fixed smart whitespace trimming bug in Mojo::DOM

mojolicious · Sep 19, 2013 · 1f82baa · 1f82baa
1 parent 8b324fa
commit 1f82baa
Show file tree

Hide file tree

Showing 4 changed files with 12 additions and 9 deletions.
diff --git a/Changes b/Changes
@@ -1,5 +1,6 @@
 
 4.40  2013-09-19
+  - Fixed smart whitespace trimming bug in Mojo::DOM.
   - Fixed table parsing bug in Mojo::DOM::HTML.
 
 4.39  2013-09-17

diff --git a/lib/Mojo/DOM.pm b/lib/Mojo/DOM.pm
@@ -329,6 +329,13 @@ sub _sibling {
 sub _text {
   my ($nodes, $recurse, $trim) = @_;
 
+  # Merge successing text nodes
+  my $i = 0;
+  while (my $next = $nodes->[$i + 1]) {
+    ++$i and next unless $nodes->[$i][0] eq 'text' && $next->[0] eq 'text';
+    splice @$nodes, $i, 2, ['text', $nodes->[$i][1] . $next->[1]];
+  }
+
   my $text = '';
   for my $n (@$nodes) {
     my $type = $n->[0];

diff --git a/lib/Mojo/DOM/HTML.pm b/lib/Mojo/DOM/HTML.pm
@@ -84,12 +84,7 @@ sub parse {
 
     # Text (and runaway "<")
     $text .= '<' if defined $runaway;
-    if (length $text) {
-      $text = html_unescape $text;
-      my $sibling = $current->[-1];
-      if (ref $sibling && $sibling->[0] eq 'text') { $sibling->[1] .= $text }
-      else { push @$current, ['text', $text] }
-    }
+    push @$current, ['text', html_unescape $text] if length $text;
 
     # DOCTYPE
     if ($doctype) { push @$current, ['doctype', $doctype] }

diff --git a/t/mojo/dom.t b/t/mojo/dom.t
@@ -1100,7 +1100,7 @@ is "$dom", <<EOF, 'right result';
 <div>D</div>works
 EOF
 $dom->at('li')->prepend_content('A3<p>A2</p>')->prepend_content('A4');
-is $dom->at('li')->text, 'A4 A3 A', 'right text';
+is $dom->at('li')->text, 'A4A3 A', 'right text';
 is "$dom", <<EOF, 'right result';
 <ul>
     24<div>A-1</div>works25<li>A4A3<p>A2</p>A</li><p>A1</p>23
@@ -1109,13 +1109,13 @@ is "$dom", <<EOF, 'right result';
 </ul>
 <div>D</div>works
 EOF
-$dom->find('li')->[1]->append_content('<p>C2</p>C3')->append_content('C4');
+$dom->find('li')->[1]->append_content('<p>C2</p>C3')->append_content(' C4');
 is $dom->find('li')->[1]->text, 'C C3 C4', 'right text';
 is "$dom", <<EOF, 'right result';
 <ul>
     24<div>A-1</div>works25<li>A4A3<p>A2</p>A</li><p>A1</p>23
     <p>B</p>
-    <li>C<p>C2</p>C3C4</li>
+    <li>C<p>C2</p>C3 C4</li>
 </ul>
 <div>D</div>works
 EOF