DJ Lucas wrote:
Just a note that there is a race with inotify in tail (coreutils-8.24)
that sometimes causes tests/tail-2/assert.sh to fail.
New patch in repo coreutils-8.24-ulink_notification_race-1.patch that
fixes tail. Patch allows the test to pass as well. IIUC, it is hardly a
serious problem, but should be fixed either by fixing tail, or killing the
test (sed /assert.sh/d -i Makefile.in). Fixing tail is obviously
preferred. I'll get around to this and others I've mentioned recently in
the next day or two. Still looking into other suggestions made over the
past week.
I saw the patch when you put it into the repo and read the email chain
about it. Honestly, how often would this come about in practice? You
would have to do a 'tail -f fn' and then delete fn. I don't think I've
ever done a 'tail -f fn' although I've certainly done 'tail -f' to stdout.
The last release of coreutils was July. The one before that was July 2014
so they may only release once a year.
In any case this does not seem to be a security bug. Does is rise to the
level of needing to be in the book?
-- Bruce
--
http://lists.linuxfromscratch.org/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
