dash has accepted a patch to remove the first argument of findvar().
It's commit e85e972 (var: move hashvar() calls into findvar()).
Apply the same change to BusyBox ash.
function old new delta
findvar 35 40 +5
mklocal 268 261 -7
exportcmd 164 157 -7
setvareq 319 310 -9
lookupvar 150 141 -9
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 1/4 up/down: 5/-32) Total: -27 bytes
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
"echo text >&0000000000002" works as you would expect,
"echo text >&9999999999" properly fails instead of creating a file
named "9999999999".
function old new delta
expredir 219 232 +13
readtoken1 3045 3053 +8
parsefname 204 201 -3
isdigit_str9 45 - -45
------------------------------------------------------------------------------
(add/remove: 0/1 grow/shrink: 2/1 up/down: 21/-48) Total: -27 bytes
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
With very large fd#, the error code path is different
from one for closed but small fd#.
Make it not abort if we are interactive:
$ echo text >&99 # this wasn't buggy
ash: dup2(9,1): Bad file descriptor
$ echo text >&9999 # this was
ash: fcntl(1,F_DUPFD,10000): Invalid argument
function old new delta
.rodata 105637 105661 +24
dup2_or_raise 35 38 +3
redirect 1084 1044 -40
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 2/1 up/down: 27/-40) Total: -13 bytes
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Commit 549deab5a (ash: move parse-time quote flag detection to
run-time) did away with the need to distinguish between backquotes
inside and outside quotes. This left a gap among the control
characters used in argument strings. Removing this gap saves a
few bytes.
function old new delta
.rodata 167346 167338 -8
cmdputs 399 388 -11
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-19) Total: -19 bytes
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
An alias expansion immediately followed by '<' and a newline is
parsed incorrectly:
~ $ alias x='echo yo'
~ $ x<
yo
~ $
sh: syntax error: unexpected newline
The echo is executed and an error is printed on the next command
submission. In dash the echo isn't executed and the error is
reported immediately:
$ alias x='echo yo'
$ x<
dash: 3: Syntax error: newline unexpected
$
The difference between BusyBox and dash is that BusyBox supports
bash-style process substitution and output redirection. These
require checking for '<(', '>(' and '&>' in readtoken1().
In the case above, when the end of the alias is found, the '<' and
the following newline are both read to check for '<('. Since
there's no match both characters are pushed back.
The next input is obtained by reading the expansion of the alias.
Once this string is exhausted the next call to __pgetc() calls
preadbuffer() which pops the string, reverts to the previous input
and recursively calls __pgetc(). This request is satisified from
the pungetc buffer. But the first __pgetc() doesn't know this:
it sees the character has come from preadbuffer() so it (incorrectly)
updates the pungetc buffer.
Resolve the issue by moving the code to pop the string and fetch
the next character up from preadbuffer() into __pgetc().
function old new delta
pgetc 28 589 +561
__pgetc 607 - -607
------------------------------------------------------------------------------
(add/remove: 0/1 grow/shrink: 1/0 up/down: 561/-607) Total: -46 bytes
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
When I planned to print the command in read_line_input, I found that after
the system started, the command printed for the first time was always
garbled.
After analysis, it is found that in the init() function of ash, the
variable basepf.buf is not initialized after applying for memory, resulting
in garbled initial data. Then assign it to the global variable
g_parsefile->buf in ash.c, and then pass g_parsefile->buf to the parameter
command of the function read_line_input in the function preadfd(), and
finally cause it to be garbled when the command is printed by
read_line_input.
The call stack is as follows:
#0 read_line_input (st=0xb6fff220, prompt=0xb6ffc910 "\\[\\033[32m\\]\\h \\w\\[\\033[m\\] \\$ ", command=command@entry=0xb6ffc230 "P\325\377\266P\325\377\266", maxsize=maxsize@entry=1024) at libbb/lineedit.c:2461
#1 0x0043ef8c in preadfd () at shell/ash.c:10812
#2 preadbuffer () at shell/ash.c:10914
#3 pgetc () at shell/ash.c:10997
#4 0x00440c20 in pgetc_eatbnl () at shell/ash.c:11039
#5 0x00440cbc in xxreadtoken () at shell/ash.c:13157
#6 0x00440f40 in readtoken () at shell/ash.c:13268
#7 0x00441234 in list (nlflag=nlflag@entry=1) at shell/ash.c:11782
#8 0x004420e8 in parsecmd (interact=<optimized out>) at shell/ash.c:13344
#9 0x00442c34 in cmdloop (top=top@entry=1) at shell/ash.c:13549
#10 0x00444e4c in ash_main (argc=<optimized out>, argv=0x444e4c <ash_main+1328>) at shell/ash.c:14747
#11 0x00407954 in run_applet_no_and_exit (applet_no=9, name=<optimized out>, argv=0xbefffd34) at libbb/appletlib.c:1024
#12 0x00407b68 in run_applet_and_exit (name=0xbefffe56 "ash", argv=0x9) at libbb/appletlib.c:1047
#13 0x00407f88 in main (argc=<optimized out>, argv=0xbefffd34) at libbb/appletlib.c:1181
Fixes: 82dd14a510 ("ash: use CONFIG_FEATURE_EDITING_MAX_LEN")
Signed-off-by: zhuyan <zhuyan34@huawei.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Has a few annoying problems:
* sleepcmd() -> sleep_main(), the parsing of bad arguments exits the shell.
* sleep_for_duration() in sleep_main() has to be interruptible for
^C traps to work, which may be a problem for other users
of sleep_for_duration().
* BUT, if sleep_for_duration() is interruptible, then SIGCHLD interrupts it
as well (try "/bin/sleep 1 & sleep 10").
* sleep_main() must not allocate anything as ^C in ash longjmp's.
(currently, allocations are only on error paths, in message printing).
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Commit daa66ed6 fixed a number of use-after-free bugs in bash pattern
substitution, however one "unguarded" STPUTC remained, which is fixed here.
function old new delta
subevalvar 1564 1576 +12
Signed-off-by: Karsten Sperling <ksperling@apple.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Since commit 9e2a5668f (ash,hush: allow builtins to be tab-completed,
closes 7532) ash and hush have supported tab completion of builtins.
Other shells, bash and ksh for example, also support tab completion
of functions and aliases.
Add such support to ash and hush.
function old new delta
ash_command_name - 92 +92
hush_command_name - 63 +63
ash_builtin_name 17 - -17
hush_builtin_name 38 - -38
------------------------------------------------------------------------------
(add/remove: 2/2 grow/shrink: 0/0 up/down: 169/-55) Total: 100 bytes
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Avi Halachmi <avihpit@yahoo.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
We can just clear mailtime_hash to zero and have the same effect.
function old new delta
changemail 8 11 +3
mail_var_path_changed 1 - -1
cmdloop 398 382 -16
------------------------------------------------------------------------------
(add/remove: 0/1 grow/shrink: 1/1 up/down: 3/-17) Total: -14 bytes
text data bss dec hex filename
1054786 559 5020 1060365 102e0d busybox_old
1054773 559 5020 1060352 102e00 busybox_unstripped
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
padvance() exit condition is return value < 0, not == 0.
After MAIL changing twice, the logic erroneously
concluded that "you have new mail".
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Mea culpa, in "Do not allocate stack string in padvance" commit
(I left an extraneous "break" statement).
function old new delta
cmdloop 329 398 +69
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
The trap and jobs builtins can be used to report information about
traps and jobs. This works when they're called from the current
shell but in a child shell the required information is usually
cleared. Special hacks allow:
- trap to work with command substitution;
- jobs to work with command substitution or in a pipeline.
Neither works with process substitution.
- Relax the test for the trap hack so it also supports pipelines.
- Pass the command to be evaluated to forkshell() in evalbackcmd()
so trap and jobs both work with process substitution.
function old new delta
forkchild 629 640 +11
argstr 1502 1496 -6
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 1/1 up/down: 11/-6) Total: 5 bytes
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Commit 1d37186fe2 (ash: add bash-compatible EPOCH variables) added
support for the EPOCHSECONDS and EPOCHREALTIME variables.
These variables are dynamic and therefore require the VDYNAMIC flag
to be non-zero. However, this is only the case if support for the
RANDOM variable is enabled.
Give VDYNAMIC a non-zero value if either EPOCH variables or RANDOM
are enabled.
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Patch by soeren@soeren-tempel.net
The idx variable points to a value in the stack string (as managed
by STPUTC). STPUTC may resize this stack string via realloc(3). If
this happens, the idx pointer needs to be updated. Otherwise,
dereferencing idx may result in a use-after free.
function old new delta
subevalvar 1562 1566 +4
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
ash and hush correctly use the value of HOME for tilde expansion.
However the line editing code in libbb obtains the user's home
directory by calling getpwuid(). Thus tildes in tab completion
and prompts may be interpreted differently than in tilde expansion.
When the line editing code is invoked from a shell make it use the
shell's interpretation of tilde. This is similar to how GNU readline
and bash collaborate.
function old new delta
get_homedir_or_NULL 29 72 +43
optschanged 119 126 +7
hush_main 1204 1211 +7
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 3/0 up/down: 57/0) Total: 57 bytes
v2: Always check for HOME before trying the password database: this
is what GNU readline does.
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Do not skip over "*p = c;" statement.
Testcase: echo ~~nouser/qwe
function old new delta
argstr 1396 1406 +10
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Without this patch, BusyBox handles bash pattern substitutions without
a terminating '/' character incorrectly.
Consider the following shell script:
_bootstrapver=5.0.211-r0
_referencesdir="/usr/${_bootstrapver/-*}/Sources"
echo $_referencesdir
This should output `/usr/5.0.211/Sources`. However, without this patch
it instead outputs `/usr/5.0.211Sources`. This is due to the fact that
BusyBox expects the bash pattern substitutions to always be terminated
with a '/' (at least in this part of subvareval) and thus reads passed
the substitution itself and consumes the '/' character which is part of
the literal string. If there is no '/' after the substitution then
BusyBox might perform an out-of-bounds read under certain circumstances.
When replacing the bash pattern substitution with `${_bootstrapver/-*/}`,
or with this patch applied, ash outputs the correct value.
Signed-off-by: Sören Tempel <soeren@soeren-tempel.net>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
The result of looking at "grep -F -B2 '*fill*' busybox_unstripped.map"
function old new delta
.rodata 108586 108460 -126
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-126) Total: -126 bytes
text data bss dec hex filename
970412 4219 1848 976479 ee65f busybox_old
970286 4219 1848 976353 ee5e1 busybox_unstripped
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Even though formally it is -s [ARGS], "sh -s" without ARGS
is the same as just "sh". And we are already over 80 chars wide
for ash --help, so make it shorter.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
- This can act as memory barrier in clang to avoid
read before assign of a const ptr
Signed-off-by: LoveSy <shana@zju.edu.cn>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
This trivial patch makes ${s:...} at least as fast as ${s#??..}
in simple tests. It's probably faster for longer substrings,
but then one wouldn't use ${s#"1024???s"} anyway -
one would switch away from sh.
function old new delta
subevalvar 1457 1503 +46
Signed-off-by: Alin Mr <almr.oss@outlook.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
While at it, change all "__asm__" to "asm"
Co-authored-by: canyie <31466456+canyie@users.noreply.github.com>
Signed-off-by: YU Jincheng <shana@zju.edu.cn>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Ron Yorston
Denys Vlasenko
zhuyan
Karsten Sperling
Shawn Landden
Sören Tempel
YU Jincheng
Alin Mr