tree 7cf2172c26930b783a5e19106b35364c8964504d
parent 0a2cb815974ea96af664fa966079966a08916722
author Elliott Hughes <enh@google.com> 1360888432 -0800
committer Elliott Hughes <enh@google.com> 1360888558 -0800

Fix the stack protector death test.

Now __stack_chk_fail calls abort(3) directly, we terminate with
SIGSEGV rather than SIGABRT. (Because of the workaround for the
debuggerd lossage in the abort(3) implementation, which was the
motivation for switching __stack_chk_fail over to abort(3).)

Also clarify the comment on the weird pthread death test, so it
doesn't get copied and pasted onto real death tests.

Change-Id: Ie832eaded61359c99e7a10db65e28f35e8f63eed