aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2009-03-31hush: make ^C work again, ^Z probably isn't workingDenis Vlasenko
function old new delta file_get 219 254 +35 check_and_run_traps 122 133 +11 run_list 2001 2010 +9 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 3/0 up/down: 55/0) Total: 55 bytes
2009-03-31hush: do not reset sighaldlers after fork - instead, prevent them from doing ↵Denis Vlasenko
any harm if they will be called. saves many sigaction calls on every fork. function old new delta hush_main 887 971 +84 sigexit 54 65 +11 run_list 2006 2001 -5 maybe_set_sighandler 76 50 -26 set_fatal_sighandler 186 - -186 ------------------------------------------------------------------------------ (add/remove: 0/1 grow/shrink: 2/2 up/down: 95/-217) Total: -122 bytes
2009-03-31hush: rework signal and trap handling. Some smaller bits are TODO,Denis Vlasenko
expect minor breakage function old new delta set_fatal_sighandler 12 186 +174 check_and_run_traps - 122 +122 maybe_set_sighandler - 76 +76 hush_main 831 887 +56 sigtimedwait - 50 +50 __GI_sigtimedwait - 50 +50 hush_exit 49 93 +44 set_mode 749 777 +28 pseudo_exec_argv 131 151 +20 static.zero_ts - 8 +8 expand_variables 1962 1970 +8 builtin_wait 172 174 +2 set_misc_sighandler 12 - -12 set_jobctrl_sighandler 12 - -12 handler_ctrl_c 16 - -16 builtin_set_mode 28 - -28 handle_trap 97 - -97 handler_ctrl_z 107 - -107 builtin_trap 545 438 -107 run_list 2149 2006 -143 ------------------------------------------------------------------------------ (add/remove: 5/6 grow/shrink: 7/2 up/down: 638/-522) Total: 116 bytes
2009-03-30implement `unset` semantics as required by POSIXMike Frysinger
2009-03-30fix DOC build output to be like kbuild standardMike Frysinger
2009-03-30update `trap` poisx linkMike Frysinger
2009-03-30do not automatically display job information when the shell is not interactiveMike Frysinger
2009-03-29enable trap testsMike Frysinger
2009-03-29first pass at `trap` support in hushMike Frysinger
2009-03-29modutils: dont pass NULL options to init_module()Denis Vlasenko
2009-03-29build system: hdparm -d is not that dangerous, remove warningDenis Vlasenko
2009-03-29sendmail: add more example in usage textDenis Vlasenko
2009-03-29module loading for 2.4: use fd-based io instead of FILE based;Denis Vlasenko
use xlseek and xread; simpler check for ELF signature. text data bss dec hex filename 824460 476 7616 832552 cb428 busybox_old 824333 476 7616 832425 cb3a9 busybox_unstripped
2009-03-29add test cases for `trap` ... but disabled for now until trap support is ↵Mike Frysinger
committed
2009-03-29sysctl: move code around to get rid of forward references.Denis Vlasenko
(incidentally it helps gcc to make code smaller)
2009-03-29sysctl: fix broken -p [file] - close bug 231Denis Vlasenko
2009-03-29mark msh as deprecatedMike Frysinger
2009-03-29use bb_strtou() in umask/wait and check errno to see if there was a problem ↵Mike Frysinger
rather than using endp
2009-03-29make sure we check job status whenever updating the promptMike Frysinger
2009-03-28test for invalid named variablesMike Frysinger
2009-03-28do not let handle_dollar() accept vars that start with a digitMike Frysinger
2009-03-28implement `wait` builtinMike Frysinger
2009-03-28remove ((expr)) bash'ismDenis Vlasenko
2009-03-28randomconfig fixDenis Vlasenko
2009-03-28update the normally disabled debug code around the syntax() func to use new ↵Mike Frysinger
maybe_die()
2009-03-28randomtest fixesDenis Vlasenko
2009-03-28add hush tests for parameter expansionMike Frysinger
2009-03-28implement most POSIX parameter expansions (~+500bytes)Mike Frysinger
2009-03-28make sure we exit based on test failure rather than always exiting with 0Mike Frysinger
2009-03-28split up feature todo by posix/bashMike Frysinger
2009-03-28unify `set` handling with command line processing, fixup `set` argv ↵Mike Frysinger
processing so it doesnt clobber argv when setting options, and barf on unhandled set options
2009-03-28handle the "-s" command line option that POSIX requiresMike Frysinger
2009-03-28rename "-f" (fake) option to "-n" to match POSIXMike Frysinger
2009-03-28tweak format modifier in debug code to fixup gcc warningMike Frysinger
2009-03-28convert some of the ENABLE_FEATURE_EDITING from CPP if to C ifMike Frysinger
2009-03-28stop lying about [[ test supportMike Frysinger
2009-03-28mkfs.vfat: fix a problem with over-estimating FAT sizeDenis Vlasenko
2009-03-28make lash deprecation more prominentMike Frysinger
2009-03-28mkfs.vfat: mkdosfs compat with choosing clyster sizeDenis Vlasenko
2009-03-28mkfs.vfat: more of pointless tweakingDenis Vlasenko
2009-03-28mkfs.vfat: small tweak to commentsDenis Vlasenko
2009-03-28mkdosfs (aka mkfs.vfat): new applet by VladimirDenis Vlasenko
text data bss dec hex filename 821202 476 7616 829294 ca76e busybox_old 823577 476 7616 831669 cb0b5 busybox_unstripped
2009-03-27ls: warning fixDenis Vlasenko
rm: accept and ignore -v (verbose)
2009-03-27tail: fix tail +N syntax not working. Closes bug 221.Denis Vlasenko
2009-03-26enable most job functions on no-mmu systemsMike Frysinger
2009-03-25tftp: when we infer local name from remote (-r [/]path/path/file),Denis Vlasenko
strip path. This mimics wget and is generally more intuitive.
2009-03-23do not make startup banner depend on job supportMike Frysinger
2009-03-23libbb: revent previous version of "concurrent history updating"Denis Vlasenko
and replace it with one which does not "snoop" history written by others. (1) it is what bug 185 needs, and (2) it is less bloaty: function old new delta load_history - 252 +252 read_line_input 3155 3287 +132 next_token 914 918 +4 qrealloc 36 33 -3 getoptscmd 713 708 -5 ------------------------------------------------------------------------------ (add/remove: 1/0 grow/shrink: 2/2 up/down: 388/-8) Total: 380 bytes
2009-03-22move parse_stream out of ENABLE_HUSH_TICK to avoid implicit decl/build errorMike Frysinger
2009-03-22libbb: make history saving/loading concurrent-safeDenis Vlasenko
* all history writers always append (not overwrite) history files * they reload history if they detect that file length has changed since last write * they trim history file only when it grows 4 times longer than MAXLINES * they do this atomically by creating new file and renaming it to old Unfortunately, this comes at a price: function old new delta load_history - 346 +346 read_line_input 3155 3358 +203 new_line_input_t 17 31 +14 ...irrelevant small jitter... ------------------------------------------------------------------------------ (add/remove: 1/0 grow/shrink: 5/5 up/down: 573/-13) Total: 560 bytes