Paul Eggert wrote on 2025-02-05:
> At first I was worried that the patch relied on Haiku errno values not
> being -1, since the Gnulib documentation said only that they can be
> negative. However, I discovered that Haiku errno values are all close to
> INT_MIN, so this isn't
Paul, Bruno, Eli,
thanks for your help.
Something went wrong installing Bruno's Haiku compatibility check to
'patch', so I've added that now.
Andreas
Thanks for catching my portability error. I installed that into 'patch'
master.
At first I was worried that the patch relied on Haiku errno values not
being -1, since the Gnulib documentation said only that they can be
negative. However, I discovered that Haiku errno values are al
On Haiku, the regex test runs into a libc-internal assertion (both from
regexec.c:1377:pop_fail_stack: num >= 0
and from glibc bug 11053. This is annoying, because it requires human
intervention for the build to continue.
Fixed like this:
2024-10-26 Bruno Haible
regex: Av
On Haiku, I see this test failure:
FAIL: test-unlockpt
===
../../gltests/test-unlockpt.c:36: assertion 'unlockpt (-1) == -1' failed
Abort
FAIL test-unlockpt (exit status: 149)
Since POSIX does not require failure here (see
<https://pubs.opengroup.org/onlinepu
On Haiku, I see these test failures:
FAIL: test-futimens
===
../../gltests/test-futimens.h:123: assertion 'get_stat_atime_ns (&st1) ==
get_stat_atime_ns (&st2)' failed
Abort
FAIL test-futimens (exit status: 149)
FAIL: test-utime
../../gltes
On Haiku, I'm seeing this test failure (in a testdir of
posix_openpt grantpt pt_chown
):
FAIL: test-posix_openpt
===
../../gltests/test-posix_openpt.c:58: assertion '0 <= slave' failed
Abort
FAIL test-posix_openpt (exit status: 149)
This patch fixes it:
On Haiku, I'm seeing a compilation error in the unit test of mbrtoc16.
The cause is that mbrtoc16 is not declared, because of this typo and
because Haiku has mbrtoc32 but not mbrtoc16.
2024-08-13 Bruno Haible
mbrtoc16: Fix compilation error on Haiku.
* lib/uchar
On Haiku, I see a test failure of test-string-desc-quotearg.
The reason is that on this platform, the C locale has UTF-8 encoding,
and therefore the quotearg.c logic uses U+2018 and U+2019 as quotation
marks, instead of the ASCII double-quote.
2023-04-05 Bruno Haible
string-desc
On Haiku, in a testdir for 'random_r', I see this compilation error:
g++ -DHAVE_CONFIG_H -DEXEEXT=\"\" -I. -I.. -DGNULIB_STRICT_CHECKING=1
-DIN_GNULIB_TESTS=1 -I. -I. -I.. -I./.. -I../gllib -I./../gllib -Wno-error -g
-O2 -MT test-stdlib-c++.o -MD -MP -MF $depbase.Tpo -c
On Haiku I see a compilation error, because EDOOFUS and EINVAL have the
same value.
This patch fixes it.
2023-04-05 Bruno Haible
strerrorname_np: Fix compilation error on Haiku.
* lib/strerrorname_np.c (strerrorname_np): Avoid conflict between
EDOOFUS and EINVAL
On Haiku (2017), I'm seeing a compilation error in a testdir:
gcc-x86 -DHAVE_CONFIG_H -DEXEEXT=\"\" -DEXEEXT=\"\" -DNO_XMALLOC -DEXEEXT=\"\"
-I. -I../../gllib -I.. -DGNULIB_STRICT_CHECKING=1
-I/boot/home/config/non-packaged/include -Wall -fvisibility=hidden
7;verify'
verify ((char32_t)(-1) >= 0);
^~
The reason is that the char16_t and char32_t types are incorrectly defined.
Reported at <https://dev.haiku-os.org/ticket/15990>. But since the bug exists
already for two years and is fatal, let me add a workaround.
2020-05-09 Bruno Hai
ate)
^~~~
The reason is that these functions are defined as static inline functions,
thus AC_CHECK_FUNCS reports that these function were not defined.
These two patches fix the problems.
2020-05-08 Bruno Haible
c32rtomb: Avoid compilation failure on Haiku.
* m4/c32rtomb.m4 (gl_F
aluated to 0
in the preprocessor directive. This is because the system's
does #include_next and thus ends up including gnulib's limits.h
recursively.
This patch fixes it.
2020-05-08 Bruno Haible
limits-h: Define LONG_BIT correctly on Haiku/x86_64.
* lib/limits.in.h: Define
On Haiku, a testdir built with --with-c++-tests produces the following errors:
../gllib/stdio.h:1736:1: error: 'vdprintf' was not declared in this scope
_GL_CXXALIASWARN (vdprintf);
^
../gllib/unistd.h: In member function 'gnulib::_gl_usleep_wrapper::operator
gnulib::_gl_uslee
On Haiku, the fcntl tests fail. Reason:
1) fcntl (oldfd, F_DUPFD_CLOEXEC, newfd)
sets the FD_CLOEXEC flag on oldfd, not on newfd (!).
2) Additionally, if newfd < 0, it fails with errno = EMFILE,
where POSIX wants errno = EINVAL.
This fixes it. I'm too lazy to write an autoconf
t status: 134)
>
> FAIL: test-unlinkat
> ===
>
> ../../gltests/test-unlinkat.c:72: assertion 'errno == EBADF' failed
> Abort
> FAIL test-unlinkat (exit status: 134)
>
> FAIL: test-unlockpt
> ===
>
> ../../
On a recent Haiku (32-bit system), a POSIX testdir shows 44 failures:
FAIL: test-areadlinkat
==
../../gltests/test-areadlinkat.c:78: assertion 'errno == EBADF' failed
Abort
FAIL test-areadlinkat (exit status: 134)
FAIL: test-cloexec
==
../../gltests/test-clo
On Haiku, the functions inet_ntop, inet_pton require linking with "-lnetwork".
But INET_NTOP_LIB and INET_PTON_LIB come out as empty. This leads to a
link error:
gcc-x86 -g -O2 -L/boot/home/config/lib -o test-inet_ntop test-inet_ntop.o
libtests.a ../gllib/libgnu.a libtests
[CCing haikuports list]
Benno Schulenberg wrote in
<https://lists.gnu.org/archive/html/bug-gnulib/2017-07/msg8.html>:
> Trying to build nano from git on Haiku failed because, among other
> things, in the gnulib module glob, a variable is declared after a
> bit of code and the
Hi,
Trying to build nano from git on Haiku failed because, among other
things, in the gnulib module glob, a variable is declared after a
bit of code and the gcc-2* compiler on Haiku does not permit that.
(This is on a 32-bit machine.)
Attached is a patch that avoids the issue. (For my build
On 01/25/2011 10:38 PM, Eric Blake wrote:
> On 01/25/2011 10:02 PM, Eric Blake wrote:
>> test-dup2 and test-fcntl are failing on Haiku when using both the dup2
>> and cloexec modules in tandem. I tested on Haiku alpha 2 (the latest
>> formal release per haiku-os.org) rather
On 01/25/2011 10:02 PM, Eric Blake wrote:
> test-dup2 and test-fcntl are failing on Haiku when using both the dup2
> and cloexec modules in tandem. I tested on Haiku alpha 2 (the latest
> formal release per haiku-os.org) rather than a nightly build, so this
> may have already b
test-dup2 and test-fcntl are failing on Haiku when using both the dup2
and cloexec modules in tandem. I tested on Haiku alpha 2 (the latest
formal release per haiku-os.org) rather than a nightly build, so this
may have already been fixed by upstream development. But it seems like
I should be
ChangeLog
index 54c4598..6cf52b7 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,11 @@
+2011-01-19 Eric Blake
+
+ binary-io: fix O_TEXT on Haiku
+ * modules/binary-io (Depends-on): Add fcntl-h.
+ * lib/binary-io.h (O_TEXT): Rely on replacement rather
+ than blindly un
the 1.4.12 header)
>>> http://ports.haiku-files.org/browser/haikuports/trunk/sys-devel/m4/patches/m4-1.4.12.diff
>>> Part of the 1.4.12 patch was upstreamed.
Thanks for the link. First, I'm going to focus on the efforts to fix
m4's use of SIGBUS:
>> Is SIGBUS really
utils from CC]
>
>> > Haiku, like BeOS have O_BINARY and O_TEXT but they have no effect:
>> > http://www.mail-archive.com/bug-gnulib@gnu.org/msg12137.html
>
> Yes, and gnulib's binary-io.h undefines O_BINARY and O_TEXT on Haiku for this
> reason.
>
>> >
Hi Scott,
> I have opened a handful of tickets in Haiku's trac to cover them:
>
> http://dev.haiku-os.org/ticket/6704
> http://dev.haiku-os.org/ticket/6705
> http://dev.haiku-os.org/ticket/6706
> http://dev.haiku-os.org/ticket/6707
Nice. For the unsetenv one, you can also attach the test program
[dropped bug-diffutils from CC]
> > Haiku, like BeOS have O_BINARY and O_TEXT but they have no effect:
> > http://www.mail-archive.com/bug-gnulib@gnu.org/msg12137.html
Yes, and gnulib's binary-io.h undefines O_BINARY and O_TEXT on Haiku for this
reason.
> > --- diffutil
scott mc wrote:
> Diffutils-3.0 builds ok on Haiku but the test suite did not. I made
> patches to three of the test files as shown here:
Thanks for the report and patches.
I'm redirecting this to bug-gnulib, since those tests
all come from that project, not diffuti
Jim Meyering wrote:
> > OK to make the test a bit weaker?
>
> Sure. It'd be nice to mention Haiku in the log,
Done.
Bruno
Bruno Haible <[EMAIL PROTECTED]> wrote:
> Hi Richard, Jim,
>
> The test-fsync test is failing on Haiku, at line 40. fsync (stdin), when
> stdin refers to a terminal input, simply succeeds on this system. I don't
> see anything in POSIX which would mandate a failure of thi
Hi Richard, Jim,
The test-fsync test is failing on Haiku, at line 40. fsync (stdin), when
stdin refers to a terminal input, simply succeeds on this system. I don't
see anything in POSIX which would mandate a failure of this call.
OK to make the test a bit weaker?
2008-11-15 Bruno H
Bruno Haible <[EMAIL PROTECTED]> writes:
> Hi Simon,
>
> I'm seeing these compilation errors on Haiku:
>
> /data/testdir1/gllib/getaddrinfo.c: In function `getaddrinfo':
> /data/testdir1/gllib/getaddrinfo.c:183: `AI_NUMERICSERV' undeclared (first
> us
Hi Simon,
I'm seeing these compilation errors on Haiku:
/data/testdir1/gllib/getaddrinfo.c: In function `getaddrinfo':
/data/testdir1/gllib/getaddrinfo.c:183: `AI_NUMERICSERV' undeclared (first use
in this function)
/data/testdir1/gllib/getaddrinfo.c:183: (Each undeclared identif
On Haiku I'm seeing these errors:
/data/testdir1/gllib/spawn.h:110: field `_sp' has incomplete type
/data/testdir1/gllib/spawnattr_setschedpolicy.c: In function
`posix_spawnattr_setschedpolicy':
/data/testdir1/gllib/spawnattr_setschedpolicy.c:28: `SCHED_OTHER' undeclared
37 matches
Mail list logo
