Date: Tue, 6 Oct 2015 11:18:04 -0700 From: NGie Cooper <yaneurabeya@gmail.com> To: Eric van Gyzen <vangyzen@FreeBSD.org> Cc: Garrett Cooper <ngie@FreeBSD.org>, "freebsd-testing@freebsd.org" <freebsd-testing@FreeBSD.org> Subject: Re: svn commit: r288906 - head/bin/ls/tests Message-ID: <0218C81A-A727-4B8D-8F7A-074E488D989C@gmail.com> In-Reply-To: <5613DE88.10207@FreeBSD.org> References: <201510060055.t960tW2V068302@repo.freebsd.org> <5613DE88.10207@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> On Oct 6, 2015, at 07:45, Eric van Gyzen <vangyzen@FreeBSD.org> wrote: >=20 > On 10/05/2015 19:55, Garrett Cooper wrote: >> Author: ngie >> Date: Tue Oct 6 00:55:31 2015 >> New Revision: 288906 >> URL: https://svnweb.freebsd.org/changeset/base/288906 >>=20 >> Log: >> Explicitly set BLOCKSIZE to 512 in the environment >=20 > I know practically nothing about atf or our usage of it, but in = principle, it seems that the framework should scrub the whole = environment before running any tests, always. Agreed. I discovered after some digging that this variable is set by the = shell with login.conf, which affects getbsize, which in turn affects = some utilities like /bin/ls. It seems that I missed it when I was = testing in my environment, probable because I use bash (my guess is the = Jenkins test runner uses either csh or sh) :/=E2=80=A6 /bin/sh (and POSIX shells for that matter) do some interesting things in = subshells. See $SHELL as an example (I don=E2=80=99t have links right = offhand to prove it, but it made my life interesting before in the = near-ish past). Thanks!=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?0218C81A-A727-4B8D-8F7A-074E488D989C>