On Thu, Dec 17, 2020 at 12:26:20AM +0100, Bruno Haible wrote:
So, it differs from 'create_temp_dir' only in the aspect that it does NOT
do automatic cleanup.
Since you are already aware of the 'clean-temp' module — you contributed to
it 8 years ago :) — what could we
* lib/tmpdir.c (create_tmp_dir): New function definition.
* lib/tmpdir.h (create_tmp_dir): New function declaration.
---
lib/tmpdir.c | 119 +++
lib/tmpdir.h | 20 +
2 files changed, 139 insertions(+)
diff --git a/lib/tmpdir.c b/lib/tmpdir.
On Sun, Sep 13, 2020 at 11:11:55PM +0200, Bruno Haible wrote:
The workaround is simple: Make sure that you include the corresponding
file -
here - before the struct definition. That is, in this case,
at the beginning of ags/object/ags_application_context.h, add
#
On Sun, Sep 13, 2020 at 01:12:28PM +0200, Bruno Haible wrote:
Hi John,
As this is your first major contribution to Gnulib, can you please make
sure
you have signed a copyright assignment with the FSF regarding 'GNULIB'?
Yes, I have.
> * lib/tmpdir.c (path_search_a
* lib/tmpdir.c (path_search_alloc): Define new function similar to
path_search, but which allocates the buffer for the result instead
of relying on the caller to preallocate it.
*lib/tmpdir.h (path_search_alloc): Declare it.
---
lib/tmpdir.c | 127 +++--
These are very valid and helpful points. I wasn't even aware of
"platform-testers".
Thanks to Ruth for reporting this problem and thanks to everyone
else for diagnosing it.
J'
On Fri, Aug 28, 2020 at 06:20:26PM -0400, Jeffrey Walton wrote:
The PSPP folks did not make an announcement
g directory `/home/ad/ad36/ad3658/Downloads/pspp-1.4.0'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/ad/ad36/ad3658/Downloads/pspp-1.4.0'
make: *** [all] Error 2
[ad3658@cus-survox1 pspp-1.4.0]$
This method is better for several reasons: 1. It avoids the danger of
cut and paste errors. 2. Naive translators can't translate the urls
(it's amazing how many do things like that!) 3. Should the urls ever
change, only this file has to be updated rather than every single
translation.
* lib/vers
I'm trying to use gnulib in the same project as ax_prog_cc_for_build.m4
from the autoconf archive[1]. But it seems that they do not play nicely
together.
Specifically, gnulibs std-gnu11.m4 seems to negate the effect of the former,
resulting in CC_FOR_BUILD always getting defined as the empty stri
In lib/error.c the function void error (int status, int errnum, const char
*message, ...)
has a va_start but no corresponding va_end.
This will leak memory.
J'
--
Avoid eavesdropping. Send strong encrypted email.
PGP Public key ID: 1024D/2DE827B3
fingerprint = 8797 A26D 0854 2EAB 0285 A2
Using gcc 4.4.5 I get *tons* of warnings from gnulib headers: Eg:
In file included from ./SOURCE/src/libpspp/argv-parser.c:26:
./SOURCE/gl/xalloc.h:176: error: no previous prototype for 'x2nrealloc'
Apparently this doesn't happen on gcc 4.7.2
The following patch fixes the problem for me. I do
On Thu, Jan 17, 2013 at 09:31:39AM -0700, Eric Blake wrote:
On 01/17/2013 08:52 AM, John Darrington wrote:
> Can gnulib's automake environment be used in a non-recursive manner?
> That is to say can I include the gnulib modules from my project's
> Makefil
Can gnulib's automake environment be used in a non-recursive manner?
That is to say can I include the gnulib modules from my project's
Makefile.am or is there no other way that using SUBDIRS=
Would a patch to make it possible be considered if I were to offer one?
J'
--
PGP Public key ID: 1024D/
On Mon, Dec 10, 2012 at 10:34:07AM -0800, Ben Pfaff wrote:
Right, I'll work on those.
What is vsnformat?
I meant to type "vsnprintf"
J'
--
PGP Public key ID: 1024D/2DE827B3
fingerprint = 8797 A26D 0854 2EAB 0285 A290 8A67 719C 2DE8 27B3
See http://keys.gnupg.net or any P
I gave it a rudimentary test and it seems to do what I want.
I will also need other related functions, though, Eg xasprintf,
vsnformat, etc.
J'
On Thu, Dec 06, 2012 at 10:59:44PM -0800, Ben Pfaff wrote:
John Darrington writes:
> Gnulib has a number of c-* variants o
Gnulib has a number of c-* variants of string processing functions,
eg c-strtod, c-strcasecmp etc But notably absent are any locale
independent printf routines. We could use some in PSPP.
So far as I can see, this could be achieved relatively easily, by
arranging for the decimal_point_char func
On Fri, Aug 24, 2012 at 02:27:35AM -0700, Paul Eggert wrote:
On 08/23/2012 11:53 PM, John Darrington wrote:
> 2. Since since that file comes from gnulib, we should recommend they do
the same upstream.
I did that just now, here:
http://git.savannah.gnu.
[Cross posting to bug-gnulib@gnu.org, bug-texi...@gnu.org and
bug-recut...@gnu.org]
On Fri, Aug 24, 2012 at 07:52:01AM +0200, Jose E. Marchesi wrote:
Hi John.
I suggest this patch.
We are following the advise of Karl Berry and not using @acronym{} in
the
From: John Darrington
Date: Sat, 28 Jul 2012 08:36:34 +0200
Subject: [PATCH] Fix memory leak in clean-temp.c
The two lists subdirs and files were never being freed. This change fixes that.
---
lib/clean-temp.c |2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/lib/clean
562fd Mon Sep 17 00:00:00 2001
From: John Darrington
Date: Tue, 19 Jun 2012 18:26:55 +0200
Subject: [PATCH] supports_delete_on_close: Initialise OSVIERSIONINFO struct per documentation
The microsoft documentation says that dwOSVersionInfoSize must be initialised
with the size of the struct. This w
On Thu, Jun 07, 2012 at 12:29:03AM -0700, Paul Eggert wrote:
On 06/07/2012 12:19 AM, John Darrington wrote:
> Surely using a symbol instead of a literal constant makes maintenance
easier not
> harder?
I don't see why. For example:
r
On Wed, Jun 06, 2012 at 01:07:41PM -0700, Paul Eggert wrote:
On 06/06/2012 12:43 PM, John Darrington wrote:
> If as you say, windows prefers \ is there a good reason NOT to use it?
Mostly to avoid the maintenance hassle.
Surely using a symbol instead of a literal const
On Wed, Jun 06, 2012 at 01:10:09PM -0600, Eric Blake wrote:
On 06/06/2012 12:54 PM, John Darrington wrote:
> I'm offering this patch which ensures that the paths of tempfiles are
> created with the appropriate separator for the machine.
What platform did you
public key.
From 7597c055a47a06a10f8f8d4e64cf2fd8131df7fe Mon Sep 17 00:00:00 2001
From: John Darrington
Date: Wed, 6 Jun 2012 20:46:45 +0200
Subject: [PATCH] Use the macro DIRECTORY_SEPARATOR instead of a literal / when building pathnames
---
lib/tempname.c |4 +++-
lib/tmpdir.c |
In PSPP we are using the path_search function from the tmpdir module to
determine the path for temporary files.
Windows users have complained that it puts temporary files in places they
consider to be inappropriate. I did a bit of research, and found that the
"normal" path on windows depends on
The comment in clean-temp.h says that temp files/directories will be
"automatically removed when the program exits either normally or through a
fatal signal" .
However, the code in clean-temp.c seems only to deal with the fatal signal case.
Consequently, in most cases, they don't get removed (ob
Cross compiling to host=i586-mingw32msvc I encountered 2 problems related to
include_next.m4.
Firstly I had to apply the attached patch to avoid configure complaining about
missing operands.
Secondly, I had to put gl_INCLUDE_NEXT in configure.ac which I was rather
surprised about.
Since it
This line in lib/siglist.h:
init_sig (SIGILL, "ILL", N_("Illegal instruction"))
contradicts the GNU Coding Standards which says:
`Please do not use the term "illegal" to refer to erroneous input to a
computer program. Please use "invalid" for this, and reserve the term
"illegal" for ac
On Tue, Nov 30, 2010 at 04:07:27AM +0100, Bruno Haible wrote:
Hi,
For a long time, we've written our test programs in *.m4 macros in such a
way
that when they fail, the return code is 1.
But often we have several tests, combined in a single program.
Example: m4
In lib/isblank.c I see the following:
/* The "blank" characters are '\t', ' ',
U+1680, U+180E, U+2000..U+2006, U+2008..U+200A, U+205F, U+3000, and none
except the first two is present in a common 8-bit encoding. Therefore
the substitute for other platforms is not more complicated
On Tue, Mar 24, 2009 at 11:18:12PM +0100, Bruno Haible wrote:
So, for numbers to strings, there are four ways to proceed:
1) Use the snprintf_l function available on MacOS X 10.5.
Drawback: Unportable, and gnulib cannot provide an easy replacement
for
lo
In pspp, we've been using the c-ctype module. Would it also be
possible to produce a snprintf function and friends which ignores
locale settings and always formats strings in the C locale?
J'
--
PGP Public key ID: 1024D/2DE827B3
fingerprint = 8797 A26D 0854 2EAB 0285 A290 8A67 719C 2DE8 27B3
The pspp build system includes the line
gnulib-tool ... --import ... unilbrk/ulc-width-linebreaks ...
My guess is that mingw is trying and failing to quote the
unilbrk/ulc-width-linebreaks string.
J'
On Wed, Oct 22, 2008 at 11:02:04AM -0300, Michel Boaventura wrote:
Hi Simon,
I
On Fri, Sep 28, 2007 at 12:59:09AM +0200, Bruno Haible wrote:
Hi,
John Darrington wrote:
> Would it be possible to add a feature to the gpl-[23].0 modules which
> provides the text of the GPL as a const char * ?
>
> This would be useful for use
Would it be possible to add a feature to the gpl-[23].0 modules which
provides the text of the GPL as a const char * ?
This would be useful for user interfaces which need to display the
licence text, such as
http://www.gtk.org/api/2.6/gtk/GtkAboutDialog.html#gtk-about-dialog-set-license
It might
On Tue, Apr 03, 2007 at 12:43:19AM +0200, Bruno Haible wrote:
Simon Josefsson wrote:
> Well, at least after I installed
> gperf, before that I got this error during build:
>
> make[2]: Entering directory `/home/jas/src/libidn/lib/gl'
> gperf -m 10 iconv_open-aix.gperf
It seems that the following patch needs to be applied, since
utf*-ucs4-unsafe.h no longer exist:
--- ../../gnulib/lib/linebreak.c2007-02-04 09:24:21.0
+0900
+++ linebreak.c 2007-03-26 18:16:47.0 +0800
@@ -26,9 +26,7 @@
#include "c-ctype.h"
#include "xsize.h"
-#includ
The instructions in progname.h say:
/* Programs using this file should do the following in main():
set_program_name (argv[0]);
*/
/* Set program_name, based on argv[0]. */
extern void set_program_name (const char *argv0);
but the GNU coding standards say:
"The program's name should be
I've been auditing our project for constness.
One that shows up from gnulib is is_basic_table from mbchar.c
This symbol should be const?
J'
--
PGP Public key ID: 1024D/2DE827B3
fingerprint = 8797 A26D 0854 2EAB 0285 A290 8A67 719C 2DE8 27B3
See http://pgp.mit.edu or any PGP keyserver for publ
39 matches
Mail list logo
