Re: [j-devel] readtable bug?

[Peter G. <pe...@ar...>]

On Wed, 24 Mar 2004 at 01:22:56 +0100, Andr=E1s_Simon wrote:
> I think this shouldn't happen:
>
> CL-USER(2): (set-macro-character #\] (get-macro-character #\)))
> T
> CL-USER(3): (read-delimited-list #\] (make-string-input-stream "a b]"))=

> Debugger invoked on condition of type END-OF-FILE:
>   #<END-OF-FILE @ #xccc588>
> Restarts:
>   0: TOP-LEVEL Return to top level.
> [1] CL-USER(4): :res
> CL-USER(5): (read-delimited-list #\] (make-string-input-stream "a b ]")=
)
> (A B)
>
> SET-MACRO-CHARACTER correctly sets ATTR_TERMINATING_MACRO as the
> attribute of #\], but I don't see where this attribute is used (other
> than in GET-MACRO-CHARACTER).

I've checked in a fix for this.

> A completely unrelated thing: I had to put back (locally) special
> handling of ".compiled" files in Load.java (from v 1.34) for ilisp to
> start up. Without it, I get backtraces like this:
>
>   54: (INVOKE-DEBUGGER #<ERROR @ #x1362012>)
>   55: (LOAD "/home/simon/LISP/ilisp/ilisp-pkg.compiled" :IF-DOES-NOT-EX=
IST NIL)
>   56: (SYSTEM::INTERACTIVE-EVAL (OR (AND (LOAD "/home/simon/LISP/ilisp/=
ilisp-pkg.compiled" :IF-DOES-NOT-EXIST NIL)
>   T) (AND (LOAD "/home/simon/LISP/ilisp/ilisp-pkg.lisp" :IF-DOES-NOT-EX=
IST NIL) T)))

I'm not sure about this one, since I don't have an ilisp setup readily
at hand to reproduce the problem.

The checked-in version of COMPILE-FILE just makes a copy of the .lisp
file with a .compiled extension, so loading the .compiled file
(provided that it exists) should be roughly equivalent to loading the
=2Elisp file. (Note that EVAL-WHEN is currently stubbed out and identical=

to PROGN, so tricks that rely on EVAL-WHEN correctly distinguishing
between compile-time and load-time will fail.)

LOAD currently ignores its keyword arguments, so it will signal an
error if the specified file does not exist, even if :IF-DOES-NOT-EXIST
NIL is specified in the call. This might be what's happening in the
situation in question.

I'll try to get around to adding support for keyword arguments to LOAD
in the next few days; maybe that will fix the problem.

Thanks for the report!

-Peter