1Busybox TODO 2 3Stuff that needs to be done. This is organized by who plans to get around to 4doing it eventually, but that doesn't mean they "own" the item. If you want to 5do one of these bounce an email off the person it's listed under to see if they 6have any suggestions how they plan to go about it, and to minimize conflicts 7between your work and theirs. But otherwise, all of these are fair game. 8 9Rob Landley <rob@landley.net>: 10 Add BB_NOMMU to platform.h and migrate __uClinux__ tests to that. 11 #if defined __UCLIBC__ && !defined __ARCH_USE_MMU__ 12 Add a libbb/platform.c 13 Implement fdprintf() for platforms that haven't got one. 14 Implement bb_realpath() that can handle NULL on non-glibc. 15 Cleanup bb_asprintf() 16 17 Migrate calloc() and bb_calloc() occurrences to bb_xzalloc(). 18 Remove obsolete _() wrapper crud for internationalization we don't do. 19 Figure out where we need utf8 support, and add it. 20 21 sh 22 The command shell situation is a big mess. We have three or four different 23 shells that don't really share any code, and the "standalone shell" doesn't 24 work all that well (especially not in a chroot environment), due to apps not 25 being reentrant. I'm writing a new shell (bbsh) to unify the various 26 shells and configurably add the minimal set of bash features people 27 actually use. The hardest part is it has to configure down as small as 28 lash while providing lash's features. The rest is easy in comparison. 29 bzip2 30 Compression-side support. 31 init 32 General cleanup (should use ENABLE_FEATURE_INIT_SYSLOG and ENABLE_FEATURE_INIT_DEBUG). 33 depmod 34 busybox lacks a way to update module deps when running from firmware without the 35 use of the depmod.pl (perl is to bloated for most embedded setups) and or orig 36 modutils. The orig depmod is rather pointless to have to add to a firmware image 37 in when we already have a insmod/rmmod and friends. 38 Unify base64 handling. 39 [done] 40 Do a SUSv3 audit 41 Look at the full Single Unix Specification version 3 (available online at 42 "http://www.opengroup.org/onlinepubs/009695399/nfindex.html") and 43 figure out which of our apps are compliant, and what we're missing that 44 we might actually care about. 45 46 Even better would be some kind of automated compliance test harness that 47 exercises each command line option and the various corner cases. 48 Internationalization 49 How much internationalization should we do? 50 51 The low hanging fruit is UTF-8 character set support. We should do this. 52 (Vodz pointed out the shell's cmdedit as needing work here. What else?) 53 54 We also have lots of hardwired english text messages. Consolidating this 55 into some kind of message table not only makes translation easier, but 56 also allows us to consolidate redundant (or close) strings. 57 58 We probably don't want to be bloated with locale support. (Not unless we 59 can cleanly export it from our underlying C library without having to 60 concern ourselves with it directly. Perhaps a few specific things like a 61 config option for "date" are low hanging fruit here?) 62 63 What level should things happen at? How much do we care about 64 internationalizing the text console when X11 and xterms are so much better 65 at it? (There's some infrastructure here we don't implement: The 66 "unicode_start" and "unicode_stop" shell scripts need "vt-is-UTF8" and a 67 --unicode option to loadkeys. That implies a real loadkeys/dumpkeys 68 implementation to replace loadkmap/dumpkmap. Plus messing with console font 69 loading. Is it worth it, or do we just say "use X"?) 70 71 Individual compilation of applets. 72 It would be nice if busybox had the option to compile to individual applets, 73 for people who want an alternate implementation less bloated than the gnu 74 utils (or simply with less political baggage), but without it being one big 75 executable. 76 77 Turning libbb into a real dll is another possibility, especially if libbb 78 could export some of the other library interfaces we've already more or less 79 got the code for (like zlib). 80 buildroot - Make a "dogfood" option 81 Busybox 1.1 will be capable of replacing most gnu packages for real world 82 use, such as developing software or in a live CD. It needs wider testing. 83 84 Busybox should now be able to replace bzip2, coreutils, e2fsprogs, file, 85 findutils, gawk, grep, inetutils, less, modutils, net-tools, patch, procps, 86 sed, shadow, sysklogd, sysvinit, tar, util-linux, and vim. The resulting 87 system should be self-hosting (I.E. able to rebuild itself from source 88 code). This means it would need (at least) binutils, gcc, and make, or 89 equivalents. 90 91 It would be a good "eating our own dogfood" test if buildroot had the option 92 of using a "make allyesconfig" busybox instead of the all of the above 93 packages. Anything that's wrong with the resulting system, we can fix. (It 94 would be nice to be able to upgrade busybox to be able to replace bash and 95 diffutils as well, but we're not there yet.) 96 97 One example of an existing system that does this already is Firmware Linux: 98 http://www.landley.net/code/firmware 99 initramfs 100 Busybox should have a sample initramfs build script. This depends on 101 bbsh, mdev, and switch_root. 102 mkdep 103 Write a mkdep that doesn't segfault if there's a directory it doesn't 104 have permission to read, isn't based on manually editing the output of 105 lexx and yacc, doesn't make such a mess under include/config, etc. 106 Group globals into unions of structures. 107 Go through and turn all the global and static variables into structures, 108 and have all those structures be in a big union shared between processes, 109 so busybox uses less bss. (This is a big win on nommu machines.) See 110 sed.c and mdev.c for examples. 111 Go through bugs.busybox.net and close out all of that somehow. 112 This one's open to everybody, but I'll wind up doing it... 113 114 115Bernhard Fischer <busybox@busybox.net> suggests to look at these: 116 New debug options: 117 -Wlarger-than-127 118 Cleanup any big users 119 -Wunused-parameter 120 Facilitate applet PROTOTYPES to provide means for having applets that 121 do a) not take any arguments b) need only one of argc or argv c) need 122 both argc and argv. All of these three options should go for the most 123 feature complete denominator. 124 Collate BUFSIZ IOBUF_SIZE MY_BUF_SIZE PIPE_PROGRESS_SIZE BUFSIZE PIPESIZE 125 make bb_common_bufsiz1 configurable, size wise. 126 make pipesize configurable, size wise. 127 Use bb_common_bufsiz1 throughout applets! 128 129As yet unclaimed: 130 131---- 132find 133 doesn't understand (), lots of susv3 stuff. 134---- 135diff 136 Make sure we handle empty files properly: 137 From the patch man page: 138 139 you can remove a file by sending out a context diff that compares 140 the file to be deleted with an empty file dated the Epoch. The 141 file will be removed unless patch is conforming to POSIX and the 142 -E or --remove-empty-files option is not given. 143--- 144patch 145 Should have simple fuzz factor support to apply patches at an offset which 146 shouldn't take up too much space. 147 148 And while we're at it, a new patch filename quoting format is apparently 149 coming soon: http://marc.theaimsgroup.com/?l=git&m=112927316408690&w=2 150--- 151ps / top 152 Add support for both RSS and VSIZE rather than just one or the other. 153 Or make it a build option. 154--- 155man 156 It would be nice to have a man command. Not one that handles troff or 157 anything, just one that can handle preformatted ascii man pages, possibly 158 compressed. This could probably be a script in the extras directory that 159 calls cat/zcat/bzcat | less 160 161 (How doclifter might work into this is anybody's guess.) 162--- 163ar 164 Write support? 165--- 166stty / catv 167 stty's visible() function and catv's guts are identical. Merge them into 168 an appropriate libbb function. 169--- 170struct suffix_mult 171 Several duplicate users of: grep -r "1024\*1024" * -B2 -A1 172 Merge to a single size_suffixes[] in libbb. 173 Users: head tail od_bloaty hexdump and (partially as it wouldn't hurt) svlogd 174--- 175tail 176 ./busybox tail -f foo.c~ TODO 177 should not print fmt=header_fmt for subsequent date >> TODO; i.e. only 178 fmt+ if another (not the current) file did change 179 180Architectural issues: 181 182bb_close() with fsync() 183 We should have a bb_close() in place of normal close, with a CONFIG_ option 184 to not just check the return value of close() for an error, but fsync(). 185 Close can't reliably report anything useful because if write() accepted the 186 data then it either went out to the network or it's in cache or a pipe 187 buffer. Either way, there's no guarantee it'll make it to its final 188 destination before close() gets called, so there's no guarantee that any 189 error will be reported. 190 191 You need to call fsync() if you care about errors that occur after write(), 192 but that can have a big performance impact. So make it a config option. 193--- 194Unify archivers 195 Lots of archivers have the same general infrastructure. The directory 196 traversal code should be factored out, and the guts of each archiver could 197 be some setup code and a series of callbacks for "add this file", 198 "add this directory", "add this symlink" and so on. 199 200 This could clean up tar and zip, and make it cheaper to add cpio and ar 201 write support, and possibly even cheaply add things like mkisofs or 202 mksquashfs someday, if they become relevant. 203--- 204Text buffer support. 205 Several existing applets (sort, vi, less...) read 206 a whole file into memory and act on it. There might be an opportunity 207 for shared code in there that could be moved into libbb... 208--- 209Memory Allocation 210 We have a CONFIG_BUFFER mechanism that lets us select whether to do memory 211 allocation on the stack or the heap. Unfortunately, we're not using it much. 212 We need to audit our memory allocations and turn a lot of malloc/free calls 213 into RESERVE_CONFIG_BUFFER/RELEASE_CONFIG_BUFFER. 214 For a start, see e.g. make EXTRA_CFLAGS=-Wlarger-than-64 215 216 And while we're at it, many of the CONFIG_FEATURE_CLEAN_UP #ifdefs will be 217 optimized out by the compiler in the stack allocation case (since there's no 218 free for an alloca()), and this means that various cleanup loops that just 219 call free might also be optimized out by the compiler if written right, so 220 we can yank those #ifdefs too, and generally clean up the code. 221--- 222Switch CONFIG_SYMBOLS to ENABLE_SYMBOLS 223 224 In busybox 1.0 and earlier, configuration was done by CONFIG_SYMBOLS 225 that were either defined or undefined to indicate whether the symbol was 226 selected in the .config file. They were used with #ifdefs, ala: 227 228 #ifdef CONFIG_SYMBOL 229 if (other_test) { 230 do_code(); 231 } 232 #endif 233 234 In 1.1, we have new ENABLE_SYMBOLS which are always defined (as 0 or 1), 235 meaning you can still use them for preprocessor tests by replacing 236 "#ifdef CONFIG_SYMBOL" with "#if ENABLE_SYMBOL". But more importantly, we 237 can use them as a true or false test in normal C code: 238 239 if (ENABLE_SYMBOL && other_test) { 240 do_code(); 241 } 242 243 (Optimizing away if() statements that resolve to a constant value 244 is known as "dead code elimination", an optimization so old and simple that 245 Turbo Pascal for DOS did it twenty years ago. Even modern mini-compilers 246 like the Tiny C Compiler (tcc) and the Small Device C Compiler (SDCC) 247 perform dead code elimination.) 248 249 Right now, busybox.h is #including both "config.h" (defining the 250 CONFIG_SYMBOLS) and "bb_config.h" (defining the ENABLE_SYMBOLS). At some 251 point in the future, it would be nice to wean ourselves off of the 252 CONFIG versions. (Among other things, some defective build environments 253 leak the Linux kernel's CONFIG_SYMBOLS into the system's standard #include 254 files. We've experienced collisions before.) 255--- 256FEATURE_CLEAN_UP 257 This is more an unresolved issue than a to-do item. More thought is needed. 258 259 Normally we rely on exit() to free memory, close files, and unmap segments 260 for us. This makes most calls to free(), close(), and unmap() optional in 261 busybox applets that don't intend to run for very long, and optional stuff 262 can be omitted to save size. 263 264 The idea was raised that we could simulate fork/exit with setjmp/longjmp 265 for _really_ brainless embedded systems, or speed up the standalone shell 266 by not forking. Doing so would require a reliable FEATURE_CLEAN_UP. 267 Unfortunately, this isn't as easy as it sounds. 268 269 The problem is, lots of things exit(), sometimes unexpectedly (xmalloc()) 270 and sometimes reliably (bb_perror_msg_and_die() or show_usage()). This 271 jumps out of the normal flow control and bypasses any cleanup code we 272 put at the end of our applets. 273 274 It's possible to add hooks to libbb functions like xmalloc() and xopen() 275 to add their entries to a linked list, which could be traversed and 276 freed/closed automatically. (This would need to be able to free just the 277 entries after a checkpoint to be usable for a forkless standalone shell. 278 You don't want to free the shell's own resources.) 279 280 Right now, FEATURE_CLEAN_UP is more or less a debugging aid, to make things 281 like valgrind happy. It's also documentation of _what_ we're trusting 282 exit() to clean up for us. But new infrastructure to auto-free stuff would 283 render the existing FEATURE_CLEAN_UP code redundant. 284 285 For right now, exit() handles it just fine. 286 287 288 289Minor stuff: 290 watchdog.c could autodetect the timer duration via: 291 if(!ioctl (fd, WDIOC_GETTIMEOUT, &tmo)) timer_duration = 1 + (tmo / 2); 292 Unfortunately, that needs linux/watchdog.h and that contains unfiltered 293 kernel types on some distros, which breaks the build. 294--- 295 use bb_error_msg where appropriate: See 296 egrep "(printf.*\([[:space:]]*(stderr|2)|[^_]write.*\([[:space:]]*(stderr|2))" 297--- 298 use bb_perror_msg where appropriate: See 299 egrep "[^_]perror" 300--- 301 Remove superfluous fmt occurances: e.g. 302 fprintf(stderr, "%s: %s not found\n", "unalias", *argptr); 303 -> fprintf(stderr, "unalias: %s not found\n", *argptr); 304--- 305 possible code duplication ingroup() and is_a_group_member() 306--- 307 Move __get_hz() to a better place and (re)use it in route.c, ash.c, msh.c 308--- 309 310 311Code cleanup: 312 313Replace deprecated functions. 314 315bzero() -> memset() 316--- 317sigblock(), siggetmask(), sigsetmask(), sigmask() -> sigprocmask et al 318--- 319vdprintf() -> similar sized functionality 320--- 321