[Ecls-list] doc strings for define-condition

[Juan Jose Garcia-Ripoll]

On Wednesday 16 July 2003 11:46, Edi Weitz wrote:
>  However, I still have issues with MK:DEFSYSTEM. I'm using
> the CLOCC snapshot from last night and MK:DEFSYSTEM compiles without a
> problem but when trying to build HTML-TEMPLATE from its system
> definition I get [...]
>   > (load "/usr/local/lisp/defsystem")
>   ;;; Loading "/usr/local/lisp/defsystem.so"
>   ;;; Loading "/usr/local/lib/ecl/cmp.so"
>   ;;; Warning: PROCLAIM is being redefined.
>   ;;; Warning: COMPILE-FILE-PATHNAME is being redefined.
>   ;;; Warning: COMPILE-FILE is being redefined.
>   ;;; Warning: COMPILE is being redefined.
>   ;;; Warning: DISASSEMBLE is being redefined.
>   ;;; Warning: WITH-COMPILATION-UNIT is being redefined.
>   ;;; Warning: REQUIRE is being redefined.
>   #P"/usr/local/lisp/defsystem.so"
>   Top level.
>
>   > (mk:compile-system "html-template")
>   ;;; Loading T
>   NIL cannot be coerced to a pathname.
>   Broken at MAKE:COMPILE-SYSTEM.

This only happens with the compiled version of MKSYSTEM. If you use 
defsystem.lisp, the bug goes aways. I am tracking this tonite.

> One more thing: Isn't ECL with recent patches supposed to load an init
> file now? I see this
>   > SI:*LISP-INIT-FILE-LIST*
>   ("~/.ecl" "~/.eclrc")
>   Top level.

Solved. This one was easy: the value of the environment variable HOME has to 
be hacked to ensure it produces a valid directory namestring in lisp. For 
instance, "/home/jlr" has to be converted to "/home/jlr/". Somehow I undid 
this change when I fixed USER-HOMEDIR-PATHNAME to avoid using getpw(), and 
use the environment variable HOME instead.

> No problem at all! I'm glad to see how fast ECL is getting more usable
> and ANSI-compliant almost every day. Without having any data to
> confirm this I'd say it has already surpassed Corman Lisp with respect
> to standards compliance and isn't far away from CLISP, maybe even
> ahead of it. Congratulations!

Thanks. :-) It's a pity that I do not have more time for working on it, but 
there are lots of things that need improvement in the usability field: the 
bytecodes debugger should have stepping, breakpoints, and function 
re-invocation; the bytecodes should have also a printed representation; ILISP 
should be adapted to work with ECL... CLISP is very good at this, but it had 
more time to mature, anyway.

Juanjo