Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!wuarchive!uunet!news.uu.net!mcsun!hp4nl!alchemy!piet
From: piet@cs.ruu.nl (Piet van Oostrum)
Newsgroups: comp.lang.perl
Subject: Memory leak??
Message-ID: <1991May22.144629.4044@cs.ruu.nl>
Date: 22 May 91 14:46:29 GMT
Sender: piet@cs.ruu.nl (Piet van Oostrum)
Reply-To: piet@cs.ruu.nl (Piet van Oostrum)
Organization: Dept of Computer Science, Utrecht University, The Netherlands
Lines: 189

I have the following program that does a grep on a directory tree. On a
large tree (like the news spool directory) it keeps growing. So I suspect a
memory leak somewhere. Of course there is a recursive subroutine in it but
after a while it should IMHO stabilize when it has processed the deepest
directory. Instead it grows to more than 3MB.

#! /local/bin/perl

die "Usage: rgrep [-iredblL] regexp filepat ...\n       rgrep -h for help\n"
    if $#ARGV < $[;

# Written by Piet van Oostrum <piet@cs.ruu.nl>
# This is really free software

$nextopt = 1;
$igncase = '';
$regpat = 0;
$links = 0;
$error = 0;
$skipbin = 1;
$debug = 0;

do { $regexp = shift (@ARGV); } while &checkopt ($regexp);
$icreg = $igncase;
$igncase = '';

eval 'sub grep_file {
	    while (<F>) {
		$ln++;
		if (/$regexp/o' . $icreg .') {
			print "$file:$ln:$_";
			print "\n" if substr($_, -1, 1) ne "\n";
		}
	    }
}';

for (@ARGV) {
    if (! &checkopt ($_)) {
        if ($igncase || $regpat || /[?*[]/ || ! -e) {
	    if ($regpat) {
		s/#/\\#/g;
		$_ = "#$_#";
	    } else { # translate File pattern into regexp
                $re = '#($|/)'; $save = $_;
	        while (/[[*?+()|.^$#]/) {
		    $re .= $`;
		    $c = $&;
		    $_ = $';
		    if ($c eq '*') { $c = '[^/]*'; }
		    elsif ($c eq '?') { $c = '[^/]'; }
		    elsif ($c eq '[') {
		    	if (/.\]/) { $c = "[$`$&"; $_ = $'; }
			else {
			    $error++;
			    printf stderr "Illegal filepattern %s\n", $save;
			}
		    } else { $c = "\\$c"; }
		    $re .= $c;
		}
		$_ = "$re$_\$#$igncase";
	    }
	    print "filepat: $_\n" if $debug;
            push (@filepat, $_);
	}
        else { push (@files, $_); print "file: $_\n" if $debug; }
    }
}

exit 1 if $errors ;

if ($#filepat < $[) {
    eval "sub in_pat {1;}" ;
}
else {
    $subtxt = 'sub in_pat { local ($f) = @_;';
    $or = "";
    for (@filepat) {
	$subtxt .= $or . '$f =~ m' . $_;
	$or = " || ";
    }
    $subtxt .= ';};1';

    if (! eval $subtxt) {
    	print $@;
    	exit 1;
    }
} 

@files = (".") if $#files < $[;

for $file (@files) {
    &do_grep ($file);
}

sub do_grep {
    local ($file) = @_;
    local (*F, $ln, $f, $g, @dirfiles);
    if (-f $file) {
	if (open (F, $file)) {
	    if (-B F) { # binary file --  may be compressed/compacted
		if (($cx1 = getc(F)) eq "\377" && (getc(F) eq "\037")) {
		    open (F, "uncompact < $file|");
		    if ($skipbin && -B F) { close (F); return; }
		}
		elsif ($cx1 eq "\037" && (getc(F) eq "\235")) {
		    open (F, "uncompress < $file|");
		    if ($skipbin && -B F) { close (F); return; }
		}
		elsif ($skipbin) {
		    close (F); return;
		}
	    }
	    print "Reading $file\n" if $debug;
	    &grep_file;
	} else {
	    print stderr "Cannot open $file\n";
	}
    }
    elsif (-d $file) {
	print "Entering $file\n" if $debug;
	if (opendir (F, $file)) {
	    @dirfiles = readdir (F);
	    closedir (F);
	    for $f (@dirfiles) {
		next if ($f eq '.' || $f eq '..');
		$g = "$file/$f";
		next if (-l $g && ($links < 1 || $links == 1 && -d $g));
		if (-f $g && &in_pat ($g) || -d _) {
		    &do_grep ($g);
		}
	    }
	} else {
	    print stderr "Can't open $file\n";
	}
    }
}

sub checkopt {
    local ($_) = $_[0];
    if (/^-/ && $nextopt) {
        $nextopt = 1;
	@opt = split (/-*/,$_); shift (@opt);
    	for $opt (@opt) {
	    if ($opt eq 'i') { $igncase = 'i'; }
	    elsif ($opt eq 'd') { $debug = 1; }
	    elsif ($opt eq 'l') { $links = 1; }
	    elsif ($opt eq 'L') { $links = 2; }
	    elsif ($opt eq 'b') { $skipbin = 0; }
	    elsif ($opt eq 'r') { $regpat = 1; }
	    elsif ($opt eq 'e') { $nextopt = 0; }
	    elsif ($opt eq 'h' || $opt eq 'H') { & help; }
	    else { $error++; printf stderr "Unknown option -%s\n", $opt; }
	}
    	return 1;
    }
    $nextopt = 1;
    return 0;
}

sub help {
    print <<'HELP'; exit 0;
Usage: rgrep [-iredblL] regexp filepat ...
  regexp = perl regular expression to search
  filepat ... = a list of files and directories to be searched or
      file patterns to match filenames.
      filepat will be interpreted as file or directory name if it exists
      as such, and does not contain the metacharacters [ ] ? or *. After
      the options -i and -r all filepats will be considered patterns.
      rgrep will search all files in any of the directories given (and its
      subdirectories) that match any of the filepats, except binary files.
      Compressed files will be searched in uncompressed form.
      Note: filepats may contain / contrary to find usage.
  -b  Don't skip binary files.
  -i  Ignore case, either in the regexp or in filename matching (depending
      on the location). Before the regexp only applies to the regexp,
      otherwise to the filepats following it.
  -r  The following filepats are treated as real perl regexps rather than
      shell style filename patterns. In this case / is not a special
      character, i.e. it is matched by . and matching is not anchored (you
      must supply ^ and $ yourself). E.g. a.b matches the file /xa/by/zz.
  -l  Do follow symbolic links only for files (default is do not follow).
  -L  Do follow symbolic links for files and directories.
  -e  Do not interpret following argument as option. Useful if regexp or
      filepat starts with a -.
  -d  Debugging: Give a lot of output on what happens.
  -h  print this message and exit.
Piet van Oostrum <piet@cs.ruu.nl>
HELP
}