Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!munnari.oz.au!frankland-river!pem
From: pem@frankland-river.aaii.oz.au (Paul E. Maisano)
Newsgroups: comp.lang.perl
Subject: Re: Operations on filehandles
Summary: Solution does not seem to work.
Message-ID: <1041@frankland-river.aaii.oz.au>
Date: 9 Feb 90 00:01:53 GMT
References: <1004@frankland-river.aaii.oz.au> <7020@jpl-devvax.JPL.NASA.GOV>
Organization: Australian AI Institute
Lines: 47

In article <7020@jpl-devvax.JPL.NASA.GOV>, lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall) writes:
> 
> I believe you can do a local on a name using the type globbing syntax:
> 
>     sub foo {
> 	local(*FILE);
> 
> 	open(FILE,'foo');
> 	etc.
> 
> I haven't exercised it heavily, but the code is supposedly there to create
> a new name if you aren't assigning another name to *FILE.  Just as normal
> variables created with a null value when you don't assign something to them.

This pretty much what I was looking for. It's neater than generating unique
filehandle names by hand, but I can't seem to get it to work in a test program.
I end up getting a segmentation fault. 

#!/usr/bin/perl
# This is just an otherwise useless program to test recursive opens.

sub read_lines {
    local($name, $n) = @_;
    local($i);
    local(*F);

    open(F, $name) || die "$name: $!\n";
    for ($i = 0; $i < $n; $i++) {
	$_ = <F>;
	print("$i ($n): ", $_) || warn("print failed: $!\n");
	if ($i == 2 && $n <= 10) { &read_lines($name, $n+5); }
    }
    close(F);
    print "finished $n\n";
}

&read_lines("/etc/passwd", 5);
exit 0;

Generating unique names for filehandles does work, though (as I expected).

------------------
Paul E. Maisano
Australian Artificial Intelligence Institute
1 Grattan St. Carlton, Vic. 3053, Australia
Ph: +613 663-7922  Fax: +613 663-7937
Email: pem@aaii.oz.au