Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!elroy.jpl.nasa.gov!jpl-devvax!lwall
From: lwall@jpl-devvax.jpl.nasa.gov (Larry Wall)
Newsgroups: comp.lang.perl
Subject: Re: better version of chat2.pl
Message-ID: <1991Apr12.070917.3248@jpl-devvax.jpl.nasa.gov>
Date: 12 Apr 91 07:09:17 GMT
References: <1991Apr11.222343.28800@iwarp.intel.com>
Reply-To: lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall)
Organization: Jet Propulsion Laboratory, Pasadena, CA
Lines: 20

In article <1991Apr11.222343.28800@iwarp.intel.com> merlyn@iwarp.intel.com (Randal L. Schwartz) writes:
: Argh.  Had some problems with filehandles not being closed.  Hmm.
: Larry, is it a bug that a failed connect() leaves a filehandle open,
: and I have to close the handle (even though the close generates an
: error)?  Or is that just the way the underlying system works?  (To
: test this, I'd have to *gasp* code in C! :-)

If the OS doesn't close a socket on connect failure, neither should Perl.
(It doesn't.)  You're probably getting the failure on close because there
are two stdio streams bound to a single socket (stdio gets confused on
some machines otherwise).  Only the first fclose() can actually close the
fd, so the second one gets an error.  That should probably be suppressed.

In fact, looking at the code, it seems to me that close shouldn't be returning
an error on that, even though $! gets set.  You aren't relying on $! without
checking the return value, are you?  The do_close() routine returns the
value from the first fclose(), not the second.  Or are you perhaps using
an older version that returned the value of the second fclose()?

Larry