Re: AIX grep limitation: coreutils build failure

Ralf Wildenhues [EMAIL PROTECTED] writes: OTOH, coreutils would not actually need the `${LIBOBJDIR}' thingies anyway, and they cause the extremely long line. That sounds like something reasonable to do in any case. I'm a bit reluctant to rely on 'tr' given its portability hassles. How about

Re: AIX tests failures

Ralf Wildenhues [EMAIL PROTECTED] wrote: With coreutils 5.96, I get three failures on AIX 4.3.3, namely: cp/fail-perm, rm/inaccessible, help-version. Verbose output is below. The last one looks funny. :-) ... ! cp: accessing `symlink': The file access permissions do not allow the

Re: AIX tests failures

Hi Jim, Thanks for fixing. Small nit: * Jim Meyering wrote on Sun, May 28, 2006 at 11:04:05AM CEST: Ralf Wildenhues [EMAIL PROTECTED] wrote: With coreutils 5.96, I get three failures on AIX 4.3.3, namely: ^^^ +# HPUX appears to fail with

Re: AIX tests failures

Ralf Wildenhues [EMAIL PROTECTED] wrote: With coreutils 5.96, I get three failures on AIX 4.3.3, namely: ^^^ +# HPUX appears to fail with EACCES rather than EPERM. This looks like an inconsistency to me. FWIW, I didn't see any

Re: AIX tests failures

Hi Jim, * Jim Meyering wrote on Sun, May 28, 2006 at 11:22:42AM CEST: Ralf Wildenhues [EMAIL PROTECTED] wrote: With coreutils 5.96, I get three failures on AIX 4.3.3, namely: ^^^ +# HPUX appears to fail with EACCES rather than EPERM.

Re: AIX tests failures

Ralf Wildenhues [EMAIL PROTECTED] wrote: With coreutils 5.96, I get three failures on AIX 4.3.3, namely: cp/fail-perm, rm/inaccessible, help-version. ... rm: cannot remove `rel': The file access permissions do not allow the specified action. --- rm: cannot remove `rel': Permission denied

Re: AIX grep limitation: coreutils build failure

Hi Paul, * Paul Eggert wrote on Sun, May 28, 2006 at 10:50:44AM CEST: I'm a bit reluctant to rely on 'tr' given its portability hassles. Well, the same can be achieved with (including M4 quoting) sed s/[[$as_cr_Letters]]//g but I don't understand the portability problems given the

5.96: some locale squashing

I need this patch for successful `make check' on a GNU/Linux with non-English locale settings. Cheers, Ralf * tests/cp/fail-perm: source lang-default. * tests/rm/inaccessible: Likewise. --- tests/cp/fail-perm 2004-06-28 20:47:24.0 +0200 +++ tests/cp/fail-perm

Re: AIX tests failures

Ralf Wildenhues [EMAIL PROTECTED] wrote: Ahh, thanks for the clarification. Speaking of trunk, I follow the :pserver:[EMAIL PROTECTED]:/cvsroot/coreutils CVS repository, both HEAD and branch b5_9x. Is that where development is happening though? Both have this as latest entry: |

Re: 5.96: some locale squashing

Ralf Wildenhues [EMAIL PROTECTED] wrote: I need this patch for successful `make check' on a GNU/Linux with non-English locale settings. Applied. Thanks. ___ Bug-coreutils mailing list Bug-coreutils@gnu.org

Re: 5.96: du test failure on Darwin

Ralf Wildenhues [EMAIL PROTECTED] wrote: On powerpc-apple-darwin8.2.0, I get one test failure: Cheers, Ralf $ cd tests/du $ make check TESTS=basic VERBOSE=yes ... out exp differ: char 1, line 1 + fail=1 + test 1 = 1 + diff -u out exp --- out 2006-05-27 15:14:12.0 -0500 +++ exp

Re: AIX tests failures

Hi Jim, * Jim Meyering wrote on Sun, May 28, 2006 at 03:48:21PM CEST: Ralf Wildenhues [EMAIL PROTECTED] wrote: With coreutils 5.96, I get three failures on AIX 4.3.3, namely: cp/fail-perm, rm/inaccessible, help-version. $ env VERBOSE=yes TESTS=help-version make check-TESTS [...] + 0

Re[2]: DF (coreutils) - bug??

PE Alex aka Parasite [EMAIL PROTECTED] writes: The problem seems to be related to well-known 137Gb LBA limitation. PE Most likely it's not a bug in df, then, and is instead a bug in the PE underyling operating system or file system. You can check this by PE invoking strace on df, and seeing

Re: AIX tests failures

Ralf Wildenhues [EMAIL PROTECTED] wrote: Hi Jim, * Jim Meyering wrote on Sun, May 28, 2006 at 03:48:21PM CEST: Ralf Wildenhues [EMAIL PROTECTED] wrote: With coreutils 5.96, I get three failures on AIX 4.3.3, namely: cp/fail-perm, rm/inaccessible, help-version. $ env VERBOSE=yes

Re: AIX tests failures

[ Cc:ing bug-gnulib ] * Jim Meyering wrote on Sun, May 28, 2006 at 07:58:24PM CEST: Ralf Wildenhues [EMAIL PROTECTED] wrote: ... AIX strndup is severely broken, at least on 4.3.3 and 5.1, similar to its strnlen; see also[1]. See for example this test: I have to confess that I wonder if

chmod

Hello, I think I have found some bug in chmod. Maybe it is more general, but I didn't manage (and didn't try hard) to reconstruct it using different utilities. The sequence (alphabetical order is important) touch a mkdir b touch c cd b chmod a-x ../* makes a segmentation fault. Though it may

Re: 5.96: du test failure on Darwin

* Jim Meyering wrote on Sun, May 28, 2006 at 03:07:08PM CEST: Ralf Wildenhues [EMAIL PROTECTED] wrote: On powerpc-apple-darwin8.2.0, I get one test failure: $ cd tests/du $ make check TESTS=basic VERBOSE=yes ... out exp differ: char 1, line 1 What type of file system was that using?

5.96: chgrp/basic failure on OpenBSD

Hi there, With 5.96, I get one test failure on i386-unknown-openbsd3.8. Cheers, Ralf $ cd tests/chgrp/ $ ls Makefile $ gmake check TESTS=basic VERBOSE=yes gmake check-TESTS gmake[1]: Entering directory `/tmp/build/tests/chgrp' + chgrp --version chgrp (GNU coreutils) 5.96 Copyright (C) 2006

Re: AIX tests failures

Ralf Wildenhues [EMAIL PROTECTED] writes: I only saw Paul's message after mostly finishing this one, so feel free to decide as you like. (I don't know whether Gnulib intends to target systems that are apparently not of interest to coreutils any more.) The general rule I use with gnulib-like

Re: gsort problem

Anthony Tiemens [EMAIL PROTECTED] writes: What version of coreutils will the patch be available in? It's in CVS, so barring some unforeseen event it'll be in the next release. When will it be released? I wouldn't venture to predict that. But you can use the patch yourself in the meantime;