Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!sdd.hp.com!hplabs!hpfcso!hpldola!hp-lsd!tbc
From: tbc@hp-lsd.COS.HP.COM (Tim Chambers)
Newsgroups: comp.lang.perl
Subject: Help a perl apprentice
Message-ID: <18840001@hp-lsd.COS.HP.COM>
Date: 8 Oct 90 23:41:58 GMT
Organization: HP Logic Systems Division - ColoSpgs, CO
Lines: 85

I have just written my first perl script.  Ok, so it's not as cryptic as TECO
(which I never did master) but it's still a pretty nonintuitive language to
me.

So I ask the perl wizards to help me get used to programming in this new
beasty.  I'm enclosing the perl script, the sh script I started with, and an
example of the program's input.  My questions are:

1.  Why doesn't the "next unless" line I've commented out do what I want it
    to?  (Trust me, it doesn't on my HP 9000 Series 300 HP-UX implementation.)

2.  Although I realize this isn't a teaching center, would anyone care to
    offer some advice on how to optimize my perl implementation of the
    algorithm?  The algorithm I use is (1) capture in an array, (2) sort the
    array, (3) format and print the array elements.  I found it clumsy to do
    all in perl, but it still runs faster than the filters I use in the shell
    script.  Are there perl-isms I am missing that make the algorithm run more
    efficiently?  (Especially my by_free sorting subroutine.)

##############################################################################
#!/usr/local/bin/perl
# $Header: mydf,v 1.3 90/10/08 17:17:19 tbc Exp $

sub by_free {
    @a = split(' ', $a);
    @b = split(' ', $b);
    $a[3] lt $b[3] ? 1 : $a[3] gt $b[3] ? -1 : 0;
}

open (BDF_PIPE, "bdf $* |");

# output
print "Filesystem                     Free (Mb) %Used\n";
print "------------------------------|---------|-----\n";
for ($i=0;<BDF_PIPE>;$i++) {
# why doesn't this work?!
#    next unless (/^.*\%.*$/);
    next unless /^[^ ]+ /;
    next if /Mounted/;
    @lines[$i] = $_;
};

close (BDF_PIPE);

@sortedlines = sort by_free @lines;

foreach $line (@sortedlines) {
    ($f1, $f2, $f3, $avail, $capacity, $dirname) = split(' ', $line);
    printf '%s', $dirname;
    for ($i = 30; $i > length($dirname); $i--) {
	printf ('.');
	}
    printf "%10.1f%6s\n", $avail / 1000, $capacity;
};

print "\n";

##############################################################################
#!/bin/sh
# $Header: mydf,v 1.2 90/09/06 08:42:15 tbc Exp $
echo "Filesystem                     Free (Mb) %Used"
echo "------------------------------|---------|-----"

# first filter to get around bdf's "pretty" format, then use awk
bdf $* | cut -c36-80 | fgrep '%' | sort -nr | \
awk '{ printf("%s", $3)
for (i = 30 ; i > length($3) ; i --) printf(".")
printf ("%10.1f%6s\n", $1 / 1000, $2) }'

echo

##############################################################################
# this is what "bdf" does, in case it's not on non-HP-UX systems
$ bdf
Filesystem           kbytes    used   avail capacity Mounted on
/dev/root            487022  405811   32508    93%   /
/dev/dsk/17s7        278082  240761   23416    91%   /dsk17.07
/dev/dsk/17s3        278082  202040   48233    81%   /users4
/dev/dsk/0s2         278082  188859   61414    75%   /disk3
/dev/dsk/0s4         119429   74980   32506    70%   /users2
/dev/dsk/0s3         230127  166056   41058    80%   /users
hp-lsd:/usr/spool/notes
                     387657  248645  131258    65%   /usr/spool/notes
hplsdry:/work2/hp64000/hptwk/twk_db
                      90607   69076   17000    80%   /usr/hp64000/hptwk/twk_db