1 add unit tests for lib/*.c
3 strip: add an option to specify the program used to strip binaries.
4 suggestion from Karl Berry
7 Address this comment: FIXME: mv's behavior in this case is system-dependent
8 Better still: fix the code so it's *not* system-dependent.
10 ls: add --format=FORMAT option that controls how each line is printed.
12 cp --no-preserve=X should not attempt to preserve attribute X
13 reported by Andreas Schwab
15 copy.c: Address the FIXME-maybe comment in copy_internal.
16 And once that's done, add an exclusion so that `cp --link'
17 no longer incurs the overhead of saving src. dev/ino and dest. filename
20 See if we can be consistent about where --verbose sends its output:
21 These all send --verbose output to stdout:
22 head, tail, rm, cp, mv, ln, chmod, chown, chgrp, install, ln
23 These send it to stderr:
27 Write an autoconf test to work around build failure in HPUX's 64-bit mode.
28 See notes in README -- and remove them once there's a work-around.
30 Integrate use of sendfile, suggested here:
31 http://mail.gnu.org/archive/html/bug-fileutils/2003-03/msg00030.html
32 I don't plan to do that, since a few tests demonstrate no significant benefit.
34 Should printf '\0123' print "\n3"?
35 per report from TAKAI Kousuke on Mar 27
36 http://mail.gnu.org/archive/html/bug-coreutils/2003-03/index.html
38 printf: consider adapting builtins/printf.def from bash
40 df: add `--total' option, suggested here http://bugs.debian.org/186007
42 seq: give better diagnostics for invalid formats:
43 e.g. no or too many % directives
44 seq: consider allowing format string to contain no %-directives
46 resolve RH report on cp -a forwarded by Tim Waugh
48 provide an ls option to do something like `--sort directory'
50 tail: don't use xlseek; it *exits*.
51 Instead, maybe use a macro and return nonzero.
53 add mktemp? Suggested by Nelson Beebe
55 df: alignment problem of `Used' heading with e.g., -mP
56 reported by Karl Berry
58 tr: support nontrivial equivalence classes, e.g. [=e=] with LC_COLLATE=fr_FR
60 lib/strftime.c: Since %N is the only format that we need but that
61 glibc's strftime doesn't support, consider using a wrapper that
62 would expand /%(-_)?\d*N/ to the desired string and then pass the
63 resulting string to glibc's strftime.
65 sort: Compress temporary files when doing large external sort/merges.
66 This improves performance when you can compress/uncompress faster than
67 you can read/write, which is common in these days of fast CPUs.
68 suggestion from Charles Randall on 2001-08-10
70 sort: Add an ordering option -R that causes 'sort' to sort according
71 to a random permutation of the correct sort order. Also, add an
72 option --random-seed=SEED that causes 'sort' to use an arbitrary
73 string SEED to select which permutations to use, in a deterministic
74 manner: that is, if you sort a permutation of the same input file
75 with the same --random-seed=SEED option twice, you'll get the same
76 output. The default SEED is chosen at random, and contains enough
77 information to ensure that the output permutation is random.
78 suggestion from Feth AREZKI, Stephan Kasal, and Paul Eggert on 2003-07-17
80 unexpand: [http://www.opengroup.org/onlinepubs/007908799/xcu/unexpand.html]
81 printf 'x\t \t y\n'|unexpand -t 8,9 should print its input, unmodified.
82 printf 'x\t \t y\n'|unexpand -t 5,8 should print "x\ty\n"
84 Let GNU su use the `wheel' group if appropriate.
85 (there are a couple patches, already)
87 sort: Investigate better sorting algorithms; see Knuth vol. 3.
89 We tried list merge sort, but it was about 50% slower than the
90 recursive algorithm currently used by sortlines, and it used more
91 comparisons. We're not sure why this was, as the theory suggests it
92 should do fewer comparisons, so perhaps this should be revisited.
93 List merge sort was implemented in the style of Knuth algorithm
94 5.2.4L, with the optimization suggested by exercise 5.2.4-22. The
95 test case was 140,213,394 bytes, 426,4424 lines, text taken from the
96 GCC 3.3 distribution, sort.c compiled with GCC 2.95.4 and running on
97 Debian 3.0r1 GNU/Linux, 2.4GHz Pentium 4, single pass with no
98 temporary files and plenty of RAM.
100 Since comparisons seem to be the bottleneck, perhaps the best
101 algorithm to try next should be merge insertion. See Knuth section
102 5.3.1, who credits Lester Ford, Jr. and Selmer Johnson, American
103 Mathematical Monthly 66 (1959), 387-389.
105 cp --recursive: perform dir traversals in source and dest hierarchy rather
106 than forming full file names. The latter (current) approach fails
107 unnecessarily when the names become very long.
109 Remove suspicious uses of alloca (ones that may allocate more than
112 Adapt these contribution guidelines for coreutils:
113 http://sources.redhat.com/automake/contribute.html
116 Changes expected to go in, someday.
117 ======================================
119 dd patch from Olivier Delhomme
121 Andreas Gruenbacher's xattr changes
123 Apply Bruno Haible's hostname changes
125 test/mv/*: clean up $other_partition_tmpdir in all cases
127 ls: when both -l and --dereference-command-line-symlink-to-dir are
128 specified, consider whether to let the latter select whether to
129 dereference command line symlinks to directories. Since -l has
130 an implicit --NO-dereference-command-line-symlink-to-dir meaning.
131 Pointed out by Karl Berry.
133 A more efficient version of factor, and possibly one that
134 accepts inputs of size 2^64 and larger.
136 dd: consider adding an option to suppress `bytes/block read/written'
137 output to stderr. Suggested here:
138 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=165045
140 Pending copyright papers:
141 ------------------------
142 ls --color: Ed Avis' patch to suppress escape sequences for
143 non-highlighted files
145 getpwnam from Bruce Korb
147 pb (progress bar) from Miika Pekkarinen
149 ------------------------------
151 Have euidaccess.m4 check for eaccess as well as euidaccess
152 If found, then do `#define euidaccess eaccess'.
154 Remove long-deprecated options. Search case-insensitive for
155 `deprecated' and `remove in '.
157 Add a distcheck-time test to ensure that every distributed
158 file is either read-only(indicating generated) or is
159 version-controlled and up to date.
161 Implement Ulrich Drepper's suggestion to use getgrouplist rather
162 than getugroups. This affects only `id', but makes a big difference
163 on systems with many users and/or groups, and makes id usable once
164 again on systems where access restrictions make getugroups fail.
165 But first we'll need a run-test (either in an autoconf macro or at
166 run time) to avoid the segfault bug in libc-2.3.2's getgrouplist.
167 In that case, we'd revert to using a new (to-be-written) getgrouplist
168 module that does most of what `id' already does.
170 remove `%s' notation:
171 grep -E "\`%.{,4}s'" src/*.c
173 remove or adjust chown's --changes option, since it
174 can't always do what it currently says it does.
176 Adapt tools like wc, tr, fmt, etc. (most of the textutils) to be
177 multibyte aware. The problem is that I want to avoid duplicating
178 significant blocks of logic, yet I also want to incur only minimal
179 (preferably `no') cost when operating in single-byte mode.
181 Remove all uses of the `register' keyword
183 pr's use of nstrftime can make it malloc a very large (up to SIZE_MAX) buffer
185 ls.c: use gettime rather than clock_gettime, gettimeofday, time
187 write rules to deal with c99->c89 patches and to automatically
188 check that they apply cleanly
195 should work the same way the underlying
196 rename ("dir", "new-name/") call does.
197 http://lists.gnu.org/archive/html/bug-coreutils/2005-11/msg00292.html