Glenn L McGrath | 90d2bff | 2004-05-01 00:49:49 +0000 | [diff] [blame] | 1 | Busybox TODO |
| 2 | |
| 3 | Stuff that needs to be done |
| 4 | |
Eric Andersen | b413a70 | 2005-02-13 22:20:35 +0000 | [diff] [blame] | 5 | tr - missing SuS3 features in busybox 1.0pre10 |
Glenn L McGrath | 90d2bff | 2004-05-01 00:49:49 +0000 | [diff] [blame] | 6 | |
| 7 | tr doesnt support [:blank:], [:digit:] or other predefined classes, [=equiv=] |
| 8 | support is also missing. |
| 9 | ---- |
Rob Landley | f4bb212 | 2005-01-24 06:56:24 +0000 | [diff] [blame] | 10 | find |
| 11 | doesn't understand () or -exec, and these are actually used out in the real |
| 12 | world. The "make uninstall" of lots of things (including busybox itself) |
| 13 | breaks because of this, and sometimes even "make install" (like udev). |
| 14 | ---- |
| 15 | comm |
| 16 | Perl needs "comm" to build. It's small and simple, but we haven't got it. |
| 17 | --- |
| 18 | sh |
| 19 | The command shell situation is a big mess. We have three or four 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 |
| 22 | being reentrant. Unifying the various shells and figuring out a configurable |
| 23 | way of adding the minimal set of bash features a given script uses is a big |
| 24 | job, but it be a big improvement. |
| 25 | --- |
| 26 | gzip |
| 27 | Can't handle compressing multiple files at once. (I don't mean making a |
| 28 | multiple file archive, I mean compressing more than one file at a time.) |
| 29 | Some global variables aren't re-initialized between runs. |
| 30 | --- |
| 31 | gunzip |
| 32 | same problem as gzip. "gunzip one.gz two.gz three.gz" doesn't work for |
| 33 | two.gz and three.gz due to global variables not getting reset. |
| 34 | --- |
| 35 | diff |
| 36 | We should have a diff -u command. We have patch, we should have diff |
| 37 | (we only need to support unified diffs though). |
| 38 | --- |
| 39 | patch |
| 40 | should have -i support, and simple fuzz factor support to apply patches |
| 41 | at an offset shouldn't take up too much space. |
| 42 | --- |
| 43 | man |
| 44 | It would be nice to have a man command. Not one that handles troff or |
| 45 | anything, just one that can handle preformatted ascii man pages, possibly |
| 46 | compressed. This could probably be a script in the extras directory that |
| 47 | calls cat/zcatbzcat | more |
| 48 | --- |
| 49 | less |
| 50 | More sucks if you're used to less. A tiny less implementation would be |
| 51 | very nice. |
| 52 | --- |
| 53 | bzip2 |
| 54 | Compression-side support. |
| 55 | |
| 56 | |
| 57 | Architectural issues: |
| 58 | |
| 59 | Do a SUSv3 audit |
| 60 | Look at the full Single Unix Specification version 3 (available online at |
| 61 | "http://www.opengroup.org/onlinepubs/009695399/nfindex.html") and |
| 62 | figure out which of our apps are compliant, and what we're missing that |
| 63 | we might actually care about. |
| 64 | |
| 65 | Even better would be some kind of automated compliance test harness that |
| 66 | exercises each command line option and the various corner cases. |
| 67 | -- |
| 68 | Unify archivers |
| 69 | Lots of archivers have the same general infrastructure. The directory |
| 70 | traversal code should be factored out, and the guts of each archiver could |
| 71 | be some setup code and a series of callbacks for "add this file", |
| 72 | "add this directory", "add this symlink" and so on. |
| 73 | |
| 74 | This could clean up tar and zip, and make it cheaper to add cpio and ar |
| 75 | write support, and possibly even cheaply add things like mkisofs someday, |
| 76 | if it becomes relevant. |
| 77 | --- |
| 78 | Text buffer support. |
| 79 | Several existing applets and potential additions (sort, vi, less...) read |
| 80 | a whole file into memory and act on it. There might be an opportunity |
| 81 | for shared code in there that could be moved into libbb... |
| 82 | --- |
| 83 | Individual compilation of applets. |
| 84 | It would be nice if busybox had the option to compile to individual applets, |
| 85 | for people who want an alternate implementation less bloated than the gnu |
| 86 | utils (or simply with less political baggage), but without it being one big |
| 87 | executable. |
| 88 | |
| 89 | Turning libbb into a real dll is another possibility, especially if libbb |
| 90 | could export some of the other library interfaces we've already more or less |
| 91 | got the code for (like zlib). |
| 92 | --- |
| 93 | buildroot - Make a "dogfood" option |
| 94 | Busybox is now capable of replacing most gnu packages for real world use, |
| 95 | such as developing software or in a live CD. A system built from busybox |
| 96 | (1.00 with updated sort.c), uclibc 0.9.27, gcc, binutils, make, and a few |
| 97 | other development tools (http://www.landley.net/code/firmware has an example |
| 98 | system using autoconf, automake, bison, flex, libtools, m4, zlib, |
| 99 | and groff: dunno what subset of that is actually necessary) is capable of |
| 100 | rebuilding itself, from scratch, under itself. |
| 101 | |
| 102 | It would be a good "eating our own dogfood" test if buildroot had the option |
| 103 | of using busybox instead of bzip2, coreutils, file, findutils, gawk, grep, |
| 104 | inetutils, modutils, net-tools, procps, sed, shadow, sysklogd, sysvinit, tar, |
| 105 | util-linux, and vim. Anything that's wrong with the resulting system, we |
| 106 | can fix. (It would be nice to be able to upgrade busybox to be able to |
| 107 | replace bash, diffutils, gzip, less, and patch as well.) |
Rob Landley | 958fa2a | 2005-06-11 22:10:42 +0000 | [diff] [blame^] | 108 | --- |
| 109 | Memory Allocation |
| 110 | We have a CONFIG_BUFFER mechanism that lets us select whether to do memory |
| 111 | allocation on the stack or the heap. Unfortunately, we're not using it much. |
| 112 | We need to audit our memory allocations and turn a lot of malloc/free calls |
| 113 | into RESERVE_CONFIG_BUFFER/RELEASE_CONFIG_BUFFER. |
| 114 | |
| 115 | And while we're at it, many of the CONFIG_FEATURE_CLEAN_UP #ifdefs will be |
| 116 | optimized out by the compiler in the stack allocation case (since there's no |
| 117 | free for an alloca()), and this means that various cleanup loops that just |
| 118 | call free might also be optimized out by the compiler if written right, so |
| 119 | we can yank those #ifdefs too, and generally clean up the code. |