Horton

Re: SIGPIPE in TAP tests

Noah Misch <noah@leadboat.com>

View thread View on postgr.es
From: Noah Misch <noah@leadboat.com>
To: Michael Paquier <michael.paquier@gmail.com>
Cc: PostgreSQL mailing lists <pgsql-hackers@postgresql.org>
Date: 2017-12-11T07:32:43Z
Lists: pgsql-hackers 
On Mon, Dec 11, 2017 at 04:19:52PM +0900, Michael Paquier wrote:
> On Sun, Dec 10, 2017 at 6:02 AM, Noah Misch <noah@leadboat.com> wrote:

> If SIGPIPE is ignored then test output just stops after generating the
> FATAL message. Oops.

You mean "If SIGPIPE is not ignored ...", right?

> > To fix the actual failures, we can cease sending "SELECT 1"; it's enough to
> > disconnect immediately.  Patch attached.
> 
> Perhaps you could use an empty string instead? I feel a bit uneasy
> about passing an undefined object to IPC::Run::run.

IPC::Run documents the equivalence of undef and '' in this context; search for
"close a child processes stdin" in
http://search.cpan.org/~rbs/IPC-Run-0.78/lib/IPC/Run.pm.  Thus, I expect both
spellings to work reliably, and I find "undef" slightly more evocative.

Commits

  1. Avoid and detect SIGPIPE race in TAP tests.