On 28/01/19 12:08 AM, Bruno Haible wrote:
> Hi Siddhesh,
>
>> Build a more general script that parses changes in a VCS and then
>> prints it in a ChangeLog format.
> Can you please give some more context?
Sorry about that, there was a pretty extensive conversation on other
lists[1][2][3] and in
Forwarded Message
Subject: bug#34220: failure to building with CompCert, patch proposed
Resent-Date: Sun, 27 Jan 2019 16:58:02 +
Resent-From: DAVID MONNIAUX
Resent-CC: bug-coreut...@gnu.org
Date: Sun, 27 Jan 2019 16:53:27 +0100 (CET)
From: DAVID MONNIAUX
To: 34...@debbug
> This gnulib-tool invocation
>
> $ ./gnulib-tool --create-testdir --dir=../testdir --single-configure \
> argv-iter unicase/u16-casecoll unicase/u32-casecoll unicase/u8-casecoll \
> uninorm/u16-normcoll uninorm/u32-normcoll uninorm/u8-normcoll \
> unistdio/u16-vasnprintf unistdio/u16-vaspri
This gnulib-tool invocation
$ ./gnulib-tool --create-testdir --dir=../testdir --single-configure \
argv-iter unicase/u16-casecoll unicase/u32-casecoll unicase/u8-casecoll \
uninorm/u16-normcoll uninorm/u32-normcoll uninorm/u8-normcoll \
unistdio/u16-vasnprintf unistdio/u16-vasprintf unistdio
On Android 4.3, I'm seeing this test failure:
FAIL: test-argv-iter
../../gltests/test-argv-iter.c:35: assertion 'fp' failed
FAIL test-argv-iter (exit status: 139)
The cause is that tmpfile() returns NULL, because /tmp does not exist.
This patch fixes it (borrowing code from
Hi Jim!
> Le 27 janv. 2019 à 17:22, Jim Meyering a écrit :
>
> Hi Akim,
> I too would start with coreutils, noting that one must be careful to
> handle signals, so that the terminal is not left in a bad state when
> output is interrupted between a pair of set/reset escape sequences.
> https://gi
> bitsetv: allow free on NULL.
>
> * lib/bitsetv.c (bitsetv_free): Do nothing when the bitsetv is NULL.
Looks good to me.
Hi Siddhesh,
> Build a more general script that parses changes in a VCS and then
> prints it in a ChangeLog format.
Can you please give some more context?
* Is it a tool meant to operate on the history and produce a (large)
ChangeLog file, like gitlog-to-changelog?
Or is it a tool meant
Hi Bruno,
> Le 27 janv. 2019 à 18:39, Bruno Haible a écrit :
>
> Hi Akim,
>
>> I'd like to have colored diagnostics in Bison, I do feel on occasion it would
>> make it easier to spot errors and warnings for instance.
>>
>> AFAICT, gnulib does not feature any module to help do this. Coreutils
I can live with the 'if' on the caller of bitsetv_free, but I felt it's
somewhat consistent with free accepting NULL.
commit 954d1e36aa815994f7f5b1fcd46ebc96a63b8df4
Author: Akim Demaille
Date: Sun Jan 27 18:49:36 2019 +0100
bitsetv: allow free on NULL.
* lib/bitsetv.c (bitsetv_f
Hi Akim,
> I'd like to have colored diagnostics in Bison, I do feel on occasion it would
> make it easier to spot errors and warnings for instance.
>
> AFAICT, gnulib does not feature any module to help do this. Coreutils (ls)
> and GCC (maybe diffutils too) are probably good sources of inspirat
From: Siddhesh Poyarekar
Build a more general script that parses changes in a VCS and then
prints it in a ChangeLog format. Currently this set of scripts has
been tested on glibc and will need a vcstocl_quirks.py for gnulib.
The scripts need to be in to make sure that they're usable for glibc
on
Build a more general script that parses changes in a VCS and then
prints it in a ChangeLog format. Currently this set of scripts has
been tested on glibc and will need a vcstocl_quirks.py for gnulib.
The scripts need to be in to make sure that they're usable for glibc
once 2.29 is released. A qui
On Sun, Jan 27, 2019 at 7:53 AM Akim Demaille wrote:
> I'd like to have colored diagnostics in Bison, I do feel on occasion it would
> make it easier to spot errors and warnings for instance.
>
> AFAICT, gnulib does not feature any module to help do this. Coreutils (ls)
> and GCC (maybe diffuti
While doing this change, I noticed that init.sh is a bit hard to grok,
because essentials of the test framework are mixed with mere convenience
functions (like 'compare'). So I took the opportunity to reorganize this
file.
2019-01-27 Bruno Haible
test-framework-sh: Improve maintainabi
Hi friends,
I'd like to have colored diagnostics in Bison, I do feel on occasion it would
make it easier to spot errors and warnings for instance.
AFAICT, gnulib does not feature any module to help do this. Coreutils (ls) and
GCC (maybe diffutils too) are probably good sources of inspiration,
Another particularity of the Android environment is that /tmp does not exist.
A variable TMPDIR is set instead.
Therefore tests that want to test /tmp should be skipped.
2019-01-27 Bruno Haible
tests: Don't assume that /tmp exists.
* tests/test-set-mode-acl-1.sh: Skip the tes
On Android 4.3, I'm seeing these test failures:
FAIL: test-set-mode-acl-1.sh
../../gltests/test-set-mode-acl-1.sh[13]: ../../gltests/test-set-mode-acl.sh:
No such file or directory
FAIL test-set-mode-acl-1.sh (exit status: 1)
FAIL: test-set-mode-acl-2.sh
==
This command fails:
$ ./gnulib-tool --test --with-tests --single-configure
unigbrk/uc-grapheme-breaks
...
gllib/Makefile.am:1706: error: LIBUNISTRING_COMPILE_UNIGBRK_U32_GRAPHEME_BREAKS
does not appear in AM_CONDITIONAL
After the type in the Makefile.am is fixed, there is another error:
../../
Jim Meyering wrote:
> You're welcome to merge those.
Done.
Bruno
In a couple of more places, it makes sense to treat Android like Linux.
2019-01-26 Bruno Haible
vma-iter: Add support for Android.
* lib/vma-iter.h (VMA_ITERATE_SUPPORTED): Define on Android as well.
* lib/vma-iter.c: Treat Android like Linux.
* lib/get-rusage-
Followup: Enable Linux specific optimizations on Android as well.
__linux__ is not defined by "gcc -mandroid" in this environment.
2019-01-26 Bruno Haible
fts: Optimize on Android.
* lib/fts.c: Treat Android like Linux.
diff --git a/lib/fts.c b/lib/fts.c
index 47de580..58232a
On Android 4.3 (on a Linux ext4 file system), I see the 'test-fts' test fail.
There are two problems:
1) fts_open apparently fails with error EINVAL. This happens because the
test program uses the 'fts_open' etc. symbols from libc instead of those
from libgnu.a - but the flag definitio
23 matches
Mail list logo
