Re: 1.7.2 cycle will open soon

From: Jeff King <peff@peff.net>
Date: May 6, 2010, 05:52 UTC
Message-ID: <20100506055236.GA16151@coredump.intra.peff.net>
In-Reply-To: <7vaaselxe8.fsf@alter.siamese.dyndns.org>

On Tue, May 04, 2010 at 10:39:59PM -0700, Junio C Hamano wrote:

> It appears that the tip of 'maint' and 'next' (but curiously not 'master')
> do not pass test t1504 on at least FreeBSD right now, but I don't have
> enough energy to hunt things down tonight.

The culprit is Jonathan's 3bf7886 (test-lib: Let tests specify commands to be run at end of test, 2010-05-02). FreeBSD's /bin/sh doesn't propagate $? over an eval:

  $ /bin/sh
  $ false; echo $?
  1
  $ false; eval 'echo $?'
  0
  $ eval 'false; echo $?' ;# but it still works inside, of course
  1

Looking at that patch, I don't see any reason that eval_ret needs to be set inside the eval. If we have multiple test_when_finished calls, we keep setting and propagating eval_ret, which doesn't make much sense to me. Why not just:

  test_run_ () {
    test_cleanup=
    eval >&3 2>&4 "$1"
    eval_ret=$?
    eval >&3 2>&4 "$test_cleanup"
    return 0
  }

  test_when_finished () {
    test_cleanup="$*; $test_cleanup"
  }

Am I missing something?

-Peff