sort's -u is Failing to Check all -k fields for Uniqueness.

Hi, This seems wrong. $ echo 1/3,1/2,1/1,2/1 | tr , \\012 | sort -nu -t / -k 1,2 1/3 2/1 $ -u should only filter out lines that compare equal on *all* key fields. I've opened an Ubuntu bug at https://launchpad.net/distros/ubuntu/+source/coreutils/+bug/56891 which has more

Re: sort's -u is Failing to Check all -k fields for Uniqueness.

This seems wrong. $ echo 1/3,1/2,1/1,2/1 | tr , \\012 | sort -nu -t / -k 1,2 1/3 2/1 $ -u should only filter out lines that compare equal on *all* key fields. I've opened an Ubuntu bug at https://launchpad.net/distros/ubuntu/+source/coreutils/+bug/56891 which has more

Re: sort's -u is Failing to Check all -k fields for Uniqueness.

Ralph Corderoy [EMAIL PROTECTED] writes: $ echo 1/3,1/2,1/1,2/1 | tr , \\012 | sort -nu -t / -k 1,2 1/3 2/1 $ -u should only filter out lines that compare equal on *all* key fields. There is only one sort key, which spans the first two fields of each line, and -n tells sort

fix for long-standing bug in cp --sparse

Alan, thanks for the report and patch. FYI, I expect all !HAVE_FTRUNCATE code to be removed sometime in 2007. 2006-08-19 Jim Meyering [EMAIL PROTECTED] * NEWS: Fix cp --sparse so that it preserves tail-end sparseness, even when the file's apparent size is not a multiple of its

FYI: minor change: tests/Makefile.am.in, and derived Makefile.am files

2006-08-19 Jim Meyering [EMAIL PROTECTED] Some of my 2006-07-03 changes to tests/*/Makefile.am were being backed out due to updates provoked by the copyright changes. * tests/Makefile.am.in (PATH): Prepend $(VG_PATH_PREFIX), so that it propagates to the derived

README: mention potential Pre-C99 build failure

A documentation change: * README: Describe potential pre-C99 build failure, and work-around. Index: README === RCS file: /fetish/cu/README,v retrieving revision 1.27 diff -u -r1.27 README --- README 17 Aug 2006 19:58:18

FYI: test for recently fixed ls -CF bug

2006-08-19 Jim Meyering [EMAIL PROTECTED] * tests/ls/stat-dtype: Test for the 2006-08-17 `ls -CF' fix. Index: tests/ls/stat-dtype === RCS file: /fetish/cu/tests/ls/stat-dtype,v retrieving revision 1.6 diff -u -r1.6

Re: Bug#355368: coreutils: FTBS if SHELL variable is not set

Sven Joachim [EMAIL PROTECTED] wrote: I hope you can remember this issue (dircolors test failing with SHELL unset) after more than five months. I'm quoting your whole mail: ... [Context for any interested readers: http://bugs.debian.org/355368 ] Thanks Sven, I've done as you suggest, with

Re: sort's -u is Failing to Check all -k fields for Uniqueness.

Hi Eric and Andreas, Andreas wrote: $ echo 1/3,1/2,1/1,2/1 | tr , \\012 | sort -nu -t / -k 1,2 There is only one sort key, which spans the first two fields of each line, and -n tells sort to only consider the numeric prefix of each key. If you want to sort on multiple keys you need to

coreutils-6.1 released

Coreutils version 6.1 has been released. If you haven't heard about the GNU coreutils, the FAQ is a good place to start: http://www.gnu.org/software/coreutils/faq/ This is a bug-fix and portability-fix release. Thanks to everyone who's helped.

PATCH: df alignment with -mP

The usage of some options (e.g., -mP) causes a bad column alignment. I have written this patch to solve this. I hope it is OK. Cheers, Gustavo diff -u -r coreutils.orig/NEWS coreutils/NEWS --- coreutils.orig/NEWS 2006-08-20 01:53:08.0 -0300 +++ coreutils/NEWS 2006-08-20