[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Problem on GNU/Hurd systems w.r.t. process substitution

From: Roland McGrath
Subject: Re: Problem on GNU/Hurd systems w.r.t. process substitution
Date: Sun, 19 Nov 2006 14:02:10 -0800 (PST)

It should be using /dev/fd.  As Chet said, figure out if the configure run
when building bash decided it wasn't there for some reason.  

Also, test that /dev/fd is working right on the Hurd, e.g.:

$ exec 5> foobar
$ ls -li /dev/fd/5 foobar
$ exec 5>&-

This won't pretend to be a symlink like Linux's /proc/self/fd (/dev/fd) does.
It will just resolve the lookup to the node, so you should see the same
info for /dev/fd/5 and foobar.

/dev/fd is implemented by /hurd/magic (trans/magic.c) and client-side magic
in glibc (same deal as /dev/tty).  If the translator is in place and
working right, then the problem may be on the glibc end.  But AFAIK nothing
in this area has changed in a very long time.

Of course, named pipes ought to work too and presumably bash uses them
correctly.  So you might also want to independently test using named pipes
by hand.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]