-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Bruno Haible on 12/16/2009 6:17 PM:
> In ANSI C, but not in C99, it is an error to typedef the same type twice:
> typedef int _gl_warn_on_use;
> typedef int _gl_warn_on_use;
>
> Therefore the last line is better changed to 'extern int
Eric Blake wrote:
> +# if 4 < __GNUC__ || (__GNUC__ == 4 && 3 <= __GNUC_MINOR__)
> +/* A compiler attribute is available in gcc versions 4.3.0 and later. */
> +# define _GL_WARN_ON_USE(function, message) \
> +extern __typeof__ (function) function __attribute__ ((__warning__ (message)))
> +
> +# e
Eric Blake byu.net> writes:
> But that appears to only work for C, since I cannot know in advance whether a
> macro name will have overloads in C++. A real case of this is for functions
> like strchr, which are reasonably typed one way in C:
>
> char *strchr (const char *, int);
>
> and two
> Sorry for the delay, and thanks again for the report. I finally took time
> on a NetBSD machine, and reproduced the problem by using CC='gcc -ansi'.
> This patch fixed the issue for me, so I'm applying it.
Eric,
Thanks a lot and Happy Christmas!
Alex
Hi Bruno.
José Marchesi reported that the 'list' and 'oset' container data structures
are not really usable in libraries, because they call xalloc_die() when
an out-of-memory condition occurs. An xalloc_die() that calls exit() is
not usable in a library. An xalloc_die() that calls set
Bruno Haible clisp.org> writes:
> There is one trick needed, though. Suppose the system has a function which may
> be declared as
> int foo (void *p);
> or int foo (char *p);
> How can we add the attribute?
> - The trick is to write
> #if HAVE_DECL_FOO
> extern __typeof__(foo)
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Eric Blake on 11/17/2009 6:12 AM:
> According to Alexander Nasonov on 11/17/2009 2:37 AM:
>> it fails here:
>
>> fflush.c(95): error: a value of type "__off_t={__int64_t={long long}}"
>> cannot be assigned to an entity of type "fpos_t"
>
Bruno Haible writes:
> Simon Josefsson wrote:
>> > If not, then take a look at gltests/Makefile. It should define
>> > ARG_NONNULL_H.
>>
>> No, there is
>>
>> stdlib.h: stdlib.in.h $(LINK_WARNING_H) $(ARG_NONNULL_H)
>>
>> but ARG_NONNULL_H is never defined in the Makefile.
>
> Did you somehow
Simon Josefsson wrote:
> > If not, then take a look at gltests/Makefile. It should define
> > ARG_NONNULL_H.
>
> No, there is
>
> stdlib.h: stdlib.in.h $(LINK_WARNING_H) $(ARG_NONNULL_H)
>
> but ARG_NONNULL_H is never defined in the Makefile.
Did you somehow pass --exclude=arg-nonnull to the g
Eric Blake wrote:
> According to Eric Blake on 12/15/2009 7:48 PM:
>> According to John Stanley on 12/15/2009 4:42 PM:
>>> Basically, what's happening is that 'touch -a ..' updated ctime in
>>> coreutils-7.6,
>>> but does not update ctime in coreutils-8.2 (hence misc/ls-time fails).
>>
>> Ouch. Th
Eric Blake wrote:
> [adding bug-gnulib]
>
> According to Eric Blake on 12/15/2009 7:48 PM:
>> According to John Stanley on 12/15/2009 4:42 PM:
>>> Basically, what's happening is that 'touch -a ..' updated ctime in
>>> coreutils-7.6,
>>> but does not update ctime in coreutils-8.2 (hence misc/ls-time
Bruno Haible writes:
> Hi Simon,
>
>> The cause seems to be a un-expanded _GL_ARG_NONNULL:
>>
>> j...@mocca:~/gnutls4win$ grep 'NONNULL is copied'
>> build/libidn-1.16/gltests/stdlib.h
>> /* The definition of _GL_ARG_NONNULL is copied here. */
>
> _GL_ARG_NONNULL should be defined in the line
12 matches
Mail list logo
