3 Stuff that needs to be done. This is organized by who plans to get around to
4 doing it eventually, but that doesn't mean they "own" the item. If you want to
5 do one of these bounce an email off the person it's listed under to see if they
6 have any suggestions how they plan to go about it, and to minimize conflicts
7 between your work and theirs. But otherwise, all of these are fair game.
9 Rob Landley <rob@landley.net>:
10 Add a libbb/platform.c
11 Implement fdprintf() for platforms that haven't got one.
12 Implement bb_realpath() that can handle NULL on non-glibc.
15 Remove obsolete _() wrapper crud for internationalization we don't do.
16 Figure out where we need utf8 support, and add it.
19 The command shell situation is a big mess. We have three different
20 shells that don't really share any code, and the "standalone shell" doesn't
21 work all that well (especially not in a chroot environment), due to apps not
23 lash is phased out. hush can be configured down to be nearly as small,
26 General cleanup (should use ENABLE_FEATURE_INIT_SYSLOG and ENABLE_FEATURE_INIT_DEBUG).
28 busybox lacks a way to update module deps when running from firmware without the
29 use of the depmod.pl (perl is to bloated for most embedded setups) and or orig
30 modutils. The orig depmod is rather pointless to have to add to a firmware image
31 in when we already have a insmod/rmmod and friends.
33 Look at the full Single Unix Specification version 3 (available online at
34 "http://www.opengroup.org/onlinepubs/009695399/nfindex.html") and
35 figure out which of our apps are compliant, and what we're missing that
36 we might actually care about.
38 Even better would be some kind of automated compliance test harness that
39 exercises each command line option and the various corner cases.
41 How much internationalization should we do?
43 The low hanging fruit is UTF-8 character set support. We should do this.
44 (Vodz pointed out the shell's cmdedit as needing work here. What else?)
46 We also have lots of hardwired english text messages. Consolidating this
47 into some kind of message table not only makes translation easier, but
48 also allows us to consolidate redundant (or close) strings.
50 We probably don't want to be bloated with locale support. (Not unless we
51 can cleanly export it from our underlying C library without having to
52 concern ourselves with it directly. Perhaps a few specific things like a
53 config option for "date" are low hanging fruit here?)
55 What level should things happen at? How much do we care about
56 internationalizing the text console when X11 and xterms are so much better
57 at it? (There's some infrastructure here we don't implement: The
58 "unicode_start" and "unicode_stop" shell scripts need "vt-is-UTF8" and a
59 --unicode option to loadkeys. That implies a real loadkeys/dumpkeys
60 implementation to replace loadkmap/dumpkmap. Plus messing with console font
61 loading. Is it worth it, or do we just say "use X"?)
63 Individual compilation of applets.
64 It would be nice if busybox had the option to compile to individual applets,
65 for people who want an alternate implementation less bloated than the gnu
66 utils (or simply with less political baggage), but without it being one big
69 Turning libbb into a real dll is another possibility, especially if libbb
70 could export some of the other library interfaces we've already more or less
71 got the code for (like zlib).
72 buildroot - Make a "dogfood" option
73 Busybox 1.1 will be capable of replacing most gnu packages for real world
74 use, such as developing software or in a live CD. It needs wider testing.
76 Busybox should now be able to replace bzip2, coreutils, e2fsprogs, file,
77 findutils, gawk, grep, inetutils, less, modutils, net-tools, patch, procps,
78 sed, shadow, sysklogd, sysvinit, tar, util-linux, and vim. The resulting
79 system should be self-hosting (I.E. able to rebuild itself from source
80 code). This means it would need (at least) binutils, gcc, and make, or
83 It would be a good "eating our own dogfood" test if buildroot had the option
84 of using a "make allyesconfig" busybox instead of the all of the above
85 packages. Anything that's wrong with the resulting system, we can fix. (It
86 would be nice to be able to upgrade busybox to be able to replace bash and
87 diffutils as well, but we're not there yet.)
89 One example of an existing system that does this already is Firmware Linux:
90 http://www.landley.net/code/firmware
92 Busybox should have a sample initramfs build script. This depends on
93 bbsh, mdev, and switch_root.
95 Write a mkdep that doesn't segfault if there's a directory it doesn't
96 have permission to read, isn't based on manually editing the output of
97 lexx and yacc, doesn't make such a mess under include/config, etc.
98 Group globals into unions of structures.
99 Go through and turn all the global and static variables into structures,
100 and have all those structures be in a big union shared between processes,
101 so busybox uses less bss. (This is a big win on nommu machines.) See
102 sed.c and mdev.c for examples.
103 Go through bugs.busybox.net and close out all of that somehow.
104 This one's open to everybody, but I'll wind up doing it...
107 Bernhard Fischer <busybox@busybox.net> suggests to look at these:
110 Cleanup any big users
112 Facilitate applet PROTOTYPES to provide means for having applets that
113 do a) not take any arguments b) need only one of argc or argv c) need
114 both argc and argv. All of these three options should go for the most
115 feature complete denominator.
116 Collate BUFSIZ IOBUF_SIZE MY_BUF_SIZE PIPE_PROGRESS_SIZE BUFSIZE PIPESIZE
117 make bb_common_bufsiz1 configurable, size wise.
118 make pipesize configurable, size wise.
119 Use bb_common_bufsiz1 throughout applets!
125 Make sure we handle empty files properly:
126 From the patch man page:
128 you can remove a file by sending out a context diff that compares
129 the file to be deleted with an empty file dated the Epoch. The
130 file will be removed unless patch is conforming to POSIX and the
131 -E or --remove-empty-files option is not given.
134 Should have simple fuzz factor support to apply patches at an offset which
135 shouldn't take up too much space.
137 And while we're at it, a new patch filename quoting format is apparently
138 coming soon: http://marc.theaimsgroup.com/?l=git&m=112927316408690&w=2
141 It would be nice to have a man command. Not one that handles troff or
142 anything, just one that can handle preformatted ascii man pages, possibly
143 compressed. This could probably be a script in the extras directory that
144 calls cat/zcat/bzcat | less
146 (How doclifter might work into this is anybody's guess.)
152 stty's visible() function and catv's guts are identical. Merge them into
153 an appropriate libbb function.
156 Several duplicate users of: grep -r "1024\*1024" * -B2 -A1
157 Merge to a single size_suffixes[] in libbb.
158 Users: head tail od_bloaty hexdump and (partially as it wouldn't hurt) svlogd
161 ./busybox tail -f foo.c~ TODO
162 should not print fmt=header_fmt for subsequent date >> TODO; i.e. only
163 fmt+ if another (not the current) file did change
165 Architectural issues:
167 bb_close() with fsync()
168 We should have a bb_close() in place of normal close, with a CONFIG_ option
169 to not just check the return value of close() for an error, but fsync().
170 Close can't reliably report anything useful because if write() accepted the
171 data then it either went out to the network or it's in cache or a pipe
172 buffer. Either way, there's no guarantee it'll make it to its final
173 destination before close() gets called, so there's no guarantee that any
174 error will be reported.
176 You need to call fsync() if you care about errors that occur after write(),
177 but that can have a big performance impact. So make it a config option.
180 Lots of archivers have the same general infrastructure. The directory
181 traversal code should be factored out, and the guts of each archiver could
182 be some setup code and a series of callbacks for "add this file",
183 "add this directory", "add this symlink" and so on.
185 This could clean up tar and zip, and make it cheaper to add cpio and ar
186 write support, and possibly even cheaply add things like mkisofs or
187 mksquashfs someday, if they become relevant.
190 Several existing applets (sort, vi, less...) read
191 a whole file into memory and act on it. There might be an opportunity
192 for shared code in there that could be moved into libbb...
195 We have a CONFIG_BUFFER mechanism that lets us select whether to do memory
196 allocation on the stack or the heap. Unfortunately, we're not using it much.
197 We need to audit our memory allocations and turn a lot of malloc/free calls
198 into RESERVE_CONFIG_BUFFER/RELEASE_CONFIG_BUFFER.
199 For a start, see e.g. make EXTRA_CFLAGS=-Wlarger-than-64
201 And while we're at it, many of the CONFIG_FEATURE_CLEAN_UP #ifdefs will be
202 optimized out by the compiler in the stack allocation case (since there's no
203 free for an alloca()), and this means that various cleanup loops that just
204 call free might also be optimized out by the compiler if written right, so
205 we can yank those #ifdefs too, and generally clean up the code.
207 Switch CONFIG_SYMBOLS to ENABLE_SYMBOLS
209 In busybox 1.0 and earlier, configuration was done by CONFIG_SYMBOLS
210 that were either defined or undefined to indicate whether the symbol was
211 selected in the .config file. They were used with #ifdefs, ala:
219 In 1.1, we have new ENABLE_SYMBOLS which are always defined (as 0 or 1),
220 meaning you can still use them for preprocessor tests by replacing
221 "#ifdef CONFIG_SYMBOL" with "#if ENABLE_SYMBOL". But more importantly, we
222 can use them as a true or false test in normal C code:
224 if (ENABLE_SYMBOL && other_test) {
228 (Optimizing away if() statements that resolve to a constant value
229 is known as "dead code elimination", an optimization so old and simple that
230 Turbo Pascal for DOS did it twenty years ago. Even modern mini-compilers
231 like the Tiny C Compiler (tcc) and the Small Device C Compiler (SDCC)
232 perform dead code elimination.)
234 Right now, busybox.h is #including both "config.h" (defining the
235 CONFIG_SYMBOLS) and "bb_config.h" (defining the ENABLE_SYMBOLS). At some
236 point in the future, it would be nice to wean ourselves off of the
237 CONFIG versions. (Among other things, some defective build environments
238 leak the Linux kernel's CONFIG_SYMBOLS into the system's standard #include
239 files. We've experienced collisions before.)
242 This is more an unresolved issue than a to-do item. More thought is needed.
244 Normally we rely on exit() to free memory, close files, and unmap segments
245 for us. This makes most calls to free(), close(), and unmap() optional in
246 busybox applets that don't intend to run for very long, and optional stuff
247 can be omitted to save size.
249 The idea was raised that we could simulate fork/exit with setjmp/longjmp
250 for _really_ brainless embedded systems, or speed up the standalone shell
251 by not forking. Doing so would require a reliable FEATURE_CLEAN_UP.
252 Unfortunately, this isn't as easy as it sounds.
254 The problem is, lots of things exit(), sometimes unexpectedly (xmalloc())
255 and sometimes reliably (bb_perror_msg_and_die() or show_usage()). This
256 jumps out of the normal flow control and bypasses any cleanup code we
257 put at the end of our applets.
259 It's possible to add hooks to libbb functions like xmalloc() and xopen()
260 to add their entries to a linked list, which could be traversed and
261 freed/closed automatically. (This would need to be able to free just the
262 entries after a checkpoint to be usable for a forkless standalone shell.
263 You don't want to free the shell's own resources.)
265 Right now, FEATURE_CLEAN_UP is more or less a debugging aid, to make things
266 like valgrind happy. It's also documentation of _what_ we're trusting
267 exit() to clean up for us. But new infrastructure to auto-free stuff would
268 render the existing FEATURE_CLEAN_UP code redundant.
270 For right now, exit() handles it just fine.
275 watchdog.c could autodetect the timer duration via:
276 if(!ioctl (fd, WDIOC_GETTIMEOUT, &tmo)) timer_duration = 1 + (tmo / 2);
277 Unfortunately, that needs linux/watchdog.h and that contains unfiltered
278 kernel types on some distros, which breaks the build.
280 use bb_error_msg where appropriate: See
281 egrep "(printf.*\([[:space:]]*(stderr|2)|[^_]write.*\([[:space:]]*(stderr|2))"
283 use bb_perror_msg where appropriate: See
286 Remove superfluous fmt occurances: e.g.
287 fprintf(stderr, "%s: %s not found\n", "unalias", *argptr);
288 -> fprintf(stderr, "unalias: %s not found\n", *argptr);
290 possible code duplication ingroup() and is_a_group_member()
292 Move __get_hz() to a better place and (re)use it in route.c, ash.c, msh.c
295 Alot of duplication that wants cleanup.
297 in_ether duplicated in network/{interface,ifconfig}.c
303 Replace deprecated functions.
306 vdprintf() -> similar sized functionality