Adding upstream version 124.0.1.upstream/124.0.1

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

2 files changed, 962 insertions, 0 deletions

diff --git a/tools/bloatview/bloatdiff.pl b/tools/bloatview/bloatdiff.pl
new file mode 100755
index 0000000000..a9bfa97226
--- /dev/null
+++ b/tools/bloatview/bloatdiff.pl
@@ -0,0 +1,372 @@
+#!/usr/bin/perl -w
+# This Source Code Form is subject to the terms of the Mozilla Public
+# License, v. 2.0. If a copy of the MPL was not distributed with this
+# file, You can obtain one at http://mozilla.org/MPL/2.0/.
+
+
+################################################################################
+
+sub usage() {
+    print <<EOUSAGE;
+# bloatdiff.pl - munges the output from
+#   XPCOM_MEM_BLOAT_LOG=1 
+#   firefox -P default resource:///res/bloatcycle.html
+# so that it does some summary and stats stuff.
+#
+# To show leak test results for a set of changes, do something like this:
+#
+#   XPCOM_MEM_BLOAT_LOG=1
+#   firefox -P default resource:///res/bloatcycle.html > a.out
+#     **make change**
+#   firefox -P default resource:///res/bloatcycle.html > b.out
+#   bloatdiff.pl a.out b.out
+
+EOUSAGE
+}
+
+$OLDFILE = $ARGV[0];
+$NEWFILE = $ARGV[1];
+#$LABEL   = $ARGV[2];
+
+if (!$OLDFILE or
+    ! -e $OLDFILE or 
+    -z $OLDFILE) {
+    print "\nError: Previous log file not specified, does not exist, or is empty.\n\n";
+    &usage();
+    exit 1;
+}
+
+if (!$NEWFILE or
+    ! -e $NEWFILE or 
+    -z $NEWFILE) {
+    print "\nError: Current log file not specified, does not exist, or is empty.\n\n";
+    &usage();
+    exit 1;
+}
+
+sub processFile {
+    my ($filename, $map, $prevMap) = @_;
+    open(FH, $filename);
+    while (<FH>) {
+        if (m{
+              ^\s*(\d+)\s          # Line number
+              ([\w:]+)\s+          # Name
+              (-?\d+)\s+           # Size
+              (-?\d+)\s+           # Leaked
+              (-?\d+)\s+           # Objects Total
+              (-?\d+)\s+           # Objects Rem
+              \(\s*(-?[\d.]+)\s+   # Objects Mean
+                 \+/-\s+
+              ([\w.]+)\)\s+        # Objects StdDev
+              (-?\d+)\s+           # Reference Total
+              (-?\d+)\s+           # Reference Rem
+              \(\s*(-?[\d.]+)\s+   # Reference Mean
+                 \+/-\s+
+              ([\w\.]+)\)          # Reference StdDev
+             }x) {
+          $$map{$2} = { name => $2,
+                        size => $3,
+                        leaked => $4,
+                        objTotal => $5,
+                        objRem => $6,
+                        objMean => $7,
+                        objStdDev => $8,
+                        refTotal => $9,
+                        refRem => $10,
+                        refMean => $11,
+                        refStdDev => $12,
+                        bloat => $3 * $5 # size * objTotal
+                      };
+        } else {
+#            print "failed to parse: $_\n";
+        }
+    }
+    close(FH);
+}
+
+%oldMap = ();
+processFile($OLDFILE, \%oldMap);
+
+%newMap = ();
+processFile($NEWFILE, \%newMap);
+
+################################################################################
+
+$inf = 9999999.99;
+
+sub getLeaksDelta {
+    my ($key) = @_;
+    my $oldLeaks = $oldMap{$key}{leaked} || 0;
+    my $newLeaks = $newMap{$key}{leaked};
+    my $percentLeaks = 0;
+    if ($oldLeaks == 0) {
+        if ($newLeaks != 0) {
+            # there weren't any leaks before, but now there are!
+            $percentLeaks = $inf;
+        }
+    }
+    else {
+        $percentLeaks = ($newLeaks - $oldLeaks) / $oldLeaks * 100;
+    }
+    # else we had no record of this class before
+    return ($newLeaks - $oldLeaks, $percentLeaks);
+}
+    
+################################################################################
+
+sub getBloatDelta {
+    my ($key) = @_;
+    my $newBloat = $newMap{$key}{bloat};
+    my $percentBloat = 0;
+    my $oldSize = $oldMap{$key}{size} || 0;
+    my $oldTotal = $oldMap{$key}{objTotal} || 0;
+    my $oldBloat = $oldTotal * $oldSize;
+    if ($oldBloat == 0) {
+        if ($newBloat != 0) {
+            # this class wasn't used before, but now it is
+            $percentBloat = $inf;
+        }
+    }
+    else {
+        $percentBloat = ($newBloat - $oldBloat) / $oldBloat * 100;
+    }
+    # else we had no record of this class before
+    return ($newBloat - $oldBloat, $percentBloat);
+}
+
+################################################################################
+
+foreach $key (keys %newMap) {
+    my ($newLeaks, $percentLeaks) = getLeaksDelta($key);
+    my ($newBloat, $percentBloat) = getBloatDelta($key);
+    $newMap{$key}{leakDelta} = $newLeaks;
+    $newMap{$key}{leakPercent} = $percentLeaks;
+    $newMap{$key}{bloatDelta} = $newBloat;
+    $newMap{$key}{bloatPercent} = $percentBloat;
+}
+
+################################################################################
+
+# Print a value of bytes out in a reasonable
+# KB, MB, or GB form.  Copied from build-seamonkey-util.pl, sorry.  -mcafee
+sub PrintSize($) {
+
+    # print a number with 3 significant figures
+    sub PrintNum($) {
+        my ($num) = @_;
+        my $rv;
+        if ($num < 1) {
+            $rv = sprintf "%.3f", ($num);
+        } elsif ($num < 10) {
+            $rv = sprintf "%.2f", ($num);
+        } elsif ($num < 100) {
+            $rv = sprintf "%.1f", ($num);
+        } else {
+            $rv = sprintf "%d", ($num);
+        }
+    }
+
+    my ($size) = @_;
+    my $rv;
+    if ($size > 1000000000) {
+        $rv = PrintNum($size / 1000000000.0) . "G";
+    } elsif ($size > 1000000) {
+        $rv = PrintNum($size / 1000000.0) . "M";
+    } elsif ($size > 1000) {
+        $rv = PrintNum($size / 1000.0) . "K";
+    } else {
+        $rv = PrintNum($size);
+    }
+}
+
+
+print "Bloat/Leak Delta Report\n";
+print "--------------------------------------------------------------------------------------\n";
+print "Current file:  $NEWFILE\n";
+print "Previous file: $OLDFILE\n";
+print "----------------------------------------------leaks------leaks%------bloat------bloat%\n";
+
+    if (! $newMap{"TOTAL"} or 
+        ! $newMap{"TOTAL"}{bloat}) {
+        # It's OK if leaked or leakPercent are 0 (in fact, that would be good).
+        # If bloatPercent is zero, it is also OK, because we may have just had
+        # two runs exactly the same or with no new bloat.
+        print "\nError: unable to calculate bloat/leak data.\n";
+        print "There is no data present.\n\n";
+        print "HINT - Did your test run complete successfully?\n";
+        print "HINT - Are you pointing at the right log files?\n\n";
+        &usage();
+        exit 1;
+    }
+
+printf "%-40s %10s %10.2f%% %10s %10.2f%%\n",
+       ("TOTAL",
+        $newMap{"TOTAL"}{leaked}, $newMap{"TOTAL"}{leakPercent},
+        $newMap{"TOTAL"}{bloat}, $newMap{"TOTAL"}{bloatPercent});
+
+################################################################################
+
+sub percentStr {
+    my ($p) = @_;
+    if ($p == $inf) {
+        return "-";
+    }
+    else {
+        return sprintf "%10.2f%%", $p;
+    }
+}
+
+# NEW LEAKS
+@keys = sort { $newMap{$b}{leakPercent} <=> $newMap{$a}{leakPercent} } keys %newMap;
+my $needsHeading = 1;
+my $total = 0;
+foreach $key (@keys) {
+    my $percentLeaks = $newMap{$key}{leakPercent};
+    my $leaks = $newMap{$key}{leaked};
+    if ($percentLeaks > 0 && $key !~ /TOTAL/) {
+        if ($needsHeading) {
+            printf "--NEW-LEAKS-----------------------------------leaks------leaks%%-----------------------\n";
+            $needsHeading = 0;
+        }
+        printf "%-40s %10s %10s\n", ($key, $leaks, percentStr($percentLeaks));
+        $total += $leaks;
+    }
+}
+if (!$needsHeading) {
+    printf "%-40s %10s\n", ("TOTAL", $total);
+}
+
+# FIXED LEAKS
+@keys = sort { $newMap{$b}{leakPercent} <=> $newMap{$a}{leakPercent} } keys %newMap;
+$needsHeading = 1;
+$total = 0;
+foreach $key (@keys) {
+    my $percentLeaks = $newMap{$key}{leakPercent};
+    my $leaks = $newMap{$key}{leaked};
+    if ($percentLeaks < 0 && $key !~ /TOTAL/) {
+        if ($needsHeading) {
+            printf "--FIXED-LEAKS---------------------------------leaks------leaks%%-----------------------\n";
+            $needsHeading = 0;
+        }
+        printf "%-40s %10s %10s\n", ($key, $leaks, percentStr($percentLeaks));
+        $total += $leaks;
+    }
+}
+if (!$needsHeading) {
+    printf "%-40s %10s\n", ("TOTAL", $total);
+}
+
+# NEW BLOAT
+@keys = sort { $newMap{$b}{bloatPercent} <=> $newMap{$a}{bloatPercent} } keys %newMap;
+$needsHeading = 1;
+$total = 0;
+foreach $key (@keys) {
+    my $percentBloat = $newMap{$key}{bloatPercent};
+    my $bloat = $newMap{$key}{bloat};
+    if ($percentBloat > 0  && $key !~ /TOTAL/) {
+        if ($needsHeading) {
+            printf "--NEW-BLOAT-----------------------------------bloat------bloat%%-----------------------\n";
+            $needsHeading = 0;
+        }
+        printf "%-40s %10s %10s\n", ($key, $bloat, percentStr($percentBloat));
+        $total += $bloat;
+    }
+}
+if (!$needsHeading) {
+    printf "%-40s %10s\n", ("TOTAL", $total);
+}
+
+# ALL LEAKS
+@keys = sort { $newMap{$b}{leaked} <=> $newMap{$a}{leaked} } keys %newMap;
+$needsHeading = 1;
+$total = 0;
+foreach $key (@keys) {
+    my $leaks = $newMap{$key}{leaked};
+    my $percentLeaks = $newMap{$key}{leakPercent};
+    if ($leaks > 0) {
+        if ($needsHeading) {
+            printf "--ALL-LEAKS-----------------------------------leaks------leaks%%-----------------------\n";
+            $needsHeading = 0;
+        }
+        printf "%-40s %10s %10s\n", ($key, $leaks, percentStr($percentLeaks));
+        if ($key !~ /TOTAL/) {
+            $total += $leaks;
+        }
+    }
+}
+if (!$needsHeading) {
+#    printf "%-40s %10s\n", ("TOTAL", $total);
+}
+
+# ALL BLOAT
+@keys = sort { $newMap{$b}{bloat} <=> $newMap{$a}{bloat} } keys %newMap;
+$needsHeading = 1;
+$total = 0;
+foreach $key (@keys) {
+    my $bloat = $newMap{$key}{bloat};
+    my $percentBloat = $newMap{$key}{bloatPercent};
+    if ($bloat > 0) {
+        if ($needsHeading) {
+            printf "--ALL-BLOAT-----------------------------------bloat------bloat%%-----------------------\n";
+            $needsHeading = 0;
+        }
+        printf "%-40s %10s %10s\n", ($key, $bloat, percentStr($percentBloat));
+        if ($key !~ /TOTAL/) {
+            $total += $bloat;
+        }
+    }
+}
+if (!$needsHeading) {
+#    printf "%-40s %10s\n", ("TOTAL", $total);
+}
+
+# NEW CLASSES
+@keys = sort { $newMap{$b}{bloatDelta} <=> $newMap{$a}{bloatDelta} } keys %newMap;
+$needsHeading = 1;
+my $ltotal = 0;
+my $btotal = 0;
+foreach $key (@keys) {
+    my $leaks = $newMap{$key}{leaked};
+    my $bloat = $newMap{$key}{bloat};
+    my $percentBloat = $newMap{$key}{bloatPercent};
+    if ($percentBloat == $inf && $key !~ /TOTAL/) {
+        if ($needsHeading) {
+            printf "--CLASSES-NOT-REPORTED-LAST-TIME--------------leaks------bloat------------------------\n";
+            $needsHeading = 0;
+        }
+        printf "%-40s %10s %10s\n", ($key, $leaks, $bloat);
+        if ($key !~ /TOTAL/) {
+            $ltotal += $leaks;
+            $btotal += $bloat;
+        }
+    }
+}
+if (!$needsHeading) {
+    printf "%-40s %10s %10s\n", ("TOTAL", $ltotal, $btotal);
+}
+
+# OLD CLASSES
+@keys = sort { ($oldMap{$b}{bloat} || 0) <=> ($oldMap{$a}{bloat} || 0) } keys %oldMap;
+$needsHeading = 1;
+$ltotal = 0;
+$btotal = 0;
+foreach $key (@keys) {
+    if (!defined($newMap{$key})) {
+        my $leaks = $oldMap{$key}{leaked};
+        my $bloat = $oldMap{$key}{bloat};
+        if ($needsHeading) {
+            printf "--CLASSES-THAT-WENT-AWAY----------------------leaks------bloat------------------------\n";
+            $needsHeading = 0;
+        }
+        printf "%-40s %10s %10s\n", ($key, $leaks, $bloat);
+        if ($key !~ /TOTAL/) {
+            $ltotal += $leaks;
+            $btotal += $bloat;
+        }
+    }
+}
+if (!$needsHeading) {
+    printf "%-40s %10s %10s\n", ("TOTAL", $ltotal, $btotal);
+}
+
+print "--------------------------------------------------------------------------------------\n";
diff --git a/tools/bloatview/bloattable.pl b/tools/bloatview/bloattable.pl
new file mode 100755
index 0000000000..e8acfabed3
--- /dev/null
+++ b/tools/bloatview/bloattable.pl
@@ -0,0 +1,590 @@
+#!/usr/bin/perl -w
+#
+# This Source Code Form is subject to the terms of the Mozilla Public
+# License, v. 2.0. If a copy of the MPL was not distributed with this
+# file, You can obtain one at http://mozilla.org/MPL/2.0/.
+
+# bloattable [-debug] [-source] [-byte n|-obj n|-ref n] <file1> <file2> ... <filen> > <html-file>
+#
+# file1, file2, ... filen should be successive BloatView files generated from the same run.
+# Summarize them in an HTML table.  Output the HTML to the standard output.
+#
+# If -debug is set, create a slightly larger html file which is more suitable for debugging this script.
+# If -source is set, create an html file that prints the html source as the output
+# If -byte n, -obj n, or -ref n is given, make the page default to showing byte, object, or reference statistics,
+#    respectively, and sort by the nth column (n is zero-based, so the first column has n==0).
+#
+# See http://lxr.mozilla.org/mozilla/source/xpcom/doc/MemoryTools.html
+
+use 5.004;
+use strict;
+use diagnostics;
+use File::Basename;
+use Getopt::Long;
+
+# The generated HTML is almost entirely generated by a script.  Only the <HTML>, <HEAD>, and <BODY> elements are explicit
+# because a <SCRIPT> element cannot officially be a direct descendant of an <HTML> element.
+# The script itself is almost all generated by an eval of a large string.  This allows the script to reproduce itself
+# when making a new page using document.write's.  Re-sorting the page causes it to regenerate itself in this way.
+
+
+
+# Return the file's modification date.
+sub fileModDate($) {
+	my ($pathName) = @_;
+	my ($dev, $ino, $mode, $nlink, $uid, $gid, $rdev, $size, $atime, $mtime, $ctime, $blksize, $blocks) =
+		stat $pathName or die "Can't stat '$pathName'";
+	return $mtime;
+}
+
+
+sub fileCoreName($) {
+	my ($pathName) = @_;
+	my $fileName = basename($pathName, "");
+	$fileName =~ s/\..*//;
+	return $fileName;
+}
+
+
+# Convert a raw string into a single-quoted JavaScript string.
+sub singleQuoteString($) {
+	local ($_) = @_;
+	s/\\/\\\\/g;
+	s/'/\\'/g;
+	s/\n/\\n/g;
+	s/<\//<\\\//g;
+	return "'$_'";
+}
+
+
+# Convert a raw string into a double-quoted JavaScript string.
+sub doubleQuoteString($) {
+	local ($_) = @_;
+	s/\\/\\\\/g;
+	s/"/\\"/g;
+	s/\n/\\n/g;
+	s/<\//<\\\//g;
+	return "\"$_\"";
+}
+
+
+# Quote special HTML characters in the string.
+sub quoteHTML($) {
+	local ($_) = @_;
+	s/\&/&amp;/g;
+	s/</&lt;/g;
+	s/>/&gt;/g;
+	s/ /&nbsp;/g;
+	s/\n/<BR>\n/g;
+	return $_;
+}
+
+
+# Write the generated page to the standard output.
+# The script source code is read from this file past the __END__ marker
+# @$scriptData is the JavaScript source for the tables passed to JavaScript.  Each entry is one line of JavaScript.
+# @$persistentScriptData is the same as @scriptData, but persists when the page reloads itself.
+# If $debug is true, generate the script directly instead of having it eval itself.
+# If $source is true, generate a script that displays the page's source instead of the page itself.
+sub generate(\@\@$$$$) {
+	my ($scriptData, $persistentScriptData, $debug, $source, $showMode, $sortColumn) = @_;
+
+	my @scriptSource = <DATA>;
+	chomp @scriptSource;
+	print <<'EOS';
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
+<HTML>
+<HEAD>
+<SCRIPT type="text/javascript">
+EOS
+
+	foreach (@$scriptData) {print "$_\n";}
+	print "\n";
+
+	print "var srcArray = [\n";
+	my @quotedScriptSource = map {
+		my $line = $_;
+		$line =~ s/^\s+//g;
+		# $line =~ s/^\/\/SOURCE\s+//g if $source;
+		$line =~ s/^\/\/.*//g;
+		$line =~ s/\s+$//g;
+		$line eq "" ? () : $line
+	} @$persistentScriptData, @scriptSource;
+	my $lastQuotedLine = pop @quotedScriptSource;
+	foreach (@quotedScriptSource) {print doubleQuoteString($_), ",\n";}
+	print doubleQuoteString($lastQuotedLine), "];\n\n";
+
+	if ($debug) {
+		push @quotedScriptSource, $lastQuotedLine;
+		foreach (@quotedScriptSource) {
+			s/<\//<\\\//g;	# This fails if a regexp ends with a '<'.  Oh well....
+			print "$_\n";
+		}
+		print "\n";
+	} else {
+		print "eval(srcArray.join(\"\\n\"));\n\n";
+	}
+	print "showMode = $showMode;\n";
+	print "sortColumn = $sortColumn;\n";
+	if ($source) {
+		print <<'EOS';
+function writeQuotedHTML(s) {
+	document.write(quoteHTML(s.toString()).replace(/\n/g, '<BR>\n'));
+}
+
+var quotingDocument = {
+  write: function () {
+		for (var i = 0; i < arguments.length; i++)
+			writeQuotedHTML(arguments[i]);
+	},
+  writeln: function () {
+		for (var i = 0; i < arguments.length; i++)
+			writeQuotedHTML(arguments[i]);
+		document.writeln('<BR>');
+	}
+};
+EOS
+	} else {
+		print "showHead(document);\n";
+	}
+	print "</SCRIPT>\n";
+	print "</HEAD>\n\n";
+	print "<BODY>\n";
+	if ($source) {
+		print "<P><TT>";
+		print quoteHTML "<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.0 Transitional//EN\" \"http://www.w3.org/TR/REC-html40/loose.dtd\">\n";
+		print quoteHTML "<HTML>\n";
+		print quoteHTML "<HEAD>\n";
+		print "<SCRIPT type=\"text/javascript\">showHead(quotingDocument);</SCRIPT>\n";
+		print quoteHTML "</HEAD>\n\n";
+		print quoteHTML "<BODY>\n";
+		print "<SCRIPT type=\"text/javascript\">showBody(quotingDocument);</SCRIPT>\n";
+		print quoteHTML "</BODY>\n";
+		print quoteHTML "</HTML>\n";
+		print "</TT></P>\n";
+	} else {
+		print "<SCRIPT type=\"text/javascript\">showBody(document);</SCRIPT>\n";
+	}
+	print "</BODY>\n";
+	print "</HTML>\n";
+}
+
+
+
+# Read the bloat file into hash table $h.  The hash table is indexed by class names;
+# each entry is a list with the following elements:
+#	bytesAlloc		Total number of bytes allocated
+#	bytesNet		Total number of bytes allocated but not deallocated
+#	objectsAlloc	Total number of objects allocated
+#	objectsNet		Total number of objects allocated but not deallocated
+#	refsAlloc		Total number of references AddRef'd
+#	refsNet			Total number of references AddRef'd but not Released
+# Except for TOTAL, all hash table entries refer to mutually exclusive data.
+# $sizes is a hash table indexed by class names.  Each entry of that table contains the class's instance size.
+sub readBloatFile($\%\%) {
+	my ($file, $h, $sizes) = @_;
+	local $_;	# Needed for 'while (<FILE>)' below.
+
+	my $readSomething = 0;
+	open FILE, $file;
+	while (<FILE>) {
+		if (my ($name, $size, $bytesNet, $objectsAlloc, $objectsNet, $refsAlloc, $refsNet) =
+			   /^\s*(?:\d+)\s+([\w:]+)\s+(\d+)\s+(-?\d+)\s+(\d+)\s+(-?\d+)\s*\([^()]*\)\s*(\d+)\s+(-?\d+)\s*\([^()]*\)\s*$/) {
+			my $bytesAlloc;
+			if ($name eq "TOTAL") {
+				$size = "undefined";
+				$bytesAlloc = "undefined";
+			} else {
+				$bytesAlloc = $objectsAlloc * $size;
+				if ($bytesNet != $objectsNet * $size) {
+					print STDERR "In '$file', class $name bytesNet != objectsNet * size: $bytesNet != $objectsNet * $size\n";
+				}
+			}
+			print STDERR "Duplicate entry $name in '$file'\n" if $$h{$name};
+			$$h{$name} = [$bytesAlloc, $bytesNet, $objectsAlloc, $objectsNet, $refsAlloc, $refsNet];
+			
+			my $oldSize = $$sizes{$name};
+			print STDERR "Mismatch of sizes of class $name: $oldSize and $size\n" if defined($oldSize) && $size ne $oldSize;
+			$$sizes{$name} = $size;
+			$readSomething = 1;
+		} elsif (/^\s*(?:\d+)\s+([\w:]+)\s/) {
+			print STDERR "Unable to parse '$file' line: $_";
+		}
+	}
+	close FILE;
+	print STDERR "No data in '$file'\n" unless $readSomething;
+	return $h;
+}
+
+
+my %sizes;			# <class-name> => <instance-size>
+my %tables;			# <file-name> => <bloat-table>; see readBloatFile for format of <bloat-table>
+
+# Generate the JavaScript source code for the row named $c.  $l can contain the initial entries of the row.
+sub genTableRowSource($$) {
+	my ($l, $c) = @_;
+	my $lastE;
+	foreach (@ARGV) {
+		my $e = $tables{$_}{$c};
+		if (defined($lastE) && !defined($e)) {
+			$e = [0,0,0,0,0,0];
+			print STDERR "Class $c is defined in an earlier file but not in '$_'\n";
+		}
+		if (defined $e) {
+			if (defined $lastE) {
+				for (my $i = 0; $i <= $#$e; $i++) {
+					my $n = $$e[$i];
+					$l .= ($n eq "undefined" ? "undefined" : $n - $$lastE[$i]) . ",";
+				}
+				$l .= " ";
+			} else {
+				$l .= join(",", @$e) . ", ";
+			}
+			$lastE = $e;
+		} else {
+			$l .= "0,0,0,0,0,0, ";
+		}
+	}
+	$l .= join(",", @$lastE);
+	return "[$l]";
+}
+
+
+
+my $debug;
+my $source;
+my $showMode;
+my $sortColumn;
+my @modeOptions;
+
+GetOptions("debug" => \$debug, "source" => \$source, "byte=i" => \$modeOptions[0], "obj=i" => \$modeOptions[1], "ref=i" => \$modeOptions[2]);
+for (my $i = 0; $i != 3; $i++) {
+	my $modeOption = $modeOptions[$i];
+	if ($modeOption) {
+		die "Only one of -byte, -obj, or -ref may be given" if defined $showMode;
+		my $nFileColumns = scalar(@ARGV) + 1;
+		die "-byte, -obj, or -ref column number out of range" if $modeOption < 0 || $modeOption >= 2 + 2*$nFileColumns;
+		$showMode = $i;
+		if ($modeOption >= 2) {
+			$modeOption -= 2;
+			$sortColumn = 2 + $showMode*2;
+			if ($modeOption >= $nFileColumns) {
+				$sortColumn++;
+				$modeOption -= $nFileColumns;
+			}
+			$sortColumn += $modeOption*6;
+		} else {
+			$sortColumn = $modeOption;
+		}
+	}
+}
+unless (defined $showMode) {
+	$showMode = 0;
+	$sortColumn = 0;
+}
+
+# Read all of the bloat files.
+foreach (@ARGV) {
+	unless ($tables{$_}) {
+		my $f = $_;
+		my %table;
+		
+		readBloatFile $_, %table, %sizes;
+		$tables{$_} = \%table;
+	}
+}
+die "No input" unless %sizes;
+
+my @scriptData;	# JavaScript source for the tables passed to JavaScript.  Each entry is one line of JavaScript.
+my @persistentScriptData;	# Same as @scriptData, but persists the page reloads itself.
+
+# Print a list of bloat file names.
+push @persistentScriptData, "var nFiles = " . scalar(@ARGV) . ";";
+push @persistentScriptData, "var fileTags = [" . join(", ", map {singleQuoteString substr(fileCoreName($_), -10)} @ARGV) . "];";
+push @persistentScriptData, "var fileNames = [" . join(", ", map {singleQuoteString $_} @ARGV) . "];";
+push @persistentScriptData, "var fileDates = [" . join(", ", map {singleQuoteString localtime fileModDate $_} @ARGV) . "];";
+
+# Print the bloat tables.
+push @persistentScriptData, "var totals = " . genTableRowSource('"TOTAL", undefined, ', "TOTAL") . ";";
+push @scriptData, "var classTables = [";
+delete $sizes{"TOTAL"};
+my @classes = sort(keys %sizes);
+for (my $i = 0; $i <= $#classes; $i++) {
+	my $c = $classes[$i];
+	push @scriptData, genTableRowSource(doubleQuoteString($c).", ".$sizes{$c}.", ", $c) . ($i == $#classes ? "];" : ",");
+}
+
+generate(@scriptData, @persistentScriptData, $debug, $source, $showMode, $sortColumn);
+1;
+
+
+# The source of the eval'd JavaScript follows.
+# Comments starting with // that are alone on a line are stripped by the Perl script.
+__END__
+
+// showMode: 0=bytes, 1=objects, 2=references
+var showMode;
+var modeName;
+var modeNameUpper;
+
+var sortColumn;
+
+// Sort according to the sortColumn.  Column 0 is sorted alphabetically in ascending order.
+// All other columns are sorted numerically in descending order, with column 0 used for a secondary sort.
+// Undefined is always listed last.
+function sortCompare(x, y) {
+	if (sortColumn) {
+		var xc = x[sortColumn];
+		var yc = y[sortColumn];
+		if (xc < yc || xc === undefined && yc !== undefined) return 1;
+		if (yc < xc || yc === undefined && xc !== undefined) return -1;
+	}
+	
+	var x0 = x[0];
+	var y0 = y[0];
+	if (x0 > y0 || x0 === undefined && y0 !== undefined) return 1;
+	if (y0 > x0 || y0 === undefined && x0 !== undefined) return -1;
+	return 0;
+}
+
+
+// Quote special HTML characters in the string.
+function quoteHTML(s) {
+	s = s.replace(/&/g, '&amp;');
+	// Can't use /</g because HTML interprets '</g' as ending the script!
+	s = s.replace(/\x3C/g, '&lt;');
+	s = s.replace(/>/g, '&gt;');
+	s = s.replace(/ /g, '&nbsp;');
+	return s;
+}
+
+
+function writeFileTable(d) {
+	d.writeln('<TABLE border=1 cellspacing=1 cellpadding=0>');
+	d.writeln('<TR>\n<TH>Name</TH>\n<TH>File</TH>\n<TH>Date</TH>\n</TR>');
+	for (var i = 0; i < nFiles; i++)
+		d.writeln('<TR>\n<TD>'+quoteHTML(fileTags[i])+'</TD>\n<TD><TT>'+quoteHTML(fileNames[i])+'</TT></TD>\n<TD>'+quoteHTML(fileDates[i])+'</TD>\n</TR>');
+	d.writeln('</TABLE>');
+}
+
+
+function writeReloadLink(d, column, s, rowspan) {
+	d.write(rowspan == 1 ? '<TH>' : '<TH rowspan='+rowspan+'>');
+	if (column != sortColumn)
+		d.write('<A href="javascript:reloadSelf('+column+','+showMode+')">');
+	d.write(s);
+	if (column != sortColumn)
+		d.write('</A>');
+	d.writeln('</TH>');
+}
+
+function writeClassTableRow(d, row, base, modeName) {
+	if (modeName) {
+		d.writeln('<TR>\n<TH>'+modeName+'</TH>');
+	} else {
+		d.writeln('<TR>\n<TD><A href="javascript:showRowDetail(\''+row[0]+'\')">'+quoteHTML(row[0])+'</A></TD>');
+		var v = row[1];
+		d.writeln('<TD class=num>'+(v === undefined ? '' : v)+'</TD>');
+	}
+	for (var i = 0; i != 2; i++) {
+		var c = base + i;
+		for (var j = 0; j <= nFiles; j++) {
+			v = row[c];
+			var style = 'num';
+			if (j != nFiles)
+				if (v > 0) {
+					style = 'pos';
+					v = '+'+v;
+				} else
+					style = 'neg';
+			d.writeln('<TD class='+style+'>'+(v === undefined ? '' : v)+'</TD>');
+			c += 6;
+		}
+	}
+	d.writeln('</TR>');
+}
+
+function writeClassTable(d) {
+	var base = 2 + showMode*2;
+
+	// Make a copy because a sort is destructive.
+	var table = classTables.concat();
+	table.sort(sortCompare);
+
+	d.writeln('<TABLE border=1 cellspacing=1 cellpadding=0>');
+
+	d.writeln('<TR>');
+	writeReloadLink(d, 0, 'Class Name', 2);
+	writeReloadLink(d, 1, 'Instance<BR>Size', 2);
+	d.writeln('<TH colspan='+(nFiles+1)+'>'+modeNameUpper+'s allocated</TH>');
+	d.writeln('<TH colspan='+(nFiles+1)+'>'+modeNameUpper+'s allocated but not freed</TH>\n</TR>');
+	d.writeln('<TR>');
+	for (var i = 0; i != 2; i++) {
+		var c = base + i;
+		for (var j = 0; j <= nFiles; j++) {
+			writeReloadLink(d, c, j == nFiles ? 'Total' : quoteHTML(fileTags[j]), 1);
+			c += 6;
+		}
+	}
+	d.writeln('</TR>');
+
+	writeClassTableRow(d, totals, base, 0);
+	for (var r = 0; r < table.length; r++)
+		writeClassTableRow(d, table[r], base, 0);
+
+	d.writeln('</TABLE>');
+}
+
+
+var modeNames = ["byte", "object", "reference"];
+var modeNamesUpper = ["Byte", "Object", "Reference"];
+var styleSheet = '<STYLE type="TEXT/CSS">\n'+
+	'BODY {background-color: #FFFFFF; color: #000000}\n'+
+	'.num {text-align: right}\n'+
+	'.pos {text-align: right; color: #CC0000}\n'+
+	'.neg {text-align: right; color: #009900}\n'+
+	'</STYLE>';
+
+
+function showHead(d) {
+	modeName = modeNames[showMode];
+	modeNameUpper = modeNamesUpper[showMode];
+	d.writeln('<TITLE>'+modeNameUpper+' Bloats</TITLE>');
+	d.writeln(styleSheet);
+}
+
+function showBody(d) {
+	d.writeln('<H1>'+modeNameUpper+' Bloats</H1>');
+	writeFileTable(d);
+	d.write('<FORM>');
+	for (var i = 0; i != 3; i++)
+		if (i != showMode) {
+			var newSortColumn = sortColumn;
+			if (sortColumn >= 2)
+				newSortColumn = sortColumn + (i-showMode)*2;
+			d.write('<INPUT type="button" value="Show '+modeNamesUpper[i]+'s" onClick="reloadSelf('+newSortColumn+','+i+')">');
+		}
+	d.writeln('</FORM>');
+	d.writeln('<P>The numbers do not include <CODE>malloc</CODE>\'d data such as string contents.</P>');
+	d.writeln('<P>Click on a column heading to sort by that column. Click on a class name to see details for that class.</P>');
+	writeClassTable(d);
+}
+
+
+function showRowDetail(rowName) {
+	var row;
+	var i;
+
+	if (rowName == "TOTAL")
+		row = totals;
+	else {
+		for (i = 0; i < classTables.length; i++)
+			if (rowName == classTables[i][0]) {
+				row = classTables[i];
+				break;
+			}
+	}
+	if (row) {
+		var w = window.open("", "ClassTableRowDetails");
+		var d = w.document;
+		d.open();
+		d.writeln('<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">');
+		d.writeln('<HTML>\n<HEAD>\n<TITLE>'+quoteHTML(rowName)+' bloat details</TITLE>');
+		d.writeln(styleSheet);
+		d.writeln('</HEAD>\n\n<BODY>');
+		d.writeln('<H2>'+quoteHTML(rowName)+'</H2>');
+		if (row[1] !== undefined)
+			d.writeln('<P>Each instance has '+row[1]+' bytes.</P>');
+
+		d.writeln('<TABLE border=1 cellspacing=1 cellpadding=0>');
+		d.writeln('<TR>\n<TH></TH>\n<TH colspan='+(nFiles+1)+'>Allocated</TH>');
+		d.writeln('<TH colspan='+(nFiles+1)+'>Allocated but not freed</TH>\n</TR>');
+		d.writeln('<TR>\n<TH></TH>');
+		for (i = 0; i != 2; i++)
+			for (var j = 0; j <= nFiles; j++)
+				d.writeln('<TH>'+(j == nFiles ? 'Total' : quoteHTML(fileTags[j]))+'</TH>');
+		d.writeln('</TR>');
+
+		for (i = 0; i != 3; i++)
+			writeClassTableRow(d, row, 2+i*2, modeNamesUpper[i]+'s');
+
+		d.writeln('</TABLE>\n</BODY>\n</HTML>');
+		d.close();
+	}
+	return undefined;
+}
+
+
+function stringSource(s) {
+	s = s.replace(/\\/g, '\\\\');
+	s = s.replace(/"/g, '\\"');
+	s = s.replace(/<\//g, '<\\/');
+	return '"'+s+'"';
+}
+
+function reloadSelf(n,m) {
+	// Need to cache these because globals go away on document.open().
+	var sa = srcArray;
+	var ss = stringSource;
+	var ct = classTables;
+	var i;
+
+	document.open();
+	// Uncomment this and comment the document.open() line above to see the reloaded page's source.
+	//var w = window.open("", "NewDoc");
+	//var d = w.document;
+	//var document = new Object;
+	//document.write = function () {
+	//	for (var i = 0; i < arguments.length; i++) {
+	//		var s = arguments[i].toString();
+	//		s = s.replace(/&/g, '&amp;');
+	//		s = s.replace(/\x3C/g, '&lt;');
+	//		s = s.replace(/>/g, '&gt;');
+	//		s = s.replace(/ /g, '&nbsp;');
+	//		d.write(s);
+	//	}
+	//};
+	//document.writeln = function () {
+	//	for (var i = 0; i < arguments.length; i++) {
+	//		var s = arguments[i].toString();
+	//		s = s.replace(/&/g, '&amp;');
+	//		s = s.replace(/\x3C/g, '&lt;');
+	//		s = s.replace(/>/g, '&gt;');
+	//		s = s.replace(/ /g, '&nbsp;');
+	//		d.write(s);
+	//	}
+	//	d.writeln('<BR>');
+	//};
+
+	document.writeln('<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">');
+	document.writeln('<HTML>\n<HEAD>\n<SCRIPT type="text/javascript">');
+
+	// Manually copy non-persistent script data
+	if (!ct.length)
+		document.writeln('var classTables = [];');
+	else {
+		document.writeln('var classTables = [');
+		for (i = 0; i < ct.length; i++) {
+			var row = ct[i];
+			document.write('[' + ss(row[0]));
+			for (var j = 1; j < row.length; j++)
+				document.write(',' + row[j]);
+			document.writeln(']' + (i == ct.length-1 ? '];' : ','));
+		}
+	}
+
+	document.writeln('var srcArray = [');
+	for (i = 0; i < sa.length; i++) {
+		document.write(ss(sa[i]));
+		if (i != sa.length-1)
+			document.writeln(',');
+	}
+	document.writeln('];');
+	document.writeln('eval(srcArray.join("\\n"));');
+	document.writeln('showMode = '+m+';');
+	document.writeln('sortColumn = '+n+';');
+	document.writeln('showHead(document);');
+	document.writeln('</SCRIPT>\n</HEAD>\n\n<BODY>\n<SCRIPT type="text/javascript">showBody(document);</SCRIPT>\n</BODY>\n</HTML>');
+	document.close();
+	return undefined;
+}