Fwd: utf-8 output in a terminal

---------- Forwarded message ---------
De : Gr=C3=A9gory Vanuxem <[email protected]>
Date: lun. 29 juil. 2024 =C3=A0 01:34
Subject: Re: [Openmcl-devel] utf-8 output in a terminal
To: R. Matthew Emerson <[email protected]>


Hello Matthew,

Le dim. 28 juil. 2024 =C3=A0 07:31, R. Matthew Emerson <[email protected]> a =
=C3=A9crit :
>
>
>
> > On Jul 27, 2024, at 8:23=E2=80=AFPM, Gr=C3=A9gory Vanuxem <g.vanuxem@gm=
ail.com> wrote:
> >
> > In FriCAS  built on top of Clozure CL I obtain this output in a termina=
l (WSL2)
> >
> > (gamma(x) =C3=83=C3=82=C2=A2=C3=83=C3=82=C3=83=C3=82 reals && Re(x) > 0=
) || (x =C3=83=C3=82=C2=A2=C3=83=C3=82=C3=83=C3=82 integers &&
> > gamma(x) =C3=83=C3=82=C2=A2=C3=83=C3=82=C3=83=C3=82 reals)
> >
> > instead of (SBCL)
> >
> > (gamma(x) =E2=88=88 reals && Re(x) > 0) || (x =E2=88=89 integers && gam=
ma(x) =E2=88=88 reals)
> >
> > if I (print ...) a string with, for example, =E2=88=88, in a pure Clozu=
re CL
> > (recently git cloned) this is correctly printed, no problem.
> >
> > So I wonder if I have to modify the output stream character encoding
> > or the output routine.
>
> Check that the external format of the FriCAS output stream is what you ex=
pect. In other words, if you pass an explicit :external-format argument to =
cl:open, make sure it matches what your terminal requires.

I don't think it is at this stage in fact, I was wrong. Better
checking what is going on, since the string comes from C (sorry I
forgot to mention that) I tend to think my problem comes from
ccl::%get-cstring. Tracing it seems to show it's not "decoding" it and
considers it as a 8859-1 string. But to test this i.e. force CCL to
use :utf8 format to "read it" I have not found. There is no option
from what I know to get-cstring like :external-format and using
with-encoded-cstrs does not seem to work.

Right now I am using some FFI macros from FriCAS to simplify my code,
I will try to revert to pure CCL code first to check my claims,
TIMTOWTDI.

> Doing (describe *terminal-io*) might show something interesting. Here=E2=
=80=99s what it shows for me:
>
> ? (describe *terminal-io*)
> #<ECHOING-TWO-WAY-STREAM input #<BASIC-CHARACTER-INPUT-STREAM UTF-8 (TTY/=
0) #x1824C7D6>, output #<BASIC-CHARACTER-OUTPUT-STREAM UTF-8 (TTY/1) #x1824=
C2A6> #x1825FBE6>
> Class: #<STANDARD-CLASS ECHOING-TWO-WAY-STREAM>
> Wrapper: #<CLASS-WRAPPER ECHOING-TWO-WAY-STREAM #x18039A16>
> Instance slots
> SHARED-RESOURCE: NIL
> OPEN-P: T
> INPUT-STREAM: #<BASIC-CHARACTER-INPUT-STREAM UTF-8 (TTY/0) #x1824C7D6>
> OUTPUT-STREAM: #<BASIC-CHARACTER-OUTPUT-STREAM UTF-8 (TTY/1) #x1824C2A6>
>
> The default external format has been #<EXTERNAL-FORMAT :UTF-8/:UNIX #x182=
49C6E> for quite a long time now.
>
> https://ccl.clozure.com/docs/ccl.html#characters-and-external-formats

This is exactly what I have when I use this function in FriCAS. I have
read the documentation and the internal utf32 representation but here
I am stuck. This is not the most important thing I need to do but it
will be imperative for some next stuff.

In anyway, thanks!

Greg

>
> >
> > And more importantly, how can I achieve that?
> >
> > - Greg
> >
> > PS:
> > (1) -> )lisp CCL:*DEFAULT-EXTERNAL-FORMAT*
> > Value =3D :UNIX
>