Pádraig Brady wrote: > I had noticed these tests were a little verbose and > had meant to simplify them. Coincidentally today > I triggered a race in tail-2/pid, so the attached > patch kills two birds with the one stone.
Good timing. I hit this today, too. >>From ba37fb2e96334b3cc784a4387d74f726be9be98d Mon Sep 17 00:00:00 2001 > From: =?utf-8?q?P=C3=A1draig=20Brady?= <p...@draigbrady.com> > Date: Thu, 3 Sep 2009 00:39:17 +0100 > Subject: [PATCH] tests: simplify and fix a race in 2 tail --follow tests > > * tests/tail-2/pid: Use the timeout command to determine process > longevity, rather than querying /proc/$pid/status. > That was racy in any case as I presume the test was copied slightly clearer: s/That/The latter/ FYI, tail -f with an unchanging file is different, now that it's based on inotify. Before, it really was always in the 'S' state. Now, it wakes up periodically. Hence this race. > from tail-2/tail-n0f and wasn't updated to handle the case > where the background process was in the R (running) state. > * tests/tail-2/wait: Likewise. > --- > tests/tail-2/pid | 37 +++------------------ > tests/tail-2/wait | 93 > ++++++++--------------------------------------------- > 2 files changed, 19 insertions(+), 111 deletions(-) > > diff --git a/tests/tail-2/pid b/tests/tail-2/pid Looks fine, and works for me. Thanks!
