[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH] add language/wisp to Guile?

From: Maxime Devos
Subject: Re: [PATCH] add language/wisp to Guile?
Date: Wed, 15 Feb 2023 00:01:50 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Thunderbird/102.7.0

> [...]
That would be nice, but would require doing changes in a critical core
part of Guile. It would change this addition from a risk-free added
feature to a risky core change.

I maintain that a new language shouldn't be merged until the Scheme-specific load path stuff is fixed/extended to work for non-Scheme things (e.g. Wisp) too -- if this requires somewhat risky (*) changes to core parts, then that just means we'll have to do some risky stuff, then.

I also expect that Guile maintainers will have the opposite opinion (i.e., ‘fixing the load path stuff isn't necessary for merging a new language implementation’).

(*) FWIW I disagree on the 'risky' assessment -- it seems like a ‘if it runs, it will work’ thing to me. That it modifies a core part of Guile, makes it less risky IMO, as it would automatically be more tested.

Aside from the (*) and the 'I also expect [...],', I don't have anything new to say about this, so I'll stop here.

> [...]
That would also enable shipping pre-compiled software without

That can already be done -- besides legalities, nothing stops people from putting [^] or [^] .scm files in $GUILE_LOAD_PATH and putting .go in $GUILE_LOAD_COMPILED_PATH.

[^]: Redacted to not give people ideas on how to circumvent stuff.
I can elaborate by non-public e-mail if you like.

so there may be strategic reasons to avoid it. Always
providing the sourcecode also makes compliance with automatic copyleft
licenses automatic.

Mm, yes, I guess. If only people weren't careless and didn't try to circumvent copyleft, then things would be easier ...

On 14-02-2023 22:24, Dr. Arne Babenhauserheide wrote:
PS: So what’s still missing here is to avoid setting the locale. Do you
     happen to have a hint how to actually do this right?

I think you might have forgotten about this:

 -- Scheme Procedure: set-port-encoding! port enc
 -- C Function: scm_set_port_encoding_x (port, enc)
     Sets the character encoding that will be used to interpret I/O to
     PORT.  ENC is a string containing the name of an encoding.  Valid
     encoding names are those defined by IANA
     (, for example
     ‘"UTF-8"’ or ‘"ISO-8859-1"’.

As such, I propose calling set-port-encoding! right in the beginning of read-one-wisp-sexp.

More concretely, replace

(define (read-one-wisp-sexp port env)
         ;; allow using "# foo" as #(foo).
         (read-hash-extend #\# (λ (chr port) #\#))
            ((eof-object? (peek-char port))
              (read-char port )); return eof: we’re done
              (let ((chunk (wisp-scheme-read-chunk port)))
                  ((not (null? chunk))
                    (car chunk))


(define (read-one-wisp-sexp port env)
  ;; Allow using "# foo" as #(foo).
  ;; Don't use the globally-acting read-hash-extend, because this
  ;; doesn't make much sense in parenthese-y (non-Wisp) Scheme.
  ;; Instead, use fluids to temporarily add the extension.
  (define %read-hash-procedures/parameter
    (fluid->parameter %read-hash-procedures))
  (parameterize ((%read-hash-procedures/parameter
                   `((#\# ,(λ (chr port) #\#))
    ;; Read Wisp files as UTF-8, to support non-ASCII characters.
    ;; TODO: would be nice to support ';; coding: whatever' lines
    ;; like in parenthese-y Scheme.
    (set-port-encoding! port "UTF-8")
    (if (eof-object? (peek-char port))
        (read-char port) ; return eof: we’re done
        (let ((chunk (wisp-scheme-read-chunk port)))
          (and (not (null? chunk)) ; <---- XXX: maybe (pair? chunk)
               (car chunk))))))


(I've also done the read-hash-extend stuff and simplified the 'cond' expressions.)


Attachment: OpenPGP_0x49E3EE22191725EE.asc
Description: OpenPGP public key

Attachment: OpenPGP_signature
Description: OpenPGP digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]