[PATCH] GCL cleanup

13 views
Skip to first unread message

Qian Yun

unread,
Dec 8, 2023, 7:07:53 AM12/8/23
to fricas-devel
Note on the system::process-some-args change:
in GCL commit bf67be1b this function was added a optional argument:

(defun process-some-args (args &optional compile &aux *load-verbose*)
......
(process-some-args args compile))

This doesn't affect FriCAS directly, but fails when try to load
fricas0 in toplevel.

Question: is it still necessary to keep fricas_gcl_rsym_hack?
Is gcl even buildable on windows right now? I didn't bother to try.
Since we can easily build on windows with sbcl/ecl/clisp, I suggest
to simply drop this hack.

- Qian
gcl-cleanup.patch

Waldek Hebisch

unread,
Dec 8, 2023, 9:33:16 AM12/8/23
to fricas...@googlegroups.com
On Fri, Dec 08, 2023 at 08:07:48PM +0800, Qian Yun wrote:
> Note on the system::process-some-args change:
> in GCL commit bf67be1b this function was added a optional argument:
>
> (defun process-some-args (args &optional compile &aux *load-verbose*)
> ......
> (process-some-args args compile))
>
> This doesn't affect FriCAS directly, but fails when try to load
> fricas0 in toplevel.

OK.

> Question: is it still necessary to keep fricas_gcl_rsym_hack?
> Is gcl even buildable on windows right now?

IIUC it is possible to build GCL on Windows. It seems that
Maxima folks offer binary based on GCL. Concerning
fricas_gcl_rsym_hack I am assuming that it is still necessary,
unless proved otherwise.

> I didn't bother to try.
> Since we can easily build on windows with sbcl/ecl/clisp, I suggest
> to simply drop this hack.

Gain from removal is tiny, IMO it is better to keep it.

--
Waldek Hebisch

Qian Yun

unread,
Dec 8, 2023, 8:32:15 PM12/8/23
to fricas...@googlegroups.com


On 12/8/23 22:33, Waldek Hebisch wrote:
>
> IIUC it is possible to build GCL on Windows. It seems that
> Maxima folks offer binary based on GCL.

It seems that they are offering clisp/sbcl/abcl for windows.
https://wolfgang.dautermann.at/maxima/nightlybuild/

> Concerning
> fricas_gcl_rsym_hack I am assuming that it is still necessary,
> unless proved otherwise.
>
>> I didn't bother to try.
>> Since we can easily build on windows with sbcl/ecl/clisp, I suggest
>> to simply drop this hack.
>
> Gain from removal is tiny, IMO it is better to keep it.
>

Status on GCL on windows:

I tried to build GCL on windows (mingw-w64), ./configure says it
doesn't recognize the system.

Web search points to "GCL 2.6.13 and Windows" by Camm:
https://lists.gnu.org/archive/html/gcl-devel/2022-12/msg00000.html

""""
2.6.13pre supports the original MinGW 32bit (of 2013) and current
cygwin 64bit.

mingw-w64 can eventually be supported
""""

I'm looking for a windows native build, so I'll not consider cygwin.
I also don't want to use outdated original mingw.

So I'll leave the hack there.

- Qian
Reply all
Reply to author
Forward
0 new messages