grep: fix echo "aa" | busybox grep -F -w "a" (should not match)

function                                             old     new   delta
grep_file                                           1461    1470      +9

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ash,hush: fix "saved" redirected fds still visible in children

Based on a patch by Mark Marshall <mark.marshall@omicronenergy.com>

function                                             old     new   delta
dup_CLOEXEC                                            -      49     +49
fcntl_F_DUPFD                                         46       -     -46

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ssl_client: fix option parsing

The wrong character was used to indicate options taking an integer
parameter.

Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

tcpsvd: fix fallout from opt_complementary removal

   text    data     bss     dec     hex filename
 933035     473    6836  940344   e5938 busybox_old
 933051     473    6836  940360   e5948 busybox_unstripped

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

udhcpd: clamp down huge auto_times to ~2M seconds, better EINTR poll handling

EINTR _should_ only happen on two signals we trap, and safe_poll
_should_ work here just fine, but there were kernel bugs where spurious EINTRs
happen (e.g. on ptrace attach). Be safe.

function                                             old     new   delta
udhcpd_main                                         1437    1468     +31

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

udhcpd: fix "not dying on SIGTERM"

Fixes:
commit 52a515d18724bbb34e3ccbbb0218efcc4eccc0a8
"udhcp: use poll() instead of select()"
Feb 16 2017

udhcp_sp_read() is meant to check whether signal pipe indeed has some data to read.
In the above commit, it was changed as follows:

- if (!FD_ISSET(signal_pipe.rd, rfds))
+ if (!pfds[0].revents)
return 0;

The problem is, the check was working for select() purely by accident.
Caught signal interrupts select()/poll() syscalls, they return with EINTR
(regardless of SA_RESTART flag in sigaction). _Then_ signal handler is invoked.
IOW: they can't see any changes to fd state caused by signal haldler
(in our case, signal handler makes signal pipe ready to be read).

For select(), it means that rfds[] bit array is unmodified, bit of signal
pipe's read fd is still set, and the above check "works": it thinks select()
says there is data to read.

This accident does not work for poll(): .revents stays clear, and we do not
try reading signal pipe as we should. In udhcpd, we fall through and block
in socket read. Further SIGTERM signals simply cause socket read to be
interrupted and then restarted (since SIGTERM handler has SA_RESTART=1).

Fixing this as follows: remove the check altogether. Set signal pipe read fd
to nonblocking mode. Always read it in udhcp_sp_read().
If read fails, assume it's EAGAIN and return 0 ("no signal seen").

udhcpd avoids reading signal pipe on every recvd packet by looping if EINTR
(using safe_poll()) - thus ensuring we have correct .revents for all fds -
and calling udhcp_sp_read() only if pfds[0].revents!=0.

udhcpc performs much fewer reads (typically it sleeps >99.999% of the time),
there is no need to optimize it: can call udhcp_sp_read() after each poll
unconditionally.

To robustify socket reads, unconditionally set pfds[1].revents=0
in udhcp_sp_fd_set() (which is before poll), and check it before reading
network socket in udhcpd.

TODO:
This might still fail: if pfds[1].revents=POLLIN, socket read may still block.
There are rare cases when select/poll indicates that data can be read,
but then actual read still blocks (one such case is UDP packets with
wrong checksum). General advise is, if you use a poll/select loop,
keep all your fds nonblocking.
Maybe we should also do that to our network sockets?

function                                             old     new   delta
udhcp_sp_setup                                        55      65     +10
udhcp_sp_fd_set                                       54      60      +6
udhcp_sp_read                                         46      36     -10
udhcpd_main                                         1451    1437     -14
udhcpc_main                                         2723    2708     -15
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 2/3 up/down: 16/-39)            Total: -23 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

tar,unzip: postpone creation of symlinks with "suspicious" targets

This mostly reverts commit bc9bbeb2b81001e8731cd2ae501c8fccc8d87cc7
"libarchive: do not extract unsafe symlinks unless $EXTRACT_UNSAFE_SYMLINKS=1"

Users report that it is somewhat too restrictive. See
https://bugs.busybox.net/show_bug.cgi?id=8411

In particular, this interferes with unpacking of busybox-based
filesystems with links like "sbin/applet" -> "../bin/busybox".

The change is made smaller by deleting ARCHIVE_EXTRACT_QUIET flag -
it is unused since 2010, and removing conditionals on it
allows commonalizing some error message codes.

function                                             old     new   delta
create_or_remember_symlink                             -      94     +94
create_symlinks_from_list                              -      64     +64
tar_main                                            1002    1006      +4
unzip_main                                          2732    2724      -8
data_extract_all                                     984     891     -93
unsafe_symlink_target                                147       -    -147
------------------------------------------------------------------------------
(add/remove: 2/1 grow/shrink: 1/2 up/down: 162/-248)          Total: -86 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

Bump version to 1.28.1

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

cp: fix option handling in non-longopt config

the patch getopt32: remove opt_complementary 22542eca18e5807b72ddc78999f5101e33f17a53

introduced a regressed in the cp command since it removed all aliases of arguments
if long_opts is not configured.

Patch by Sebastian Gottschall <s.gottschall@dd-wrt.com>

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

svlogd: improve --help text

function                                             old     new   delta
packed_usage                                       32278   32367     +89

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ip: fix crash in "ip neigh show"

parse_rtattr() was using tb[] array without initializing it.

Based on patch by Balaji Punnuru <balaji_punnuru@cable.comcast.com>

function                                             old     new   delta
parse_rtattr                                          85     107     +22
print_route                                         1630    1617     -13
print_linkinfo                                       807     794     -13
iproute_get                                          835     822     -13
print_rule                                           680     665     -15
ll_remember_index                                    263     248     -15
print_addrinfo                                      1223    1197     -26
ipaddr_list_or_flush                                1253    1223     -30
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 1/7 up/down: 22/-125)          Total: -103 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

tls: fix hash calculations if client cert is requested and sent

Symptoms: connecting to
    openssl s_server -cert vsftpd.pem -port 990 -debug -cipher AES128-SHA
works, but with "-verify 1" option added it does not.

function                                             old     new   delta
tls_xread_record                                     474     499     +25
tls_handshake                                       1582    1607     +25
bad_record_die                                        98     110     +12
tls_run_copy_loop                                    282     293     +11
tls_xread_handshake_block                             58      51      -7
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 4/1 up/down: 73/-7)              Total: 66 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

cp: implement -T

Implement "cp -T". Some Linux kernel Makefiles started using this recently,
so allow also building on systems using busybox cp.

function                                             old     new   delta
cp_main                                              360     428     +68
copy_file                                           1678    1676      -2
packed_usage                                       32290   32259     -31
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 1/2 up/down: 76/-39)             Total: 35 bytes

Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

tar: accomodate non-terminated tar.chksum fields as seen from github.com

function                                             old     new   delta
get_header_tar                                      1783    1696     -87

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

udhcpc6: fix ipv6prefix[_lease] envvar value in script invocation

Based on a patch by DannyAAM <danny@saru.moe>.

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

umount: ignore -c

"-c, --no-canonicalize: Do not canonicalize paths."

As busybox doesn't canonicalize paths in the first place it is safe to ignore
this option.

See https://github.com/systemd/systemd/issues/7786

Signed-off-by: Shawn Landden <slandden@gmail.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ash: fail if 'shift' operand is out of range

If the numeric argument passed to ash's 'shift' built-in is greater than
'$#' the command performs no operation and exits successfully. It should
return a non-zero exit code instead:

http://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html#shift

This is consistent with bash and hush.

function                                             old     new   delta
shiftcmd                                             122     120      -2

Signed-off-by: Ingo van Lil <inguin@gmx.de>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

Bump version to 1.28.0

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

scripts/randomtest: do not try building static libbysubox

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

chown: fix a mistake in opt_complementary change

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

randomconfig fixes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ntpd: do run the script at leat once in 11 minutes

function                                             old     new   delta
ntpd_main                                           1197    1226     +29

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

env: -u option fails due to typo

The -u option is supposed to be allowed to appear multiple times; the
option string supplied to getopt32long requires it to be followed by a
nonnegative integer.

Reported-by: Keith Maxwell <keith.maxwell@gmail.com>
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

tar: code shrink

function                                             old     new   delta
packed_usage                                       31863   31853     -10
tar_main                                            1013    1002     -11
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-21)             Total: -21 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

tar: code shrink, better help text

function                                             old     new   delta
tar_main                                             994    1013     +19
packed_usage                                       31893   31863     -30
writeTarFile                                         250     207     -43
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 1/2 up/down: 19/-73)            Total: -54 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

tar: improve help text

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

inetd: fix for running by non-root

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

unshare: -r should map root to user, not the other way around

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

unshare: -r implies -U, not -u

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

unzip: add missing -j to trivial usage

Signed-off-by: Eugene Rudoy <gene.devel@gmail.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

unzip: fix content listing and filtering when -j is used

Original Info-ZIP's unzip uses unstripped filenames
while doing content listing and filtering, i.e.
 - in content listing mode -j is ignored completely
 - filtering is applied to non-stripped names, -j
   takes effect first while extracting the files

997ad2c64abbe931dffa3598b015c5de04e515cf strips path
components a little bit too early resulting in behavior
deviations.

Fix it by doing stripping after listing/filtering.

p.s. Info-ZIP's unzip behavior is the same as
     that of tar in --strip-components=NUM mode

Signed-off-by: Eugene Rudoy <gene.devel@gmail.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

lineedit: get terminal width before printing prompt

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

lineedit: do not tab-complete any strings which have control characters

function                                             old     new   delta
add_match                                             41      68     +27

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ash: fix nofork bug where environment is not properly passed to a command

function                                             old     new   delta
listvars                                             144     252    +108
evalcommand                                         1500    1546     +46
showvars                                             142     147      +5
shellexec                                            242     245      +3
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 4/0 up/down: 162/0)             Total: 162 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

init: reduce the window when init can lose reboot/poweroff signals

function                                             old     new   delta
init_main                                            695     712     +17

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ash: retain envvars with bad names in initial environment. Closes 10231

Reworks "ash: [VAR] Sanitise environment variable names on entry"
commit.

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ntpd: skip over setting next DNS resolution attempt if it is not needed

function                                             old     new   delta
ntpd_main                                           1177    1197     +20
resolve_peer_hostname                                127     129      +2

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

Fix build failures if MAXHOSTNAMELEN or MAXPATHLEN is not defined

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

grep: skip grepping symlinks to directories

When grep is passed -r, recursive_action will treat any symlinks to
directories not in the root as normal files, since it lstat's them and
is therefore told they are not directories. However, file_action_grep
will still try to fopen and read from them to see whether they match,
which varies in behaviour across platforms. Linux will give EISDIR and
thus grep will not find any matching lines, but FreeBSD will give the
raw contents of the directory itself, which may match the given pattern.
Also, if grep is passed -c, it will even print a count for these
symlinks, even on Linux.

Since this recursive_action behaviour is required for the correct
functioning of other applets, such as tar, grep should handle this
special case and skip any such symlinks.

function                                             old     new   delta
file_action_grep                                      80     161     +81

Signed-off-by: James Clarke <jrtc27@jrtc27.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ntpd: mention in help text that -d can be repeated

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ntpd: improve treatment of DNS resolution failures

function                                             old     new   delta
ntpd_main                                           1106    1177     +71
resolve_peer_hostname                                122     127      +5
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 2/0 up/down: 76/0)               Total: 76 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

udp_io, traceroute: Standardise IPv6 PKTINFO handling to be portable

The current standard (RFC 3542) is for IPV6_RECVPKTINFO to be given to
setsockopt, and IPV6_PKTINFO to be used as the packet type. Previously,
RFC 2292 required IPV6_PKTINFO to be used for both, but RFC 3542
re-purposed IPV6_PKTINFO when given to setsockopt. The special
Linux-specific IPV6_2292PKTINFO has the same semantics as IPV6_PKTINFO
in RFC 2292, but was introduced at the same time as IPV6_RECVPKTINFO.

Therefore, if we have IPV6_RECVPKTINFO available, we can use the RFC
3542 style, and if not, we assume that only the RFC 2292 API is
available, using IPV6_PKTINFO for both.

Signed-off-by: James Clarke <jrtc27@jrtc27.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

df: Use statvfs instead of non-standard statfs

Platforms differ on what their implementations of statfs include.
Importantly, FreeBSD's does not include a f_frsize member inside struct
statfs. However, statvfs is specified by POSIX and includes everything
we need, so we can just use that instead.

Signed-off-by: James Clarke <jrtc27@jrtc27.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

xfuncs: Handle missing non-POSIX termios constants

Signed-off-by: James Clarke <jrtc27@jrtc27.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

blkdiscard: Only build on Linux

Signed-off-by: James Clarke <jrtc27@jrtc27.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

networking: Fall back on IPPROTO_RAW when SOL_RAW is not defined

Signed-off-by: James Clarke <jrtc27@jrtc27.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ntpd: perform DNS resolution out of send/receive loop - closes 10466

Bad case: send request to server1good.com; then try to resolve server2bad.com -
this fails, and failure takes ~5 secs; then receive server1's
response 5 seconds later. We'll never sync up in this case...

function                                             old     new   delta
ntpd_main                                           1079    1106     +27

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

iproute/iprule: support toolchains without RTA_TABLE routing attribute

 iproute.c: In function 'print_route':
 iproute.c:85:9: error: 'RTA_TABLE' undeclared (first use in this function)
 iproute.c:85:9: note: each undeclared identifier is reported only once for each function it appears in
 iproute.c: In function 'iproute_modify':
 iproute.c:467:36: error: 'RTA_TABLE' undeclared (first use in this function)

Fix it by partially #ifdef'ing the code added in b42107f21538e39d9a344376372f8261aed589b2

Signed-off-by: Eugene Rudoy <gene.devel@gmail.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

time: fix build for toolchains without O_CLOEXEC

Based on patch by Eugene Rudoy <gene.devel@gmail.com>

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

unlzma: fix SEGV, closes 10436

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

bunzip2: fix runCnt overflow from bug 10431

This particular corrupted file can be dealth with by using "unsigned".
If there will be cases where it genuinely overflows, there is a disabled
code to deal with that too.

function                                             old     new   delta
get_next_block                                      1678    1667     -11

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hush: fix comment parsing in `cmd`, closes 10421

function                                             old     new   delta
parse_stream                                        2692    2690      -2

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

lsscsi: fix xchdir("..") from symlink in /sys/bus/scsi/devices

Signed-off-by: Markus Gothe <nietzsche@lysator.liu.se>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

whitespace fixes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

typo fix

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ash: survive failures in $PS1 expansion. Closes 10371

function                                             old     new   delta
expandstr                                            120     209     +89

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

lineedit: improve multiline PS1 - redraw using last PS1 line. Closes 10381

This patch only affects prompts with newlines.

We redraw the prompt [+ input] occasionally, e.g. during tab completion,
history browsing or search, etc, and we expect it to align with prior
redraws, such that the visible effect is that only the input changes.

With multi-line PS1, redraw always printed the prompt some lines below
the old one, which resulted in terminal scroll during every redraw.

Now we only redraw the last PS1 line, so vertical alignment is easier to
manage (we already calculated it using only the last line, but re-drew
all lines - that was the culprit), which fixes those extra scrolls.

Notes:
- We now use the full prompt for the initial draw, after clear-screen (^L),
  and after tab-completion choices are displayed. Everything else now
  redraws using the last/sole prompt line.

- During terminal resize we now only redraw the last[/sole] prompt line,
  which is arguably better because it's hard to do right (and we never did).

- Good side effect for reverse-i-search: its prompt now replaces only the
  last line of the original prompt - like other shells do.

function                                             old     new   delta
put_prompt_custom                                      -      66     +66
draw_custom                                            -      66     +66
parse_and_put_prompt                                 766     806     +40
read_line_input                                     3867    3884     +17
input_tab                                           1069    1076      +7
cmdedit_setwidth                                      61      63      +2
redraw                                                59      47     -12
put_prompt                                            46       -     -46
------------------------------------------------------------------------------
(add/remove: 2/1 grow/shrink: 4/1 up/down: 198/-58)           Total: 140 bytes

Signed-off-by: Avi Halachmi <avihpit@yahoo.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

sendmail allow "=" symbol in recipient, closes 10241

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

whitespace and comment format fixes, no code changes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

whitespace and comment format fixes, no code changes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

do not include <sys/stat.h> just before "libbb.h", it's there already

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

whitespace and comment format fixes, no code changes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

free: no longer include common_bufsiz.h

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

free: make it NOFORK

function                                             old     new   delta
parse_cached_kb                                        -      85     +85
free_main                                            580     537     -43
------------------------------------------------------------------------------
(add/remove: 1/0 grow/shrink: 0/1 up/down: 85/-43)             Total: 42 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

date: maybe_set_utc only once

It's not clear at all why we do it twice. git archaeology did not help.

function                                             old     new   delta
date_main                                            836     839      +3
maybe_set_utc                                         16       -     -16

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

i2cset: fix 'i' & 's' modes (3 extra bytes were sent)

When 'i' or 's' modes are selected block array is filled from offset 3
(blen = 3) but copied to data.block buffer from offset 0 so first 3 bytes
contains garbage from stack.
The buffer that is sent is also 3 bytes too long due to those extra 3
bytes.

Signed-off-by: Avi Fishman <AviFishman70@gmail.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

use %m printf specifier where appropriate

function                                             old     new   delta
out                                                   85      75     -10
udhcpd_main                                         1472    1461     -11
open_stdio_to_tty                                     98      85     -13
init_exec                                            245     232     -13
udhcpc_main                                         2763    2749     -14
do_cmd                                              4771    4755     -16
status_line_bold_errno                                32      14     -18
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 0/7 up/down: 0/-95)             Total: -95 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

udhcpc[6]: make log2 messages for chaddr field indented like the rest

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

udhcpc: paranoia when using kernel UDP mode for sending renew: server ID may be bogus

With new code, we request that target IP (server ID) must be directly reachable.
If it's not, this happens:

udhcpc: waiting 2000 seconds
udhcpc: entering listen mode: kernel
udhcpc: opening listen socket on *:68 wlan0
udhcpc: entering renew state
udhcpc: sending renew to 1.1.1.1
udhcpc: send: Network is unreachable
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 1.1.1.1 needs routing, this is fishy!
udhcpc: entering rebinding state
udhcpc: entering listen mode: raw
udhcpc: created raw socket
udhcpc: sending renew to 0.0.0.0
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ going to use broadcast

which is the desired behavior. Before the patch, packet to 1.1.1.1 was routed
over eth0 (!) and maybe even into Internet (!!!).

function                                             old     new   delta
udhcpc_main                                         2752    2763     +11
udhcp_send_kernel_packet                             295     301      +6
send_renew                                            82      84      +2
send_packet                                          166     168      +2
bcast_or_ucast                                        23      25      +2
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 5/0 up/down: 23/0)               Total: 23 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

udhcpc[6]: initialize entire sockaddr_ll

I see random field values like sll_hatype=0x267 when I strace.
They seem to not matter, but just in case they sometimes do,
let's at least have deterministic values (via memset(0)).

function                                             old     new   delta
change_listen_mode                                   308     322     +14

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

regularize format of source file headers, no code changes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

ip,ip*: make them NOEXEC

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

update NOFORK_NOEXEC.lst

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

httpd: fix handling of range requests

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

stty: add cmspar, flusho, extproc attributes

function                                             old     new   delta
mode_info                                            688     704     +16
mode_name                                            502     516     +14
wrapf                                                162     157      -5
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 2/1 up/down: 30/-5)              Total: 25 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

netcat: net applet (alias to nc)

function                                             old     new   delta
packed_usage                                       31807   31856     +49
applet_names                                        2701    2708      +7
applet_main                                         1560    1564      +4
applet_install_loc                                   195     196      +1
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 4/0 up/down: 61/0)               Total: 61 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

Clarify OPOST bit meaning

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

less,microcom,lineedit: use common routine to set raw termios

function                                             old     new   delta
get_termios_and_make_raw                               -     139    +139
xget1                                                 39       8     -31
read_line_input                                     3912    3867     -45
less_main                                           2525    2471     -54
set_termios_to_raw                                   116      36     -80
------------------------------------------------------------------------------
(add/remove: 1/0 grow/shrink: 0/4 up/down: 139/-210)          Total: -71 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexedit: code shrink

function                                             old     new   delta
hexedit_main                                        1171    1170      -1

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexedit: fixes to "goto address" code

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexedit: position in correct column on "goto" command

function                                             old     new   delta
hexedit_main                                        1169    1202     +33

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexedit: code shrink

function                                             old     new   delta
hexedit_main                                        1176    1169      -7

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexdump: code shrink

function                                             old     new   delta
remap                                                173     178      +5
move_mapping_lower                                   107     109      +2
move_mapping_further                                 141     143      +2
hexedit_main                                        1191    1176     -15
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 3/1 up/down: 9/-15)              Total: -6 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexedit: implement "[enter] goto offset" key

This is a must if you need to edit sector 123456789999 on your /dev/disk.

   text    data     bss     dec     hex filename
 922745     481    6832  930058   e310a busybox_old
 923023     481    6832  930336   e3220 busybox_unstripped

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexedit: restore screen on exit

function                                             old     new   delta
hexedit_main                                         998    1082     +84
restore_term                                           -      29     +29
remap                                                168     173      +5
------------------------------------------------------------------------------
(add/remove: 1/0 grow/shrink: 2/0 up/down: 118/0)             Total: 118 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexedit: optimize output buffering

function                                             old     new   delta
hexedit_main                                         970     998     +28

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexedit: implement page up/down

function                                             old     new   delta
hexedit_main                                         924     970     +46

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

*: use ESC define instead of "\033"; use ESC[m instead of ESC[0m

   text    data     bss     dec     hex filename
 922535     481    6832  929848   e3038 busybox_old
 922534     481    6832  929847   e3037 busybox_unstripped

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexedit: smarter redraw

function                                             old     new   delta
format_line                                          197     209     +12
hexedit_main                                         930     924      -6
redraw_cur_line                                      104      94     -10
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 1/2 up/down: 12/-16)             Total: -4 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

hexedit: new applet

function                                             old     new   delta
hexedit_main                                           -     930    +930
format_line                                            -     197    +197
remap                                                  -     168    +168
move_mapping_further                                   -     141    +141
move_mapping_lower                                     -     107    +107
redraw_cur_line                                        -     104    +104
packed_usage                                       31802   31812     +10
applet_names                                        2688    2696      +8
applet_main                                         1552    1556      +4
applet_suid                                           97      98      +1
applet_install_loc                                   194     195      +1
------------------------------------------------------------------------------
(add/remove: 7/0 grow/shrink: 5/0 up/down: 1671/0)           Total: 1671 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

networking/interface.c: remove superfluous "& 0377"

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

setfattr: new applet

function                                             old     new   delta
setfattr_main                                          -     189    +189
packed_usage                                       31516   31588     +72
setxattr                                               -      53     +53
lsetxattr                                              -      53     +53
removexattr                                            -      37     +37
lremovexattr                                           -      37     +37
applet_names                                        2649    2658      +9
find_applet_by_name                                  124     128      +4
applet_main                                         1532    1536      +4
------------------------------------------------------------------------------
(add/remove: 7/0 grow/shrink: 4/0 up/down: 458/0)             Total: 458 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

build system: fix a compiler warning

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

build system: fix "allnoconfig" to clear all options. Closes 10296

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

chat: do not die when HANGUP param is missing. Closes 10016

function                                             old     new   delta
chat_main                                           1356    1379     +23

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

httpd: skip "Status: " from CGI, including space. Closes 10291

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

klibc-utils: add ipconfig.c work-in-progress

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

xargs: fix a thinko in last commit

IF_FEATURE_XARGS_SUPPORT_PARALLEL(char *opt_a = NULL;)
is replaced with
IF_FEATURE_XARGS_SUPPORT_ARGS_FILE

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

xargs: support -a FILE

The GNU-specific option -a lets xargs read the arguments from a file
rather than from stdin.

This is particularly convenient when debugging in gdb interactively,
and it might be of more general use.

function                                             old     new   delta
xargs_main                                           788     823     +35
packed_usage                                       31683   31671     -12

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

klibc-utils: remove accidental "or later" licensing wording.

I (slightly) prefer my code to be under "GPLv2 only" license.

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>