[Asdf-devel] converting asdf build & test to Lisp

Robert P. Goldman rpgoldman at sift.info
Sun Jun 1 16:36:58 UTC 2014

Faré wrote:
> Robert, can you test this branch again (and the syntax-control branch)
> and tell me what needs be done (if anything) before they can be merged
> in? I believe the minimakefile is now stable (I haven't made
> significant changes in over a week) and has feature parity with the
> master branch (and many new features made possible by rewriting from
> shell to CL).

This week I will do some tests on systems with non-trivial readtables to
try to figure out what exactly is going on with the syntax-control branch.

Once I understand what's going on there, we can probably merge it.

minimakefile will have to wait.  I don't see a pressing need to merge
it.  It provides no new functionality, only an experiment with using
ASDF to enable CL to be used as a scripting langauge.  That's nice, but
until it *eases* my development, instead of complicating it, I don't
expect to merge this branch.  The existing makefile isn't broken in a
way that the minimakefile fixes.  I have already specified changes that
need to be made in this branch before it can be merged:

1. It must provide bash completion that is *at least as* good as make does.

2. It must have more documentation.

Note that these are lower bounds.  My ASDF development infrastructure is
not a place for experimentation. Experiments that provide clear benefits
*to me* are likely to be accepted. Experiments that further unrelated
goals (develop CL as a scripting environment) will not: novel CL
technology development is not my job as ASDF maintainer.

As the delays in reviewing the syntax-control branch illustrate, I don't
have a lot of spare cycles for ASDF right now (a number of projects
entering critical state this month). So ASDF items that don't provide a
compelling benefit (bugfix or significantly easier development) are
going to be back-burnered till more attention becomes available.

Sorry, but it's better that I restrict myself to doing what I can do
competently, rather than overextend and do a crummy job on everything.

In general, I think this should be acceptable, because the heroic days
of fixing the big bugs in ASDF are over.  We are entering a stage of
maintaining mature and successful technology.


More information about the asdf-devel mailing list