3 Stuff that needs to be done
6 doesn't understand (), lots of susv3 stuff.
9 The command shell situation is a big mess. We have three or four different
10 shells that don't really share any code, and the "standalone shell" doesn't
11 work all that well (especially not in a chroot environment), due to apps not
12 being reentrant. Unifying the various shells and figuring out a configurable
13 way of adding the minimal set of bash features a given script uses is a big
14 job, but it be a big improvement.
16 Note: Rob Landley (rob@landley.net) is working on this one, but very slowly...
19 We should have a diff -u command. We have patch, we should have diff
20 (we only need to support unified diffs though).
23 Would be nice. The basic susv3 options, plus fuser -k.
26 Should have simple fuzz factor support to apply patches at an offset which
27 shouldn't take up too much space.
29 And while we're at it, a new patch filename quoting format is apparently
30 coming soon: http://marc.theaimsgroup.com/?l=git&m=112927316408690&w=2
33 It would be nice to have a man command. Not one that handles troff or
34 anything, just one that can handle preformatted ascii man pages, possibly
35 compressed. This could probably be a script in the extras directory that
36 calls cat/zcat/bzcat | less
39 Compression-side support.
45 Look at the full Single Unix Specification version 3 (available online at
46 "http://www.opengroup.org/onlinepubs/009695399/nfindex.html") and
47 figure out which of our apps are compliant, and what we're missing that
48 we might actually care about.
50 Even better would be some kind of automated compliance test harness that
51 exercises each command line option and the various corner cases.
54 Lots of archivers have the same general infrastructure. The directory
55 traversal code should be factored out, and the guts of each archiver could
56 be some setup code and a series of callbacks for "add this file",
57 "add this directory", "add this symlink" and so on.
59 This could clean up tar and zip, and make it cheaper to add cpio and ar
60 write support, and possibly even cheaply add things like mkisofs someday,
61 if it becomes relevant.
64 Several existing applets (sort, vi, less...) read
65 a whole file into memory and act on it. There might be an opportunity
66 for shared code in there that could be moved into libbb...
68 Individual compilation of applets.
69 It would be nice if busybox had the option to compile to individual applets,
70 for people who want an alternate implementation less bloated than the gnu
71 utils (or simply with less political baggage), but without it being one big
74 Turning libbb into a real dll is another possibility, especially if libbb
75 could export some of the other library interfaces we've already more or less
76 got the code for (like zlib).
78 buildroot - Make a "dogfood" option
79 Busybox 1.1 will be capable of replacing most gnu packages for real world use,
80 such as developing software or in a live CD. It needs wider testing.
82 Busybox should now be able to replace bzip2, coreutils, e2fsprogs, file,
83 findutils, gawk, grep, inetutils, less, modutils, net-tools, patch, procps,
84 sed, shadow, sysklogd, sysvinit, tar, util-linux, and vim. The resulting
85 system should be self-hosting (I.E. able to rebuild itself from source code).
86 This means it would need (at least) binutils, gcc, and make, or equivalents.
88 It would be a good "eating our own dogfood" test if buildroot had the option
89 of using a "make allyesconfig" busybox instead of the all of the above
90 packages. Anything that's wrong with the resulting system, we can fix. (It
91 would be nice to be able to upgrade busybox to be able to replace bash and
92 diffutils as well, but we're not there yet.)
94 One example of an existing system that does this already is Firmware Linux:
95 http://www.landley.net/code/firmware
98 We have a CONFIG_BUFFER mechanism that lets us select whether to do memory
99 allocation on the stack or the heap. Unfortunately, we're not using it much.
100 We need to audit our memory allocations and turn a lot of malloc/free calls
101 into RESERVE_CONFIG_BUFFER/RELEASE_CONFIG_BUFFER.
103 And while we're at it, many of the CONFIG_FEATURE_CLEAN_UP #ifdefs will be
104 optimized out by the compiler in the stack allocation case (since there's no
105 free for an alloca()), and this means that various cleanup loops that just
106 call free might also be optimized out by the compiler if written right, so
107 we can yank those #ifdefs too, and generally clean up the code.
109 Switch CONFIG_SYMBOLS to ENABLE_SYMBOLS
111 In busybox 1.0 and earlier, configuration was done by CONFIG_SYMBOLS
112 that were either defined or undefined to indicate whether the symbol was
113 selected in the .config file. They were used with #ifdefs, ala:
121 In 1.1, we have new ENABLE_SYMBOLS which are always defined (as 0 or 1),
122 meaning you can still use them for preprocessor tests by replacing
123 "#ifdef CONFIG_SYMBOL" with "#if ENABLE_SYMBOL". But more importantly, we
124 can use them as a true or false test in normal C code:
126 if (ENABLE_SYMBOL && other_test) {
130 (Optimizing away if() statements that resolve to a constant value
131 is known as "dead code elimination", an optimization so old and simple that
132 Turbo Pascal for DOS did it twenty years ago. Even modern mini-compilers
133 like the Tiny C Compiler (tcc) and the Small Device C Compiler (SDCC)
134 perform dead code elimination.)
136 Right now, busybox.h is #including both "config.h" (defining the
137 CONFIG_SYMBOLS) and "bb_config.h" (defining the ENABLE_SYMBOLS). At some
138 point in the future, it would be nice to wean ourselves off of the
139 CONFIG versions. (Among other things, some defective build environments
140 leak the Linux kernel's CONFIG_SYMBOLS into the system's standard #include
141 files. We've experienced collisions before.)
144 This is more an unresolved issue than a to-do item. More thought is needed.
146 Normally we rely on exit() to free memory, close files, and unmap segments
147 for us. This makes most calls to free(), close(), and unmap() optional in
148 busybox applets that don't intend to run for very long, and optional stuff
149 can be omitted to save size.
151 The idea was raised that we could simulate fork/exit with setjmp/longjmp
152 for _really_ brainless embedded systems, or speed up the standalone shell
153 by not forking. Doing so would require a reliable FEATURE_CLEAN_UP.
154 Unfortunately, this isn't as easy as it sounds.
156 The problem is, lots of things exit(), sometimes unexpectedly (xmalloc())
157 and sometimes reliably (bb_perror_msg_and_die() or show_usage()). This
158 jumps out of the normal flow control and bypasses any cleanup code we
159 put at the end of our applets.
161 It's possible to add hooks to libbb functions like xmalloc() and bb_xopen()
162 to add their entries to a linked list, which could be traversed and
163 freed/closed automatically. (This would need to be able to free just the
164 entries after a checkpoint to be usable for a forkless standalone shell.
165 You don't want to free the shell's own resources.)
167 Right now, FEATURE_CLEAN_UP is more or less a debugging aid, to make things
168 like valgrind happy. It's also documentation of _what_ we're trusting
169 exit() to clean up for us. But new infrastructure to auto-free stuff would
170 render the existing FEATURE_CLEAN_UP code redundant.
172 For right now, exit() handles it just fine.