Paul Eggert [Tue, 13 Sep 2005 22:08:41 +0000 (22:08 +0000)]
(TIME_SPEC_DATE): No longer needs to be nonzero, so
remove the "=1".
(TIME_SOEC_HOURS, TIME_SPEC_MINUTES): Must be at end now, so put
them there.
(time_spec_string, time_spec): Hours and minutes must be at
start now, so put them there.
(rfc_2822_format): Now a string constant, not a boolean. All uses
changed.
(iso_8601_format, rfc_format): Remove.
(RFC_3339_OPTION): New constant.
(long_options): Add --rfc-3339.
(usage): Add --rfc-3339. Don't mention --iso-8601.
Mention %:z, %::z, %:::z.
(main): Simplify calculation of 'format'; it was getting too hairy
to follow. Add --rfc-3339.
(show_date): Assume format arg is not NULL, which is the case
now. The default code is moved to 'main'. This simplifies things
and allows the default to be calculated just once.
Paul Eggert [Tue, 13 Sep 2005 22:07:58 +0000 (22:07 +0000)]
(Time conversion specifiers, Options for date):
Document date --rfc-3339 and new specifiers %:z, %::z, %:::z. Use
"date and time" consistently; the old version sometimes said "time
and date". Fix a minor bug in the documentation for --rfc-2822:
it claimed day-of-month < 10 had leading space, not leading zero.
Use a consistent format for terms like "RFC".
Jim Meyering [Sat, 10 Sep 2005 13:56:45 +0000 (13:56 +0000)]
csplit could produce corrupt output, given input lines longer than 8KB
(load_buffer): Don't read from free'd memory
when handling lines longer than the initial buffer length.
(save_to_hold_area): Don't leak the previous hold_area buffer.
Reported by Tristan Miller and Luke Kendall.
* NEWS: Mention this.
(load_buffer): Avoid integer overflow in buffer
size calculations for very long lines.
Paul Eggert [Fri, 9 Sep 2005 21:11:07 +0000 (21:11 +0000)]
(SKIP_SOMETHING, find_occurs_in_text):
Store match length in regoff_t,
not int. Assume that negative return values less than -2
represent regoff_t overflow.
(matcher_error): New function.
(SKIP_SOMETHING, find_occurs_in_text): Use it to report matcher errors.
(alloc_and_compile_regex): No longer any need to worry about
int versus size_t mismatch.
Paul Eggert [Fri, 9 Sep 2005 21:09:48 +0000 (21:09 +0000)]
(proc_text): Store match length in regoff_t,
not int. Assume that negative return values less than -2
represent regoff_t overflow.
(build_type_arg): Check for size_t overflow.
Paul Eggert [Thu, 8 Sep 2005 06:52:09 +0000 (06:52 +0000)]
Use the phrase "niceness" instead of "nice value" to describe
the biased nice value that can go negative. This corrects
a discrepancy with POSIX, which states that nice values are
nonnegative.
Jim Meyering [Mon, 5 Sep 2005 17:13:09 +0000 (17:13 +0000)]
(setuid-etc): New test, for today's addition of colorization for
setuid, setgid files and sticky, other-writable, sticky-and-other-writable
directories.
Jim Meyering [Mon, 5 Sep 2005 17:12:07 +0000 (17:12 +0000)]
Colorize set-user-ID and set-group-ID files and sticky,
other-writable, and sticky-and-other-writable directories.
(indicator_no[]): Add new symbols.
(indicator_name[]): Add corresponding mode strings.
(color_indicator[]): Add an entry for each new mode string.
(print_color_indicator): Honor new types.
From Mike Frysinger, based on a patch from Fedora.
Jim Meyering [Mon, 5 Sep 2005 17:11:15 +0000 (17:11 +0000)]
Colorize set-user-ID and set-group-ID files and sticky,
other-writable, and sticky-and-other-writable directories.
(slack_codes): Add new dircolors mode names.
(ls_codes): Add corresponding two-letter ls mode strings.
From Mike Frysinger, based on a patch from Fedora.
Jim Meyering [Mon, 5 Sep 2005 16:57:40 +0000 (16:57 +0000)]
(__builtin_constant_p) [__GNUC__ <= 2]: Define to 1.
(verify_type__) [verify_error_if_non_const__]: New member/test,
to help detect when verify or verify_expr is mistakenly passed
a non-constant argument within a function.
Jim Meyering [Sat, 27 Aug 2005 19:05:24 +0000 (19:05 +0000)]
Go ahead and leave the patch in (i.e., revert
today's change). It doesn't cause a problem after all, since
--exclude-from=- is always handled before --files0-from=F.
Jim Meyering [Sat, 27 Aug 2005 18:12:54 +0000 (18:12 +0000)]
Revert the du.c part of the change from 2005-07-02:
That change (to reopen stdin on F for --files0-from=F) made it so
--exclude-from=- and --files0-from=F would not work together.
Jim Meyering [Wed, 24 Aug 2005 15:25:29 +0000 (15:25 +0000)]
(stdopen): Revert to iterating over descriptors, but
rather than using fstat, use fcntl with its F_GETFD flag, which
is more efficient. Unlike the 2-syscall dup-close approach,
this one doesn't apply any unnecessary pressure on the file
descriptor pool -- thus there is one fewer way to fail.