title_of

cho45 · Dec 1, 2011 · 2fd7ab7 · 2fd7ab7
1 parent 07a834c
commit 2fd7ab7
Show file tree

Hide file tree

Showing 4 changed files with 75 additions and 15 deletions.
diff --git a/benchmark/benchmark.pl b/benchmark/benchmark.pl
@@ -2,7 +2,8 @@
 use strict;
 use warnings;
 use utf8;
-use Benchmark qw(:all) ;
+use Time::HiRes;
+use Benchmark qw(:all :hireswallclock) ;
 
 my $text = <<EOS;
 * Xatena
@@ -68,7 +69,7 @@
 my $markdown = Text::Markdown->new;
 my $textile  = Text::Textile->new;
 
-cmpthese(-1, {
+my $results = timethese(-1, {
 	"Text::Xatena $Text::Xatena::VERSION" => sub {
 		my $html = $thx->format($text);
 	},
@@ -83,6 +84,8 @@
 	},
 });
 
+cmpthese($results);
+
 __END__
 MacBook Air 1.8GHz Intel Core i7 / 4GB RAM
                           Rate Text::Hatena 0.20 Text::Markdown 1.000031 Text::Textile 2.12 Text::Xatena 0.14

diff --git a/lib/Text/Xatena/Inline.pm b/lib/Text/Xatena/Inline.pm
@@ -75,8 +75,7 @@ match qr<\[((?:https?|ftp)://[^\s:]+(?::\d+)?[^\s:]+)(:(?:title(?:=([^[]+))?|bar
                 $title = $self->cache->get($uri);
                 if (not defined $title) {
                     eval {
-                        my $res = $self->ua->get($uri);
-                        ($title) = ($res->decoded_content =~ qr|<title[^>]*>([^<]*)</title>|i);
+                        $title = $self->title_of($uri);
                         $self->cache->set($uri, $title, "30 days");
                     };
                     if ($@) {

diff --git a/lib/Text/Xatena/Inline/Aggressive.pm b/lib/Text/Xatena/Inline/Aggressive.pm
@@ -11,10 +11,21 @@ sub cache { $_[0]->{cache} }
 sub ua    { $_[0]->{ua} || $ua }
 
 sub new {
-	my ($class, @args) = @_;
-	my $self = $class->SUPER::new(@args);
-	$self->{aggressive} = 1;
-	$self;
+    my ($class, @args) = @_;
+    my $self = $class->SUPER::new(@args);
+    $self->{aggressive} = 1;
+    $self;
+}
+
+sub title_of {
+    my ($self, $uri) = @_;
+    my $res = $self->ua->get($uri);
+    if ($res->is_success) {
+        my ($title) = ($res->decoded_content =~ qr|<title[^>]*>([^<]*)</title>|i);
+        $title;
+    } else {
+        $res->status_line;
+    }
 }
 
 1;

diff --git a/t/04_inline_aggressive.t b/t/04_inline_aggressive.t
@@ -1,6 +1,8 @@
+use utf8;
 use strict;
 use warnings;
 use lib 't/lib';
+use Encode;
 use Text::Xatena::Test;
 use Cache::MemoryCache;
 use LWP::UserAgent;
@@ -11,12 +13,27 @@ local $Text::Xatena::Test::INLINE_ARGS = [ cache => Cache::MemoryCache->new ];
     no warnings 'redefine';
     *LWP::UserAgent::get = sub {
         my ($self, $uri) = @_;
-        local $_ = $uri;
-        if (qr|http://example.com/|) {
-            HTTP::Response->new(200, "OK", [], "<title>Example Web Page</title>");
-        } else {
-            die "unknown url";
-        }
+
+        my $res = {
+            'http://example.com/' => sub {
+                HTTP::Response->new(200, "OK", ['Content-Type' => 'text/html'], "<title>Example Web Page</title>");
+            },
+            'http://example.com/utf-8' => sub {
+                HTTP::Response->new(200, "OK", ['Content-Type' => 'text/html'], encode("utf-8", "<title>エグザンプルウェブページ</title>"));
+            },
+            'http://example.com/shift_jis' => sub {
+                HTTP::Response->new(200, "OK", ['Content-Type' => 'text/html'], encode("shift_jis", "<meta charset='shift_jis'><title>エグザンプルウェブページ</title>"));
+            },
+            'http://example.com/euc-jp' => sub {
+                HTTP::Response->new(200, "OK", ['Content-Type' => 'text/html'], encode("euc-jp", "<meta charset='euc-jp'><title>エグザンプルウェブページ</title>"));
+            },
+            'http://example.com/shift_jis_ct' => sub {
+                HTTP::Response->new(200, "OK", ['Content-Type' => 'text/html; charset=shift_jis'], encode("shift_jis", "<title>エグザンプルウェブページ</title>"));
+            },
+        };
+
+        my $sub = $res->{$uri} or BAIL_OUT("Unexpected web access");
+        $sub->();
     };
 };
 
@@ -117,7 +134,7 @@ http://example.com/
 </q>
 </p>
 
-=== test
+=== http title
 --- input
 [http://example.com/:title]
 --- expected
@@ -126,3 +143,33 @@ http://example.com/
 Example Web Page
 </a>
 </p>
+
+=== http title
+--- input
+[http://example.com/utf-8:title]
+--- expected
+<p>
+<a href="http://example.com/utf-8">
+エグザンプルウェブページ
+</a>
+</p>
+
+=== http title
+--- input
+[http://example.com/shift_jis:title]
+--- expected
+<p>
+<a href="http://example.com/shift_jis">
+エグザンプルウェブページ
+</a>
+</p>
+
+=== http title
+--- input
+[http://example.com/euc-jp:title]
+--- expected
+<p>
+<a href="http://example.com/euc-jp">
+エグザンプルウェブページ
+</a>
+</p>