Re: kern/44950: fcntl(F_MAXFD) returns bogus values

[Masao Uebayashi <uebayasi%gmail.com@localhost>]

The following reply was made to PR kern/44950; it has been noted by GNATS.

From: Masao Uebayashi <uebayasi%gmail.com@localhost>
To: Christos Zoulas <christos%zoulas.com@localhost>
Cc: Julio Merino <jmmv%julipedia.org@localhost>, 
gnats-bugs%netbsd.org@localhost, kern-bug-people%netbsd.org@localhost, 
        gnats-admin%netbsd.org@localhost, netbsd-bugs%netbsd.org@localhost, 
martin%netbsd.org@localhost
Subject: Re: kern/44950: fcntl(F_MAXFD) returns bogus values
Date: Wed, 11 May 2011 23:46:30 +0900

 On Wed, May 11, 2011 at 11:43 PM, Christos Zoulas 
<christos%zoulas.com@localhost> wro=
 te:
 > On May 11, 10:20am, jmmv%julipedia.org@localhost (Julio Merino) wrote:
 > -- Subject: Re: kern/44950: fcntl(F_MAXFD) returns bogus values
 >
 > | On 5/11/11 8:53 AM, Christos Zoulas wrote:
 > | >
 > | > Fix the test so that it runs a loop to close file descriptors from 0 =
 to the
 > | > descriptor soft limit
 > |
 > | I think atf-run should be doing this before execing the test program, i=
 f
 > | only for consistency reasons. =A0You can open a PR against me if you ag=
 ree
 > | so I don't forget!
 >
 > I am not sure which way to go there. Make the startup environment
 > more complicated in the name of consistency, or have tests that
 > care for a particular setup make sure that the setup they care
 > about is consistent?
 >
 > I.e. a file descriptor test should make sure that descriptors are
 > consistent, a resource limit test should do the same for resources,
 > etc. Otherwise the startup can end up being expensive, plus it
 > might be exercising the same functions that test is supposed to.
 >
 > And where does it stop? Do we reset process group, tty process
 > groups, current working directory, environment, signal handlers,
 > sysv ipc stuff, resource limits, timers, ...
 
 That's a place for RUMP too?