Re: poll test failure on Cygwin

On Sun, Mar 14, 2010 at 23:33, Bruno Haible wrote: > Hi Paolo, Eric, > > On Cygwin 1.5.x the 'poll' test fails like this: > >  Unconnected socket test... passed >  Connected sockets test... passed >  General socket test with fork... passed >  Pipe test... failed (expecting POLLHUP after shutdown)

Re: cosl test failure

On Sun, Mar 14, 2010 at 23:18, Bruno Haible wrote: > Hi Paolo, > > On 2010-01-18, in > , > I observed that gnulib's cosl replacement function does not have the > necessary accuracy. This was due to a wrong formula: The term >  sin

Re: iconv_open: Build failure with gperf 2.7.2

Hi Ludo, Ludovic Courtès wrote: > The ‘-m’ gperf option used by the ‘iconv_open’ module isn’t available in > the ancient gperf 2.7.2: It is documented in gnulib's DEPENDENCIES file: * GNU gperf 3.0.1 or newer. + Recommended. Needed if you use the 'iconv_open' module. + Homepage: http

flock test failures on Solaris and MacOS X

Hi, In and it was reported that the 'flock' tests fail on MacOS X 10.5. This is still the case today: test-flock.c:80: assertion failed /bin/sh: line 1: 62

iconv_open: Build failure with gperf 2.7.2

Hello, The ‘-m’ gperf option used by the ‘iconv_open’ module isn’t available in the ancient gperf 2.7.2: --8<---cut here---start->8--- Making all in lib rm -f alloca.h-t alloca.h && { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; cat ../../lib/alloca.i

unlink test failure on MacOS X

Hi Jim, Eric, On MacOS X 10.5, the unlink() and unlinkat() tests fail: test-unlink.h:49: assertion failed /bin/sh: line 1: 17670 Abort trap EXEEXT='' srcdir='.' ${dir}$tst FAIL: test-unlink test-unlink.h:49: assertion failed /bin/sh: line 1: 63034 Abort trap E

Re: link-warning usage improvements

Eric Blake wrote on 2009-12-31: > Date: Thu, 24 Dec 2009 12:06:48 -0700 > Subject: [PATCH 4/4] link-warning: always build headers with link warnings > > Replacement headers must be built unconditionally if they use > GL_LINK_WARNING, if the warning is to ever trigger on a > glibc system during CFL

poll test failure on Cygwin

Hi Paolo, Eric, On Cygwin 1.5.x the 'poll' test fails like this: Unconnected socket test... passed Connected sockets test... passed General socket test with fork... passed Pipe test... failed (expecting POLLHUP after shutdown) FAIL: test-poll.exe What do you think? Is Cygwin broken? Or

Re: cosl test failure

Hi Paolo, On 2010-01-18, in , I observed that gnulib's cosl replacement function does not have the necessary accuracy. This was due to a wrong formula: The term sincosl_table [index + SINCOSL_COS_HI] * cos_l_m1 was being subtrac

posix_spawn fix

Compiling a testdir for 'spawn' on mingw, I see this error: ../gllib/spawn.h:406: error: `POSIX_SPAWN_RESETIDS' undeclared here (not in a function) ../gllib/spawn.h:407: error: `POSIX_SPAWN_SETPGROUP' undeclared here (not in a function) ../gllib/spawn.h:407: error: `POSIX_SPAWN_SETSIGDEF' undecl

time_r fix

Compiling a testdir for 'time' on mingw, I see this error: ../gllib/time.h:379: error: `::localtime_r' has not been declared ../gllib/time.h:394: error: `::gmtime_r' has not been declared This fixes it. 2010-03-14 Bruno Haible Fix compilation error on mingw when module 'time_r' is n

Re: LONG_LONG_MIN

Jim Meyering wrote: > Which macros from would be better? I was thinking of INT64_MIN, INT64_MAX, UINT64_MAX. But using these would assume that 'long long' has 64 bits, which is not future-proof... > Using LLONG_MIN etc. sounds fine. OK, I'm using this. If on some platforms, LLONG_MIN is undefin

Re: lib-ignore: link error with Sun C++ compiler

Ralf Wildenhues wrote: > > The reason is that the flag -Wl,-z,ignore is accepted by the C compiler > > (cc), but > > leads to an error with the C++ compiler (CC). > > Sun CC needs '-Qoption ld ' as prefix to linker flags (mind the trailing > space) rather than '-Wl,'. Thanks for this info. So,

Re: [PATCH] (x)memcoll: performance improvement when input is known to be NUL delimited.

Hi Chen, > ... it was consistently (as in every run for 20 runs) faster, and thus I've > included it in the patch. Thanks. I'll wait for the FSF notification that the legal papers have arrived. > This is not what I expected at all, and I'm having a hard time coming up with > a reason why this i

Re: [PATCH] (x)memcoll: performance improvement when input is known to be NUL delimited.

Hi Chen, > > It would be good to start the function with a safety check: > > > > if (!(s1len > 0 && s1[s1len - 1] == '\0')) > >abort (); > > if (!(s2len > 0 && s2[s2len - 1] == '\0')) > >abort (); > > If the whole point was performance, but we introduce 4 conditiona

Re: bootstrap and git submodules

Hi Jim and Eric, Thanks for your explanations how 'bootstrap' is to be used with git. But I still don't find the current state reasonable, because: * There are five use cases of the 'bootstrap' script: 1) A normal user who wants to check out coreutils with the version of gnulib w

bootstrap: allow more than one submodule

Hi Eric, As I understand it, 'bootstrap' currently updates all submodules when it wants to update only the gnulib submodule. What about packages that will have other submodules? Here is a proposed fix: 2010-03-14 Bruno Haible * build-aux/bootstrap: Apply "git submodule" commands only