[asdf-devel] 2.013 later this week

[Faré]

Dear ASDF users,

unless you have requests for urgent bugs and/or object to some of the
changes I just made (see below), I intend to release 2.013 later this
week.

Changes since 2.012:
* cleanup in use of features. Introducing #+asdf-unix and
#+asdf-window. (2.012.1)
* cleanup in support for clisp. Implementation identifier vs processor
name (2.012.2)
* support for Digitool's legacy system (R)MCL (2.012.3) - as requested
by p-cos to test his library.
* cleanup when using portable pathnames (2.012.4)
* cleanup in support for ECL (2.012.5, 2.012.6)
* parse the pathname of a defsystem portably when it's a string. (2.012.7)
* support for Symbolics' legacy system Genera (2.012.8) - proof of
concept, completes legacy support and allowed me to test my libraries.
* support for .. as :back in our portable pseudo-namestring language
and in merge-pathnames* - should make quite a few pathnames things
less ugly. (2.012.8)
* Remember source-registry and output-translations parameters
(2.012.8) - should provide a portable solution to the need many people
have expressed to be able to tweak a configuration after the fact.
* some documentation updates

Possibly questionable change:
* To support Genera, I removed ~@< from our error message format
strings. Frankly, I think cuteness matters that much when printing
error.

Feedback requested:
* are the (currently unexported) variables *source-registry-parameter*
and *output-translations-parameter* acceptable APIs for
reconfigurability of ASDF? If so, I'll export them before release. If
not, I'd appreciate API proposals, if possible from people who
understand how current code works.

ASDF is pretty stable, and we're really polishing corner cases, but
also making some

[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
Give up all hope for a better yesterday, even a better just now.
Never give up hope for a better tomorrow. — Patri Friedman