public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
From: LIU Hao <lh_mouse@126.com>
To: xclaesse@gmail.com, gcc@gcc.gnu.org
Subject: Re: Issue with dllexport/dllimport inline function
Date: Wed, 12 Oct 2022 01:42:54 +0800	[thread overview]
Message-ID: <13cbf2ec-4027-3df7-aa76-3d680afc075b@126.com> (raw)
In-Reply-To: <0c24f08964acc66a402d94c1b6a389dee4efa163.camel@gmail.com>


[-- Attachment #1.1: Type: text/plain, Size: 1235 bytes --]

在 2022-10-11 22:26, xclaesse@gmail.com 写道:
> #ifdef GLIB_COMPILATION
> #  define _GLIB_API _GLIB_EXPORT
> #  define GLIB_INLINE __attribute__((__dllexport__)) __attribute__((__gnu_inline__)) extern inline

This is not correct. Typically, `dllexport` indicates that 'I want an external definition' but 
`gnu_inline` indicates that 'I do not want an external definition', so it's a contradiction: you get 
either multiple definitions, or none (which causes undefined references).


To user code which imports such a function, it should be seen as

    ```
    __attribute__((__gnu_inline__))
    extern inline
    int g_strcmp0(const char*, const char*)
    ```

This has the desired effect: The function can be inlined where appropriate. And if not, a call to 
the external thunk is emitted; no weak definition is emitted in either case.


To the library itself which exports it, it should be seen as

    ```
    __attribute__((__dllexport__))
    extern inline
    int g_strcmp0(const char*, const char*)
    ```

There is no `gnu_inline` in this case. GCC always generates a COMDAT/weak/linkonce definition, which 
will not cause multiple definition errors.



-- 
Best regards,
LIU Hao


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 840 bytes --]

  reply	other threads:[~2022-10-11 17:42 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-10 16:39 xclaesse
2022-10-11  5:00 ` LIU Hao
2022-10-11 13:10   ` xclaesse
2022-10-11 13:28     ` LIU Hao
2022-10-11 13:35       ` LIU Hao
2022-10-11 14:26         ` xclaesse
2022-10-11 17:42           ` LIU Hao [this message]
2022-10-11 18:37             ` xclaesse
2022-10-12  3:03               ` LIU Hao
2022-11-01 15:38                 ` xclaesse
2022-11-01 16:06                   ` LIU Hao
2022-11-01 18:11                     ` xclaesse
2022-11-02  4:44                       ` LIU Hao
  -- strict thread matches above, loose matches on Subject: below --
2022-10-10 16:32 xclaesse

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=13cbf2ec-4027-3df7-aa76-3d680afc075b@126.com \
    --to=lh_mouse@126.com \
    --cc=gcc@gcc.gnu.org \
    --cc=xclaesse@gmail.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).