* tests/test-glob.c (main): Enhance test.
* doc/posix-functions/glob.texi (glob): Document the fix.
Signed-off-by: Eric Blake
---
ChangeLog |6 +++
doc/posix-functions/glob.texi |4 ++
tests/test-glob.c | 78 ++--
3 fi
* lib/glob.c (glob): Lose register modifier. Drop redundant
code.
(globfree): Drop useless if-before-free.
(collated_compare): Add const-safety.
Signed-off-by: Eric Blake
---
ChangeLog |8
lib/glob.c | 13 +
2 files changed, 13 insertions(+), 8 deletions(-)
diff --g
From: Andreas Schwab
* posix/glob.c (glob): Match only directories when pattern ends in
a slash. Return no match when pattern is empty.
Reported via Paolo Bonzini.
Signed-off-by: Eric Blake
---
ChangeLog |7 +++
lib/glob.c | 47 ---
2 fil
Updating glob is turning out to be trickier than I thought.
For starters, m4/glob.m4 claims that the use of the glob module
implies that you want GNU semantics; which means that we should
really be creating glob-posix and glob-gnu, then making glob
an alias for glob-gnu.
Next, m4/glob.m4 claims th
* lib/unistd.in.h (includes): Only worry about if we
are overriding gethostname.
Suggested by Bruno Haible.
Signed-off-by: Eric Blake
---
> Would it be possible to conditionalize this with
> #if @GNULIB_GETHOSTNAME@
> like it was before? In a package that has nothing to do with
> sockets and
Eric Blake wrote:
> POSIX does not specify openpty, so this comment is misleading.
Oops, you're right. Fixed:
2010-03-21 Bruno Haible
Fix comments.
* lib/forkpty.c (rpl_forkpty): Fix comment.
* lib/openpty.c (rpl_openpty): Likewise.
Reported by Eric Blake.
--
Eric Blake wrote:
> +/* Get all possible declarations of gethostname(). */
> +#if @UNISTD_H_HAVE_WINSOCK2_H@ && !defined _GL_INCLUDING_WINSOCK2_H
> +# define _GL_INCLUDING_WINSOCK2_H
> +# include
> +# undef _GL_INCLUDING_WINSOCK2_H
> +#endif
>
Would it be possible to conditionalize this with
2010/3/22 Eric Blake :
> Since commit 4e86671a, gethostname failed to compile on mingw.
> Gnulib's sys/socket.h includes winsock2.h, which then includes
> unistd.h prior to declaring gethostname. The fix is to ensure
> that our replacement unistd.h does not declare any replacements
> until we are
On 03/22/2010 11:28 AM, Paolo Bonzini wrote:
> I think this should go in gnulib too, right?
>
> Paolo
>
> -- Forwarded message --
> From: Andreas Schwab
> Date: Mon, Mar 22, 2010 at 16:13
> Subject: [PATCH] Fix glob with empty pattern and patterns ending in slash
> To: libc-hac..
On 03/22/2010 11:56 AM, Simon Josefsson wrote:
>> Typo on line 26.
>
> Thanks, now it works. Any reason not to push this?
Only that I was not set up to reproduce the failure (cross-compiling
from cygwin was running into its own set of problems, because cygwin's
gcc-3 compiler mistakenly picks up
Eric Blake writes:
> On 03/22/2010 11:39 AM, Simon Josefsson wrote:
>> Eric Blake writes:
>>
>>> Since commit 4e86671a, gethostname failed to compile on mingw.
>>> Gnulib's sys/socket.h includes winsock2.h, which then includes
>>> unistd.h prior to declaring gethostname. The fix is to ensure
>
On 03/22/2010 11:39 AM, Simon Josefsson wrote:
> Eric Blake writes:
>
>> Since commit 4e86671a, gethostname failed to compile on mingw.
>> Gnulib's sys/socket.h includes winsock2.h, which then includes
>> unistd.h prior to declaring gethostname. The fix is to ensure
>> that our replacement unist
Eric Blake writes:
> Since commit 4e86671a, gethostname failed to compile on mingw.
> Gnulib's sys/socket.h includes winsock2.h, which then includes
> unistd.h prior to declaring gethostname. The fix is to ensure
> that our replacement unistd.h does not declare any replacements
> until we are su
I think this should go in gnulib too, right?
Paolo
-- Forwarded message --
From: Andreas Schwab
Date: Mon, Mar 22, 2010 at 16:13
Subject: [PATCH] Fix glob with empty pattern and patterns ending in slash
To: libc-hac...@sourceware.org
* posix/glob.c (glob): Match only directorie
Since commit 4e86671a, gethostname failed to compile on mingw.
Gnulib's sys/socket.h includes winsock2.h, which then includes
unistd.h prior to declaring gethostname. The fix is to ensure
that our replacement unistd.h does not declare any replacements
until we are sure that winsock2.h is completel
On 03/21/2010 01:45 PM, Matthias Bolte wrote:
> 2010/3/20 Eric Blake :
>> Since commit 4e86671a, gethostname failed to compile on mingw.
>> Gnulib's sys/socket.h includes winsock2.h, which then includes
>> unistd.h prior to declaring gethostname. Therefore, unistd.h
>> cannot replace gethostname u
On 03/21/2010 06:17 PM, Bruno Haible wrote:
> Here's a proposed patch to use a name gl_FUNC_XYZ for macros that deal with
> the replacement of a single function.
>
> Objections?
>
>
> 2010-03-21 Bruno Haible
>
> * m4/getlogin.m4 (gl_FUNC_GETLOGIN): Renamed from gl_GETLOGIN.
> * m
On 03/21/2010 05:10 PM, Bruno Haible wrote:
> -#if HAVE_DECL_OPENPTY
> +#if HAVE_OPENPTY
> +
> +/* Provider a wrapper with the precise POSIX prototype. */
> # undef openpty
> int
> -rpl_openpty (int *amaster, int *aslave, char *name, struct termios const
> *termp,
> - struct winsize con
On 03/21/2010 05:56 PM, Bruno Haible wrote:
> As an addendum to this patch, let's make the naming of macros more consistent:
>
>
> 2010-03-21 Bruno Haible
>
> pty: Consistent macro naming.
> * m4/pty_h.m4 (gl_PTY_H): Renamed from gl_PTY.
> * m4/pty.m4 (gl_FORKPTY, gl_OPENPTY
19 matches
Mail list logo
