Path: csiph.com!goblin1!goblin.stu.neva.ru!usenet.stanford.edu!not-for-mail
From: Harald van Dijk <harald@gigawatt.nl>
Newsgroups: gnu.bash.bug
Subject: Re: Bizarre interaction bug involving bash w/ lastpipe + Almquist 'wait'
Date: Thu, 6 Feb 2020 19:29:41 +0000
Lines: 51
Approved: bug-bash@gnu.org
Message-ID: <mailman.390.1581022060.2412.bug-bash@gnu.org>
References: <10e3756b-5e8f-ba00-df0d-b36c93fa2281@inlv.org> <f8b210f5-dd59-2c7f-05d4-be0a89316d3d@gigawatt.nl>
NNTP-Posting-Host: lists.gnu.org
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: quoted-printable
To: Martijn Dekker <martijn@inlv.org>, DASH shell mailing list <dash@vger.kernel.org>, busybox <busybox@busybox.net>, Bug reports for the GNU Bourne Again SHell <bug-bash@gnu.org>, Robert Elz <kre@munnari.OZ.AU>, Jilles Tjoelker <jilles@stack.nl>
Envelope-to: bug-bash@gnu.org
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:73.0) Gecko/20100101 Thunderbird/73.0
In-Reply-To: <10e3756b-5e8f-ba00-df0d-b36c93fa2281@inlv.org>
Content-Language: en-US
Precedence: list
Xref: csiph.com gnu.bash.bug:15869

On 06/02/2020 16:12, Martijn Dekker wrote:
> This is probably the strangest bug (or maybe pair of bugs) I've run int=
o=20
> in nearly five years of breaking shells by developing modernish.
>=20
> I've traced it to an interaction between bash >=3D 4.2 (i.e.: bash with=
=20
> shopt -s lastpipe) and variants of the Almquist shell, at least: dash,=20
> gwsh, Busybox ash, FreeBSD sh, and NetBSD 9.0rc2 sh.
>=20
> Symptom: if 'return' is invoked on bash in the last element of a pipe=20
> executed in the main shell environment, then if you subsequently 'exec'=
=20
> an Almquist shell variant so that it has the same PID, its 'wait'=20
> builtin breaks.
>=20
> I can consistently reproduce this on Linux, macOS, FreeBSD, NetBSD=20
> 9.0rc2, OpenBSD, and Solaris.
>=20
> To reproduce this, you need bash >=3D 4.2, some Almquist shell variant,=
=20
> and these two test scripts:
>=20
> ---begin test.bash---
> fn() {
>  =C2=A0=C2=A0=C2=A0=C2=A0: | return
> }
> shopt -s lastpipe || exit
> fn
> exec "${1:-dash}" test.ash
> ---end test.bash---
>=20
> ---begin test.ash---
> echo '*ash-begin'
> : &
> echo '*ash-middle'
> wait "$!"
> echo '*ash-end'
> ---end test.ash---
>=20
> When executing test.bash with dash, gwsh, Busybox ash, or FreeBSD sh,=20
> then test.ash simply waits forever on executing 'wait "$!"'.

Nice test. bash leaves the process in a state where SIGCHLD is blocked,=20
and the various ash-based shells do not unblock it. Because of that,=20
they do not pick up on the fact that the child process has terminated. I=20
would consider this a bug both in bash and in the ash-based shells.

Cheers,
Harald van Dijk