I found a small issue with the :depends-on form for :drei-mcclim when trying to load the latest mcclim on clisp 2.41. Drei uses #. in :depends-on to try to define a dependency on swank if it's loaded. Unfortunately on clisp, this results in nil rather than no value. I'm seeing this error:
component NIL not found, required by #1=#<SYSTEM "drei-mcclim" #x19F581E9> [Condition of type ASDF:MISSING-DEPENDENCY]
There was discussion on c.l.l. about the behavior of the reader macro function in this regard, see the thread started by Pascal Bourguignon:
http://groups.google.com/group/comp.lang.lisp/msg/6d137dca962cfaf4?hl=en&...
Note that I'm not actually trying to run drei on clisp, just trying to load mcclim.asd.
"Jack Unrue" jdunrue@gmail.com writes:
I found a small issue with the :depends-on form for :drei-mcclim when trying to load the latest mcclim on clisp 2.41. Drei uses #. in :depends-on to try to define a dependency on swank if it's loaded. Unfortunately on clisp, this results in nil rather than no value. I'm seeing this error:
Argh, I guess I need to change that to pushing a mcclim.system::use-swank symbol onto *FEATURES*, which is what I was trying to avoid in the first place. Or perhaps the (values) can replaced with something that will be loaded anyway, such as :flexichain? The best solution is, of course, to get Swank (it should work on CLISP). :-)
There was discussion on c.l.l. about the behavior of the reader macro function in this regard, see the thread started by Pascal Bourguignon:
http://groups.google.com/group/comp.lang.lisp/msg/6d137dca962cfaf4?hl=en&...
I think that my interpretation of how #.(values) should be treated is the most useful one, but I acknowledge that it's undefined by the spec, and we thus shouldn't use it.