[8588] in athena10
[Debathena] #1101: Grand plan for fixing build-all and da and
daemon@ATHENA.MIT.EDU (Debathena Trac)
Fri Oct 21 16:19:40 2011
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
From: "Debathena Trac" <debathena@MIT.EDU>
Cc: debathena@MIT.EDU
To: geofft@MIT.EDU
Date: Fri, 21 Oct 2011 20:19:37 -0000
Reply-To:
Message-ID: <042.48fd593332c90b4210756c6bb3285899@mit.edu>
Content-Transfer-Encoding: 8bit
#1101: Grand plan for fixing build-all and da and sbuildhack
----------------------------+--------------------------------
Reporter: geofft | Owner:
Type: enhancement | Status: new
Priority: normal | Milestone: The Distant Future
Component: development | Keywords:
Upstream bug: |
----------------------------+--------------------------------
Our current largest issue is that it's hard to upload a fixed version of a
package for a new release without confusing everything (both build-all and
the person who comes along and runs da sbuildhack later). There are some
other suboptimal things about the build infrastructure that can be fixed
along the way.
Here's a proposal that seems reasonably sane:
* The stamps directory should specify versions of packages, so instead of
e.g. stamps/lucid/debathena-reactivate.done, you have stamps/debathena-
reactivate_2.0~ubuntu10.04.done. This way you know which versions have
been built, you automatically no longer have the stamp if there's a new
version of a package, you automatically no longer have the stamp if
there's a bump of your tag or a new distro upstream, etc.
* We should have a script that synchronizes the stamps directory out of
the repo, and call it at the top of do-build. (So, more or less, do-build
depends on the actual state of the repo, not stamps files.)
* da sbuildhack should also use the stamps files, so that it doesn't
attempt to build packages that already exist.
* da sbuildhack (and build-all) should respect X-Build-For: control fields
instead of the unversioned
* Rename sbuildhack to something less stupid.
It occurs to me that da and build-all are the same type of thing, just
perpendicular, and maybe they should be the same program at some point.
--
Ticket URL: <http://athena10.mit.edu/trac/ticket/1101>
Debathena <http://debathena.mit.edu>
MIT Debathena Project
