Re: (in)stability / parallel installs / platform-2.7.93

[Andreas Rottmann]

Jan Nieuwenhuizen <address@hidden> writes:

> Andy Wingo writes:
>
>> The bits about strict stability scared me at first. I mean, we don't
>> have a stable product, and I don't see a time in the future when we can
>> call it stable.
>
> Do we have a policy for handling &multiple &returns from a function?
>
In G-Wrap, there is support for output arguments, which is excercised
via so-called type rules for the .defs in the wrapset specification,
e.g. (from glib-spec.scm):

(add-type-rule! ws '(("gint*" "*")) '(int out))

This means all arguments that have a gint* type, regardless of
argument name ("*"), are to be treated as integer output arguments
"(int out)". Multiple return values are (naturally) communicated with
(values ...).

> In my patch wrapper fro g_io_channel_read_line, I return a list,
> combining GIOStatus and &line, and ignoring GError (see below
>
>     +  result = g_io_channel_read_line (channel, &line, NULL, NULL, NULL);
>     +
>     +  return scm_list_2 (scm_long2num (result), scm_makfrom0str (line));
>
I think it should be possible to solve this via type rules, without
the need for an extra wrapper.

> I can't commit to gna yet, maybe you can include my patch for the next
> release?  I also like to wrap gtk-menu-popup to allow menu positioning
> function.
>
I'll have look at merging your changesets.

Regards, Rotty
-- 
Andreas Rottmann         | address@hidden      | address@hidden | address@hidden
http://yi.org/rotty      | GnuPG Key: http://yi.org/rotty/gpg.asc
Fingerprint              | DFB4 4EB4 78A4 5EEE 6219  F228 F92F CFC5 01FD 5B62

To iterate is human; to recurse, divine.