problems with guile?

problems with guile?

[Dmitry Morozhnikov]

hello. 

i want to play with cgen, but can`t get it run. what version of guile must be used for cgen? i try 1.3.4 and 1.4.1, but they both complain in map procedure:

guile> (load "dev.scm")
guile> (load-opc)
guile> (cload #:arch "ia32")

Backtrace:
 9  [dynamic-wind #<procedure ()> #<procedure ()> #<procedure ()>]
10* [#<procedure ()>]
11* (letrec ((loop (lambda # #))) (loop (read)))
12  [loop (include "simplify.inc")]
    ...
13  (begin (reader-process entry) (loop (read)))
14* [reader-process (dsmf f-mod-r/m "entire mod-r/m byte" ...)]
15  (let ((expansion (if # entry #))) (-reader-process-expanded expansion))
    ...
16  [-reader-process-expanded-1 #]
17* (let ((command #)) (if command (let* # #) ...))
18  (if command (let* (# # #) (if # # ...)) ...)
    ...
19  [define-full-multi-ifield f-mod-r/m "entire mod-r/m byte" ...]
20  (let ((f #)) (current-ifld-add! f) f)
21* [-multi-ifield-parse "define-full-multi-ifield" f-mod-r/m ...]
22  (let (# # #) (elm-xset! result # name) (elm-xset! result # #) ...)
23* (if insert (elm-xset! result (quote insert) ...) ...)
24  [elm-xset! #(#("object" # f-mod-r/m ...) (# #f # ...)) insert ...
25*  [-multi-ifield-make-default-insert f-mod-r/m (#(# #) #(# #) #(# #))]
26   (let* ((lengths #) (shifts #)) (let (#) (apply rtx-make #)))
27   (let ((exprs #)) (apply rtx-make (cons # #)))
28*  [map #<procedure (f length shift)> (#(# #) #(# #) #(# #)) (2 3 3) ...]

./ifield.scm:824:18: In procedure map in expression (map (lambda # #) subfields ...):
./ifield.scm:824:18: Argument out of range: (0 0 2 5)
ABORT: (out-of-range)

actually, lists 'shifts' in strings 827 and 846 are larger then lists 'subfields' and 'lengths'. when i remove one '(cons 0 ...' from 'shifts' constructor -- cgen starts, but also can`t produce anything from ia32.cpu (which is interesting for me). 

tried with cgen-1.0 and cvs version.

-- 
DM, Dmitry Morozhnikov, PGPKEY 0xB6C14D95,
music at http://mp3.com/DMIceman

Re: problems with guile?

[Frank Ch. Eigler]

[-- Attachment #1: Type: text/plain, Size: 772 bytes --]

Hi -

> i want to play with cgen, but can`t get it run. 

> what version of guile must be used for cgen?
> i try 1.3.4 and 1.4.1, but they both complain in map procedure:

We've had success with several versions of guile, those two included.
The problem you are seeing:

> guile> (load "dev.scm")
> guile> (load-opc)
> guile> (cload #:arch "ia32")
> [...]

is probably due to starting cgen in something other than batch mode
(as in the "make" of sid/sim/opcodes, the only mode that we tend to
run it under), or if you're trying to process incomplete .cpu files.
The ia32, ia64, and several cgen cpu models are (or rather, used
to be) works-in-progress.  Working ports are those for which an
actual generated sid/sim/opcodes target exists.


- FChE

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: problems with guile?

[Dmitry Morozhnikov]

On Mon, 24 Jun 2002 13:08:11 -0400
"Frank Ch. Eigler" <fche@redhat.com> wrote:

FCE> is probably due to starting cgen in something other than batch mode
FCE> (as in the "make" of sid/sim/opcodes, the only mode that we tend to
FCE> run it under), or if you're trying to process incomplete .cpu files.
FCE> The ia32, ia64, and several cgen cpu models are (or rather, used
FCE> to be) works-in-progress.  Working ports are those for which an
FCE> actual generated sid/sim/opcodes target exists.

yes, thanks. can i hope what ia32 be completed someday? :-)

BTW: i`m sure what -multi-ifield-make-default-insert and -multi-ifield-make-default-extract has error in 'shifts' initialize. can`t understand what they do exactly, but length of 'shifts' list are always longer than 'lengths'. when processing other cpu (like m32r) this routines are simple avoided.

-- 
DM, Dmitry Morozhnikov, PGPKEY 0xB6C14D95,
music at http://mp3.com/DMIceman