From: Hannes Domani <ssbssa@yahoo.de>
To: Tom Tromey <tom@tromey.com>
Cc: "gdb-patches@sourceware.org" <gdb-patches@sourceware.org>,
"blarsen@redhat.com" <blarsen@redhat.com>
Subject: Re: [PATCH v2] Allow calling of user-defined function call operators
Date: Fri, 3 May 2024 20:35:29 +0000 (UTC) [thread overview]
Message-ID: <829989170.10609936.1714768529914@mail.yahoo.com> (raw)
In-Reply-To: <87msp6d5ik.fsf@tromey.com>
Am Freitag, 3. Mai 2024 um 22:06:46 MESZ hat Tom Tromey <tom@tromey.com> Folgendes geschrieben:
> >>>>> "Hannes" == Hannes Domani <ssbssa@yahoo.de> writes:
>
> Hannes> The change in operation::evaluate_funcall is to make sure the type
> Hannes> fields are only used for function types, only they use them as the
> Hannes> argument types.
>
> IIRC the evaluation operations are all kind of complicated and
> hairy... but it seems to me that the type of the chosen overload of
> operator() would supply the type here?
Here the overload of operator() is chosen based on the argument values,
not the other way round.
> Hannes> + type *ftype = callee->type ();
> Hannes> +
> Hannes> + /* If the callee is a struct, there might be a user-defined function call
> Hannes> + operator that should be used instead. */
> Hannes> + std::vector<value *> vals;
> Hannes> + if (overload_resolution
> Hannes> + && exp->language_defn->la_language == language_cplus
> Hannes> + && check_typedef (ftype)->code () == TYPE_CODE_STRUCT)
> Hannes> + {
>
> One question to consider is whether this should be done in the
> expression node or elsewhere. Other operator overloads are handled in
> the value API instead.
>
> I'm not sure which is better. It would probably be cleaner to do it in
> the expression nodes, like you've done. Actually the best would
> probably be to make a new operation subclass and avoid the need for a
> language check.
I've now moved the logic into evaluate_subexp_do_call so all operation
subclasses would profit from this, so I don't understand how a new
operation subclass would work for this.
> However, the value API is convenient to use -- for example, this is what
> makes operator overloading work in the Python API.
>
> You can see the distinction with this patch by trying to call a
> struct-with-operator() object from Python.
Calling a struct-with-operator() object from Python does not work, because
valpy_call directly calls call_function_by_hand.
It would maybe be possible to also call evaluate_subexp_do_call there.
Hannes
next prev parent reply other threads:[~2024-05-03 20:35 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20240427163606.1780-1-ssbssa.ref@yahoo.de>
2024-04-27 16:36 ` Hannes Domani
2024-05-03 18:29 ` Guinevere Larsen
2024-05-03 18:51 ` Hannes Domani
2024-05-06 12:29 ` Guinevere Larsen
2024-05-03 20:06 ` Tom Tromey
2024-05-03 20:35 ` Hannes Domani [this message]
2024-05-06 16:31 ` Tom Tromey
2024-05-15 19:53 ` Tom Tromey
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=829989170.10609936.1714768529914@mail.yahoo.com \
--to=ssbssa@yahoo.de \
--cc=blarsen@redhat.com \
--cc=gdb-patches@sourceware.org \
--cc=tom@tromey.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).