| Age | Commit message (Collapse) | Author |
|---|
|
hosts with 64-bit time_t and 32-bit int.
Include "inttostr.h".
(time_string): New function, copied from who.c.
(print_entry): Use it.
|
|
hosts with 64-bit time_t and 32-bit int.
Include "inttostr.h".
(time_string): If ctime fails, print the raw time as an integer
instead of dumping core.
|
|
Problem reported by Guido Leenders in:
http://lists.gnu.org/archive/html/bug-coreutils/2004-06/msg00056.html
|
|
PACKAGE) once again. This restores `GNU' to the parenthesized package
name in --version output. Before, the first argument from AC_INIT,
`GNU coreutils', would be propagated to the PACKAGE variable. Now,
`GNU ' is trimmed. Reported by Richard Stallman.
|
|
would clash with a typedef in Tru64's <sys/types.h>. From Albert Chin.
|
|
|
|
|
|
|
|
POSIXLY_CORRECT. POSIX does not specify any options, but it
does not prohibit options either, so "unlink" is like "expr" here.
|
|
even if POSIXLY_CORRECT, since POSIX does not require a diagnostic.
|
|
since POSIX says the behavior is unspecified here.
(main): Recognize --help, --version even if
POSIXLY_CORRECT. POSIX does not specify any options, but it
does not prohibit options either, so "printf" is like "expr" here.
|
|
(main): When in a non-POSIX locale, ignore POSIXLY_CORRECT, since
POSIX specifies the behavior only in the POSIX locale.
|
|
is set. POSIX reserves upper-case environment variables to the
implementation, so it's OK for ls to depend on TABSIZE.
|
|
Instead, include "euidaccess.h".
|
|
(UT_ID) [!HAVE_STRUCT_XTMP_UT_ID]: Remove bogus comment,
as (sizeof "??") reliably returns 3.
(print_line): Guard against idle and pid being too long
(which is possible when printing headers).
(print_user): Allocate enough bytes for idlestr. Use IDLESTR_LEN.
Avoid unnecessary cast of sizeof to int.
(make_id_equals_comment): Do not assume that UT_ID returns
a string; it might return a non-null-terminated array.
Use strncat instead. It's not very often where strncat is
exactly what you want, but this is one of those rare cases.
|
|
|
|
|
|
old_uid and old_gid to required_uid and required_gid respectively.
|
|
particular, a no-op chown is no longer reported as a change; this
reverts to previous behavior. Also, fix both commands so that -v
report failures even if the failure is not due to the chmod or
chown syscalls.
* src/chown-core.c (change_file_owner): Return -1 on error, not
1 sometimes and -1 on others. Our caller ORs together our results,
and (-1 | 1) == 0 on ones-complement hosts.
With -v report all errors verbosely, not just some.
Fix bug when chopt->root_dev_ino && !chopt->affect_symlink_referent:
file_stats wasn't set properly in that case.
|
|
|
|
particular, a no-op chown is no longer reported as a change; this
reverts to previous behavior. Also, fix both commands so that -v
report failures even if the failure is not due to the chmod or
chown syscalls.
* src/chmod.c (CH_NOT_APPLIED): New constant.
(describe_change): Handle it.
(process_file): Use it, if a symlink wasn't changed.
(mode_changed): Return bool, not int. Accept new argument
NEW_MODE; all callers changed. This lets us avoid statting the
file unless the new mode has unusual bits.
(process_file): Return -1 on error. With -v, report all errors
verbosely, not just some.
|
|
|
|
|
|
needed. Chown a directory only after chowning its children; this
avoids problems if the new directory ownership doesn't permit
access to the children. Dereference symlinks before doing
ROOT_DEV_INO_CHECK, not after, so that we catch symlinks to /.
Do not optimize away the chown() system call when the file's owner
and group already have the desired value. POSIX does not permit
this optimization. Rely on chown and lchown to do the right
thing with symlinks and/or -1 arguments, now that we have wrappers
to do this. Use ENOTSUPP not ENOSYS, and ignore all ENOTSUPP
errors, not just command-line errors.
(chown_files): Pass FTS_STAT to xfts_open if we don't need file status.
|
|
requires either -H or -L, and -R -h requires -P. If -H, specify
FTS_PHYSICAL as well as FTS_COMFOLLOW; this is faster. Make this
file as much like chown.c as possible.
|
|
requires either -H or -L, and -R -h requires -P. If -H, specify
FTS_PHYSICAL as well as FTS_COMFOLLOW; this is faster. Make this
file as much like chgrp.c as possible.
|
|
|
|
prevented a diagnostic of any operands.
|
|
|
|
|
|
Prompted by a suggestion from Kalle Olavi Niemitalo
in http://bugs.debian.org/207035.
(direct_mode): New function.
(do_wipefd): Turn on direct-mode I/O.
(dopass): If a file's first write fails with EINVAL,
turn off direct-mode I/O and retry the write.
|
|
|
|
The POSIX SYNOPSIS does not allow this option combination.
|
|
a write error. Problem reported by Jon Peatfield in:
http://lists.gnu.org/archive/html/bug-coreutils/2004-06/msg00020.html
|
|
<http://mail.gnu.org/archive/html/bug-coreutils/2003-08/msg00105.html>
where 'dd' created a file that was too large. The bug was that dd
assumed that the input file offset does not advance after a failed
read; but POSIX says that the input file offset is undefined after
a failed read.
(MAX_BLOCKSIZE): New macro.
(input_seekable, input_seek_errno, input_offset,
input_offset_overflow): New vars.
(scanargs): Reject block sizes greater than MAX_BLOCKSIZE.
(advance_input_offset): New function.
(skip_via_lseek): Set errno to zero when reporting our failure,
so that we don't report based on garbage errno.
(skip): If fdesc is standard input, advance the input offset.
Do not quit if reading, and if noerror was specified;
POSIX seems to require this.
If read fails on output file, report the earlier lseek failure
instead; this fixes a FIXME in dd_copy.
(advance_input_after_read_error): New function.
(dd_copy): Use it, instead of assuming that failed reads
do not advance the file pointer. Advance input offset
after nonfailed reads. Advance only a partial block if
the previous read (before the failed read) succeeded, and
do not generate an output block of zeros in this case.
(main): Determine initial input offset, seekability of input,
and error if it wasn't seekable.
|
|
remove a symlink to a file on an off-line NFS-mounted partition.
Reported by David Howells in https://bugzilla.redhat.com/124699.
(write_protected_non_symlink): New function.
Don't invoke euidaccess on symlinks.
(prompt): Use write_protected_non_symlink rather than using
euidaccess directly, being careful not to call lstat twice for a file.
|
|
abutting byte or character ranges. Reported by David Krider in
http://lists.gnu.org/archive/html/bug-coreutils/2004-05/msg00132.html
(print_kth): Remove special case for open-ended range.
(set_fields): Record the range start index for an interval even
when it abuts another interval on its low side.
Also record the range start index of the longest right-open-interval.
|
|
we need to test this in just one place now.
(usage): Mention -C.
(unquote): Note that \055, \n, etc are escaped.
Do not worry about POSIXLY_CORRECT when warning about ambiguous
escape sequences.
\ at end of string stands for itself.
Do not diagnose invalid backslash escapes: POSIX says the behavior
is unspecified in this case, so we don't need to diagnose it.
(main): Add support for -C (currently an alias for -c).
Do not diagnose 'tr [:upper:] [:upper:], as POSIX does not require
a diagnostic here.
|
|
|
|
|
|
|
|
|
|
royally if the link length is long (e.g., GNU/Hurd). Use
xreadlink instead, it's safer. Don't bother to read the link if
it's the wrong size. Add a FIXME because this area is a bit murky
and undocumented.
|
|
Prefer the notation `STREQ (a, b)' over `!strcmp (a, b)'
and `STREQ (a, b)' over `strcmp (a, b) == 0'.
|
|
|
|
Prefer the notation `STREQ (a, b)' over `strcmp (a, b) == 0'.
|
|
Prefer the notation `STREQ (a, b)' over `strcmp (a, b) == 0'.
|
|
|
|
|
|
|
