Re: setproctitle()

Peter Seebach wrote: > Is there any interest in attempting to provide a moderately portable > setproctitle()? (I think it would be reasonable for it to, on some > systems, simply fail to do anything...) > > Apparently, this functionality exists in "util-linux-ng", and sendmail has > an implementat

Re: [PATCH] bootstrap: don't ignore download_po_files failure.

Eric Blake wrote: > On 05/03/2010 02:28 PM, Giuseppe Scrivano wrote: >> I have noticed this error today while I was on a network with a silly >> firewall that blocked rsync. I think `bootstrap' should fail in this >> case. > > Even nicer would be to have a fallback to wget that goes through http i

Re: [PATCH] bootstrap: don't ignore download_po_files failure.

Giuseppe Scrivano wrote: > I have noticed this error today while I was on a network with a silly > firewall that blocked rsync. I think `bootstrap' should fail in this > case. Thank you. I pushed that.

setproctitle()

Is there any interest in attempting to provide a moderately portable setproctitle()? (I think it would be reasonable for it to, on some systems, simply fail to do anything...) Apparently, this functionality exists in "util-linux-ng", and sendmail has an implementation that works on Linux, but it

gnu-web-doc-update failed for diffutils

Normally after a release, I run build-aux/gnu-web-doc-update. For diffutils, it failed like this, and I nearly failed to notice the failure, because this diagnostic was followed by a lot more output and the final exit status was 0: make[1]: Nothing to be done for `all-am'. make[1]: Leaving dir

Re: [PATCH] bootstrap: don't ignore download_po_files failure.

On 05/03/2010 02:28 PM, Giuseppe Scrivano wrote: > Hello, > > I have noticed this error today while I was on a network with a silly > firewall that blocked rsync. I think `bootstrap' should fail in this > case. Even nicer would be to have a fallback to wget that goes through http in the case whe

[PATCH] bootstrap: don't ignore download_po_files failure.

Hello, I have noticed this error today while I was on a network with a silly firewall that blocked rsync. I think `bootstrap' should fail in this case. Cheers, Giuseppe >From 59fbac2381583a3946adb7e3f2299f8143159b2d Mon Sep 17 00:00:00 2001 From: Giuseppe Scrivano Date: Mon, 3 May 2010 22:05

Re: [PATCH] maint.mk, announce-gen: allow project-specific announcement mail headers

Jim Meyering wrote: > This helps factor some bits out of README-release, > and thus helps reduce the number of manual steps required > to make a release. > > I'm about to commit a related patch for coreutils. Whoops. I didn't re-test after adding the --help delta. >From e6addf84d6331d634b5d76db03

[PATCH] maint.mk, announce-gen: allow project-specific announcement mail headers

FYI, This helps factor some bits out of README-release, and thus helps reduce the number of manual steps required to make a release. I'm about to commit a related patch for coreutils. >From 8c2ede4827e6294ed53a4017d5e8fb5d853ac1b3 Mon Sep 17 00:00:00 2001 From: Jim Meyering Date: Mon, 3 May 201

[PATCH] test-xalloc-die: avoid unwarranted test failure on OpenSolaris 5.11

Without this patch, diffutils' "make check" would fail on OpenSolaris 5.11 like this: (note the unexpected "1> out" line in "err2"): ... + create_exe_shims_ ... + return 0 + shift + test 0 != 0 + export PATH + test-xalloc-die + 2> err + tr -d '\015' + 0< err

Re: test results on Cygwin

Simon Josefsson writes: > I'm running another build now. We are down to two failed self-tests under cygwin, see: http://autobuild.josefsson.org/gnulib/log-201005030918958478000.txt The failures are: file_has_acl("tmpdir0") returned yes, expected no FAIL: test-file-has-acl.sh Unconnected sock

Re: maint: extend the always_defined_macros syntax-check

Jim Meyering wrote: > Three files in coreutils defined WIFSIGNALED and WTERMSIG macros > to accommodate systems that lack them. However, now that gnulib's > sys/wait.h guarantees those, there's no point. I've removed them. > > maint: remove now-redundant definitions provided by sys/wait.h >

maint: extend the always_defined_macros syntax-check

Three files in coreutils defined WIFSIGNALED and WTERMSIG macros to accommodate systems that lack them. However, now that gnulib's sys/wait.h guarantees those, there's no point. I've removed them. maint: remove now-redundant definitions provided by sys/wait.h * src/timeout.c (WIFSIGNALED

[PATCH] maint.mk: prohibit unwarranted use of

I think we all agree that uses of bcmp, bcopy, bzero, index and rindex should have been removed from our code at least 10 years ago, so unless you use strcasecmp, strncasecmp, ffs or ffsll, you have no reason to include strings.h. >From 0bbf667efe89bf63421ea7f0c4566eec261199fd Mon Sep 17 00:00:00

[PATCH] maint.mk: remove obsolete comments

FYI, I've just pushed this. >From 030b04575ac679b62eacf247543a451d02d0e319 Mon Sep 17 00:00:00 2001 From: Jim Meyering Date: Mon, 3 May 2010 09:49:58 +0200 Subject: [PATCH] maint.mk: remove obsolete comments * top/maint.mk: Remove stale, commented-out rules. --- ChangeLog|5 + top/m

wctob: small correction

This fixes a wctob replacement bug on weird platforms. 2010-05-03 Bruno Haible wctob: Fix for weird platforms. * lib/wctob.c (wctob): When wint_t is larger than wchar_t, check the argument value. --- lib/wctob.c.origMon May 3 10:06:00 2010 +++ lib/wctob.c Mon May