Re: [PATCH] Allow specifying the socket style for open-socket-for-uri.

[Maxime Devos]

Op 20-07-2023 om 16:08 schreef Christopher Baines:

Since this allows specifying additional behaviours for the socket
through using SOCK_CLOEXEC and/or SOCK_NONBLOCK (when bitwise or'ed with
SOCK_STREAM).

Note that Guile/guile-gnutls currently doesn't support performing the
TLS handshake on a non-blocking socket, so this currently won't work.

I would think that it currently works for _http_. (https is another matter of course.)

* module/web/client.scm (open-socket-for-uri): Allow specifying the
socket style.
---
  module/web/client.scm | 10 +++++++---
  1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/module/web/client.scm b/module/web/client.scm
index 6c54c5021..a5405d17f 100644
--- a/module/web/client.scm
+++ b/module/web/client.scm
@@ -317,9 +317,11 @@ host name without trailing dot."
    (read-response port))

-                              #:key (verify-certificate? #t))
+                              #:key (verify-certificate? #t)
+                              (socket-style SOCK_STREAM))
    "Return an open input/output port for a connection to URI-OR-STRING.
-When VERIFY-CERTIFICATE? is true, verify HTTPS server certificates."
+When VERIFY-CERTIFICATE? is true, verify HTTPS server certificates.
+SOCKET-STYLE is passed as the second argument to the socket procedure."

(define* (open-socket-for-uri uri-or-string

I think it should mention that socket-style doesn't work yet for https.

    (define uri
      (ensure-uri-reference uri-or-string))
    (define https?
@@ -346,7 +348,9 @@ When VERIFY-CERTIFICATE? is true, verify HTTPS server 
certificates."
        (let* ((ai (car addresses))
               (s  (with-fluids ((%default-port-encoding #f))
                     ;; Restrict ourselves to TCP.
-                   (socket (addrinfo:fam ai) SOCK_STREAM IPPROTO_IP))))
+                   (socket (addrinfo:fam ai)
+                           socket-style
+                           IPPROTO_IP))))
          (catch 'system-error
            (lambda ()
              (connect s (addrinfo:addr ai))

Best regards,
Maxime.

I don't think that #:socket-style SOCK_SEQPACKET makes much sense, so I would propose having a #:socket-flags argument instead. Also, describing the new argument in terms of how it is used in the implementation doesn't seem great to me -- I would instead propose doing what the documentation 'accept' does with its 'flags' argument.

OpenPGP_0x49E3EE22191725EE.asc
Description: OpenPGP public key

OpenPGP_signature
Description: OpenPGP digital signature