busybox/shell
Ron Yorston 4a36ef11ac ash: regressions in process substitution
Stacy Harper reports that this script:

   test() { . /tmp/bb_test; }
   echo "export TEST=foo" >/tmp/bb_test
   test 2>/dev/null
   echo "$TEST"

correctly prints 'foo' in BusyBox 1.33 but hangs in 1.34.

Bisection suggested the problem was caused by commit a1b0d3856 (ash: add
process substitution in bash-compatibility mode).  Removing the call to
unwindredir() in cmdloop() introduced in that commit makes the script
work again.

Additionally, these examples of process substitution:

   while true; do cat <(echo hi); done
   f() { while true; do cat <(echo hi); done }
   f

result in running out of file descriptors.  This is a regression from
v5 of the process substitution patch caused by changes to evalcommand()
not being transferred to v6.

function                                             old     new   delta
static.pushredir                                       -      99     +99
evalcommand                                         1729    1750     +21
exitreset                                             69      86     +17
cmdloop                                              372     365      -7
unwindredir                                           28       -     -28
pushredir                                            112       -    -112
------------------------------------------------------------------------------
(add/remove: 1/2 grow/shrink: 2/1 up/down: 137/-147)          Total: -10 bytes

Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2021-09-02 17:49:00 +02:00
..
ash_test shell: update psubst testcases 2021-07-27 04:20:32 +02:00
hush_test shell: update psubst testcases 2021-07-27 04:20:32 +02:00
ash_doc.txt
ash_ptr_hack.c
ash_remove_unnecessary_code_in_backquote_expansion.patch ash: save Ron's patch from oblivion 2021-06-06 13:01:25 +02:00
ash.c ash: regressions in process substitution 2021-09-02 17:49:00 +02:00
brace.txt
Config.src nologin: make it possible to build it as single applet 2020-06-24 15:05:22 +02:00
cttyhack.c config: update size information 2018-12-28 03:20:17 +01:00
hush_doc.txt
hush_leaktool.sh
hush.c hush: fix "error: invalid preprocessing directive ##" 2021-08-15 20:08:53 +02:00
Kbuild.src
match.c hush: optimize ${var/pattern/repl} for trivial patterns 2021-07-27 17:53:55 +02:00
match.h
math.c shell: remove FAST_FUNC from a static function 2020-10-01 00:57:05 +02:00
math.h shell: move all definitions of strto_arith_t() together 2019-05-26 14:02:10 +02:00
random.c whitespace fixes 2018-07-17 15:04:17 +02:00
random.h
README
README.job
shell_common.c shells: do not allow bare "read" in non-bash compat configs 2021-06-06 12:08:43 +02:00
shell_common.h hush: fix "export PS1=xyz" and "local PS1=xyz" messing up prompt 2019-05-14 18:56:04 +02:00

http://www.opengroup.org/onlinepubs/9699919799/
Open Group Base Specifications Issue 7


http://www.opengroup.org/onlinepubs/9699919799/utilities/V3_chap01.html
Shell & Utilities

It says that any of the standard utilities may be implemented
as a regular shell built-in. It gives a list of utilities which
are usually implemented that way (and some of them can only
be implemented as built-ins, like "alias"):

alias
bg
cd
command
false
fc
fg
getopts
jobs
kill
newgrp
pwd
read
true
umask
unalias
wait


http://www.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html
Shell Command Language

It says that shell must implement special built-ins. Special built-ins
differ from regular ones by the fact that variable assignments
done on special builtin are *PRESERVED*. That is,

VAR=VAL special_builtin; echo $VAR

should print VAL.

(Another distinction is that an error in special built-in should
abort the shell, but this is not such a critical difference,
and moreover, at least bash's "set" does not follow this rule,
which is even codified in autoconf configure logic now...)

List of special builtins:

. file
: [argument...]
break [n]
continue [n]
eval [argument...]
exec [command [argument...]]
exit [n]
export name[=word]...
export -p
readonly name[=word]...
readonly -p
return [n]
set [-abCefhmnuvx] [-o option] [argument...]
set [+abCefhmnuvx] [+o option] [argument...]
set -- [argument...]
set -o
set +o
shift [n]
times
trap n [condition...]
trap [action condition...]
unset [-fv] name...

In practice, no one uses this obscure feature - none of these builtins
gives any special reasons to play such dirty tricks.

However. This section also says that *function invocation* should act
similar to special built-in. That is, variable assignments
done on function invocation should be preserved after function invocation.

This is significant: it is not unthinkable to want to run a function
with some variables set to special values. But because of the above,
it does not work: variable will "leak" out of the function.