1If you're interested in helping, here are some tasks that we've considered
2over the years.  Beware: some are quite old and no longer valid.  To avoid
3wasting your time by duplicating work or by working on a task that is no
4longer pertinent, please search the mailing list and post your intent
5before embarking on a big project.
6
7==================================================
8Modify chmod so that it does not change an inode's st_ctime
9  when the selected operation would have no other effect.
10  First suggested by Hans Ecke <http://hans.ecke.ws> in
11    http://thread.gmane.org/gmane.comp.gnu.coreutils.bugs/2920
12  Discussed more recently on <http://bugs.debian.org/497514>.
13
14document the following in coreutils.texi:
15  [
16  pinky
17
18Suggestion from Paul Eggert:
19  More generally, there's not that much use for imaxtostr nowadays,
20  since the inttypes module and newer versions of gettext allow things
21  like _("truncating %s at %" PRIdMAX " bytes") to work portably.
22  I suspect that (if someone cares to take the time) we can remove
23  all instances of imaxtostr and umaxtostr in coreutils and gnulib.
24
25cp --recursive: use fts and *at functions to perform directory traversals
26  in source and destination hierarchy rather than forming full file names.
27  The latter (current) approach fails unnecessarily when the names
28  become very long, and requires space and time that is quadratic in the
29  depth of the hierarchy.  [Bo Borgerson is working on this]
30
31printf:
32  Now that gnulib supports *printf("%a"), import one of the
33  *printf-posix modules so that printf(1) will support %a even on
34  platforms where the native *printf(3) is deficient.
35  Suggestion from Eric Blake.
36
37renice: POSIX utility, needs implementing.
38  suggestion from Karl Berry (among others).
39  Bob Proulx is working on this.
40
41doc/coreutils.texi:
42  Address this comment: FIXME: mv's behavior in this case is system-dependent
43  Better still: fix the code so it's *not* system-dependent.
44
45ls: add --format=FORMAT option that controls how each line is printed.
46
47cp --no-preserve=X should not attempt to preserve attribute X
48  reported by Andreas Schwab
49
50copy.c: Address the FIXME-maybe comment in copy_internal.
51And once that's done, add an exclusion so that `cp --link'
52no longer incurs the overhead of saving src. dev/ino and dest. filename
53in the hash table.
54
55Write an autoconf test to work around build failure in HPUX's 64-bit mode.
56See notes in README -- and remove them once there's a work-around.
57
58Integrate use of sendfile, suggested here:
59  http://mail.gnu.org/archive/html/bug-fileutils/2003-03/msg00030.html
60I don't plan to do that, since a few tests demonstrate no significant benefit.
61
62printf: consider adapting builtins/printf.def from bash
63
64tail: don't use xlseek; it *exits*.
65  Instead, maybe use a macro and return nonzero.
66
67tr: support nontrivial equivalence classes, e.g. [=e=] with LC_COLLATE=fr_FR
68
69lib/strftime.c: Since %N is the only format that we need but that
70  glibc's strftime doesn't support, consider using a wrapper that
71  would expand /%(-_)?\d*N/ to the desired string and then pass the
72  resulting string to glibc's strftime.
73
74unexpand: [http://www.opengroup.org/onlinepubs/007908799/xcu/unexpand.html]
75  printf 'x\t \t y\n'|unexpand -t 8,9 should print its input, unmodified.
76  printf 'x\t \t y\n'|unexpand -t 5,8 should print "x\ty\n"
77
78Let GNU su use the `wheel' group if appropriate.
79  (there are a couple patches, already)
80
81sort: Investigate better sorting algorithms; see Knuth vol. 3.
82
83  We tried list merge sort, but it was about 50% slower than the
84  recursive algorithm currently used by sortlines, and it used more
85  comparisons.  We're not sure why this was, as the theory suggests it
86  should do fewer comparisons, so perhaps this should be revisited.
87  List merge sort was implemented in the style of Knuth algorithm
88  5.2.4L, with the optimization suggested by exercise 5.2.4-22.  The
89  test case was 140,213,394 bytes, 426,4424 lines, text taken from the
90  GCC 3.3 distribution, sort.c compiled with GCC 2.95.4 and running on
91  Debian 3.0r1 GNU/Linux, 2.4GHz Pentium 4, single pass with no
92  temporary files and plenty of RAM.
93
94  Since comparisons seem to be the bottleneck, perhaps the best
95  algorithm to try next should be merge insertion.  See Knuth section
96  5.3.1, who credits Lester Ford, Jr. and Selmer Johnson, American
97  Mathematical Monthly 66 (1959), 387-389.
98
99shred: Update shred as described here to conform to DoD 5220 rules:
100http://lists.gnu.org/archive/html/bug-coreutils/2007-05/msg00075.html
101
102Remove suspicious uses of alloca (ones that may allocate more than
103   about 4k)
104
105Adapt these contribution guidelines for coreutils:
106  http://sources.redhat.com/automake/contribute.html
107
108Improve test coverage.
109  See HACKING for instructions on generating an html test coverage report.
110  Find a program that has poor coverage and improve.
111
112Changes expected to go in, someday.
113======================================
114
115  dd patch from Olivier Delhomme
116
117  test/mv/*: clean up $other_partition_tmpdir in all cases
118
119  ls: when both -l and --dereference-command-line-symlink-to-dir are
120  specified, consider whether to let the latter select whether to
121  dereference command line symlinks to directories.  Since -l has
122  an implicit --NO-dereference-command-line-symlink-to-dir meaning.
123  Pointed out by Karl Berry.
124
125  dd: consider adding an option to suppress `bytes/block read/written'
126  output to stderr.  Suggested here:
127    http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=165045
128
129  Pending copyright papers:
130  ------------------------
131  getpwnam from Bruce Korb
132
133  pb (progress bar) from Miika Pekkarinen
134
135  ------------------------------
136
137Remove long-deprecated options.  Search case-insensitive for
138`deprecated' and `remove in '.  Automate this.
139
140Add a distcheck-time test to ensure that every distributed
141file is either read-only(indicating generated) or is
142version-controlled and up to date.
143
144remove `%s' notation (now that they're all gone, add a maint.mk sc_
145    rule to ensure no new ones are added):
146  grep -E "\`%.{,4}s'" src/*.c
147
148remove all uses of the `register' keyword: Done.  add a maint.mk rule
149  for this, too.
150
151remove or adjust chown's --changes option, since it
152  can't always do what it currently says it does.
153
154Support arbitrary-precision arithmetic in those tools for which it
155makes sense.  Factor and expr already support this via libgmp.
156The "test" program is covered via its string-based comparison of
157integers.  To be converted: seq.
158
159Adapt tools like wc, tr, fmt, etc. (most of the textutils) to be
160  multibyte aware.  The problem is that I want to avoid duplicating
161  significant blocks of logic, yet I also want to incur only minimal
162  (preferably `no') cost when operating in single-byte mode.
163
164pr's use of nstrftime can make it malloc a very large (up to SIZE_MAX) buffer
165
166-----
167
168Copyright (C) 2002-2010 Free Software Foundation, Inc.
169
170This program is free software: you can redistribute it and/or modify
171it under the terms of the GNU General Public License as published by
172the Free Software Foundation, either version 3 of the License, or
173(at your option) any later version.
174
175This program is distributed in the hope that it will be useful,
176but WITHOUT ANY WARRANTY; without even the implied warranty of
177MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
178GNU General Public License for more details.
179
180You should have received a copy of the GNU General Public License
181along with this program.  If not, see <http://www.gnu.org/licenses/>.
182