Re: [Chicken-users] can you believe it? compile time vs. run time issues, again!

[felix winkelmann]

On 5/29/05, Michele Simionato <address@hidden> wrote:
> How many times I have been bitten by compile-time vs. run-time issues?
> 10, 12? One would think I had understood the lesson, however ... :-(
> 
> My motivation here was a nifty trick to generate dynamics records
> without using macros:
> 

Very nifty. But the use of `eval' at runtime usually indicates
problems when compiling.

Ok, the correct way is this:

(declare (run-time-macros))      ; make sure macro expanders are
available at run-time

(require-extension match) ; make sure match-support (is needed by
match expanders) is linked

(include "chicken-match-macros.scm")  ; include the macro-defs for match...
(include "chicken-more-macros.scm")  ; include the macro-defs for cut

(define (dynamic-record ...) ...)


The syntax available for `eval' is deliberately restricted: usuall one doesn't
need all the extra macro-expanders. But if you need eval with all the extra
stuff, you might want to put the above code (minus def dynamic-record)
into an extra-file, say bigeval.scm, and install it as an extension:

chicken-setup bigeval

>From now on you can just say `(require-extension bigeval)' and you
will have all the extra macros available at run-time.

Actually, this could be turned into an egg...

> 
> $ csc -run-time-macros dynamic-record.scm
> /usr/lib/gcc-lib/i486-linux/3.3.5/../../../crt1.o(.text+0x18): In
> function `_start':
> ../sysdeps/i386/elf/start.S:98: undefined reference to `main'
> collect2: ld returned 1 exit status
> *** Shell command terminated with exit status 1: gcc -o dynamic-record
> dynamic-record.o -lchicken -lsrfi-chicken -lstuffed-chicken
> -L/usr/local/lib -Wl,-R/usr/local/lib -ldl -lpcre -lm  -ldl -lpcre

I didn't get this error. Did you add a unit declaration somewhere?


cheers,
felix