1 restore djgpp, eventually
3 add unit tests for lib/*.c
5 strip: add an option to specify the program used to strip binaries.
6 suggestion from Karl Berry
9 Address this comment: FIXME: mv's behavior in this case is system-dependent
10 Better still: fix the code so it's *not* system-dependent.
12 implement --target-directory=DIR for install (per texinfo documentation)
14 ls: add --format=FORMAT option that controls how each line is printed.
16 cp --no-preserve=X should not attempt to preserve attribute X
17 reported by Andreas Schwab
19 copy.c: Address the FIXME-maybe comment in copy_internal.
20 And once that's done, add an exclusion so that `cp --link'
21 no longer incurs the overhead of saving src. dev/ino and dest. filename
24 See if we can be consistent about where --verbose sends its output:
25 These all send --verbose output to stdout:
26 head, tail, rm, cp, mv, ln, chmod, chown, chgrp, install, ln
27 These send it to stderr:
31 Write an autoconf test to work around build failure in HPUX's 64-bit mode.
32 See notes in README -- and remove them once there's a work-around.
34 Integrate use of sendfile, suggested here:
35 http://mail.gnu.org/archive/html/bug-fileutils/2003-03/msg00030.html
36 I don't plan to do that, since a few tests demonstrate no significant benefit.
38 Should printf '\0123' print "\n3"?
39 per report from TAKAI Kousuke on Mar 27
40 http://mail.gnu.org/archive/html/bug-coreutils/2003-03/index.html
42 printf: consider adapting builtins/printf.def from bash
44 df: add `--total' option, suggested here http://bugs.debian.org/186007
46 seq: give better diagnostics for invalid formats:
47 e.g. no or too many % directives
48 seq: consider allowing format string to contain no %-directives
50 m4: rename all macros that start with AC_ to start with another prefix
52 resolve RH report on cp -a forwarded by Tim Waugh
54 Martin Michlmayr's patch to provide ls with `--sort directory' option
56 tail: don't use xlseek; it *exits*.
57 Instead, maybe use a macro and return nonzero.
59 add mktemp? Suggested by Nelson Beebe
61 df: alignment problem of `Used' heading with e.g., -mP
62 reported by Karl Berry
64 tr: support nontrivial equivalence classes, e.g. [=e=] with LC_COLLATE=fr_FR
66 lib/strftime.c: Since %N is the only format that we need but that
67 glibc's strftime doesn't support, consider using a wrapper that
68 would expand /%(-_)?\d*N/ to the desired string and then pass the
69 resulting string to glibc's strftime.
71 sort: Compress temporary files when doing large external sort/merges.
72 This improves performance when you can compress/uncompress faster than
73 you can read/write, which is common in these days of fast CPUs.
74 suggestion from Charles Randall on 2001-08-10
76 sort: Add an ordering option -R that causes 'sort' to sort according
77 to a random permutation of the correct sort order. Also, add an
78 option --random-seed=SEED that causes 'sort' to use an arbitrary
79 string SEED to select which permutations to use, in a deterministic
80 manner: that is, if you sort a permutation of the same input file
81 with the same --random-seed=SEED option twice, you'll get the same
82 output. The default SEED is chosen at random, and contains enough
83 information to ensure that the output permutation is random.
84 suggestion from Feth AREZKI, Stephan Kasal, and Paul Eggert on 2003-07-17
86 unexpand: [http://www.opengroup.org/onlinepubs/007908799/xcu/unexpand.html]
87 printf 'x\t \t y\n'|unexpand -t 8,9 should print its input, unmodified.
88 printf 'x\t \t y\n'|unexpand -t 5,8 should print "x\ty\n"
90 Let GNU su use the `wheel' group if appropriate.
91 (there are a couple patches, already)
93 sort: Investigate better sorting algorithms; see Knuth vol. 3.
95 We tried list merge sort, but it was about 50% slower than the
96 recursive algorithm currently used by sortlines, and it used more
97 comparisons. We're not sure why this was, as the theory suggests it
98 should do fewer comparisons, so perhaps this should be revisited.
99 List merge sort was implemented in the style of Knuth algorithm
100 5.2.4L, with the optimization suggested by exercise 5.2.4-22. The
101 test case was 140,213,394 bytes, 426,4424 lines, text taken from the
102 GCC 3.3 distribution, sort.c compiled with GCC 2.95.4 and running on
103 Debian 3.0r1 GNU/Linux, 2.4GHz Pentium 4, single pass with no
104 temporary files and plenty of RAM.
106 Since comparisons seem to be the bottleneck, perhaps the best
107 algorithm to try next should be merge insertion. See Knuth section
108 5.3.1, who credits Lester Ford, Jr. and Selmer Johnson, American
109 Mathematical Monthly 66 (1959), 387-389.
111 cp --recursive: perform dir traversals in source and dest hierarchy rather
112 than forming full file names. The latter (current) approach fails
113 unnecessarily when the names become very long.
115 tail --p is now ambiguous
117 Remove suspicious uses of alloca (ones that may allocate more than
120 Adapt these contribution guidelines for coreutils:
121 http://sources.redhat.com/automake/contribute.html
124 Changes expected to go in, post-5.2.1:
125 ======================================
127 wc: add an option, --files0-from [as for du] to make it read NUL-delimited
128 file name arguments from a file.
130 dd patch from Olivier Delhomme
132 Apply Andreas Gruenbacher's ACL and xattr changes
134 Apply Bruno Haible's hostname changes
136 test/mv/*: clean up $other_partition_tmpdir in all cases
138 ls: when both -l and --dereference-command-line-symlink-to-dir are
139 specified, consider whether to let the latter select whether to
140 dereference command line symlinks to directories. Since -l has
141 an implicit --NO-dereference-command-line-symlink-to-dir meaning.
142 Pointed out by Karl Berry.
144 A more efficient version of factor, and possibly one that
145 accepts inputs of size 2^64 and larger.
147 dd: consider adding an option to suppress `bytes/block read/written'
148 output to stderr. Suggested here:
149 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=165045
151 Pending copyright papers:
152 ------------------------
153 ls --color: Ed Avis' patch to suppress escape sequences for
154 non-highlighted files
156 getpwnam from Bruce Korb
158 pb (progress bar) from Miika Pekkarinen
160 ------------------------------
161 Look into improving the performance of md5sum.
162 `openssl md5' is consistently about 30% faster than md5sum on an idle
163 AMD 2000-XP system with plenty of RAM and a 261 MB input file.
164 openssl's md5 implementation is in assembly, generated by a perl script.
166 On an AMD-64 system, using a 700MB file on a tmpfs file system
167 (and enough RAM so that no actual disk reads were performed),
168 GNU md5sum is slightly faster than `openssl md5', e.g.:
170 2.38s user 0.38s system 100% cpu 2.756 total (gnu md5sum)
172 2.52s user 0.34s system 100% cpu 2.869 total
174 However, `openssl sha1' is about 5% faster than GNU sha1sum:
176 3.32s user 0.33s system 99% cpu 3.653 total (openssl sha1)
177 3.45s user 0.39s system 99% cpu 3.843 total (gnu sha1sum)
179 The above are using the debian-sid (amd_64 alioth) binaries from
180 coreutils-5.2.1. When I compile the latest (coreutils-cvs) with
181 gcc-4.0 -O3, I get slightly (2-3%) better sha1sum performance,
182 and a ~7% *decrease* in performance for md5sum. I suspect that
183 with the right compiler options you can do much better.
184 ------------------------------
186 Have euidaccess.m4 check for eaccess as well as euidaccess
187 If found, then do `#define euidaccess eaccess'.
189 Remove long-deprecated options like tail's --allow-missing
191 Add a distcheck-time test to ensure that every distributed
192 file is either read-only(indicating generated) or is
193 version-controlled and up to date.
195 Implement Ulrich Drepper's suggestion to use getgrouplist rather
196 than getugroups. This affects only `id', but makes a big difference
197 on systems with many users and/or groups, and makes id usable once
198 again on systems where access restrictions make getugroups fail.
199 But first we'll need a run-test (either in an autoconf macro or at
200 run time) to avoid the segfault bug in libc-2.3.2's getgrouplist.
201 In that case, we'd revert to using a new (to-be-written) getgrouplist
202 module that does most of what `id' already does.
204 remove `%s' notation:
205 grep -E "\`%.{,4}s'" src/*.c
207 remove.c should never exit, yet may do so (see uses of EXIT_FAILURE)
209 remove or adjust chown's --changes option, since it
210 can't always do what it currently says it does.
212 Adapt tools like wc, tr, fmt, etc. (most of the textutils) to be
213 multibyte aware. The problem is that I want to avoid duplicating
214 significant blocks of logic, yet I also want to incur only minimal
215 (preferably `no') cost when operating in single-byte mode.
217 Remove all uses of the `register' keyword