The Tor Browser / file comparison
comparison: gfx/tests/process-textruns.pl
gfx/tests/process-textruns.pl
- changeset 0
- 6474c204b198
equal
deleted
inserted
replaced
| -1:000000000000 | 0:0c501c72900f |
|---|---|
| 1 #!/usr/bin/perl -w | |
| 2 | |
| 3 # This Source Code Form is subject to the terms of the Mozilla Public | |
| 4 # License, v. 2.0. If a copy of the MPL was not distributed with this | |
| 5 # file, You can obtain one at http://mozilla.org/MPL/2.0/. | |
| 6 | |
| 7 # This script is a bunch of utilities for computing statistics about the textruns | |
| 8 # created during a Gecko run. | |
| 9 # | |
| 10 # Usage: | |
| 11 # 1) Uncomment #define DUMP_TEXT_RUNS in gfxAtsuiFonts.cpp | |
| 12 # 2) Build | |
| 13 # 3) Run over some test set, redirecting stdout to a file | |
| 14 # 4) Pipe that file through this script | |
| 15 | |
| 16 # --exclude-spaces-only: ignore all textruns that consistent of zero or more | |
| 17 # spaces | |
| 18 my $exclude_spaces = grep(/^--exclude-spaces-only$/, @ARGV); | |
| 19 | |
| 20 # --dump-runs: process textruns into a format that can be used by | |
| 21 # gfxTextRunPerfTest, print that on standard output, and do nothing else | |
| 22 my $dump_runs = grep(/^--dump-runs$/, @ARGV); | |
| 23 | |
| 24 # --obfuscate: ROTL13 the textrun text | |
| 25 my $obfuscate = grep(/^--obfuscate$/, @ARGV); | |
| 26 | |
| 27 my @textruns = (); | |
| 28 | |
| 29 while (<STDIN>) { | |
| 30 if (/^0x(\w+)\((.*)\) TEXTRUN "(.*)" ENDTEXTRUN$/) { | |
| 31 my %tr = ( fontgroup => $1, | |
| 32 families => $2, | |
| 33 text => $3 ); | |
| 34 push(@textruns, \%tr); | |
| 35 } elsif (/^0x(\w+)\((.*)\) TEXTRUN "(.*)$/) { | |
| 36 my %tr = ( fontgroup => $1, | |
| 37 families => $2 ); | |
| 38 my $text = $3."\n"; | |
| 39 while (<STDIN>) { | |
| 40 if (/^(.*)" ENDTEXTRUN$/) { | |
| 41 $text .= $1; | |
| 42 last; | |
| 43 } | |
| 44 $text .= $_; | |
| 45 } | |
| 46 $tr{text} = $text; | |
| 47 push(@textruns, \%tr); | |
| 48 } | |
| 49 } | |
| 50 | |
| 51 my %quote = ( "\\" => 1, "\"" => 1 ); | |
| 52 | |
| 53 sub quote_str { | |
| 54 my ($text) = @_; | |
| 55 my @chars = split(//, $text); | |
| 56 my @strs = (); | |
| 57 foreach my $c (@chars) { | |
| 58 if (ord($c) >= 0x80) { | |
| 59 $c = "\\x".sprintf("%x",ord($c)).'""'; | |
| 60 } elsif ($quote{$c}) { | |
| 61 $c = "\\$c"; | |
| 62 } elsif ($c eq "\n") { | |
| 63 $c = " "; | |
| 64 } | |
| 65 push(@strs, $c); | |
| 66 } | |
| 67 return '"'.join("", @strs).'"'; | |
| 68 } | |
| 69 | |
| 70 if ($dump_runs) { | |
| 71 foreach my $tr (@textruns) { | |
| 72 print "{ ", "e_str($tr->{families}), ",\n"; | |
| 73 my $text = $tr->{text}; | |
| 74 if ($obfuscate) { | |
| 75 $text =~ tr/a-mA-Mn-zN-Z/n-zN-Za-mA-M/; | |
| 76 } | |
| 77 print " ", "e_str($text), " },\n"; | |
| 78 } | |
| 79 exit(0); | |
| 80 } | |
| 81 | |
| 82 my %trs_by_text = (); | |
| 83 my %trs_by_text_and_fontgroup = (); | |
| 84 my %trs_by_trimmed_text_and_fontgroup = (); | |
| 85 my @tr_lengths = (); | |
| 86 | |
| 87 $trs_by_text{" "} = []; | |
| 88 $trs_by_text{""} = []; | |
| 89 | |
| 90 sub trim { | |
| 91 my ($s) = @_; | |
| 92 $s =~ s/^ *//g; | |
| 93 $s =~ s/ *$//g; | |
| 94 return $s; | |
| 95 } | |
| 96 | |
| 97 my $total_textruns = 0; | |
| 98 | |
| 99 foreach my $tr (@textruns) { | |
| 100 if ($exclude_spaces && $tr->{text} =~ /^ *$/) { | |
| 101 next; | |
| 102 } | |
| 103 ++$total_textruns; | |
| 104 push(@{$trs_by_text{$tr->{text}}}, $tr); | |
| 105 push(@{$trs_by_text_and_fontgroup{$tr->{fontgroup}.$tr->{text}}}, $tr); | |
| 106 push(@{$trs_by_trimmed_text_and_fontgroup{$tr->{fontgroup}.&trim($tr->{text})}}, $tr); | |
| 107 if (1 < scalar(@{$trs_by_trimmed_text_and_fontgroup{$tr->{fontgroup}.&trim($tr->{text})}})) { | |
| 108 $tr_lengths[length($tr->{text})]++; | |
| 109 } | |
| 110 } | |
| 111 | |
| 112 print "Number of textruns:\t$total_textruns\n"; | |
| 113 print "Number of textruns which are one space:\t", scalar(@{$trs_by_text{" "}}), "\n"; | |
| 114 print "Number of textruns which are empty:\t", scalar(@{$trs_by_text{""}}), "\n"; | |
| 115 | |
| 116 my $count = 0; | |
| 117 foreach my $k (keys(%trs_by_text)) { | |
| 118 if ($k =~ /^ *$/) { | |
| 119 $count += @{$trs_by_text{$k}}; | |
| 120 } | |
| 121 } | |
| 122 print "Number of textruns which are zero or more spaces:\t$count\n"; | |
| 123 | |
| 124 print "Number of unique textruns by text and fontgroup:\t", scalar(keys(%trs_by_text_and_fontgroup)), "\n"; | |
| 125 print "Number of unique textruns by trimmed text and fontgroup:\t", scalar(keys(%trs_by_trimmed_text_and_fontgroup)), "\n"; | |
| 126 | |
| 127 my $sum = 0; | |
| 128 my $weighted_sum = 0; | |
| 129 if (1) { | |
| 130 print "Textrun length distribution:\n"; | |
| 131 for my $i (0..(scalar(@tr_lengths)-1)) { | |
| 132 my $amount = defined($tr_lengths[$i])?$tr_lengths[$i]:0; | |
| 133 $sum += $amount; | |
| 134 $weighted_sum += $i*$amount; | |
| 135 print "$i\t$sum\t$weighted_sum\n"; | |
| 136 } | |
| 137 } |
