source: branches/3.2/mindi-busybox/TODO @ 3186

Last change on this file since 3186 was 2725, checked in by Bruno Cornec, 9 years ago
  • Update mindi-busybox to 1.18.3 to avoid problems with the tar command which is now failing on recent versions with busybox 1.7.3
File size: 11.8 KB
1Busybox TODO
3Harvest patches from
6Stuff that needs to be done.  This is organized by who plans to get around to
7doing it eventually, but that doesn't mean they "own" the item.  If you want to
8do one of these bounce an email off the person it's listed under to see if they
9have any suggestions how they plan to go about it, and to minimize conflicts
10between your work and theirs.  But otherwise, all of these are fair game.
12Rob Landley suggested this:
13  Implement bb_realpath() that can handle NULL on non-glibc.
15  sh
16    The command shell situation is a mess.  We have two different
17    shells that don't really share any code, and the "standalone shell" doesn't
18    work all that well (especially not in a chroot environment), due to apps not
19    being reentrant.
21  Do a SUSv3 audit
22    Look at the full Single Unix Specification version 3 (available online at
23    "") and
24    figure out which of our apps are compliant, and what we're missing that
25    we might actually care about.
27    Even better would be some kind of automated compliance test harness that
28    exercises each command line option and the various corner cases.
30  Internationalization
31    How much internationalization should we do?
33    The low hanging fruit is UTF-8 character set support.  We should do this.
34    See TODO_unicode file.
36    We also have lots of hardwired english text messages.  Consolidating this
37    into some kind of message table not only makes translation easier, but
38    also allows us to consolidate redundant (or close) strings.
40    We probably don't want to be bloated with locale support.  (Not unless we
41    can cleanly export it from our underlying C library without having to
42    concern ourselves with it directly.  Perhaps a few specific things like a
43    config option for "date" are low hanging fruit here?)
45    What level should things happen at?  How much do we care about
46    internationalizing the text console when X11 and xterms are so much better
47    at it?  (There's some infrastructure here we don't implement: The
48    "unicode_start" and "unicode_stop" shell scripts need "vt-is-UTF8" and a
49    --unicode option to loadkeys.  That implies a real loadkeys/dumpkeys
50    implementation to replace loadkmap/dumpkmap.  Plus messing with console font
51    loading.  Is it worth it, or do we just say "use X"?)
53  Individual compilation of applets.
54    It would be nice if busybox had the option to compile to individual applets,
55    for people who want an alternate implementation less bloated than the gnu
56    utils (or simply with less political baggage), but without it being one big
57    executable.
59    Turning libbb into a real dll is another possibility, especially if libbb
60    could export some of the other library interfaces we've already more or less
61    got the code for (like zlib).
63  buildroot - Make a "dogfood" option
64    Busybox 1.1 will be capable of replacing most gnu packages for real world
65    use, such as developing software or in a live CD.  It needs wider testing.
67    Busybox should now be able to replace bzip2, coreutils, e2fsprogs, file,
68    findutils, gawk, grep, inetutils, less, modutils, net-tools, patch, procps,
69    sed, shadow, sysklogd, sysvinit, tar, util-linux, and vim.  The resulting
70    system should be self-hosting (I.E. able to rebuild itself from source
71    code).  This means it would need (at least) binutils, gcc, and make, or
72    equivalents.
74    It would be a good "eating our own dogfood" test if buildroot had the option
75    of using a "make allyesconfig" busybox instead of the all of the above
76    packages.  Anything that's wrong with the resulting system, we can fix.  (It
77    would be nice to be able to upgrade busybox to be able to replace bash and
78    diffutils as well, but we're not there yet.)
80    One example of an existing system that does this already is Firmware Linux:
83  initramfs
84    Busybox should have a sample initramfs build script.  This depends on
85    shell, mdev, and switch_root.
87  mkdep
88    Write a mkdep that doesn't segfault if there's a directory it doesn't
89    have permission to read, isn't based on manually editing the output of
90    lexx and yacc, doesn't make such a mess under include/config, etc.
92  Group globals into unions of structures.
93    Go through and turn all the global and static variables into structures,
94    and have all those structures be in a big union shared between processes,
95    so busybox uses less bss.  (This is a big win on nommu machines.)  See
96    sed.c and mdev.c for examples.
98  Go through and close out all of that somehow.
99    This one's open to everybody, but I'll wind up doing it...
101Bernhard Reutner-Fischer <> suggests to look at these:
102  New debug options:
103    -Wlarger-than-127
104    Cleanup any big users
106    make bb_common_bufsiz1 configurable, size wise.
107    make pipesize configurable, size wise.
108    Use bb_common_bufsiz1 throughout applets!
110As yet unclaimed:
114  Make sure we handle empty files properly:
115    From the patch man page:
117    you can remove a file by sending out a context diff that compares
118    the file to be deleted with an empty file dated the Epoch.  The
119    file will be removed unless patch is conforming to POSIX and the
120    -E or --remove-empty-files option is not given.
123  Should have simple fuzz factor support to apply patches at an offset which
124  shouldn't take up too much space.
126  And while we're at it, a new patch filename quoting format is apparently
127  coming soon:
129stty / catv
130  stty's visible() function and catv's guts are identical. Merge them into
131  an appropriate libbb function.
133struct suffix_mult
134  Several duplicate users of: grep -r "1024\*1024" * -B2 -A1
135  Merge to a single size_suffixes[] in libbb.
136  Users: head tail od_bloaty hexdump and (partially as it wouldn't hurt) svlogd
139  ./busybox tail -f foo.c~ TODO
140  should not print fmt=header_fmt for subsequent date >> TODO; i.e. only
141  fmt+ if another (not the current) file did change
143Architectural issues:
145bb_close() with fsync()
146  We should have a bb_close() in place of normal close, with a CONFIG_ option
147  to not just check the return value of close() for an error, but fsync().
148  Close can't reliably report anything useful because if write() accepted the
149  data then it either went out to the network or it's in cache or a pipe
150  buffer.  Either way, there's no guarantee it'll make it to its final
151  destination before close() gets called, so there's no guarantee that any
152  error will be reported.
154  You need to call fsync() if you care about errors that occur after write(),
155  but that can have a big performance impact.  So make it a config option.
157Unify archivers
158  Lots of archivers have the same general infrastructure.  The directory
159  traversal code should be factored out, and the guts of each archiver could
160  be some setup code and a series of callbacks for "add this file",
161  "add this directory", "add this symlink" and so on.
163  This could clean up tar and zip, and make it cheaper to add cpio and ar
164  write support, and possibly even cheaply add things like mkisofs or
165  mksquashfs someday, if they become relevant.
167Text buffer support.
168  Several existing applets (sort, vi, less...) read
169  a whole file into memory and act on it.  Use open_read_close().
171Memory Allocation
172  We have a CONFIG_BUFFER mechanism that lets us select whether to do memory
173  allocation on the stack or the heap.  Unfortunately, we're not using it much.
174  We need to audit our memory allocations and turn a lot of malloc/free calls
176  For a start, see e.g. make EXTRA_CFLAGS=-Wlarger-than-64
178  And while we're at it, many of the CONFIG_FEATURE_CLEAN_UP #ifdefs will be
179  optimized out by the compiler in the stack allocation case (since there's no
180  free for an alloca()), and this means that various cleanup loops that just
181  call free might also be optimized out by the compiler if written right, so
182  we can yank those #ifdefs too, and generally clean up the code.
185  This is more an unresolved issue than a to-do item.  More thought is needed.
187  Normally we rely on exit() to free memory, close files and unmap segments
188  for us.  This makes most calls to free(), close(), and unmap() optional in
189  busybox applets that don't intend to run for very long, and optional stuff
190  can be omitted to save size.
192  The idea was raised that we could simulate fork/exit with setjmp/longjmp
193  for _really_ brainless embedded systems, or speed up the standalone shell
194  by not forking.  Doing so would require a reliable FEATURE_CLEAN_UP.
195  Unfortunately, this isn't as easy as it sounds.
197  The problem is, lots of things exit(), sometimes unexpectedly (xmalloc())
198  and sometimes reliably (bb_perror_msg_and_die() or show_usage()).  This
199  jumps out of the normal flow control and bypasses any cleanup code we
200  put at the end of our applets.
202  It's possible to add hooks to libbb functions like xmalloc() and xopen()
203  to add their entries to a linked list, which could be traversed and
204  freed/closed automatically.  (This would need to be able to free just the
205  entries after a checkpoint to be usable for a forkless standalone shell.
206  You don't want to free the shell's own resources.)
208  Right now, FEATURE_CLEAN_UP is more or less a debugging aid, to make things
209  like valgrind happy.  It's also documentation of _what_ we're trusting
210  exit() to clean up for us.  But new infrastructure to auto-free stuff would
211  render the existing FEATURE_CLEAN_UP code redundant.
213  For right now, exit() handles it just fine.
216Minor stuff:
217  watchdog.c could autodetect the timer duration via:
218    if(!ioctl (fd, WDIOC_GETTIMEOUT, &tmo)) timer_duration = 1 + (tmo / 2);
219  Unfortunately, that needs linux/watchdog.h and that contains unfiltered
220  kernel types on some distros, which breaks the build.
222  use bb_error_msg where appropriate: See
223  egrep "(printf.*\([[:space:]]*(stderr|2)|[^_]write.*\([[:space:]]*(stderr|2))"
225  use bb_perror_msg where appropriate: See
226  egrep "[^_]perror"
228  possible code duplication ingroup() and is_a_group_member()
230  Move __get_hz() to a better place and (re)use it in route.c, ash.c
232  See grep -r strtod
233  Alot of duplication that wants cleanup.
235  in_ether duplicated in network/{interface,ifconfig}.c
237  unify progress_meter. wget, flash_eraseall, pipe_progress, fbsplash, setfiles.
239  support start-stop-daemon -d <chdir-path>
241vdprintf() -> similar sized functionality
244(TODO list after discussion 11.05.2009)
246* shrink tc/brctl/ip
247  tc/brctl seem like fairly large things to try and tackle in your timeframe,
248  and i think people have posted attempts in the past. Adding additional
249  options to ip though seems reasonable.
251* add tests for some applets
253* implement POSIX utilities and audit them for POSIX conformance. then
254  audit them for GNU conformance. then document all your findings in a new
255  doc/conformance.txt file while perhaps implementing some of the missing
256  features.
257  you can find the latest POSIX documentation (1003.1-2008) here:
259  and the complete list of all utilities that POSIX covers:
261  The first step would to generate a file/matrix what is already archived
262  (also IPV6)
264* implement 'at'
266* rpcbind (former portmap) or equivalent
267  so that we don't have to use -o nolock on nfs mounts
269* check IPV6 compliance
271* generate a mini example using kernel+busybox only (+libc) for example
273* more support for advanced linux 2.6.x features, see: iotop
274  most likely there is more
276* even more support for statistics: mpstat, iostat, powertop....
Note: See TracBrowser for help on using the repository browser.