[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-dejagnu] Leaked fd into spawned children
From: |
Jan Kratochvil |
Subject: |
Re: [Bug-dejagnu] Leaked fd into spawned children |
Date: |
Sat, 31 Aug 2013 08:36:22 +0200 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
On Sat, 31 Aug 2013 07:45:52 +0200, Ben Elliston wrote:
> > Actual results:
> > lrwx------ 1 jkratoch jkratoch 64 Aug 26 21:15 0 -> /dev/pts/5
> > lrwx------ 1 jkratoch jkratoch 64 Aug 26 21:15 1 -> /dev/pts/5
> > lrwx------ 1 jkratoch jkratoch 64 Aug 26 21:15 2 -> /dev/pts/5
> > lrwx------ 1 jkratoch jkratoch 64 Aug 26 21:15 3 -> /dev/pts/2
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > lr-x------ 1 jkratoch jkratoch 64 Aug 26 21:15 4 -> /proc/25522/fd
>
> Do you know what fd 3 is, in this case?
Not sure if you ask the easy part but it is tty of runtest, the user's tty.
fds 0/1/2 are pty created by expect.
> I've tried a few things with Expect and a Dejagnu .exp script that doesn't
> use remote_exec and can't reproduce it there.
IIRC I GDBed it a bit a fd 3 was created from Tcl bytecode. And as I do not
have effective Tcl debugging knowledge I just filed this bug.
Thanks,
Jan