oweals/busybox.git
19 years agoFEATURE->ENABLE_FEATURE loses one for 'top' applet
"Vladimir N. Oleynik" [Wed, 14 Sep 2005 16:08:06 +0000 (16:08 -0000)]
FEATURE->ENABLE_FEATURE loses one for 'top' applet

19 years agoPatch from Berhnard Fischer to keep duplicate objects out of the library.
Rob Landley [Wed, 14 Sep 2005 15:41:18 +0000 (15:41 -0000)]
Patch from Berhnard Fischer to keep duplicate objects out of the library.

19 years agoremove unrequired dependences
"Vladimir N. Oleynik" [Wed, 14 Sep 2005 15:36:56 +0000 (15:36 -0000)]
remove unrequired dependences

19 years agoCleanup patch from tito: migrate option parsing to the bb_ functions and fix
Rob Landley [Wed, 14 Sep 2005 15:36:52 +0000 (15:36 -0000)]
Cleanup patch from tito: migrate option parsing to the bb_ functions and fix
types.

19 years agoCleanup patch from Tito.
Rob Landley [Wed, 14 Sep 2005 15:28:15 +0000 (15:28 -0000)]
Cleanup patch from Tito.

19 years agobb_mkdep version 2.0. speed up *2, remove problem of find e2fsprogs/uu*.h, spelling...
"Vladimir N. Oleynik" [Wed, 14 Sep 2005 15:09:06 +0000 (15:09 -0000)]
bb_mkdep version 2.0. speed up *2, remove problem of find e2fsprogs/uu*.h, spelling corrections by Bernhard Fischer

19 years agoOur config should write out each symbol once and only once, but still write
Rob Landley [Wed, 14 Sep 2005 14:40:01 +0000 (14:40 -0000)]
Our config should write out each symbol once and only once, but still write
out all symbols in all sub-menus.  I think this finally does it right.

19 years agoTest full susv3 spec for uniq (except internationalization).
Rob Landley [Wed, 14 Sep 2005 14:36:40 +0000 (14:36 -0000)]
Test full susv3 spec for uniq (except internationalization).

I think this covers it.  We fail two corner cases, both of which are explicit
violations of the spec, and both of which gnu passes.

19 years agocommitting bug #7:
Paul Fox [Wed, 14 Sep 2005 14:08:38 +0000 (14:08 -0000)]
committing bug #7:
     0000007: which and wd-located files
     which doesn't search $PATH when there's a file in the WD with
     the same name of the 'filename' parameter...

19 years agoinitial ppc64 support
Mike Frysinger [Wed, 14 Sep 2005 00:07:26 +0000 (00:07 -0000)]
initial ppc64 support

19 years agoallow either backspace or DEL, in addition to the user's erase
Paul Fox [Tue, 13 Sep 2005 22:20:37 +0000 (22:20 -0000)]
allow either backspace or DEL, in addition to the user's erase
char, to be used in get_input_line()

19 years agobb_mkdep release. speed up *3, clearing, more comments
"Vladimir N. Oleynik" [Tue, 13 Sep 2005 16:50:53 +0000 (16:50 -0000)]
bb_mkdep release. speed up *3, clearing, more comments

19 years agowhitespace/syntax cleanups, dont mind me :p
Mike Frysinger [Tue, 13 Sep 2005 02:29:39 +0000 (02:29 -0000)]
whitespace/syntax cleanups, dont mind me :p

19 years agotouch up usage for swap utils
Mike Frysinger [Tue, 13 Sep 2005 02:26:24 +0000 (02:26 -0000)]
touch up usage for swap utils

19 years agoCleanup patch from Tito.
Rob Landley [Tue, 13 Sep 2005 01:30:19 +0000 (01:30 -0000)]
Cleanup patch from Tito.

19 years agosysctl was seriously broken. And since in some cases 'CONFIG_SYSCTL'
Eric Andersen [Mon, 12 Sep 2005 21:42:48 +0000 (21:42 -0000)]
sysctl was seriously broken.  And since in some cases 'CONFIG_SYSCTL'
could be included into application space via the kernel config option
of the same name, nasty compilation problems could result.

19 years agoKunihiko IMAI writes:
Eric Andersen [Mon, 12 Sep 2005 19:16:11 +0000 (19:16 -0000)]
Kunihiko IMAI writes:

We're using busybox and thank the authors.

I found a short buffer allocation at busybox/sysklod/syslod.c.  It
mis-declares length of filenames, seems to forget for '\0' or two
column numbered suffix of backup files (ex. messages.10).

Here is a patch for the problem.

Thanks.

19 years agobb_mkdep speed up * 10!
"Vladimir N. Oleynik" [Mon, 12 Sep 2005 16:39:47 +0000 (16:39 -0000)]
bb_mkdep speed up * 10!

19 years agonew my scripts/mm_mkdep, dependences work now
"Vladimir N. Oleynik" [Mon, 12 Sep 2005 12:33:27 +0000 (12:33 -0000)]
new my scripts/mm_mkdep, dependences work now

19 years agoClean up memory allocation.
Rob Landley [Sun, 11 Sep 2005 23:45:28 +0000 (23:45 -0000)]
Clean up memory allocation.

19 years agoThis combines a patch from Bernhard Fischer (moving the clean of
Rob Landley [Sun, 11 Sep 2005 01:16:47 +0000 (01:16 -0000)]
This combines a patch from Bernhard Fischer (moving the clean of
scripts/config from distclean to clean) with a sed consolidation that's
been in my tree for a bit, and switching the GPL boilerplate to just point
at LICENSE.

19 years agoCleanup patch by Bernhard Fischer, removing unnecessary includes of
Rob Landley [Sun, 11 Sep 2005 01:05:30 +0000 (01:05 -0000)]
Cleanup patch by Bernhard Fischer, removing unnecessary includes of
getopt.h, whitespace changes, typos, etc.

19 years agouse CONFIG_BUFFER macros like cow suggested and shrink code size some more by using...
Mike Frysinger [Sat, 10 Sep 2005 04:10:18 +0000 (04:10 -0000)]
use CONFIG_BUFFER macros like cow suggested and shrink code size some more by using less variables and more gotos ;)

19 years agoPatch from Dirk Clemens so df doesn't run numbers together on drives >100gig.
Rob Landley [Sat, 10 Sep 2005 02:59:35 +0000 (02:59 -0000)]
Patch from Dirk Clemens so df doesn't run numbers together on drives >100gig.

19 years agosome tweaks by cow to shrink a little
Mike Frysinger [Sat, 10 Sep 2005 02:47:19 +0000 (02:47 -0000)]
some tweaks by cow to shrink a little

19 years agoPut the menu_is_visible test back because it avoids printing spurious
Rob Landley [Thu, 8 Sep 2005 11:50:48 +0000 (11:50 -0000)]
Put the menu_is_visible test back because it avoids printing spurious
(albeit commented out) garbage.

19 years agoRob, I don't know why you feel some unexplainable compulsion to get rid of
Manuel Novoa III [Thu, 8 Sep 2005 06:02:49 +0000 (06:02 -0000)]
Rob, I don't know why you feel some unexplainable compulsion to get rid of
a meaningless warning.  But I do expect you preserve the coding style
and variable names when all you're doing is tweaking some of my code.

I repeat... do NOT change whitespace, variable names, or coding style in
any of my code simply to conform to your coding style.

19 years agoA bit overdue...
Rob Landley [Thu, 8 Sep 2005 03:48:47 +0000 (03:48 -0000)]
A bit overdue...

19 years agoTito sent more size tweaks.
Rob Landley [Thu, 8 Sep 2005 03:27:06 +0000 (03:27 -0000)]
Tito sent more size tweaks.

19 years agoFix the warning by rewriting the function to be smaller and simpler.
Rob Landley [Thu, 8 Sep 2005 03:22:09 +0000 (03:22 -0000)]
Fix the warning by rewriting the function to be smaller and simpler.

I'd appreciate somebody on a __BIG_ENDIAN platform testing this out; I haven't
got the hardware...

19 years agoTito sent in a few patches converting instances of CONFIG_FEATURE_CLEAN_UP to
Rob Landley [Thu, 8 Sep 2005 03:11:58 +0000 (03:11 -0000)]
Tito sent in a few patches converting instances of CONFIG_FEATURE_CLEAN_UP to
ENABLE_FEATURE_CLEAN_UP.

19 years agocommitting:
Paul Fox [Wed, 7 Sep 2005 16:56:02 +0000 (16:56 -0000)]
committing:
       Summary      0000242: ash: read -t broken
this also implements -n and -s options to read.  (they're configured
together because most of their code is in common, and separating them
seemed silly.

19 years agoYeah, Linus is the original author of three of our apps, but A) he doesn't
Rob Landley [Wed, 7 Sep 2005 07:12:19 +0000 (07:12 -0000)]
Yeah, Linus is the original author of three of our apps, but A) he doesn't
work at transmeta anymore so that email's bogus, B) hands up everybody who
thinks emailing him about a problem in the busybox version of those apps
would be productive?

19 years agoFix warnings.
Rob Landley [Wed, 7 Sep 2005 07:04:56 +0000 (07:04 -0000)]
Fix warnings.

19 years agoThis code exists to compensate for ancient broken build environments which,
Rob Landley [Wed, 7 Sep 2005 06:39:02 +0000 (06:39 -0000)]
This code exists to compensate for ancient broken build environments which,
as far as I can tell, are no longer relevant.  Modern busybox refuses to
build under libc5 (there's a specific test and #error for that), and
I'm not sure building against 2.1 kernel headers on Alpha was ever relevant.

I'm happy to put any of this back if anybody can point to a real need for it,
but if so we need to specifically document what environment is being
compensated for.  (And we should quarrantine the build environment code
into one place, anyway.  Maybe "quirks.h" for known compiler and
libc quirks?)

19 years agoDon't comment warnings. Don't put #warnings about warnings. _FIX_ warnings.
Rob Landley [Wed, 7 Sep 2005 04:18:36 +0000 (04:18 -0000)]
Don't comment warnings.  Don't put #warnings about warnings.  _FIX_ warnings.

This fixes the warning, and makes the binary smaller out of sheer pique.
(Yes, since Manuel did this one it's nice tight code that took several
attempts to shrink, but I was ticked.)

Add the start of a test for uniq; this is about the first 1/3 of the
tests we need for full susv3 coverage of uniq.

19 years agoAmir Shalem writes:
Mike Frysinger [Tue, 6 Sep 2005 21:53:21 +0000 (21:53 -0000)]
Amir Shalem writes:
this patch fix syscall names mix-up in libb/syscalls.c

19 years ago- sync traceroute with Slackware-10.1
"Vladimir N. Oleynik" [Tue, 6 Sep 2005 16:08:33 +0000 (16:08 -0000)]
- sync traceroute with Slackware-10.1
  (support -FIl -g gw -i if -z pt now).
- libbb/getopt_ulflags.c support bb_opt_complementally="x-x" as trigger now

19 years agotypo, thanks, Bernhard Fischer
"Vladimir N. Oleynik" [Tue, 6 Sep 2005 13:21:39 +0000 (13:21 -0000)]
typo, thanks, Bernhard Fischer

19 years agoWhitespace-level changes. Replace s0,s1,e0,e1 with real variable names and
Rob Landley [Tue, 6 Sep 2005 01:57:11 +0000 (01:57 -0000)]
Whitespace-level changes.  Replace s0,s1,e0,e1 with real variable names and
clean up whitespace and curly brackets a bit.  Resulting binary should be
identical.

19 years agoget_terminal_width_height.c was in there twice.
Rob Landley [Mon, 5 Sep 2005 18:42:18 +0000 (18:42 -0000)]
get_terminal_width_height.c was in there twice.

19 years agoBernhard Fischer pointed out that CONFIG_FEATURE_MOUNT_FORCE no longer exists.
Rob Landley [Mon, 5 Sep 2005 17:52:47 +0000 (17:52 -0000)]
Bernhard Fischer pointed out that CONFIG_FEATURE_MOUNT_FORCE no longer exists.

19 years agoRegarding 11330: when you fix a bug that the testsuite doesn't catch, add
Rob Landley [Mon, 5 Sep 2005 16:27:19 +0000 (16:27 -0000)]
Regarding 11330: when you fix a bug that the testsuite doesn't catch, add
a test case to the test suite.

19 years agodestroy bug [0000381]
"Vladimir N. Oleynik" [Mon, 5 Sep 2005 15:50:56 +0000 (15:50 -0000)]
destroy bug [0000381]

19 years agodestroy bug [0000404]
"Vladimir N. Oleynik" [Mon, 5 Sep 2005 15:46:26 +0000 (15:46 -0000)]
destroy bug [0000404]

19 years ago- usage fresh bb_getopt_ulflags for ls -- very size reduce.
"Vladimir N. Oleynik" [Mon, 5 Sep 2005 15:06:57 +0000 (15:06 -0000)]
- usage fresh bb_getopt_ulflags for ls -- very size reduce.
- small indent corrections
- remove unecessary malloc and free

19 years ago1) bb_opt_complementaly -> bb_opt_complementally
"Vladimir N. Oleynik" [Mon, 5 Sep 2005 14:46:07 +0000 (14:46 -0000)]
1) bb_opt_complementaly -> bb_opt_complementally
2) better support long options
3) new flag '!' for bb_opt_complementally: produce bb_show_usage() if BB_GETOPT_ERROR internally

19 years ago1) sync with dash_0.5.2-7
"Vladimir N. Oleynik" [Mon, 5 Sep 2005 13:25:11 +0000 (13:25 -0000)]
1) sync with dash_0.5.2-7
2) but expand PS# as config option
3) correct kill error message again
4) remove show "line number" for interactive run (patch pending for dash)

19 years agoremove infinity loop ./busybox --help busybox
"Vladimir N. Oleynik" [Mon, 5 Sep 2005 11:25:27 +0000 (11:25 -0000)]
remove infinity loop ./busybox --help busybox

19 years agoThe previous commit left confdata writing out:
Rob Landley [Mon, 5 Sep 2005 11:04:30 +0000 (11:04 -0000)]
The previous commit left confdata writing out:

CONFIG_NUMERIC_CONSTANT=

And on reading it back in, it would complain that '' was an invalid value for
that field.  I.E. "make allnoconfig && make" worked fine, but
"make allnoconfig && make menuconfig" barfed reading in the config file.

So now I have it write out "0" as the blank value.  (It's initialized to the
default value when the menu becomes visible anyway; I checked.)  That seems
to work.

19 years agoNow that "make sizes" is in, the faq should mention it...
Rob Landley [Mon, 5 Sep 2005 10:25:51 +0000 (10:25 -0000)]
Now that "make sizes" is in, the faq should mention it...

19 years agoFix the warnings, and fix the following two obvious segfaults:
Rob Landley [Mon, 5 Sep 2005 06:16:53 +0000 (06:16 -0000)]
Fix the warnings, and fix the following two obvious segfaults:
./busybox getopt -n one -n two woot
./busybox getopt -o one -o two woot

This entire applet is still an enormous pile of garbage, which I can't clean
up because I really have no idea what it's for.  (Both "man getopt" and trying
it out on the command line a bit fail to enlighten me.  Reading the code, the
fact half of it seems to be special cases for bash vs tcsh does not fill me
with confidence.)

19 years agoActually returning ret might be good.
Rob Landley [Mon, 5 Sep 2005 05:36:15 +0000 (05:36 -0000)]
Actually returning ret might be good.

19 years agoFixes so "make allnoconfig" works again.
Rob Landley [Mon, 5 Sep 2005 04:13:33 +0000 (04:13 -0000)]
Fixes so "make allnoconfig" works again.

The configure system's save function edited out sub-menus that wouldn't be
displayed in the current configuration, meaning config.h wouldn't have #udef
entries for those symbols, meaning bb_config.h would have the relevant
ENABLE_ missing instead of defined to 0.  This broke the build.

So I fixed it, and then reorganized the applets.c and busybox.c to take
away the warnings this revealed (code that would be optimized out was making
calls to functions that hadn't been prototyped.  So I added an #else case
to those #ifdefs to #define the relevant functions to empty macros to
placate the warnings.

I also reorganized the applets.c code to make adding such an #else case less
of a pain (and make the need for prototyping go away by moving the functions
up before they were used, and generally wind up with fewer #ifdefs in
the code by putting all the logic in one place).  This resulted in a huge
seeming patch, when most if it just moves code from one place to another
without touching it...

Upside: make allyesconfig and make allnoconfig should both work now.

19 years agoRemove trailing crud from busybox.tests. (Sleep deprivation strikes again...)
Rob Landley [Sun, 4 Sep 2005 11:13:10 +0000 (11:13 -0000)]
Remove trailing crud from busybox.tests.  (Sleep deprivation strikes again...)

19 years agoGeneral cleanup of command line parsing to allow "busybox" to work as a prefix.
Rob Landley [Sun, 4 Sep 2005 11:10:37 +0000 (11:10 -0000)]
General cleanup of command line parsing to allow "busybox" to work as a prefix.
(I.E. any argv[0] that starts with "busybox" winds up in busybox_main().)

Added testing/busybox.tests which tests the following permutations:

./busybox
./busybox-suffix
./busybox cat
./busybox-suffix cat
./busybox --help
./busybox-suffix --help
./busybox --help cat
./busybox-suffix --help cat
./busybox --help unknown
./busybox-suffix --help unknown
./unknown

Also repair the test suite so ./runtest calls the ".tests" scripts properly.

Note: you can now go "busybox busybox busbox ls -l" and it'll take it.  The
new code is pretty generic.  I can block that if anybody can come up with a
good reason to...

19 years agotouchup some grammar, formatting, and remove some outdated notes
Mike Frysinger [Fri, 2 Sep 2005 23:06:30 +0000 (23:06 -0000)]
touchup some grammar, formatting, and remove some outdated notes

19 years agoupdate the verbage describing what is needed for developer access
Eric Andersen [Fri, 2 Sep 2005 14:10:47 +0000 (14:10 -0000)]
update the verbage describing what is needed for developer access

19 years agoThis goes along with the new testsuite code...
Rob Landley [Fri, 2 Sep 2005 01:55:58 +0000 (01:55 -0000)]
This goes along with the new testsuite code...

19 years agoWorking on a new test harness. Moved the sort tests into it.
Rob Landley [Fri, 2 Sep 2005 00:41:53 +0000 (00:41 -0000)]
Working on a new test harness.  Moved the sort tests into it.

19 years agoMissing break was screwing up 'y//' command. Bug #248.
Rob Landley [Fri, 2 Sep 2005 00:10:06 +0000 (00:10 -0000)]
Missing break was screwing up 'y//' command.  Bug #248.

19 years agoFix for bug 383: attempting to "tar c /" would error out because stripping
Rob Landley [Thu, 1 Sep 2005 11:36:21 +0000 (11:36 -0000)]
Fix for bug 383: attempting to "tar c /" would error out because stripping
trailing / turns that into an empty string.

19 years agoAccording to bug #63, crond is unhappy with crontab lines that don't end in a
Rob Landley [Thu, 1 Sep 2005 10:23:57 +0000 (10:23 -0000)]
According to bug #63, crond is unhappy with crontab lines that don't end in a
newline, or lines that have trailing spaces.

19 years agoTracking system but #4: php needs the environment variable SCRIPT_FILENAME.
Rob Landley [Thu, 1 Sep 2005 09:38:32 +0000 (09:38 -0000)]
Tracking system but #4: php needs the environment variable SCRIPT_FILENAME.

19 years agoBernhard Fischer trimmed down dos2unix a bit.
Rob Landley [Thu, 1 Sep 2005 03:11:19 +0000 (03:11 -0000)]
Bernhard Fischer trimmed down dos2unix a bit.

19 years agoPatch -i support from Berhnard Fischer.
Rob Landley [Thu, 1 Sep 2005 03:02:23 +0000 (03:02 -0000)]
Patch -i support from Berhnard Fischer.

19 years agoBernhard Fischer suggested that get_terminal_width_height() should return
Rob Landley [Thu, 1 Sep 2005 02:57:45 +0000 (02:57 -0000)]
Bernhard Fischer suggested that  get_terminal_width_height() should return
the result of the ioctl so callers can tell if we have a tty.  (0 means
we have a tty, nonzero means the ioctl couldn't find size info and we
fake 80x24.  Really we should fake 80x25, but oh well...)

19 years agoOff by one error in max_leases sanity check. Bug 349, apparently.
Rob Landley [Thu, 1 Sep 2005 02:43:39 +0000 (02:43 -0000)]
Off by one error in max_leases sanity check.  Bug 349, apparently.

19 years agoBernhard Fischer sent a patch to make "make sizes" work when building in
Rob Landley [Thu, 1 Sep 2005 02:40:21 +0000 (02:40 -0000)]
Bernhard Fischer sent a patch to make "make sizes" work when building in
another output directory.

19 years agoBusybox should only multiplex based on argv[1] when it's called as "busybox".
Rob Landley [Thu, 1 Sep 2005 00:02:56 +0000 (00:02 -0000)]
Busybox should only multiplex based on argv[1] when it's called as "busybox".
Otherwise if you build busybox without a given applet you get the wrong error
message when you call it via a symlink to that applet.

(You also get the wrong behavior; it tries to use argv[1] as the command
name just like busybox does for _any_ unknown, and although I doubt
"echo rm -rf *" is common usage there's no upside and enough downside to
make me nervous.)

This fixes it.

19 years agoAnand Avati hit an integer overflow problem in our unzip code.
Rob Landley [Wed, 31 Aug 2005 22:03:15 +0000 (22:03 -0000)]
Anand Avati hit an integer overflow problem in our unzip code.

19 years agoDon't comment warnings, _FIX_ warnings. (And putting in #warnings about
Rob Landley [Tue, 30 Aug 2005 20:26:17 +0000 (20:26 -0000)]
Don't comment warnings, _FIX_ warnings.  (And putting in #warnings about
other warnings is just gross.)

On a side note, while I was there, I made the code slightly smaller.

19 years agoWhy have a separate CONFIG_INIT_SWAPON when we already have CONFIG_SWAPONOFF?
Rob Landley [Tue, 30 Aug 2005 18:50:37 +0000 (18:50 -0000)]
Why have a separate CONFIG_INIT_SWAPON when we already have CONFIG_SWAPONOFF?

19 years agoBug system entry #1 says that when busybox doesn't have swapoff,
Rob Landley [Tue, 30 Aug 2005 18:17:05 +0000 (18:17 -0000)]
Bug system entry #1 says that when busybox doesn't have swapoff,
init shouldn't call swapoff.

19 years agoAmir Shalem found some bugs in the new mount code; unknown options didn't get
Rob Landley [Tue, 30 Aug 2005 17:07:49 +0000 (17:07 -0000)]
Amir Shalem found some bugs in the new mount code; unknown options didn't get
added to the list, and my assumption that nfsmount() actually called
mount() was incorrect (and I coded it wrong anyway; I hate having to touch
codepaths I can't personally test).

19 years agoDirk Clemens pointed out how easy it is to support bzip2 compression, since we
Rob Landley [Tue, 30 Aug 2005 03:40:03 +0000 (03:40 -0000)]
Dirk Clemens pointed out how easy it is to support bzip2 compression, since we
shell out to an external program to handle gzip anyway...

19 years agopoint to busybox.net
Eric Andersen [Sun, 28 Aug 2005 12:47:39 +0000 (12:47 -0000)]
point to busybox.net

19 years agoDumb little hack, "make sizes". Try it.
Rob Landley [Sun, 28 Aug 2005 08:24:21 +0000 (08:24 -0000)]
Dumb little hack, "make sizes".  Try it.

19 years agomake sure __NF_sysfs exists before trying to use it
Mike Frysinger [Sat, 27 Aug 2005 18:19:01 +0000 (18:19 -0000)]
make sure __NF_sysfs exists before trying to use it

19 years agosome portability fixes by rmh in Bug 395
Mike Frysinger [Sat, 27 Aug 2005 18:18:06 +0000 (18:18 -0000)]
some portability fixes by rmh in Bug 395

19 years agoBernhard Fischer submitted a couple of Makefile patches:
Rob Landley [Wed, 24 Aug 2005 00:41:52 +0000 (00:41 -0000)]
Bernhard Fischer submitted a couple of Makefile patches:
- Fix building out-of-tree
- remove duplicate rule in toplevel Makefile
- peruse make's builtin notion of `dirname $@'

19 years agoRemove something we did, not that I'm doing something, and add something we
Rob Landley [Tue, 23 Aug 2005 23:08:17 +0000 (23:08 -0000)]
Remove something we did, not that I'm doing something, and add something we
haven't done.  Woo.

19 years agoWhen compiling with FEATURE_MTAB_SUPPORT disabled, the call to erase_mtab()
Rob Landley [Tue, 23 Aug 2005 20:03:17 +0000 (20:03 -0000)]
When compiling with FEATURE_MTAB_SUPPORT disabled, the call to erase_mtab()
can never be made because useMtab is initialized to 0, and all the other
assignments of that variable assign 0 to it.  Any compiler that can perform
simple constant propogation on local variables will optimize away if statements
testing against that variable, thus the call to erase_mtab() will never be
made.

When compiling for arm using gcc 3.3.3 with FEATURE_MTAB_SUPPORT disabled,
the linker complains that it can't find erase_mtab().  The arm optimizer isn't
exactly the brightest member of the family, and apparently needs to be hit over
the head with a hammer to get its' attention...

19 years agoTwo FEATURE_CLEAN_UP patches from Shaun Jackman, closing filehandles.
Rob Landley [Mon, 22 Aug 2005 15:57:50 +0000 (15:57 -0000)]
Two FEATURE_CLEAN_UP patches from Shaun Jackman, closing filehandles.

19 years agoBernhard Fischer provided a mountpoint(1) applet. This is apparently something
Rob Landley [Sat, 20 Aug 2005 05:07:08 +0000 (05:07 -0000)]
Bernhard Fischer provided a mountpoint(1) applet.  This is apparently something
sysvinit provides, and which is used by the debian init scripts.

19 years agoAnnounce busybox 1.01.
Rob Landley [Wed, 17 Aug 2005 01:55:58 +0000 (01:55 -0000)]
Announce busybox 1.01.

19 years agoTito (among others) writes: previous patch was wrong, please stop screwing things up
Mike Frysinger [Wed, 17 Aug 2005 01:20:10 +0000 (01:20 -0000)]
Tito (among others) writes: previous patch was wrong, please stop screwing things up

19 years agoWhen doing 'make release' nuke the .svn directories, not the
Eric Andersen [Wed, 17 Aug 2005 00:03:23 +0000 (00:03 -0000)]
When doing 'make release' nuke the .svn directories, not the
CVS dirs (since they are no longer present)

19 years agothe () around isspace actually matters
Mike Frysinger [Tue, 16 Aug 2005 05:39:07 +0000 (05:39 -0000)]
the () around isspace actually matters

19 years agoBernhard Fischer says: use xmalloc() instead of malloc()
Mike Frysinger [Tue, 16 Aug 2005 05:32:42 +0000 (05:32 -0000)]
Bernhard Fischer says: use xmalloc() instead of malloc()

19 years agoBernhard Fischer says: fix comment typo
Mike Frysinger [Tue, 16 Aug 2005 05:27:51 +0000 (05:27 -0000)]
Bernhard Fischer says: fix comment typo

19 years agoBernhard Fischer writes: fix devfsd usage
Mike Frysinger [Mon, 15 Aug 2005 22:13:21 +0000 (22:13 -0000)]
Bernhard Fischer writes: fix devfsd usage

19 years agoIf we goto singlemount, do _not_ try to continue through the loop we jumped
Rob Landley [Sun, 14 Aug 2005 19:26:14 +0000 (19:26 -0000)]
If we goto singlemount, do _not_ try to continue through the loop we jumped
into.  (That means "mount -t ext2 /dev/thingy thingy" would segfault if
it failed instead of giving us an error message.)

19 years agoPatch from Stephane Billiart to improve ash error message.
Rob Landley [Sun, 14 Aug 2005 19:16:51 +0000 (19:16 -0000)]
Patch from Stephane Billiart to improve ash error message.

19 years agoStephane Billiart noticed that the return code of the new mount was horked.
Rob Landley [Sun, 14 Aug 2005 18:46:34 +0000 (18:46 -0000)]
Stephane Billiart noticed that the return code of the new mount was horked.
(Cleaned up the return code path to consistently use numbers so I don't keep
confusing EXIT_SUCCESS with TRUE and EXIT_FAILURE with FALSE...)

19 years agoVery minor tweak to tail.
Rob Landley [Sat, 13 Aug 2005 00:35:00 +0000 (00:35 -0000)]
Very minor tweak to tail.

19 years agomake sure distclean punts include/bbconfigopts.h
Mike Frysinger [Thu, 11 Aug 2005 21:59:52 +0000 (21:59 -0000)]
make sure distclean punts include/bbconfigopts.h

19 years agoMajor rewrite of mount, umount, losetup. Untangled lots of code, shrunk
Rob Landley [Wed, 10 Aug 2005 20:35:54 +0000 (20:35 -0000)]
Major rewrite of mount, umount, losetup.  Untangled lots of code, shrunk
things down a bit, fixed a number of funky corner cases, added support for
several new features (things like mount --move, mount --bind, lazy unounts,
automatic detection of loop mounts, and so on).  Probably broke several
other things, but it's fixable.  (Bang on it, tell me what doesn't work for
you...)

Note: you no longer need to say "-o loop".  It does that for you when
necessary.

Still need to add "user mount" support, which involves making mount suid.  Not
too hard to do under the new infrastructure, just haven't done it yet...

The previous code had the following notes, that belong in the version
control comments:

- * 3/21/1999   Charles P. Wright <cpwright@cpwright.com>
- *             searches through fstab when -a is passed
- *             will try mounting stuff with all fses when passed -t auto
- *
- * 1999-04-17  Dave Cinege...Rewrote -t auto. Fixed ro mtab.
- *
- * 1999-10-07  Erik Andersen <andersen@codepoet.org>.
- *              Rewrite of a lot of code. Removed mtab usage (I plan on
- *              putting it back as a compile-time option some time),
- *              major adjustments to option parsing, and some serious
- *              dieting all around.
- *
- * 1999-11-06  mtab support is back - andersee
- *
- * 2000-01-12   Ben Collins <bcollins@debian.org>, Borrowed utils-linux's
- *              mount to add loop support.
- *
- * 2000-04-30  Dave Cinege <dcinege@psychosis.com>
- *             Rewrote fstab while loop and lower mount section. Can now do
- *             single mounts from fstab. Can override fstab options for single
- *             mount. Common mount_one call for single mounts and 'all'. Fixed
- *             mtab updating and stale entries. Removed 'remount' default.
- *

19 years agoimplemented a builtin echo command in ash. moved the guts of the
Paul Fox [Tue, 9 Aug 2005 19:38:05 +0000 (19:38 -0000)]
implemented a builtin echo command in ash.  moved the guts of the
echo applet into libbb, and now call bb_echo() from both echo.c
and ash.c

19 years agocommiting:
Paul Fox [Thu, 4 Aug 2005 19:04:46 +0000 (19:04 -0000)]
commiting:
    0000025: vi-editing mode for ash