William Pursell wrote:
> I recall having an issue with git 1.4 in which
> git log did not produce correct chronology,
> and rpm would reject ChangeLogs generated from
> git log because they were out of order. I cannot
> duplicate that problem with modern git, but I
> notice in the mailing list ar
William Pursell wrote:
> This is a simple modification to gitlog-to-changelog
> that allows arguments to be passed to git-log
> to allow finer control over construction
> of the ChangeLog. The intended use case is to allow
> a ChangeLog to be built for a branch without requiring
> that branch to
When gnulib redefines mbstate_t, it also needs to override the system's
mbsrtowcs implementation. Needed in particular on AIX.
2008-12-20 Bruno Haible
* lib/wchar.in.h (mbsrtowcs): Redefine if REPLACE_MBSRTOWCS is set.
* m4/mbsrtowcs.m4 (gl_FUNC_MBSRTOWCS): Invoke gl_MBSTATE_T
I recall having an issue with git 1.4 in which
git log did not produce correct chronology,
and rpm would reject ChangeLogs generated from
git log because they were out of order. I cannot
duplicate that problem with modern git, but I
notice in the mailing list archive some concern
about gitlog-to-c
This is a simple modification to gitlog-to-changelog
that allows arguments to be passed to git-log
to allow finer control over construction
of the ChangeLog. The intended use case is to allow
a ChangeLog to be built for a branch without requiring
that branch to be checked out.
diff --git a/build-
The unit test for mbrtowc uncovered four different bugs in various systems of
the system-provided mbrtowc function.
mbrtowc was introduced to allow traversing strings that are stored in
non-contiguous memory blocks. This essential functionality does not
work on AIX 5.1, HP-UX 11.11, OSF/1 5.1, Sol
> New module 'wctob'.
> * lib/wchar.in.h (wctob): New declaration.
> * lib/wctob.c: New file.
> * m4/wctob.m4: New file.
> * modules/wctob: New file.
wctob does not work on Solaris 9 and older: It simply fails to recognize
single-byte characters, even in a simple ISO-
Tom G. Christensen wrote:
> + rm -f t-select-in.tmp
> + ./test-select-fd r 0 t-select-in.tmp
> + cat t-select-in.tmp
> + test 0 = 1
> + exit 1
Thanks. So apparently Solaris 2.6 has a problem with select() of /dev/null.
I verified that the test passes on Solaris 7.
It's probably not worth wor
> > Indeed. The mbstate_t type and mbsinit etc. functions were not part of
> > C89 + appendices, they were introduced in C99. Therefore, the
> > -D_XOPEN_SOURCE=500 option should better be added to the detection of flags
> > for C99, not C89.
Ben Pfaff corrected me: mbstate_t, mbsinit etc. are par
Bruno Haible writes:
>> Also, your proposed patch only touched the C89 tests; what about C99?
>
> Indeed. The mbstate_t type and mbsinit etc. functions were not part of
> C89 + appendices, they were introduced in C99. Therefore, the
> -D_XOPEN_SOURCE=500 option should better be added to the detec
On Sat, Dec 20, 2008 at 03:47:41PM +0100, Bruno Haible wrote:
> > > Unconnected socket test... passed
> > > Connected sockets test... failed (expecting POLLHUP after shutdown)
> > > General socket test with fork... failed (expecting POLLHUP after
> > > shutdown)
> > > Pipe test... passed
> > > FAIL
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Bruno Haible on 12/20/2008 7:40 AM:
> Indeed. The mbstate_t type and mbsinit etc. functions were not part of
> C89 + appendices, they were introduced in C99. Therefore, the
> -D_XOPEN_SOURCE=500 option should better be added to the detecti
> 2008-12-18 Bruno Haible
>
> * m4/locale-ja.m4: New file, from GNU gettext.
This test exhibited a false positive: It reported that Cygwin supports a
ja_JP.EUC-JP locale. But this locale has MB_CUR_MAX = 1, leading to subsequent
test failures. This fixes it:
2008-12-20 Bruno Haible
Simon Josefsson wrote:
> > Unconnected socket test... passed
> > Connected sockets test... failed (expecting POLLHUP after shutdown)
> > General socket test with fork... failed (expecting POLLHUP after
> > shutdown)
> > Pipe test... passed
> > FAIL: test-poll
> > FAIL: test-select-in.sh
>
> I reca
Hi Eric,
> > On HP-UX 11.11, mbstate_t is not defined by unless _XOPEN_SOURCE
> > is
> > set to 500. (Even when _HPUX_SOURCE is set!)
>
> What happens if _XOPEN_SOURCE is set to something larger than 500,
> implying a newer standards version?
Then it is as if the flag was not present at all. O
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Bruno Haible on 12/20/2008 5:43 AM:
> On HP-UX 11.11, mbstate_t is not defined by unless _XOPEN_SOURCE is
> set to 500. (Even when _HPUX_SOURCE is set!)
What happens if _XOPEN_SOURCE is set to something larger than 500,
implying a newer
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Simon Josefsson on 12/20/2008 6:00 AM:
>> ./test-parse-duration.sh: -u3: is not an identifier
>> FAIL: test-parse-duration.sh
>
> The problem seems to be:
>
>while read -u3 line
>
> Possibly the -u parameter to read isn't portable?
"Tom G. Christensen" writes:
> test-flock.c:72: assertion failed, errno = 22
> FAIL: test-flock
Reported before, I think.
> ./test-parse-duration.sh: -u3: is not an identifier
> FAIL: test-parse-duration.sh
The problem seems to be:
while read -u3 line
Possibly the -u parameter to read isn
On HP-UX 11.11, mbstate_t is not defined by unless _XOPEN_SOURCE is
set to 500. (Even when _HPUX_SOURCE is set!)
I have submitted a patch so that autoconf sets this flag by default. But there
are so many INSTALL files which only recommend CC="cc -Ae", that IMO this
needs to be handled by the gnul
> New module 'wctob'.
> * lib/wchar.in.h (wctob): New declaration.
IRIX 6.5 has the wctob function but lacks its declaration. This patch provides
the missing declaration.
2008-12-20 Bruno Haible
Ensure wctob is declared on IRIX 6.5.
* lib/wchar.in.h (wctob): Decla
The compiler used was GCC 4.3.2 configured for native as/ld.
GNU gettext 0.17 and GNU libiconv 1.12 was installed and available.
Build flags set in env:
CPPFLAGS="-I/usr/tgcware/include"
LDFLAGS="-L/usr/tgcware/lib -Wl,-R,/usr/tgcware/lib"
Using Simons daily snapshot the results are:
21 matches
Mail list logo
