OSDN Git Service

android-x86/external-toybox.git
8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Fri, 27 May 2016 20:57:40 +0000 (13:57 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agoAdd -o TNAME for android, fix display of last field (retain left justifification
Rob Landley [Fri, 27 May 2016 19:54:27 +0000 (14:54 -0500)]
Add -o TNAME for android, fix display of last field (retain left justifification
when trimmed for screen width), make ARGS path trimming logic work with
spaces in path.

8 years agoStabilize another sort.
Elliott Hughes [Wed, 25 May 2016 21:43:56 +0000 (14:43 -0700)]
Stabilize another sort.

Unstable sorting means confusing diffs if you're checking in the generated
files. (Which I shouldn't be doing, but getting this bundle of random scripts
into Android's build system isn't going to be easy...)

8 years agoStop building toybox route.
Elliott Hughes [Wed, 25 May 2016 21:47:37 +0000 (14:47 -0700)]
Stop building toybox route.

From the commit that stopped creating the symlink:

    We should remove the "route" command because its output is irrelevant and
    misleading. This is because it only looks at the "main" routing table,
    which is a last resort routing table that is a) trumped by all other
    networks, b) empty except for directly-connected routes c) cannot even be
    used by non-root users.

    The proper way to understand device routing is to look at the output of
    "ip rule" and "ip route", both of which are already available.

    Example of how this is confusing:

      $ adb shell route
      Kernel IP routing table
      Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
      160.249.218.72  *               255.255.255.248 U     0      0        0 rmnet_data0
      192.168.144.0   *               255.255.240.0   U     0      0        0 wlan0

    These aren't guaranteed to be used. For example, according to the output
    above 160.249.218.73 should go through rmnet_data0, but it doesn't:

      $ adb shell ip route get 160.249.218.73
      160.249.218.73 via 192.168.159.254 dev wlan0  src 192.168.159.164  uid 0
          cache

    Also, there is no default route. But the device has networking just fine:

      $ adb shell ip route get 8.8.8.8
      8.8.8.8 via 192.168.159.254 dev wlan0  src 192.168.159.164  uid 0
          cache

Bug: http://b/27603033
Change-Id: I40be0a7e2841144797a67f9aeb82b9097cec5400

8 years agoStand alone builds of OLDTOY should show help text for corresponding NEWTOY.
Rob Landley [Wed, 25 May 2016 17:56:56 +0000 (12:56 -0500)]
Stand alone builds of OLDTOY should show help text for corresponding NEWTOY.

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Tue, 24 May 2016 22:27:49 +0000 (15:27 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agoThe "show parent command path" commit changed the criteria for active nodes,
Rob Landley [Tue, 24 May 2016 21:17:55 +0000 (16:17 -0500)]
The "show parent command path" commit changed the criteria for active nodes,
and the collate logic wasn't updated. (Oops.)

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Tue, 24 May 2016 16:13:35 +0000 (09:13 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agoRegenerate generated files.
Elliott Hughes [Tue, 24 May 2016 16:12:04 +0000 (09:12 -0700)]
Regenerate generated files.

Change-Id: I2014135856d6316a53c22cace2cc935cf65434f6

8 years agoFix remaining pkill tests.
Elliott Hughes [Mon, 23 May 2016 17:46:47 +0000 (10:46 -0700)]
Fix remaining pkill tests.

I don't know that anyone cares, but these failures made me think I'd
broken something while testing the -SIGNAL patch.

The -s parsing wasn't taking into account that -s 0 is a special case,
and the -o test was assuming that pkill can tell the difference between
two processes started at roughly the same time. Hopefully there's
higher-resolution data available that can avoid the need for yet another
sleep in the tests.

8 years agoFix pkill -9.
Elliott Hughes [Mon, 23 May 2016 17:34:01 +0000 (10:34 -0700)]
Fix pkill -9.

This fixes the existing test. Internal Android bug 28877702.

I've left -l as-is, even though the desktop doesn't support that.

8 years agoDon't truncate number fields for anything but right edge of screen, instead
Rob Landley [Mon, 23 May 2016 00:44:10 +0000 (19:44 -0500)]
Don't truncate number fields for anything but right edge of screen, instead
let them overflow and try to reclaim extra space from later short fields.

8 years agoMake -o COMMAND show parent command path for threads, and use saved length
Rob Landley [Sun, 22 May 2016 17:16:06 +0000 (12:16 -0500)]
Make -o COMMAND show parent command path for threads, and use saved length
calculations instead of doing strlen() on strings again to store lengths.

8 years agoComment and help text tweaks.
Rob Landley [Sun, 22 May 2016 17:00:16 +0000 (12:00 -0500)]
Comment and help text tweaks.

8 years agoAdd bufgetgrgid()
Rob Landley [Fri, 20 May 2016 19:28:13 +0000 (14:28 -0500)]
Add bufgetgrgid()

8 years agoMake build dependencies more granular. This should fix the problem where
Rob Landley [Fri, 20 May 2016 12:32:51 +0000 (07:32 -0500)]
Make build dependencies more granular. This should fix the problem where
"make top; make ps" produces a ps that can't do -A because generated/obj/ps.o
didn't get rebuilt.

8 years agoAdd bufgetpwuid. (Repeated calls to getpwuid() are really expensive.)
Rob Landley [Tue, 17 May 2016 17:10:07 +0000 (12:10 -0500)]
Add bufgetpwuid. (Repeated calls to getpwuid() are really expensive.)

8 years agoAdd top -O
Rob Landley [Mon, 16 May 2016 19:51:37 +0000 (14:51 -0500)]
Add top -O

8 years agoAdd stat -tL and -c %m%t%T.
Rob Landley [Mon, 16 May 2016 17:24:42 +0000 (12:24 -0500)]
Add stat -tL and -c %m%t%T.

Suggested by Sameer Pradhan, or possibly Bilal Qureshi.

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Tue, 10 May 2016 22:03:41 +0000 (15:03 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agoUse CFG_TOYBOX_NORECURSE to avoid
Rob Landley [Tue, 10 May 2016 08:26:02 +0000 (03:26 -0500)]
Use CFG_TOYBOX_NORECURSE to avoid
<strike>all that tedious mucking about with hyperspace</strike>
stack measuring in ways that confuse security stuff.

8 years agoTurn on CONFIG_TOYBOX_NORECURSE for SafeStack.
Elliott Hughes [Sat, 7 May 2016 18:09:44 +0000 (11:09 -0700)]
Turn on CONFIG_TOYBOX_NORECURSE for SafeStack.

Bug: http://b/27729263
Change-Id: Ibe1f530198045e4133960af61800b40d4348c59a

8 years agoWork around increasingly insane compiler developers wanting to make everything
Rob Landley [Sat, 7 May 2016 05:21:34 +0000 (00:21 -0500)]
Work around increasingly insane compiler developers wanting to make everything
undefined behavior so the optimizer can silently eliminate your entire program.

8 years agoRegenerate generated files.
Elliott Hughes [Fri, 6 May 2016 19:49:08 +0000 (12:49 -0700)]
Regenerate generated files.

Change-Id: I781e947ca3cf323262b0acb4269727357d7f9347

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Fri, 6 May 2016 18:06:20 +0000 (11:06 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agoFix static linkage of toybox binary.
Alistair Strachan [Thu, 5 May 2016 23:11:35 +0000 (16:11 -0700)]
Fix static linkage of toybox binary.

If toybox was compiled statically e.g. for use in a recovery ramdisk,
it would not build after b66a29a. This is because libselinux already
exports a selinux_log_callback() function which getprop.c is trying
to override.

This change simply makes the second symbol static, since it is only
used in a function table so we do not actually need to export it
from toybox.

Change-Id: I603258877ae2286660df35738bbf4f4285df5b54
Signed-off-by: Alistair Strachan <alistair.strachan@imgtec.com>
8 years agoAdd top -H.
Elliott Hughes [Thu, 5 May 2016 17:27:56 +0000 (10:27 -0700)]
Add top -H.

Also enable the same behavior we have with ps, where adding a thread-related
field implies that you're interested in threads.

Also clean up the help text slightly --- we're still not including the
TOP_COMMON help, but at least when we do, we'll have each option covered with
no duplicates.

8 years agoAndroid roadmap update.
Elliott Hughes [Thu, 5 May 2016 16:23:07 +0000 (09:23 -0700)]
Android roadmap update.

I removed iftop a while back but didn't remove it everywhere in the
documentation. I switched over to toybox ps yesterday.

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Thu, 5 May 2016 16:34:17 +0000 (09:34 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years ago"top -b -n1" should end with a newline.
Elliott Hughes [Thu, 5 May 2016 01:25:41 +0000 (18:25 -0700)]
"top -b -n1" should end with a newline.

The use of TT.time to determine whether to use \r or \n is a little odd,
but let's stick with that for now. Correct the spelling of millitime and
add a test.

8 years agoElliott says -T should fetch threads even when not displaying thread fields.
Rob Landley [Wed, 4 May 2016 23:43:19 +0000 (18:43 -0500)]
Elliott says -T should fetch threads even when not displaying thread fields.

8 years agoSimplify ps android scheduling policy fetch slightly.
Rob Landley [Wed, 4 May 2016 23:37:50 +0000 (18:37 -0500)]
Simplify ps android scheduling policy fetch slightly.

8 years agoSwitch to toybox ps.
Elliott Hughes [Wed, 4 May 2016 22:18:27 +0000 (15:18 -0700)]
Switch to toybox ps.

Change-Id: I1f0307a8c9511d02a173555b92ebd0f83df229c9

8 years agoRegenerate generated files.
Elliott Hughes [Wed, 4 May 2016 20:42:18 +0000 (13:42 -0700)]
Regenerate generated files.

Change-Id: I8d1590a61a7339f65cd1bdf28601f85768563b57

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Wed, 4 May 2016 20:36:49 +0000 (13:36 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agoHostname cleanup, as described on the list.
Rob Landley [Wed, 4 May 2016 10:14:35 +0000 (05:14 -0500)]
Hostname cleanup, as described on the list.

8 years agoAdd -b and -F arguments to hostname
Paul Barker [Sun, 1 May 2016 14:42:57 +0000 (15:42 +0100)]
Add -b and -F arguments to hostname

These arguments are required to correctly set the hostname at boot time. They
are used by the '/etc/init.d/hostname.sh' init script in an OpenEmbedded system.

8 years agoFluff up README.
Rob Landley [Mon, 2 May 2016 23:59:57 +0000 (18:59 -0500)]
Fluff up README.

8 years agoAdd a ps "PCY" field for Android scheduling policy.
Elliott Hughes [Sat, 30 Apr 2016 01:04:20 +0000 (18:04 -0700)]
Add a ps "PCY" field for Android scheduling policy.

8 years agoRegenerate generated files.
Elliott Hughes [Sat, 30 Apr 2016 16:52:51 +0000 (09:52 -0700)]
Regenerate generated files.

Change-Id: Id500eb735af13344c65dfce96f37199e65d4fe53

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Sat, 30 Apr 2016 16:26:12 +0000 (09:26 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agoFirst guess at showing thread names properly.
Rob Landley [Sat, 30 Apr 2016 00:45:18 +0000 (19:45 -0500)]
First guess at showing thread names properly.

8 years agoShow "-" rather 0 if we don't know whether a process is 32- or 64-bit.
Elliott Hughes [Fri, 29 Apr 2016 21:04:40 +0000 (14:04 -0700)]
Show "-" rather 0 if we don't know whether a process is 32- or 64-bit.

8 years agoFix ps -O help output alignment.
Elliott Hughes [Fri, 29 Apr 2016 20:44:01 +0000 (13:44 -0700)]
Fix ps -O help output alignment.

8 years agoRegenerate generated files.
Elliott Hughes [Fri, 29 Apr 2016 21:10:40 +0000 (14:10 -0700)]
Regenerate generated files.

Change-Id: Ifbb9056f6fa700c3426012d4ec6c43b70a4947c4

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Fri, 29 Apr 2016 18:57:24 +0000 (11:57 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agops thread supportand 32/64 bit detection.
Rob Landley [Mon, 25 Apr 2016 22:46:26 +0000 (17:46 -0500)]
ps thread supportand 32/64 bit detection.

Add ps -o BIT,TID,TCNT, and make -T display "PID,TID" for default output types
(adding TCNT to -f)

8 years agoFix find bug reported by Tom Marshall, add test for it, and while we're at it
Rob Landley [Mon, 25 Apr 2016 05:55:01 +0000 (00:55 -0500)]
Fix find bug reported by Tom Marshall, add test for it, and while we're at it
fix two tests looking for a too-specific error message (so TEST_HOST failed).

8 years agoAdd "infrastructure in search of a user" to cleanup.html.
Rob Landley [Sun, 24 Apr 2016 19:30:31 +0000 (14:30 -0500)]
Add "infrastructure in search of a user" to cleanup.html.

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Sun, 24 Apr 2016 16:19:49 +0000 (09:19 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agoAdd the non-tty more(1) test.
Elliott Hughes [Sat, 23 Apr 2016 21:20:38 +0000 (14:20 -0700)]
Add the non-tty more(1) test.

(This was supposed to be in the earlier patch, but I was fooled by
"git commit -a" yet again...)

8 years agoFix more to not append an extra newline.
Elliott Hughes [Fri, 22 Apr 2016 01:18:05 +0000 (18:18 -0700)]
Fix more to not append an extra newline.

More's sigatexit handler needs to distinguish between normal exit and exit
due to receipt of a signal.

Change tty_sigreset to look at the signal number too, so that pressing 'q'
to exit top doesn't cause its exit status to be 128.

8 years agoFix more to a non-tty.
Elliott Hughes [Fri, 22 Apr 2016 00:41:40 +0000 (17:41 -0700)]
Fix more to a non-tty.

And add a test.

8 years agoRemove inappropriate bzip tests (we're not comparing with random "host version")
Rob Landley [Sat, 23 Apr 2016 10:26:17 +0000 (05:26 -0500)]
Remove inappropriate bzip tests (we're not comparing with random "host version")
replace with testing 2 known files out of blkid tests, and add badcrc test.

8 years agoFiz bzcat segfault reported by John Regehr (bad crc didn't set error message).
Rob Landley [Sat, 23 Apr 2016 10:04:37 +0000 (05:04 -0500)]
Fiz bzcat segfault reported by John Regehr (bad crc didn't set error message).

8 years agoAllow EVAL= to run test under a different shell (for sh.test)
Rob Landley [Sat, 23 Apr 2016 00:08:56 +0000 (19:08 -0500)]
Allow EVAL= to run test under a different shell (for sh.test)

8 years agoRegenerate generated files.
Elliott Hughes [Fri, 22 Apr 2016 15:34:27 +0000 (08:34 -0700)]
Regenerate generated files.

Change-Id: I2557e28527c58f25c814a961c003371b1a3214fc

8 years agoHave dirtree_recurse() take the new dirfd as an argument.
Rob Landley [Thu, 21 Apr 2016 22:46:25 +0000 (17:46 -0500)]
Have dirtree_recurse() take the new dirfd as an argument.

8 years agoFix touch -t seconds parsing
Davis Mosenkovs [Wed, 20 Apr 2016 21:44:13 +0000 (14:44 -0700)]
Fix touch -t seconds parsing

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Thu, 21 Apr 2016 01:22:50 +0000 (18:22 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agoStart of shell tests: $(()) and $'' expansion.
Rob Landley [Wed, 20 Apr 2016 06:56:10 +0000 (01:56 -0500)]
Start of shell tests: $(()) and $'' expansion.

8 years agoReplace list_working with just list, "make list list_pending" shows combined list.
Rob Landley [Wed, 20 Apr 2016 05:47:05 +0000 (00:47 -0500)]
Replace list_working with just list, "make list list_pending" shows combined list.

8 years agops and top are implemented from the android todo. Record enh's wishlist items.
Rob Landley [Wed, 20 Apr 2016 04:03:12 +0000 (23:03 -0500)]
ps and top are implemented from the android todo. Record enh's wishlist items.

8 years agoDust off toysh, remove ancient config debris, add start of prompt logic.
Rob Landley [Sun, 17 Apr 2016 21:34:47 +0000 (16:34 -0500)]
Dust off toysh, remove ancient config debris, add start of prompt logic.

8 years agoFix tail -NUM again.
Elliott Hughes [Sat, 16 Apr 2016 15:19:23 +0000 (08:19 -0700)]
Fix tail -NUM again.

This time with a test.

8 years agoFix "ssh top".
Elliott Hughes [Sat, 16 Apr 2016 01:48:12 +0000 (18:48 -0700)]
Fix "ssh top".

The ps.c change allows us to start. The interestingtimes.c change allows
us to clean up properly afterwards if you ^C out (which you usually do).

Tested with both ssh and "adb shell" (the latter being where I noticed
the problem).

8 years agoFix top -b.
Elliott Hughes [Sat, 16 Apr 2016 01:38:53 +0000 (18:38 -0700)]
Fix top -b.

Batch mode should never output terminal escape sequences, should ignore
the keyboard, and should include a gap between datasets.

8 years agoRename gratuitous references to Roger Zelazny's amber series, add comments.
Rob Landley [Sat, 16 Apr 2016 19:03:36 +0000 (14:03 -0500)]
Rename gratuitous references to Roger Zelazny's amber series, add comments.

8 years agoStop creating a symlink for route.
Elliott Hughes [Fri, 15 Apr 2016 18:57:05 +0000 (11:57 -0700)]
Stop creating a symlink for route.

We should remove the "route" command because its output is irrelevant and
misleading. This is because it only looks at the "main" routing table,
which is a last resort routing table that is a) trumped by all other
networks, b) empty except for directly-connected routes c) cannot even be
used by non-root users.

The proper way to understand device routing is to look at the output of
"ip rule" and "ip route", both of which are already available.

Example of how this is confusing:

  $ adb shell route
  Kernel IP routing table
  Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
  160.249.218.72  *               255.255.255.248 U     0      0        0 rmnet_data0
  192.168.144.0   *               255.255.240.0   U     0      0        0 wlan0

These aren't guaranteed to be used. For example, according to the output
above 160.249.218.73 should go through rmnet_data0, but it doesn't:

  $ adb shell ip route get 160.249.218.73
  160.249.218.73 via 192.168.159.254 dev wlan0  src 192.168.159.164  uid 0
      cache

Also, there is no default route. But the device has networking just fine:

  $ adb shell ip route get 8.8.8.8
  8.8.8.8 via 192.168.159.254 dev wlan0  src 192.168.159.164  uid 0
      cache

This is a first step. If no device breaks, I'll stop building it too. In the
meantime, it's still available as "toybox route".

Bug: http://b/27603033
Change-Id: Ie444bfa9083e1f11b332d5d050a07cec48436e11

8 years agoFix a buffer overflow in diff -r.
Andy Chu [Sun, 20 Mar 2016 06:11:30 +0000 (23:11 -0700)]
Fix a buffer overflow in diff -r.

We were doing two 32-byte memset()s instead of two 16-byte memset()s.
'dir' referred to the instance (array of 2) and not the struct type.

Add some test coverage for diff, including a case that hit this bug.

The bug was found by running cp.test under AddressSanitizer, since it
happens to use diff.

8 years agoAndy Chu pointed out an out of bounds access for zero length lines.
Rob Landley [Fri, 15 Apr 2016 02:00:54 +0000 (21:00 -0500)]
Andy Chu pointed out an out of bounds access for zero length lines.

While we're at it, use unsigned for the line length.

8 years agosed -f - should read from stdin.
Rob Landley [Mon, 11 Apr 2016 16:32:36 +0000 (11:32 -0500)]
sed -f - should read from stdin.

8 years agoDidn't check in all my local basename_r->getbasename changes. (Oops.)
Rob Landley [Sun, 10 Apr 2016 19:35:51 +0000 (14:35 -0500)]
Didn't check in all my local basename_r->getbasename changes. (Oops.)

8 years agoJakob Flierl pointed out a broken URL in the README.
Rob Landley [Sat, 9 Apr 2016 16:04:16 +0000 (11:04 -0500)]
Jakob Flierl pointed out a broken URL in the README.

8 years agoRedefining basename_r to mean something random seems popular (bionic and freebsd
Rob Landley [Fri, 8 Apr 2016 23:25:59 +0000 (18:25 -0500)]
Redefining basename_r to mean something random seems popular (bionic and freebsd
both did it) so use getbasename instead.

8 years agoMerge remote-tracking branch 'toybox/master' into HEAD
Elliott Hughes [Wed, 30 Mar 2016 17:22:06 +0000 (10:22 -0700)]
Merge remote-tracking branch 'toybox/master' into HEAD

8 years agoRegenerate generated files.
Elliott Hughes [Wed, 30 Mar 2016 17:20:45 +0000 (10:20 -0700)]
Regenerate generated files.

Change-Id: I5323d381d9dd2e10c90fc10d16ef21dbffdd716b

8 years agoWhitespace and comment.
Rob Landley [Wed, 30 Mar 2016 08:20:13 +0000 (03:20 -0500)]
Whitespace and comment.

8 years agoImplement mv -n / cp -n (no clobber).
Andy Chu [Sat, 19 Mar 2016 07:34:42 +0000 (00:34 -0700)]
Implement mv -n / cp -n (no clobber).

This fixes a failing test case in mv.test.

Test changes:
- Add coverage for -i (interactive).
- Better descriptions, better formatting, and removed some redundant
  cases.

8 years agoFix build when wchar_t is unsigned.
Elliott Hughes [Sat, 26 Mar 2016 17:54:23 +0000 (10:54 -0700)]
Fix build when wchar_t is unsigned.

  external/toybox/lib/linestack.c:84:5: error: conflicting types for 'crunch_str'
  int crunch_str(char **str, int width, FILE *out, char *escmore,
      ^
  external/toybox/lib/lib.h:228:5: note: previous declaration is here
  int crunch_str(char **str, int width, FILE *out, char *escmore,
      ^

I'm not sure whether the interface was intended to be wchar_t or int, but
switching to int is the smaller change...

8 years agoFix exit code of things like sed -i that use tempfile_handler().
Rob Landley [Tue, 29 Mar 2016 08:59:20 +0000 (03:59 -0500)]
Fix exit code of things like sed -i that use tempfile_handler().

(The sigatexit() handler list code now calls the handlers on the normal
exit path, so _exit(1) from the handler is wrong; exit_signal() modifies
toys.exitval if necessary.)

8 years agobasename: fix segfault on null input; add tests
Samuel Holland [Sun, 20 Mar 2016 18:13:21 +0000 (11:13 -0700)]
basename: fix segfault on null input; add tests

When passed an empty string, glibc's basename() returns a pointer to the
string "." in read-only memory. If an empty suffix is given, it fits
the condition of being shorter than the path, so we try to overwrite the
null byte and crash. Fix this by just ignoring empty suffixes; they
don't do anything anyway.

8 years agoAdd od -w.
Elliott Hughes [Sat, 19 Mar 2016 03:12:04 +0000 (20:12 -0700)]
Add od -w.

I'm not an od user myself, but I copy & pasted an 'od' invocation from a bug
and it didn't work. Which was embarrassing because I'd just claimed that the
od user didn't need to run od on their desktop (as they had done) and could
have run it on the device instead...

8 years agotoolbox prlimit is no more.
Elliott Hughes [Mon, 7 Mar 2016 22:51:10 +0000 (14:51 -0800)]
toolbox prlimit is no more.

Replaced by toybox ulimit -p.

8 years agoDebian bug 635570 did something horribly nonstandard (depending on a side
Rob Landley [Thu, 24 Mar 2016 21:20:36 +0000 (16:20 -0500)]
Debian bug 635570 did something horribly nonstandard (depending on a side
effect of "sed -e 'a\'" with no trailing line). But there's an actual user,
and it's not hard to implement, and it's not hard to implement, and there
isn't an obvious _other_ way to do it, so add the behavior and a test for it.

Fix some bad/missing comments while I was there, and add a couple TODOs.

8 years agoFix bzcat.test (as noted by Andy Chu), wean tests off $TOPDIR and supply $FILES
Rob Landley [Wed, 23 Mar 2016 08:44:51 +0000 (03:44 -0500)]
Fix bzcat.test (as noted by Andy Chu), wean tests off $TOPDIR and supply $FILES
instead, move tests/blkid into tests/files/blkid.

8 years agoThe test infrastructure prints the command name now, so file.test shouldn't.
Rob Landley [Wed, 23 Mar 2016 08:30:23 +0000 (03:30 -0500)]
The test infrastructure prints the command name now, so file.test shouldn't.

8 years agoMinor blkid cleanup. (There was a while(ptr[-1]==' ') ptr--; that could fall
Rob Landley [Wed, 23 Mar 2016 08:27:39 +0000 (03:27 -0500)]
Minor blkid cleanup. (There was a while(ptr[-1]==' ') ptr--; that could fall
off the start of the string. I pulled on the thread...)

8 years agoRedo build stuff in response to Andy Chu's suggestions.
Rob Landley [Wed, 23 Mar 2016 08:25:37 +0000 (03:25 -0500)]
Redo build stuff in response to Andy Chu's suggestions.

Toybox single binaries are now made directly with the new name instead of
stomping toybox and getting renamed.

Unstripped files now live in generated/unstripped.

Target to run all tests is now "make tests" to avoid conflict with "make test".

.singleconfig now has .PHONY: entries for all test_$NAME targets.

Default install location changed to /usr/toybox, code.html now says how
to change it (set $PREFIX).

scripts/install.sh --uninstall works now. (And you get to keep the pieces!)

8 years agoblkid: Handle short/empty vfat labels; update tests
Samuel Holland [Sun, 20 Mar 2016 18:13:22 +0000 (11:13 -0700)]
blkid: Handle short/empty vfat labels; update tests

vfat labels have a fixed space allocated for them. In the case of a
shorter label, the remaining bytes are padded with spaces. A vfat
filesystem with no label (i.e. that will show up in Windows as "Local
Disk (X:)" or "Removable Disk (X:)") is stored as "NO NAME    ".

Both of these changes match behavior from util-linux.

8 years agoPosix requires "" to be false, and 00 should be false, and -0 should be false.
Rob Landley [Mon, 21 Mar 2016 01:39:27 +0000 (20:39 -0500)]
Posix requires "" to be false, and 00 should be false, and -0 should be false.
Fix memory leaks in regex. Simplify regex check (if we matched, subexpression
had to match). Don't exit() from a command, set exitval and return.

8 years agoMake lsof 10x faster by caching /proc/net socket info.
Elliott Hughes [Sat, 19 Mar 2016 15:56:09 +0000 (08:56 -0700)]
Make lsof 10x faster by caching /proc/net socket info.

I deliberately didn't do this first time round because for me our lsof
is already 10x faster than traditional lsof, and caching means we potentially
give less information about a socket that's created while we're running.
It turns out that traditional lsof caches anyway, so I guess nobody cares.

This also fixes a mistake where lsof used CFG_FREE instead of CFG_TOYBOX_FREE.

8 years agoFix last commit (had newline before first ==> name <==) and add -f tests.
Rob Landley [Fri, 18 Mar 2016 19:12:53 +0000 (14:12 -0500)]
Fix last commit (had newline before first ==> name <==) and add -f tests.

8 years agoFix "tail -f single_file".
Josh Gao [Wed, 16 Mar 2016 22:41:13 +0000 (15:41 -0700)]
Fix "tail -f single_file".

TT.file_no was being incorrectly calculated as 0 when tail -f was passed
a single argument.

8 years agoexpr cleanup: simplify get_str() and free returned allocations.
Rob Landley [Thu, 17 Mar 2016 20:19:16 +0000 (15:19 -0500)]
expr cleanup: simplify get_str() and free returned allocations.

8 years agoRemove 'exit' from xargs test too.
Andy Chu [Thu, 17 Mar 2016 19:02:55 +0000 (12:02 -0700)]
Remove 'exit' from xargs test too.

8 years agoFix bug where all tests aren't being run with 'make test'.
Andy Chu [Thu, 17 Mar 2016 07:06:32 +0000 (00:06 -0700)]
Fix bug where all tests aren't being run with 'make test'.

The tests/*.test files shouldn't explicitly exit, because they are
sourced in scripts/test.sh.  No tests after sed were being run.

8 years agoexpr cleanup: traverse optargs with tok instead of incrementing optargs and
Rob Landley [Thu, 17 Mar 2016 02:55:19 +0000 (21:55 -0500)]
expr cleanup: traverse optargs with tok instead of incrementing optargs and
copying itinto a second variable. Also add error message for empty ( )

8 years agoMove whitespace and curly brackets around, move reference URLs to top of file,
Rob Landley [Wed, 16 Mar 2016 17:33:32 +0000 (12:33 -0500)]
Move whitespace and curly brackets around, move reference URLs to top of file,
eliminate syntax_error() function (just call error_exit() directly),
eliminate advance() wrapper, pipe expected error messages in test to /dev/null

8 years agoFix type coercion bugs in expr.
Andy Chu [Tue, 15 Mar 2016 20:52:25 +0000 (13:52 -0700)]
Fix type coercion bugs in expr.

All tests pass now; this fixes the 2 remaining failures, including a
segfault.

The structure of the code has changed a lot -- instead of having a tiny
function per operator, we have eval_op() which does common type coercion
and then evaluates the operator.  I tried writing it a couple different
ways, and this was the cleanest.

The OPS table now contains the operator string, precedence level,
signature for type coercion, and operator ID.