improved encode/decode performance in Mojo::Util with a cache

mojolicious · Nov 25, 2012 · ad9c717 · ad9c717
1 parent cf0855f
commit ad9c717
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 10 deletions.
diff --git a/Changes b/Changes
@@ -1,6 +1,7 @@
 
 3.62  2012-11-26
   - Improved compatibility with IO::Socket::SSL 1.79.
+  - Improved encode/decode performance in Mojo::Util by using a cache.
   - Improved tests.
   - Fixed clone bugs in Mojo::URL.
 

diff --git a/lib/Mojo/Util.pm b/lib/Mojo/Util.pm
@@ -4,7 +4,7 @@ use Mojo::Base 'Exporter';
 use Carp 'croak';
 use Digest::MD5 qw(md5 md5_hex);
 use Digest::SHA qw(sha1 sha1_hex);
-use Encode ();
+use Encode 'find_encoding';
 use File::Basename 'dirname';
 use File::Spec::Functions 'catfile';
 use MIME::Base64 qw(decode_base64 encode_base64);
@@ -37,6 +37,9 @@ $REVERSE{$ENTITIES{$_}} //= $_
   for sort { @{[$a =~ /[A-Z]/g]} <=> @{[$b =~ /[A-Z]/g]} }
   sort grep {/;/} keys %ENTITIES;
 
+# Encoding cache
+my %CACHE;
+
 our @EXPORT_OK = (
   qw(b64_decode b64_encode camelize class_to_file class_to_path decamelize),
   qw(decode encode get_line hmac_md5_sum hmac_sha1_sum html_escape),
@@ -88,11 +91,11 @@ sub decamelize {
 sub decode {
   my ($encoding, $bytes) = @_;
   return undef
-    unless eval { $bytes = Encode::decode($encoding, $bytes, 1); 1 };
+    unless eval { $bytes = _encoding($encoding)->decode("$bytes", 1); 1 };
   return $bytes;
 }
 
-sub encode { Encode::encode(shift, shift) }
+sub encode { _encoding($_[0])->encode("$_[1]") }
 
 sub get_line {
 
@@ -370,6 +373,10 @@ sub _encode {
   return exists $REVERSE{$_[0]} ? "&$REVERSE{$_[0]}" : "&#@{[ord($_[0])]};";
 }
 
+sub _encoding {
+  $CACHE{$_[0]} //= find_encoding($_[0]) // croak "Unknown encoding '$_[0]'";
+}
+
 sub _hmac {
   my ($hash, $string, $secret) = @_;
 

diff --git a/t/mojo/util.t b/t/mojo/util.t
@@ -65,21 +65,29 @@ is b64_encode('foobar$%^&3217'), "Zm9vYmFyJCVeJjMyMTc=\n",
 is b64_decode("Zm9vYmFyJCVeJjMyMTc=\n"), 'foobar$%^&3217',
   'right base64 decoded result';
 
-# UTF-8 b64_encode
+# b64_encode (UTF-8)
 is b64_encode(encode 'UTF-8', "foo\x{df}\x{0100}bar%23\x{263a}"),
   "Zm9vw5/EgGJhciUyM+KYug==\n", 'right base64 encoded result';
 
-# UTF-8 b64_decode
+# b64_decode (UTF-8)
 is decode('UTF-8', b64_decode "Zm9vw5/EgGJhciUyM+KYug==\n"),
   "foo\x{df}\x{0100}bar%23\x{263a}", 'right base64 decoded result';
 
 # b64_encode (custom line ending)
 is b64_encode('foobar$%^&3217', ''), 'Zm9vYmFyJCVeJjMyMTc=',
   'right base64 encoded result';
 
-# Decode invalid UTF-8
+# decode (invalid UTF-8)
 is decode('UTF-8', "\x{1000}"), undef, 'decoding invalid UTF-8 worked';
 
+# decode (invalid encoding)
+is decode('does_not_exist', ''), undef,
+  'decoding with invalid encoding worked';
+
+# encode (invalid encoding)
+eval { encode('does_not_exist', '') };
+like $@, qr/Unknown encoding 'does_not_exist'/, 'right error';
+
 # url_escape
 is url_escape('business;23'), 'business%3B23', 'right url escaped result';
 
@@ -128,11 +136,11 @@ is html_unescape('foobar'), 'foobar', 'right html unescaped result';
 is html_unescape('&Ltf&amp&0oo&nbspba;&ltr'), "&Ltf&&0oo\x{00a0}ba;<r",
   'right html unescaped result';
 
-# UTF-8 html_escape
+# html_escape (UTF-8)
 is html_escape("fo\nobar<baz>&\"\x{152}\x{02ae4}"),
   "fo\nobar&lt;baz&gt;&amp;&quot;&OElig;&Dashv;", 'right html escaped result';
 
-# UTF-8 html_unescape
+# html_unescape (UTF-8)
 is html_unescape(decode 'UTF-8', 'foo&lt;baz&gt;&#x26;&#34;&OElig;&Foo;'),
   "foo<baz>&\"\x{152}&Foo;", 'right html unescaped result';
 
@@ -149,10 +157,10 @@ is xml_escape(qq{la<f>\nbar"baz"'yada\n'&lt;la}),
   "la&lt;f&gt;\nbar&quot;baz&quot;&#39;yada\n&#39;&amp;lt;la",
   'right xml escaped result';
 
-# UTF-8 xml_escape with nothing to escape
+# xml_escape (UTF-8 with nothing to escape)
 is xml_escape('привет'), 'привет', 'right xml escaped result';
 
-# UTF-8 xml_escape
+# xml_escape (UTF-8)
 is xml_escape('привет<foo>'), 'привет&lt;foo&gt;',
   'right xml escaped result';