David Eisner [EMAIL PROTECTED] writes:
I was using du to get a sense of whether an rsync operation between
two hosts was successful.
I don't see how that can work in general, even if we make the changes
suggested, since in general directory sizes won't be the same on
different machines.
Gerard Beekmans [EMAIL PROTECTED] writes:
I tried out the new 4.5.8 release and it doesn't exhibit the same
symptoms. It runs the check and determines that I don't have a
working C stack overflow, but without using up all the RAM.
Is this normal for a Linux system?
Yes, it's expected for
[EMAIL PROTECTED] writes:
But -u -d is clearly valid, since POSIX says it's valid.
have you a link to the POSIX docs?
http://www.opengroup.org/onlinepubs/007904975/utilities/uniq.html
In order to understand that page fully, you need to read a lot of the
other part of the spec. But for
Bernard Giroud [EMAIL PROTECTED] writes:
Index: configure.ac
Wouldn't it be simpler to put this:
#ifndef initialize_main
# define initialize_main(argcp, argvp)
#endif
into system.h rather than modifying configure.ac?
That is how diffutils does it.
[EMAIL PROTECTED] writes:
I have found a linux-specific use for this in my boot scripts, and would
like to be able to see other errors from ln without cluttering up the
boot process with this one.
How about if we just remove the warning instead? It annoys me too.
I often use hard links to
Dan Jacobson [EMAIL PROTECTED] writes:
$ man test
-t [FD]
file descriptor FD (stdout by default) is opened on a terminal
I.e. argument optional
$ help test
-t FD True if FD is opened on a terminal.
Not so optional with bash then.
You guys should
[EMAIL PROTECTED] (Paul Jarc) writes:
Paul Eggert [EMAIL PROTECTED] wrote:
POSIX long ago decided that FD is not optional with test -t. GNU
'test' conforms to POSIX in this respect.
bash's does, but coreutils' doesn't.
Good point. I looked at coreutils/src/test.c and noticed some other
Jim Meyering [EMAIL PROTECTED] writes:
However, if you can find a few other people who say
they'd like that functionality and make a good case
for it, I'll reconsider.
If you go that route, I'd suggest adding an option
--suffix-alphabet=STRING, so that the user can specify
an arbitrary
gregory mott [EMAIL PROTECTED] writes:
for example, en_IN repeatably produces proper results, but en_AU
repeatably fails to handle some special characters properly.
I reproduced your results on my host (Debian GNU/Linux 3.0r1).
However, on my host what you were doing was a user error, as en_IN
Dan Jacobson [EMAIL PROTECTED] writes:
The du and df commands give differing results. Perhaps df can't see
into multitrack CDROMs.
df just reports what the underlying system call tells it (statfs on
GNU/Linux), so most likely the problem is not in coreutils.
Possibly there are multiple
gregory mott [EMAIL PROTECTED] writes:
can you point me to an appropriate RTFM that ideally would layout what
encodings are used by what locales, or how to tell what encoding you
have/need, etc usw?
Sorry, no; this stuff tends to be scattered around all over the place.
On my Debian GNU/Linux
gregory mott [EMAIL PROTECTED] writes:
when i pass textual input to sort, how does sort come to decide or infer
the encoding?
From the locale. It looks at the LC_ALL environment variable; if that
isn't set it looks at LC_CTYPE; if that isn't set it looks at LANG;
otherwise the default locale
That doesn't look like a bug to me. For example, the first line
of hints.txt contains
1 . SPACE TAB T
so there are three columns before the tab, which means that expand -t1
and expand -t2 should behave the same. Perhaps you didn't notice the
SPACE?
Jesse Kornblum [EMAIL PROTECTED] writes:
2. The files we are working with will become evidence to be used in
criminal proscecutions.
If I were a criminal defense lawyer attacking your use of 'split' on
technical grounds, I'd ask why you require a particular implementation
of 'split' with
[EMAIL PROTECTED] writes:
Yes, I knew that the space was actually there.
If you use the very first line of hints.txt file, as if you had a
column ruleguide, you may notice that the output (expand-
bug.txt) never show the start of the sentence below the
first five of the ruleguide.
Sorry,
Tim Waugh [EMAIL PROTECTED] writes:
With Linux it is possible to have a single mountpoint mounted several
times; the last mounted fs is the one whose stats are visible, and
they are stacked.
What should df do in this case? Currently it shows the stats for the
last mounted fs for each fs
specification. But the GNU
help string does not need to be so pedantic.
It might be helpful to document this issue better in the coreutils
manual, to help forestall user confusion in the future. I notice a
similar discrepancy with uname -m. Here's a proposed patch.
2003-07-14 Paul Eggert [EMAIL
Parimi, Venkateshwara Rao [EMAIL PROTECTED] writes:
On a solaris 2.6 machine configure dieing with segmentation fault.
This appears to be a bug in your shell.
I can't reproduce the bug in the following environment:
coreutils 5.0 (This is the latest version of textutils.)
Jim Meyering [EMAIL PROTECTED] writes:
Is NEWS-OS 4.2R something you care about? Make variables
Do people still use that OS?
Sony NEWS-OS 4.2R was released in 1992, so I highly doubt it.
However, NEWS-OS 4.2.1a+ (released 1995) is probably still in use by a
diehard or two, since it was
Andreas Schwab [EMAIL PROTECTED] writes:
LC_CTYPE=de_DE.UTF-8
LC_TIME=POSIX
LC_COLLATE=POSIX
This is fine-tuned to DTRT(tm) for my needs. Maybe the LC_CTYPE and/or
LC_COLLATE settings are confusing glibc. I'll check that, and see if
setting them to en_US.UTF-8 helps.
POSIX says:
If
input file
+ with the same --random-seed=SEED option twice, you'll get the same
+ output. The default SEED is chosen at random, and contains enough
+ information to ensure that the output permutation is random.
+ suggestion from Feth AREZKI, Stephan Kasal, and Paul Eggert on 2003-07-17
Parimi, Venkateshwara Rao [EMAIL PROTECTED] writes:
With the same shells it works fine on another machine with Solaris 2.6.
What is the difference between the two machines? That would be
one way to track down the problem.
I still suggest that you install the Sun-recommended patches in
Andreas Schwab [EMAIL PROTECTED] writes:
2003-07-18 Andreas Schwab [EMAIL PROTECTED]
* memcoll.c (memcoll) [!HAVE_STRCOLL]: Clear errno.
--- coreutils-5.0.1/lib/memcoll.c.~1~ 2003-06-06 22:11:58.0 +0200
+++ coreutils-5.0.1/lib/memcoll.c 2003-07-18 11:58:37.0
.
2003-07-18 Paul Eggert [EMAIL PROTECTED]
* jm-macros.m4 (jm_MACROS): Invoke gl_MEMCOLL.
* memcoll.m4: New file, from gnulib.
Index: m4/jm-macros.m4
===
RCS file: /cvsroot/coreutils/coreutils/m4/jm-macros.m4,v
9,999,999 bytes. The fix handles only the case of regular
files, but that's an important case.
This patch fixes a couple of other minor related bugs that I noticed
while implementing this change: there was a confusion between and NULL,
and a possible arithmetic overflow.
2003-07-20 Paul Eggert [EMAIL
Wendy Palm [EMAIL PROTECTED] writes:
Copyright \(co 2003 Free Software Foundation, Inc.
\(co gives a copyright symbol on my linux box, but
results in the curser going to the top of the window
and overwriting on all my other machines
sun blade (5.8 Generic_108528-19 sun4u sparc
until
it's all ready.
Anyway, I noticed on glitch in the merged version; here's a patch.
2003-07-20 Paul Eggert [EMAIL PROTECTED]
* src/wc.c (get_input_fstatus): Fix typo: `stat' was being
invoked with a null pointer when there were no file arguments.
--- wc.c.~1.88.~Sun
[EMAIL PROTECTED] (Bob Proulx) writes:
But wouldn't someone need 'col -b'? I have always used it that way
and now can't say if it is require or not.
: -b Do not output any backspaces, printing only the last character
: written to each column position.
Otherwise the
coreutils into gnulib after applying the following patch to the
coreutils version. (The long list of copyright years is for
conformance to the copyright advice in the gnu maintainer's manual.)
2003-07-22 Paul Eggert [EMAIL PROTECTED]
* xalloc.h (XCALLOC, XREALLOC, CCLONE): Fix under
Maciek Olczak [EMAIL PROTECTED] writes:
# echo ~${file}~
~~
# [ -d ${file} ]; echo $?
0
POSIX requires this behavior. POSIX says that the shell command
`[ FOO ]' exits true (status 0) if FOO is not null, false (status
1) otherwise. In this case FOO is -d, which is not null, so
`[ -d ]'
Alfred M. Szmidt [EMAIL PROTECTED] writes:
Because the same syntax is used in other places, e.g.:
install -m %-rw-rw-r-- file directory
mkdir -m %drwxr-xr-x directory
and similarly for mkfifo and mknod.
Cool, I didn't know this.
Sorry, I didn't explain myself clearly
: ## ##
Here is a patch. (Perhaps all of coreutils should invoke
AC_CHECK_HEADERS in the new form? But I wasn't that ambitious here.)
2003-08-05 Paul Eggert [EMAIL PROTECTED]
* m4/prereq.m4 (jm_PREREQ_PHYSMEM, jm_PREREQ_STAT): Ignore headers
that are present but cannot be compiled
is not using some wide-char functions that it could use on Solaris
2.5.1. However, these problems don't prevent the coreutils build from
succeeding, so I didn't investigate this further.
2003-08-08 Paul Eggert [EMAIL PROTECTED]
* m4/vasnprintf.m4 (gl_PREREQ_VASNPRINTF):
Solaris 2.5.1
Ed Avis [EMAIL PROTECTED] writes:
On a system where local time is +0100:
% (unset TZ; date +%z)
+0100
% (TZ= date +%z)
+
Are you sure this is correct?
This is a C library issue, not a coreutils issue, since coreutils
simply repeats what the C library tells it.
In many C libraries,
4a
===
-8a
+12 a
I happen to have a Network Appliance and Solaris 8 box, and looked
into this. Here is a patch.
2003-08-08 Paul Eggert [EMAIL PROTECTED]
* tests/du/basic: Ensure that a/b/F has at least 65 bytes too.
--- basic.~1.10.~ Fri Mar 28 04:55:04 2003
Alfred M. Szmidt [EMAIL PROTECTED] writes:
Why not fix these problems by disambiguating the syntax? You can put
a new character in front of the new-format mode strings. E.g.,
chmod %-rw-rw-r-- file
Why not fix it with a seperate option instead (which this basicly is)?
that don't have
stdlib.h? If so, you probably wouldn't like that part of the
following patch:
2003-08-18 Paul Eggert [EMAIL PROTECTED]
* lib/same.c: Include stdlib.h and string.h unconditionally,
as we're now assuming that part of hosted C89.
(free) [!HAVE_DECL_FREE]: Remove
Wendy Palm [EMAIL PROTECTED] writes:
with the change i proposed, the only error i have with the tests is
date-tests and i haven't had time to check what that problem is yet.
I doubt whether the tests exercise the md5 and sha code a lot, and I
suspect that there will be latent bugs, unless
a patch for these problems.
2003-09-02 Paul Eggert [EMAIL PROTECTED]
* NEWS: sort -t '\0' now uses a NUL tab.
sort option order no longer matters, unless POSIX requires it.
* doc/coreutils.texi (sort invocation): -d now overrides -i.
whitespace - blanks; whitespace
Lute Kamstra [EMAIL PROTECTED] writes:
10550 is written as 10k
Thanks for the bug report. I installed the following patch into gnulib.
2003-09-03 Paul Eggert [EMAIL PROTECTED]
* human.c (human_readable): Fix bug that rounded 10501 to 10k.
Bug reported by Lute Kamstra
Robert Millan [EMAIL PROTECTED] writes:
Please apply this patch for uname -o to cope with that. It also adds
knetbsd-gnu, which is the Glibc-based/KNetBSD-based variant of GNU.
Thanks, but your email contained no patch; see
http://mail.gnu.org/archive/html/bug-coreutils/2003-09/msg00017.html.
Koblinger Egmont [EMAIL PROTECTED] writes:
Currently the form ``install -D some files /here/'' (with trailing slash)
is unusable for anything
Unfortunately POSIX standard says that standard utilities must ignore
any trailing slashes on existing directory names. POSIX does not
standardize
Buciuman Adrian [EMAIL PROTECTED] writes:
I've post a message some weeks ago about a possible dd bug.
I assume that you're talking about this one?
http://mail.gnu.org/archive/html/bug-coreutils/2003-08/msg00105.html
What do you think about ? Is that indeed a bug ?
It's not easy to say. I
Buciuman Adrian [EMAIL PROTECTED] writes:
I have dozens of faulty diskettes , so I can snail- mail some to you.
Thanks, but I'd rather figure it out directly. Possibly my floppy
drive reader will behave differently with those diskettes.
But I am now almost sure something is wrong in GNU dd ,
Jim covered most of the points nicely, but I thought I'd fill in some
corners:
Aron Griffis [EMAIL PROTECTED] writes:
I don't see a motivation for their removal, since the older syntax
is unambiguous and in heavy use by thousands of UNIX scripts already
in existence.
As Jim said, the
it's your problem.
Here is a patch to CVS coreutils. I can't easily test this patch,
though, since I don't have your hardware. Can you try this patch out?
2003-09-22 Paul Eggert [EMAIL PROTECTED]
Fix bug that may be related to Buciuman Adrian's bug report in
http://mail.gnu.org
[EMAIL PROTECTED] (Bob Proulx) writes:
Could you describe what that option does for those of us not having
access to a solaris machine?
On Solaris 9, nohup -p pid ... arranges for the referenced processes
to become immune to hangups.
Similiarly, nohup -g pgid ... arranges for the referenced
erno palonheimo [EMAIL PROTECTED] writes:
i noticed that -x option of du, which is supposed to make it stay
inside one filesystem, doesn't apparently do anything.
It works for me, with CVS coreutils and
Linux kernel 2.4.18-bf2.4, when I use du -x /.
Possibly the bug has been fixed since
Sorry, but it looks to me like your host has been rootkitted. If so,
it's probably time to reinstall from scratch, and to do a better job
of securing it next time.
___
Bug-coreutils mailing list
[EMAIL PROTECTED]
JGraham [EMAIL PROTECTED] writes:
tell me you've never run a process that needed to run, then realized
that you had to log off?
I can't tell you that. OK, you're starting to convince me.
http://www.chiark.greenend.org.uk/~peterb/linux/interfereproc/
suggests that it'll be a bit tricky to get
Jim Meyering [EMAIL PROTECTED] writes:
So this would work like bash's `disown -h'?
I think it's stronger than disown -h. Not that I'm an expert, but my
impression is that disown -h merely arranges for the subprocess to
continue undisturbed even if Bash is HUPped. But Solaris 'nohup -p
27'
; but they aren't checked in under CVS,
so all the 'sort' tests fail.
2003-10-12 Paul Eggert [EMAIL PROTECTED]
* tests/du/no-x: Change wording of diagnostic to match latest du.c.
* tests/sort/sort-tests: Remove from CVS; assume that people
brave enough to check coreutils out
into coreutils soon.
2003-10-27 Paul Eggert [EMAIL PROTECTED]
* host-os.m4 (UTILS_HOST_OS): Identify GNU/KFreeBSD and
GNU/KNetBSD using their new names (which have a K in front of
the kernel name). Requested by Richard Stallman.
Index: host-os.m4
Robert Millan [EMAIL PROTECTED] writes:
http://mail.gnu.org/archive/html/bug-coreutils/2003-09/msg00032.html
Thanks. But I didn't intend to change the netbsd*-gnu rule. That one
corresponds to the system based on NetBSD's kernel and libc (as opposed to
knetbsd*-gnu which is NetBSD's
I don't get the results that you do. What is the output of the following
commands on your host?
sort --version
locale
[EMAIL PROTECTED] writes:
[EMAIL PROTECTED]ls -l | sort -k 5
total 3165056
-rw-r--r-- 1 robertd users 678 Jul 10 15:15
rejected_tms_inbox.dmp.gz
-rw-r--r-- 1
Patricio Baya [EMAIL PROTECTED] writes:
[EMAIL PROTECTED] controles]$ date
vie oct 31 13:41:49 ART 2003
[EMAIL PROTECTED] controles]$ date -d 1 month ago
mi oct 1 13:42:04 ART 2003
The re is a big mistake. If today is 31/10/2003 , 1 Month Ago must return or
30/09 or someting else, but not
Stefanos Harhalakis [EMAIL PROTECTED] writes:
I believe that it should have an option to compare
source and destination files before replacing the destination.
I like the basic idea, and have been using a wrapper like yours for many years.
But doesn't the implementation have to be more
ari [EMAIL PROTECTED] writes:
'POSIXLY_CORRECT' may be specified if strict posix compliance must
be followed. I see no reason why this cannot be the case with the
'head' and 'tail' functions.
My feeling is that POSIXLY_CORRECT is not something that users would
normally want to set, whereas
ari [EMAIL PROTECTED] writes:
[EMAIL PROTECTED] said this stuff:
After 10 years of being merely `obsolescent', head -N has finally
been officially declared to be `obsolete'.
I have yet to see a pointer to where the historic usage has been
declared obsolete, outside of personal
[EMAIL PROTECTED] (Paul Jarc) writes:
I think it ought to be possible for the user building coreutils to say
what version of conformance they want - say, with a ./configure
argument - and this should override the unistd.h definition.
That might be reasonable. The coreutils maintainer should
[EMAIL PROTECTED] writes:
From an organizational standpoint, is it appropriate to re-use the
lcm routine residing in od.c in the following manner?
I'd put it into the lib directory, so that the code can be shared.
Also, it should take and return size_t, not unsigned int.
Paul Eggert [EMAIL PROTECTED]
* man/Makefile.am (check-programs-vs-x):
Work even if $(programs) contains '$'.
Work even if 'missing=1' in environment.
Don't report an error simply because $(programs) outputs nothing.
--- man/Makefile.am.~1.27.~ Wed Sep 10 02
I went through cut.c and found some problems on hosts where size_t is
wider than int. While I was at it, I fixed the widths of all the
integer types that I could find. Here's a proposed patch.
2003-11-05 Paul Eggert [EMAIL PROTECTED]
Fix 'cut' problems with size_t overflow
Sorry to be a pain, but I now remember that there's similar code in
diffutils/lib/cmpbuf.c's buffer_lcm function. It attempts to yield
reasonable values if buffer sizes are zero, and it does a sanity check
against maximum reasonable values. This all came about because of
freaky buffer sizes in
, awk and
printf) require that the values must be parsed in the C locale, not in
the user-specified locale.
Clearly coreutils printf is nonconforming now, since it doesn't do
that. I vote that other utilities be consistent with POSIX. Here's a
proposed patch.
2003-11-24 Paul Eggert [EMAIL
Jim Meyering [EMAIL PROTECTED] writes:
Wouldn't it be better (in case xstrtod is called with a
modified LC_NUMERIC locale) if c_strtod restored the original
setting for LC_NUMERIC, rather than the environment-derived one?
I was lazy and assumed that c_strtod would be called
only in
potential buffer overruns,
sigh). Rather than fix the problems piecemeal, how about if we adjust
'ls' so that it adjusts all the column widths to fit the data, not
just the file sizes? Here's a proposed patch, relative to CVS
coreutils.
2003-12-03 Paul Eggert [EMAIL PROTECTED]
* NEWS: ls
.
2003-12-03 Paul Eggert [EMAIL PROTECTED]
* getgroups.c (getgroups): xmalloc takes one argument, not two.
Bug reported by Alfred M. Szmidt.
Index: getgroups.c
===
RCS file: /cvsroot/gnulib/gnulib/lib/getgroups.c,v
Jim Meyering [EMAIL PROTECTED] writes:
Please call the new option --output=FILE (no short-named option).
If you do this, please also update the documentation:
I see one detail here that needs to be nailed down.
One might think that the user could easily implement the equivalent of:
nohup
Michael Rasmussen [EMAIL PROTECTED] writes:
The -b option is missing. ??
Still present in v4.5.3
What is touch -b supposed to do? It's not in the standard; see:
http://www.opengroup.org/onlinepubs/007904975/utilities/touch.html
___
Bug-coreutils
Stephen Newman [EMAIL PROTECTED] writes:
For years the option on tail was tail -number, now all of a sudden
it's tail -n (number)
is this a bug, if not why was it changed?
This is covered in the coreutils manual; look for _POSIX2_VERSION.
Perhaps your distribution provider changed the version
Gianni Ciolli [EMAIL PROTECTED] writes:
The output of
$ uniq --help
contains the line
-W, --check-fields=N compare no more than N fields in lines
This isn't true of the latest coreutils release. Perhaps you have
a version of uniq that is modified in your GNU/Linux distribution?
If
Chris Van Nuys [EMAIL PROTECTED] writes:
An example of the output (of course, it all lines up in the Unix
shell, doesn't paste well into Eudora though):
This seems to be a problem with your patch as well. Can you please
send your patches using a mechanism that doesn't mess up tabs or white
Thierry Vignaud [EMAIL PROTECTED] writes:
# unpatched coreutils-5.0.91:
$ perl -e 'system(/usr/bin/[, 1, =, 1, ])'
/usr/bin/[: too many arguments
I don't get that behavior with unpatched coreutils 5.1.0;
please see the transcript below.
One theory is that your packaging somehow created a
Jim Meyering [EMAIL PROTECTED] writes:
So here's a better spec:
If canonicalize_file_name (FILE) succeeds, work exactly as now.
else if lstat (FILE) succeeds,
linkname = readlink (FILE)
if linkname starts with `/', print linkname
otherwise, print $(readlink -f $(dirname
Paolo Montrasio [EMAIL PROTECTED] writes:
Paul's suggestion can be modified in this way:
1) read the timestamp of the existing destination file (if it doesn't
exist yet we don't have any problem and we can resume normal
processing)
2) write the current time to destination file's atime
3)
zone indication
that isn't documented as being supported. Here's an untested patch.
2004-01-19 Paul Eggert [EMAIL PROTECTED]
* tests/touch/relative: Use TZ=UTC0, not TZ=utc (which isn't
portable). Problem reported by Christian Krackowizer. Also, use
+ rather than +0
Warren L Dodge [EMAIL PROTECTED] writes:
On the two solaris systems the following command works.
tail -1 file
On linux is complains and says to use
tail -n 1 file
Any idea what caused this?
It's because your GNU/Linux systems have unistd.h that claim
conformance to POSIX 1003.1-2001,
[EMAIL PROTECTED] (Paul Jarc) writes:
Paul Eggert [EMAIL PROTECTED] wrote:
+ * tests/touch/relative: For 'ls' use TZ=UTC0, not TZ=utc (which isn't
+ portable).
OOC, is TZ=UTC portable?
TZ=UTC is not portable, but TZ=UTC0 has been standardized by POSIX for
many years, and is portable
[EMAIL PROTECTED] (Paul Jarc) writes:
This is unclear to me. It sounds like what you mean is that the
*spec* change predates 4.5.1, but the actually-exhibited *behavior*
change is only a few months old. Is that right?
I meant that the coreutils code itself introduced these changes before
Bernard Giroud [EMAIL PROTECTED] writes:
Does POSIX specify all exit codes that might be returned from utilities ?
It specifies them in some cases, but not all. For example, for diff
0 means success, 1 means differences, 1 means error. For env 0
means success, 127 means not found, 126 means
the ball rolling by installing the
following into Autoconf.
2004-01-26 Paul Eggert [EMAIL PROTECTED]
* doc/autoconf.texi (Default Includes): Include stdint.h even if
HAVE_INTTYPES_H is defined. This is needed on Tru64 5.1b with
Compac C V6.5-207 (dtk), which defines uintmax_t
Paul Eggert [EMAIL PROTECTED]
* src/stat.c: Include timespec.h.
(human_time): Accept nanoseconds arg. Also, accept time
rather than pointer-to-const-time, for clarity. All callers changed.
(print_stat): Pass nanoseconds arg to human_time.
Problem reported
John D. Ballentine III [EMAIL PROTECTED] writes:
the HP make failed with yet another interesting series of messages:
-g -O2 -c `test -f 'human.c' || echo './'`human.c
human.c: In function `adjust_value':
human.c:111: parse error before `l'
This is strange, since there isn't an identifier
Jeff Martens [EMAIL PROTECTED] writes:
wc: /mnt/floppy/teaching.pdf:2447: Invalid or incomplete multibyte or wide character
I think this bug is fixed in the latest coreutils
ftp://alpha.gnu.org/gnu/coreutils/coreutils-5.1.2.tar.gz;
if not, please let us know the details.
the same as saying that ` and ' are appropriate
for the locale.
Here's the proposed patch to the coreutils documentation, which
reflects your other comments:
2004-02-09 Paul Eggert [EMAIL PROTECTED]
* doc/coreutils.texi (Formatting the file names):
Improve wording for --quoting
[EMAIL PROTECTED] (Paul Jarc) writes:
rm --as-is foo would be equivalent to unlink foo, and
I didn't suggest adding this option to rm. rm's determination of
whether to treat the path as a directory is already determined
entirely from the command line, from -r/-R.
Not exactly: if foo is
[EMAIL PROTECTED] (Peter Seebach) writes:
BLOCKSIZEThe size of the block units used by several commands,
most notably df(1), du(1) and ls(1).
Can you find a complete list of BSD programs that use getbsize,
and which contexts they use it in?
Here's why I'd
Reuben Thomas [EMAIL PROTECTED] writes:
What I expect when I run with LANG=en_GB.UTF-8 is either for uniq to
return an error (because the file is not valid text), or to print
one single line (if it's being lenient).
Can you please try coreutils-5.2.0? It has some patches in this area.
of
POSIXLY_CORRECT. This conforms, is because POSIX allows the chown
user.group file behavior as an upward-compatible extension, so long
as it comes into play only where a vanilla chown would report an error
(which is what the code already does).
Here is a proposed patch.
2004-02-23 Paul Eggert [EMAIL
I ran into some make check failures when I set BLOCK_SIZE in my
environment before running the coreutils tests. Here's a patch.
2004-02-25 Paul Eggert [EMAIL PROTECTED]
* tests/du/deref-args, tests/du/exclude, tests/du/slash,
tests/du/trailing-slash: Run envvar-check in case
Thanks for the bug report. 'configure' is generated from 'configure.ac',
so I assume the following patch would be better.
2004-03-01 Paul Eggert [EMAIL PROTECTED]
* configure.ac: Include signal.h when checking for strsignal,
sys_siglist, and friends. Problem reported by Toney
--version
sort (coreutils) 5.2.0
Written by Mike Haertel and Paul Eggert.
Copyright (C) 2004 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
$ uname -a
Linux penguin
Dan Jacobson [EMAIL PROTECTED] writes:
The join man page should put the obsolescent options last.
The join Info page should say which options are obsolescent.
You don't mention which version you're talking about, but I think it's
fixed in coreutils 5.2.0, where the man page doesn't mention the
Tim Waugh [EMAIL PROTECTED] writes:
-du -kD slink-to-64k | sed 's/^6[0-9]/64/' out
+du -kD slink-to-64k | sed 's/^6[0-9]\|72/64/' out
Alas, that \| isn't portable POSIX code. Perhaps better would be
du -kD slink-to-64k | sed 's/^[67][0-9]/64/' out
At Wed, 10 Mar 2004 15:30:46 -0300, Marcelo Ricardo Leitner [EMAIL PROTECTED] writes:
I got here some bash scripts that broke with this change and I'm
wondering why change such a thing, used for many years?.
My original motivation for proposing the patch was that ls's output
lines were getting
optimizations.
2004-03-27 Paul Eggert [EMAIL PROTECTED]
* NEWS: cp -pu and mv -u (when copying) now take the destination
file system time stamp resolution into account.
* doc/coreutils.texi (mv invocation): Document this.
(cp invocation): Document -u (it was missing
Piotr Pitkowski [EMAIL PROTECTED] writes:
But having 'date +%y%m%d -d yesterday' called just after midnight on
march 28 and 29, I've got the same date (040327) due to daylight saving
time shift.
I couldn't reproduce this bug with coreutils 5.2.1. I set
TZ=Europe/Warsaw in the environment and
Nicolas BENOIT [EMAIL PROTECTED] writes:
$df /dev/hda7 -h -T
FilesystemTypeSize Used Avail Use% Mounted on
/dev/hda7 vfat126G 35G 91G 28% /mnt/prtg
$strace df /dev/hda7 -h -T 21 | grep statfs
statfs(/mnt/prtg, {f_type=MSDOS_SUPER_MAGIC, f_bsize=32768,
for reporting that. Here is a proposed patch to Automake's
install-sh to handle this case.
2004-04-01 Paul Eggert [EMAIL PROTECTED]
* lib/install-sh: If mv -f works, use it, and fall back to
the old test -f + rm -f + mv method only if mv -f does
not work. This improves
1 - 100 of 3249 matches
Mail list logo