fix some gcc warnings

Hi, When I compile a gnulib testdir with "gcc -Wall", I see the following warnings (among others): test-ffsl.c: In function 'main': test-ffsl.c:44:12: warning: unused variable 'x' [-Wunused-variable] test-posix_spawn3.c: In function 'parent_main': test-posix_spawn3.c:78:11: warning: ignoring

some gcc warnings

Hi Paul, Simon, When I compile a gnulib testdir on a glibc system with "gcc -Wall", I see the following warnings (among others): rijndael-api-fst.c: In function 'rijndaelBlockEncrypt': rijndael-api-fst.c:234:11: warning: dereferencing type-punned pointer will break strict-aliasing rules

Re: what platforms are you testing on?

On 16/10/16 12:43, Bruno Haible wrote: > Hi all, > > Noticing that the new 'getprogname' module probably exhibits a compilation > error on HP-UX, IRIX, OSF/1 (error "getprogname module not ported to this > OS"), > I would like to update the list of supported platforms at >

fix obsolete uses of AC_MSG_NOTICE

Hi, In the configure output of a full gnulib testdir, we have thousands of "checking ..." messages but also 2 times "configure: checking ..." ... checking whether fflush works on input streams... (cached) no checking for _fseeki64... no checking whether fstatat (..., 0) works... yes checking for

Re: getprogname: comments and test failure on Cygwin

On Sun, Oct 16, 2016 at 5:15 AM, Pádraig Brady wrote: > On 16/10/16 12:55, Bruno Haible wrote: >> Hi, >> >> The 'getprogname' module test fails on Cygwin 2.6, because the returned >> value is "test-getprogname", not "test-getprogname.exe". (On mingw, on the >> other hand, it

Re: fix some gcc warnings

On 16/10/16 16:01, Bruno Haible wrote: > Hi, > > When I compile a gnulib testdir with "gcc -Wall", I see the following warnings > (among others): > > test-ffsl.c: In function 'main': > test-ffsl.c:44:12: warning: unused variable 'x' [-Wunused-variable] > > test-posix_spawn3.c: In function

Re: fix obsolete uses of AC_MSG_NOTICE

On 16/10/16 15:09, Bruno Haible wrote: > Hi, > > In the configure output of a full gnulib testdir, we have thousands of > "checking ..." messages but also 2 times "configure: checking ..." > > ... > checking whether fflush works on input streams... (cached) no > checking for _fseeki64... no >

a bad use of strerror_r

Hi, When I compile a gnulib testdir on a glibc system with "gcc -Wall", I see the following warnings (among others): argp-help.c: In function 'argp_failure': argp-help.c:1878:17: warning: assignment makes pointer from integer without a cast [enabled by default] This is a dangerous one, because

Re: getprogname: comments and test failure on Cygwin

On 16/10/16 12:55, Bruno Haible wrote: > Hi, > > The 'getprogname' module test fails on Cygwin 2.6, because the returned > value is "test-getprogname", not "test-getprogname.exe". (On mingw, on the > other hand, it really is "test-getprogname.exe".) > > Also, while at it, I'd like to add

what platforms are you testing on?

Hi all, Noticing that the new 'getprogname' module probably exhibits a compilation error on HP-UX, IRIX, OSF/1 (error "getprogname module not ported to this OS"), I would like to update the list of supported platforms at https://www.gnu.org/software/gnulib/manual/html_node/Target-Platforms.html

The non-recursive-gnulib-prefix-hack module

Hi Jim, all, When I try to use the gnulib-tool --create-testdir option to produce a testdir for all modules, it encounters errors and produces no toplevel Makefile.in: $ ./gnulib-tool --create-testdir --with-tests --dir=/tmp/testdir1 gnulib-tool: warning: module euidaccess depends on a module

Re: license change of futimens was incomplete

Aurelien Jarno wrote: > I actually only added a single line of code :-). Anyway you have my > agreement to relicense it to LGPL. Well, it was 1 line of code with a 7-lines git commit entry :) Thanks for all the approvals, I've pushed this change. Bruno -- In memoriam Thet Win Aung

getprogname: comments and test failure on Cygwin

Hi, The 'getprogname' module test fails on Cygwin 2.6, because the returned value is "test-getprogname", not "test-getprogname.exe". (On mingw, on the other hand, it really is "test-getprogname.exe".) Also, while at it, I'd like to add comments: 1. The declaration in getprogname.h does not state

quotearg test failures

Hi Pádraig, On a glibc system (glibc 2.15, Linux 3.8) I get these test failures from a gnulib testdir: 1) For module 'quotearg': FAIL: test-quotearg.sh $ LOCALE=fr_FR.UTF-8 LOCALEDIR=locale ./test-quotearg test-quotearg.h:53: assertion 'la == lb' failed (gdb) where #0 0x77a52035 in

fix test failure in test-u8-strtok

Hi Daiki, On a glibc system (glibc 2.15, Linux 3.8) I get this test failure from a gnulib testdir: ./../build-aux/test-driver: line 107: 13312 Aborted "$@" > $log_file 2>&1 FAIL: test-u8-strtok unistr/test-u-strtok.h:80: assertion 'result == first_ptr' failed FAIL

freelocale woes

Hi Paul, You wrote in Don't call freelocale on a locale that was the base of a successful newlocale, as that results in a double free. On which system was this? I'm asking because - In the glibc implementation

Re: what platforms are you testing on?

On Sun, Oct 16, 2016 at 4:43 AM, Bruno Haible wrote: > Hi all, > > Noticing that the new 'getprogname' module probably exhibits a compilation > error on HP-UX, IRIX, OSF/1 (error "getprogname module not ported to this > OS"), > I would like to update the list of supported

fix module qsort_r

Hi Paul, The module 'qsort_r' does not work at all on systems that don't have this function. For example, on AIX 7.1, with a gnulib testdir: test-qsort_r.c: In function 'main': test-qsort_r.c:40:3: warning: implicit declaration of function 'qsort_r' [-Wimplicit-function-declaration] qsort_r

test-duplocale crashes

On a glibc system (glibc 2.15, Linux 3.8) I get this test failure from a gnulib testdir: ./../build-aux/test-driver: line 107: 6269 Segmentation fault "$@" > $log_file 2>&1 FAIL: test-duplocale FAIL test-duplocale (exit status: 139) (gdb) where #0 0x77a4aa6f in freelocale ()

Re: The non-recursive-gnulib-prefix-hack module

On Sun, Oct 16, 2016 at 5:13 AM, Bruno Haible wrote: > Hi Jim, all, > > When I try to use the gnulib-tool --create-testdir option to produce a testdir > for all modules, it encounters errors and produces no toplevel Makefile.in: > > $ ./gnulib-tool --create-testdir --with-tests

Re: what platforms are you testing on?

Hello Bruno and all, On Sun, Oct 16, 2016 at 4:43 AM, Bruno Haible wrote: > Can each of you please give a list of platforms to which you have access and > on which you frequently or occasionally test gnulib modules? I run automated testing for tarballs posted on

Re: quotearg test failures

On 16/10/16 17:49, Bruno Haible wrote: > Hi Pádraig, > > On a glibc system (glibc 2.15, Linux 3.8) I get these test failures > from a gnulib testdir: > > 1) For module 'quotearg': > > FAIL: test-quotearg.sh > > $ LOCALE=fr_FR.UTF-8 LOCALEDIR=locale ./test-quotearg > test-quotearg.h:53:

Re: test-duplocale crashes

On 16/10/16 19:10, Bruno Haible wrote: > On a glibc system (glibc 2.15, Linux 3.8) I get this test failure > from a gnulib testdir: > > ./../build-aux/test-driver: line 107: 6269 Segmentation fault "$@" > > $log_file 2>&1 > FAIL: test-duplocale > > FAIL test-duplocale (exit status: 139) >

Re: quotearg test failures

Hi Pádraig, > The attached should fix up those tests. I disagree with the removal of the '%' checks in tests/test-sh-quote.c and tests/test-system-quote-main.c. These checks are there to guarantee that the functions won't write past the allocated buffer. For example, sh-quote.h specifies how

[PATCH] dfa: addition of new state on demand

When dfa builds a state, generates all next states. However, I believe most of them are not used. This patch changes as that when dfa builds a state, generates a next state including next input character only. The following test was improved from 2.52s to 0.67s by the patch in my machine. $

fix a warning on AIX

On AIX 7.1 with gcc, I see this compilation warning: test-system-quote-child.c: In function 'main': test-system-quote-child.c:48:5: warning: implicit declaration of function 'fopen' [-Wimplicit-function-declaration] FILE *fp = fopen (EXPECTED_DATA_FILE, "rb"); ^