On Fri, Sep 02, 2005 at 09:35:05AM -0500, Albert Chin wrote:
> On Fri, Sep 02, 2005 at 09:26:16AM +0200, Simon Josefsson wrote:
> > Albert Chin <[EMAIL PROTECTED]> writes:
> >
> > > On Thu, Sep 01, 2005 at 01:28:39PM +0200, Simon Josefsson wrote:
> > >&
tch below fixes this. Of course the 11.23/IA C compiler has a c99
mode. Should I juse use 'c99' instead?
--
albert chin ([EMAIL PROTECTED])
-- snip snip
2005-11-23 Albert Chin-A-Young <[EMAIL PROTECTED]>
lib/stdbool_.h: Don't define _Bool if compiler has both
todetect AC_PROG_CXX and build with the
C++ compiler only if AC_PROG_CXX has been called (AC_PROVIDE_IFELSE?).
--
albert chin ([EMAIL PROTECTED])
-- snip snip
--- mbstate_t.m4.orig 2005-01-26 00:38:40.0 -0600
+++ mbstate_t.m42005-11-26 00:36:42.285226000 -0600
@@ -13,7
On Fri, Nov 25, 2005 at 11:11:05PM -0800, Paul Eggert wrote:
> +#if defined __cplusplus || defined __BEOS__
> typedef bool _Bool;
> +#else
This will break on C++ compilers with bool _and_ _Bool, like the HP-UX
11.23/IA-64 aCC6 compiler.
--
albert chin ([EMAIL
On Sun, Nov 27, 2005 at 10:08:44PM -0800, Paul Eggert wrote:
> Albert Chin <[EMAIL PROTECTED]> writes:
>
> > We've currently solved it by implementing
> > separate defines depending on the language.
>
> That doesn't sound quite right. If you compile wit
On Mon, Nov 28, 2005 at 01:04:10PM -0800, Paul Eggert wrote:
> Albert Chin <[EMAIL PROTECTED]> writes:
>
> > I'm talking about _one_ project with two languages, C and C++. In
> > such a project, if human.h includes because of
> > HAVE_STDINT_H, how should
On Mon, Nov 28, 2005 at 04:43:06PM -0800, Paul Eggert wrote:
> Albert Chin <[EMAIL PROTECTED]> writes:
>
> > We tried this with gl_STDINT_H. However, the custom stdint_.h defines
> > uintmax_t which conflicts with gl_AC_TYPE_UINTMAX_T.
>
> That sounds like a pr
ed in /usr/include.dtk but not available in the shared
library.
--
albert chin ([EMAIL PROTECTED])
ory leak
IY67450 2 hostmibd dumped core on _findiop function
IY73547 2 _load_secmethod does not unlock _sec_rmutex for KRB5A module
IY72109 2 syslog_r() does not set errno to EINVAL
IY75013 2 Problem with wordexp() when SIGCHLD is caught
IY74171 1 libmemops.a doesn't support 64 bit
rte.libc 5.1.0.55C Flibc Library
>
> So they fixed strnlen between these library versions.
Looking at the bos.rte.libc fixed bugs list:
IY62952 2 strnlen can return results longer than the limit
I have no idea which revision of bos.rte.libc this is in or if the
above is the fix.
--
albert chin ([EMAIL PROTECTED])
ally used (this is the '-z
> > ignore' option of the Solaris linker).
> >
> > At some point I was going to propose to move this into gnulib.
>
> Does the GNU linker support this feature too? ld-2.16.1 does not document
> a "-z ignore" option.
--as-needed.
--
albert chin ([EMAIL PROTECTED])
#endif
The best way to test for getaddrinfo() is to compile a small program
using the function and include .
--
albert chin ([EMAIL PROTECTED])
isn't available on HP-UX 10.20 nor 11.00.
2006-07-09 Albert Chin-A-Young <[EMAIL PROTECTED]>
* lib/mbchar.h: Use only if available, not
unconditonally.
--
albert chin ([EMAIL PROTECTED])
-- snip snip
Index: l
On Sun, Jul 09, 2006 at 09:41:40PM -0700, Paul Eggert wrote:
> Albert Chin <[EMAIL PROTECTED]> writes:
>
> > isn't available on HP-UX 10.20 nor 11.00.
>
> modules/mbchar says that you're supposed to include mbchar.h this way:
>
> #if HAVE_WCHAR_H &&a
n of the same name defined
in UnixWare 7.1.1 . We need this #include before the #define
below.
However, we don't want to #include on all platforms because
- Tru64 with Desktop Toolkit C has a bug: must be included before
.
- BSD/OS 4.1 has a bug: and must be included before
. */
#if HAVE_DECL_MBSWIDTH_IN_WCHAR_H
# include
#endif
--
albert chin ([EMAIL PROTECTED])
/include/va_list.h"
typedef char *va_list;
#line 3 "b.c"
va_copy(a,b)
So, if we #define va_copy() before , it's obvious that
is redefining it. But, to what? It looks like the only way
to get the builtin va_copy() to work on AIX 5.3 is to add
-qlanglvl=extc99 to CFLAGS?
--
albert chin ([EMAIL PROTECTED])
On Sun, Feb 04, 2007 at 07:52:08AM -0700, Eric Blake wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> According to Albert Chin on 2/4/2007 7:47 AM:
> > The IBM C compiler on AIX 4.3.3 (v6) doesn't like signed bit types:
> > gmake[1]: Entering direc
On Sat, Feb 03, 2007 at 06:35:53PM -0800, Paul Eggert wrote:
> Albert Chin <[EMAIL PROTECTED]> writes:
>
> > It looks like the only way
> > to get the builtin va_copy() to work on AIX 5.3 is to add
> > -qlanglvl=extc99 to CFLAGS?
>
> For what it's wo
On Mon, Feb 05, 2007 at 10:49:52AM -0600, Matthew Woehlke wrote:
> Paul Eggert wrote:
> >Albert Chin <[EMAIL PROTECTED]> writes:
> >>It looks like the only way
> >>to get the builtin va_copy() to work on AIX 5.3 is to add
> >>-qlanglvl=extc99 to CFLAGS
_FUNC_SNPRINTF (in m4/snprintf.m4) so
> that it detects the bug and replaces the broken snprintf with gnulib
> version.
glib, which is LGPL, has a AC_FUNC_SNPRINTF_C99 in acinclude.m4 you
might want to check out.
--
albert chin ([EMAIL PROTECTED])
Any reason all uses of #include aren't wrapped with #ifdef
HAVE_CONFIG_H like so:
#ifdef HAVE_CONFIG_H
# include
#endif
Some .c files have the above and some don't. Seems to make sense for
all to have the form above.
--
albert chin ([EMAIL PROTECTED])
ot
have a C99-compliant vsnprintf(). The attached program returns -1 on
Solaris 6/SPARC thru Solaris 9/SPARC which has vsnprintf(), for
example.
--
albert chin ([EMAIL PROTECTED])
#include
#include
#include
int
doit(char * s, ...)
{
char buffer[32];
va_list args;
int r;
/* AIX 5.1 and S
On Mon, Feb 19, 2007 at 03:53:10PM -0800, Ben Pfaff wrote:
> Albert Chin <[EMAIL PROTECTED]> writes:
>
> > Huh? So, if vsnprintf() exists, you assume it's C99-compliant? glib
> > has a check for this, AC_FUNC_VSNPRINTF_C99, because that's not the
> >
On Mon, Feb 19, 2007 at 04:47:47PM -0800, Ben Pfaff wrote:
> Albert Chin <[EMAIL PROTECTED]> writes:
>
> > On Mon, Feb 19, 2007 at 03:53:10PM -0800, Ben Pfaff wrote:
> >> Albert Chin <[EMAIL PROTECTED]> writes:
> >>
> >> > Huh? So, if v
On Fri, Feb 23, 2007 at 12:26:00AM +0100, Jim Meyering wrote:
> Albert Chin <[EMAIL PROTECTED]> wrote:
> > On Thu, Feb 22, 2007 at 01:22:31AM +0100, Jim Meyering wrote:
> >>
> >> Bad timing. I was about to make a test release.
> >> If anyone else
LD_LIBRARY_PATH.
--
albert chin ([EMAIL PROTECTED])
files in the current directory?
> Because if yes, that would prevent parallel 'make install'.
> Or it could leave behind object files? Sorry, so far I wasn't able
> to see easily in which directory this compilation command would be
> executed.
And, if $compile_command can be libtool, you'll definitely run into
problems, especially on AIX with -export-dynamic.
--
albert chin ([EMAIL PROTECTED])
-UX 10.20: # # . . . . #
HP-UX 10.20's libc has snprintf() but there is _no_ function prototype
for it.
--
albert chin ([EMAIL PROTECTED])
d is already POSIX compliant.
> gl_cv_func_vsnprintf_posix=yes
> fi
> fi
> if test $gl_cv_func_vsnprintf_posix = no; then
> if ! expr "$gl_cv_func_printf_directive_a" : ".*yes" > /dev/null; then
> AC_DEFINE([NEED_PRINTF_DIRECTIVE_A], 1,
>
fix this? There is a -nodtk option to the commercial C
compiler which reverts to the system cc but that would need to be done
for _most_ gnulib-using programs, something that is not desirable.
--
albert chin ([EMAIL PROTECTED])
On Wed, Apr 04, 2007 at 09:14:12PM -0600, Eric Blake wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> According to Albert Chin on 4/4/2007 8:59 PM:
> >
> > Unfortunately, "#include_next " doesn't include
> > /usr/include/stdio.h. I
On Wed, Apr 04, 2007 at 10:45:44PM -0500, Albert Chin wrote:
> On Wed, Apr 04, 2007 at 09:14:12PM -0600, Eric Blake wrote:
> > -BEGIN PGP SIGNED MESSAGE-
> > Hash: SHA1
> >
> > According to Albert Chin on 4/4/2007 8:59 PM:
> > >
> > > Un
header file
#line 92 "/usr/include/math.h"
* The ANSI standard requires that certain values be in math.h.
* ANSI required entries in math.h
* prototype in math.h.
#line 1324 "/usr/include/math.h"
#line 1358 "/usr/vac/include/math.h"
This results in:
S["NEXT_MATH_H"]="\"///usr/vac/include/math.h\""
But this is wrong. What we want is:
S["NEXT_MATH_H"]="\"///usr/include/math.h\""
else HUGE_VAL isn't found when lib/strtod.c is compiled.
Maybe we should use -C only when -E produces empty output?
--
albert chin ([EMAIL PROTECTED])
On Sun, Nov 09, 2008 at 03:15:22PM +0100, Bruno Haible wrote:
> Albert Chin wrote:
> > > The fix that I found is to use "xlc -E -C" instead of "xlc -E". It has the
> > > effect of turning the config.status lines
> > >
> > > S[&
ach of these functions.
>
> I'll therefore extend the modules in a way that they work also on older
> systems without HAVE_MBRTOWC.
>
> Something like this. Probably also adding replacements for btowc and mbrtowc.
>
> [ snip snip ]
Any plans to commit the patch?
--
albert chin ([EMAIL PROTECTED])
t;///usr/include/wchar.h"
# 33 "///usr/include/wchar.h"
# 1 "./wchar.h"
# 1 "./wchar.h"
# 42 "///usr/include/wchar.h"
# 58 "./wchar.h"
$ cc -DHAVE_CONFIG_H -I. -I.. -I../intl -g -E btowc.c | \
grep wint_t | head -2
typedef
The Sun C++ compiler fails to compile the test case in m4/mktime.m4
correctly because of the K&R style function prototypes. Patch attached.
--
albert chin (ch...@thewrittenword.com)
diff --git a/m4/mktime.m4 b/m4/mktime.m4
index 5faf393..75e55c5 100644
--- a/m4/mktime.m4
+++ b/m4/mktim
On Wed, Jan 14, 2009 at 07:42:50PM +0100, Jim Meyering wrote:
> Albert Chin wrote:
> > The Sun C++ compiler fails to compile the test case in m4/mktime.m4
> > correctly because of the K&R style function prototypes. Patch attached.
>
> Hi Albert,
>
> That patch
some platforms, this test will
fail.
On RHEL4, g++ returns:
error: invalid conversion from `unsigned char*' to `char*'
because the `translate' member is char *. RHEL5 work fine as the
`translate' member is unsigned char *.
--
albert chin (ch...@thewrittenword.com)
lib/arpa_inet.in.h needs the usual:
#ifdef __cplusplus
extern "C" {
#endif
...
#ifdef __cplusplus
}
#endif
to work under C++. Patch attached.
--
albert chin (ch...@thewrittenword.com)
>From 36f93105aace962d70703887bfa633ca0c0f0638 Mon Sep 17 00:00:00 2001
From:
t expects a function or variable name
#pragma optional wcrtomb
^
1 error detected in the compilation of "getdate.c".
gmake[2]: *** [getdate.o] Error 2
$ cd gllib
$ cc -DHAVE_CONFIG_H -I. -I.. -I../intl -g -E getdate.c | \
grep mbstate_t | head -2
extern size_t rpl_wcrtomb (char *s, wchar_t wc, mbstate_t *ps);
typedef charmbstate_t;
How about adding something like "|| (defined __sgi && defined
_GL_JUST_INCLUDE_SYSTEM_WCHAR_H)" to your patch?
--
albert chin (ch...@thewrittenword.com)
INTF_SIZE1 will have the same problem.
--
albert chin (ch...@thewrittenword.com)
b/2009-01/msg00098.html
--
albert chin (ch...@thewrittenword.com)
Patch attached to update wording for m4/include_next.m4 to reveal recent
IBM PTF's that fix the #include_next bug in IBM C 9.0/10.1.
--
albert chin (ch...@thewrittenword.com)
>From 9095347bf98b43fce0268d72bef47e74bca1eec1 Mon Sep 17 00:00:00 2001
From: Albert Chin-A-Young
Date: Fri, 13
.. -g -c dummy.c
source='strtoull.c' object='strtoull.o' libtool=no \
DEPDIR=.deps depmode=hp /bin/sh ../build-aux/depcomp \
cc -Ae -DHAVE_CONFIG_H -I. -I.. -g -c strtoull.c
cpp: "stdlib.h", line 52: error 4036: Can't open include file 'stdint.h
On Tue, Mar 24, 2009 at 02:27:20PM +0100, Simon Josefsson wrote:
> Eric Blake writes:
>
> > According to Simon Josefsson on 3/24/2009 1:48 AM:
> >> Albert Chin writes:
> >>
> >>> With latest gnulib on HP-UX 10.20:
> >>> $ ./gnulib
doesn't work.
So, do we need to compensate for this. I hacked up something to work
around this and have attached it.
--
albert chin ([EMAIL PROTECTED])
Index: lib/string_.h
===
--- lib/string_.h.orig 2007-09-01 16:44:18.80505143
xhibit different behavior. I think gnulib should
accommodate this. But, I think it should be supported on a
case-by-case basis. That is, if someone reports a problem building
some mixed C/C++ project against gnulib, we fix that part of gnulib
rather than some major overhaul. This would be the least intrusive and
require the least amount of work.
--
albert chin ([EMAIL PROTECTED])
rinfo
#else
#define getaddrinfo ogetaddrinfo
#endif
The correct method to test for getaddrinfo on this system is to
include and try linking a program using getaddrinfo(). Patch
attached.
--
albert chin ([EMAIL PROTECTED])
Index: m4/getaddrin
AC_TRY_LINK([
#ifdef HAVE_WS2TCPIP_H
#include
#endif
], [getaddrinfo(0, 0, 0, 0);], gl_cv_w32_getaddrinfo=yes)
So, the above should be fixed as well.
--
albert chin ([EMAIL PROTECTED])
ed earlier this year by Jim Meyering:
http://lists.gnu.org/archive/html/bug-gnulib/2007-01/msg00374.html
The attached patch addresses the problem in the same way.
2007-10-17 Albert Chin-A-Young <[EMAIL PROTECTED]>
Jim Meyering <[EMAIL PROTECTED]>
* lib/str
uot;, line 249: warning 2012: Unrepresentable preprocessor
number 1LL.
cpp: "test-stdint.c", line 249: error 4038: Bad syntax for #if condition.
*** Error exit code 1
The problem is the following:
#if INTMAX_MIN && INTMAX_MAX && UINTMAX_MAX
/* ok */
#else
err o
ing
division by zero.
cc: "../../gltests/test-vasprintf-posix.c", line 3307: warning 548: Floating
division by zero.
tests/test-vasprintf-posix.c:518: has:
{ /* Negative zero. */
char *result;
int retval =
my_asprintf (&result, "%La %d", -0.0L, 33, 44, 55);
ASSERT (result != NULL);
if (have_minus_zero ())
ASSERT (strcmp (result, "-0x0p+0 33") == 0);
ASSERT (retval == strlen (result));
free (result);
}
The value of result is "0x0p+0 33".
--
albert chin ([EMAIL PROTECTED])
dir/test-fseeko.sh"
FAIL: test-fseeko.sh
../../gltests/test-ftello.c:94: assertion failed
../../gltests/test-ftello.sh: line 3: 10806 Aborted
./test-ftello${EXEEXT} 1 <"$srcdir/test-ftello.sh"
FAIL: test-ftello.sh
PASS: test-lseek.sh
PASS: test-stdio
PASS: test-unist
ong).
Sounds ok to me. The HP 10.20 preprocessor will never support long long.
So, we either abandon support for it or remove the tests. The HP C
compiler is fine with long longs, just not the preprocessor.
--
albert chin ([EMAIL PROTECTED])
| #include_next
...
Compiler bug? It seems like #include_next isn't working. Oddly, if I
modify fcntl.h so it's just:
#include_next
then everything works. But, if I precede the #include_next with other
#include directives, the error above occurs. Definitely seems like a
compile
] ... [-f Script_file] ... [File ...]
executing aclocal -I glm4
...
If I make sure GNU sed as 'sed' is in $PATH, so /bin/sed is not used,
it works. Haven't yet found the culprit though.
--
albert chin ([EMAIL PROTECTED])
On Thu, Sep 04, 2008 at 12:26:43PM -0500, Albert Chin wrote:
> [[ snip snip ]]
>
> Compiler bug? It seems like #include_next isn't working. Oddly, if I
> modify fcntl.h so it's just:
> #include_next
> then everything works. But, if I precede the #include_next with
On Thu, Sep 04, 2008 at 08:40:42PM +0200, Ralf Wildenhues wrote:
> * Albert Chin wrote on Thu, Sep 04, 2008 at 07:33:36PM CEST:
> > $ ./gnulib-tool --test --with-tests fcntl
> [...]
> > sed: 0602-429 No editing script was provided.
> > Usage: sed [-n] Script [File ..
ocale);
or:
r = strtold_l (nptr, endptr, locale);
Where is strtod_l/strtold_l defined? I can't find it in any source file.
--
albert chin (ch...@thewrittenword.com)
which are supposed to be
> defined by the system.
>
> Here's a proposed fix:
Thanks. The patch works.
--
albert chin (ch...@thewrittenword.com)
-e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \
-e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)' \
128:< $(srcdir)/wchar.in.h; \
So, 101 lines which causes sed to fail.
I suppose we need to require AC_PROG_SED and use $SED instead of sed
accordingly?
--
albert chin (ch...@thewrittenword.com)
On Tru64 UNIX 4.0D and 5.1, with the commercial Compaq Desktop
Toolkit, needs .
Patch below.
--
albert chin ([EMAIL PROTECTED])
-- snip snip
Index: lib/regex.c
===
RCS file: /cvsroot/gnulib/gnulib/lib/regex.c,v
retrieving
f all three tests terminate OK,
> dnl 0.5 if the first test terminates OK but the second one loops endlessly,
> dnl 0.0 if the first test already loops endlessly.
> ===
Does it make sens to add the compiler y
x27;ve added support for once-only execution (? la pthread_once) to
> lock.h and lock.c.
I'd review this:
http://mailman.mit.edu/pipermail/kerberos/2005-May/007723.html
It details how the weak referenced stubs in libc on some systems work
with pthread_once().
--
a
a.google.com/group/comp.sys.sgi.admin/browse_thread/thread/4586067cd05bb6cb/d971cc2459ec0ded?q=pthread+group:comp.sys.sgi.admin+author:albert+author:chin&rnum=1&hl=en#d971cc2459ec0ded
--
albert chin ([EMAIL PROTECTED])
___
bug-gnulib mailing li
quot;hidden"))) int hiddenfunc (void);
extern __attribute__((__visibility__("default"))) int exportedfunc (void);
$ xlc -c vis.c
[no error]
--
albert chin ([EMAIL PROTECTED])
___
bug-gnulib mailing list
bug-gnulib@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnulib
t which isn't available on
HP-UX. However, wget has the OpenSSL exception to the GPL and I don't
want to import strtoll and invalidate the wget license.
--
albert chin ([EMAIL PROTECTED])
___
bug-gnulib mailing list
bug-gnulib@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnulib
On Thu, Aug 11, 2005 at 08:20:42PM +0200, Simon Josefsson wrote:
> Albert Chin <[EMAIL PROTECTED]> writes:
>
> > I'm in need of this for strtoll() in wget which isn't available on
> > HP-UX. However, wget has the OpenSSL exception to the GPL and I don't
&g
echo '#define GTLS_SSIZE_T long' >>$outfile
fi
echo '#endif /* __GNUTLSCONFIG_H__ */' >>$outfile
])
--
albert chin ([EMAIL PROTECTED])
___
bug-gnulib mailing list
bug-gnulib@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnulib
break
])
done
done
if test "x$socklen_t_equiv" = x; then
AC_MSG_ERROR([Cannot find a type to use in place of socklen_t])
fi
])
AC_MSG_RESULT($socklen_t_equiv)
AC_DEFINE_UNQUOTED(socklen_t, $socklen_t_equiv,
On Thu, Sep 01, 2005 at 01:28:39PM +0200, Simon Josefsson wrote:
> Albert Chin <[EMAIL PROTECTED]> writes:
>
> > We created the following macro for curl. It's been tested on the
> > following systems:
>
> Has the copyright been assigned to the FSF?
No. I&#
On Fri, Sep 02, 2005 at 09:26:16AM +0200, Simon Josefsson wrote:
> Albert Chin <[EMAIL PROTECTED]> writes:
>
> > On Thu, Sep 01, 2005 at 01:28:39PM +0200, Simon Josefsson wrote:
> >> Albert Chin <[EMAIL PROTECTED]> writes:
> >>
> >> > We cre
0400
#define _IOUNGETC 01000
#define _IOINUSE02000 /* new flag for _THREAD_SAFE */
#define _IONONSTD 04000
#define _IOCLOSE01
#define _IOWINIT02
#define _IOWSET 04
...
What's the best way to fix this?
--
albert chin (ch...@thewrittenword.com)
On Sun, Jun 03, 2007 at 02:53:22PM +0100, Gary V. Vaughan wrote:
> [[Cc:ing Albert Chin: Albert, why do we need +Onofltacc in the
> standard hpux11.23
> flags, when it clearly causes NaN regressions?]]
>From the cc(1) man page:
+O[no]fltacc Disable [enable] fl
On Mon, Jun 04, 2007 at 04:24:26AM +0200, Bruno Haible wrote:
> Albert Chin wrote:
> > From the cc(1) man page:
> >+O[no]fltacc Disable [enable] floating-point optimizations that
> > can result in numerical differences.
> > ...
76 matches
Mail list logo
