From: Bruno Haible <bruno@clisp.org>
To: bug-gnulib@gnu.org
Cc: Eli Zaretskii <eliz@gnu.org>
Subject: Re: swab wrapper fails mingw.org's MinGW compilation of GDB
Date: Sun, 20 Jun 2021 17:22:47 +0200 [thread overview]
Message-ID: <4157125.b07XYjAfJy@omega> (raw)
In-Reply-To: <83czvimkq1.fsf@gnu.org>
Eli Zaretskii wrote in
<https://lists.gnu.org/archive/html/bug-gnulib/2021-03/msg00135.html>:
> The following snippet from Gnulib's unistd.h causes a compilation
> error when building the current development version of GDB 11:
>
> #if @GNULIB_MDA_SWAB@
> /* On native Windows, map 'swab' to '_swab', so that -loldnames is not
> required. In C++ with GNULIB_NAMESPACE, avoid differences between
> platforms by defining GNULIB_NAMESPACE::creat always. */
> # if defined _WIN32 && !defined __CYGWIN__
> # if !(defined __cplusplus && defined GNULIB_NAMESPACE)
> # undef swab
> # define swab _swab
> # endif
> _GL_CXXALIAS_MDA (swab, void, (char *from, char *to, int n));
> # else
> _GL_CXXALIAS_SYS (swab, void, (const void *from, void *to, ssize_t n));
> # endif
> _GL_CXXALIASWARN (swab);
> #endif
>
> The problem is that mingw.org's MinGW uses a slightly different
> prototype of _swab:
>
> _CRTIMP __cdecl __MINGW_NOTHROW void _swab (const char *, char *, size_t);
>
> So the difference between the prototypes causes this compilation error
> in C++ programs:
>
> CXX unittests/string_view-selftests.o
> In file included from ./../gdbsupport/common-defs.h:86,
> from ./defs.h:28,
> from unittests/string_view-selftests.c:26:
> ./../gnulib/import/unistd.h: In member function 'gnulib::_gl_swab_wrapper::operator gnulib::_gl_swab_wrapper::type() const':
> ./../gnulib/import/unistd.h:2543:1: error: invalid conversion from 'void (__attribute__((cdecl)) *)(const char*, char*, size_t)' {aka 'void (__attribute__((cdecl)) *)(const char*, char*, unsigned int)'} to 'gnulib::_gl_swab_wrapper::type' {aka 'void (*)(char*, char*, int)'} [-fpermissive]
> 2543 | _GL_CXXALIAS_MDA (swab, void, (char *from, char *to, int n));
> | ^~~~~~~~~~~~~~~~
> | |
> | void (__attribute__((cdecl)) *)(const char*, char*, size_t) {aka void (__attribute__((cdecl)) *)(const char*, char*, unsigned int)}
Thanks for the report. A similar problem also exists on Solaris and HP-UX.
The patch below should fix it. Tested on Solaris with CC and g++.
> The suggested fix is as follows (__MINGW32_VERSION is defined by
> mingw.org's MinGW, but not by MinGW64):
It is good to know how to distinguish older mingw from mingw64; thanks
for this info. But here we don't need to make the distinction, since we
have a _GL_CXXALIAS_MDA_CAST macro for these situations.
2021-06-20 Bruno Haible <bruno@clisp.org>
unistd: Avoid compilation error in C++ mode on Solaris, HP-UX, mingw.
Reported by Eli Zaretskii <eliz@gnu.org> in
<https://lists.gnu.org/archive/html/bug-gnulib/2021-03/msg00135.html>.
* lib/unistd.in.h (swab): Consider different declarations on Solaris,
HP-UX, and old mingw.
diff --git a/lib/unistd.in.h b/lib/unistd.in.h
index d4d4ba7..73c882f 100644
--- a/lib/unistd.in.h
+++ b/lib/unistd.in.h
@@ -2034,9 +2034,17 @@ _GL_WARN_ON_USE (sleep, "sleep is unportable - "
# undef swab
# define swab _swab
# endif
-_GL_CXXALIAS_MDA (swab, void, (char *from, char *to, int n));
-# else
+/* Need to cast, because in old mingw the arguments are
+ (const char *from, char *to, size_t n). */
+_GL_CXXALIAS_MDA_CAST (swab, void, (char *from, char *to, int n));
+# else
+# if defined __hpux /* HP-UX */
+_GL_CXXALIAS_SYS (swab, void, (const char *from, char *to, int n));
+# elif defined __sun && !defined _XPG4 /* Solaris */
+_GL_CXXALIAS_SYS (swab, void, (const char *from, char *to, ssize_t n));
+# else
_GL_CXXALIAS_SYS (swab, void, (const void *from, void *to, ssize_t n));
+# endif
# endif
_GL_CXXALIASWARN (swab);
#endif
prev parent reply other threads:[~2021-06-20 15:22 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-29 8:23 swab wrapper fails mingw.org's MinGW compilation of GDB Eli Zaretskii
2021-06-20 15:22 ` Bruno Haible [this message]
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
List information: https://lists.gnu.org/mailman/listinfo/bug-gnulib
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4157125.b07XYjAfJy@omega \
--to=bruno@clisp.org \
--cc=bug-gnulib@gnu.org \
--cc=eliz@gnu.org \
/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).