd make g-common a very useful tool
-proper metadata output for g-common, where possible with help from
the repo drivers
-move common.ebuild code around a bit (nothing substantial)
-writing a layman driver for g-common (this is necessary for )
-documentation (!)
This basically means that the true g-cran driver is basically done
now, and the rest is just making it usable.

To make up for last report's size, I'll hereby end this report. Should
you have any questions or concerns, I'll be in #gentoo-soc as always
(nick: tulcod), or you can email me.