Fix CSS selector parsing with tab characters

TFBW · TFBW · commit 0c6f4fd9203c · 2025-08-02T00:55:49.000+10:00
diff --git a/lib/Mojo/DOM/CSS.pm b/lib/Mojo/DOM/CSS.pm
@@ -90,13 +90,13 @@ sub _compile {
     if ($css =~ /\G\s*,\s*/gc) { push @$group, [] }
 
     # Combinator
-    elsif ($css =~ /\G\s*([ >+~])\s*/gc) {
+    elsif ($css =~ /\G\s*([ \t>+~])\s*/gc) {
       push @$last,      ['pc', 'scope'] unless @$last;
       push @$selectors, $1;
     }
 
     # Class or ID
-    elsif ($css =~ /\G([.#])((?:$ESCAPE_RE\s|\\.|[^,.#:[ >~+])+)/gco) {
+    elsif ($css =~ /\G([.#])((?:$ESCAPE_RE\s|\\.|[^,.#:[ \t>~+])+)/gco) {
       my ($name, $op) = $1 eq '.' ? ('class', '~') : ('id', '');
       push @$last, ['attr', _name($name), _value($op, $2)];
     }
@@ -124,7 +124,7 @@ sub _compile {
     }
 
     # Tag
-    elsif ($css =~ /\G((?:$ESCAPE_RE\s|\\.|[^,.#:[ >~+])+)/gco) {
+    elsif ($css =~ /\G((?:$ESCAPE_RE\s|\\.|[^,.#:[ \t>~+])+)/gco) {
       my $alias = (my $name = $1) =~ s/^([^|]*)\|// && $1 ne '*' ? $1                                  : undef;
       my $ns    = length $alias                                  ? $ns{$alias} // return [['invalid']] : $alias;
       push @$last, ['tag', $name eq '*' ? undef : _name($name), _unescape($ns)];
diff --git a/t/mojo/dom.t b/t/mojo/dom.t
@@ -3100,4 +3100,15 @@ subtest 'Unknown CSS selector' => sub {
   like $@, qr/Unknown CSS selector: p\[/, 'right error';
 };
 
+subtest 'Handle tab in selector' => sub {
+  my $dom = Mojo::DOM->new(<<EOF);
+<!DOCTYPE html>
+<ul> <li>Ax1</li> </ul>
+EOF
+  for my $selector ("ul li", "ul\tli", "ul \tli", "ul\t li") {
+    is_deeply $dom->find($selector)->map(sub { $_->to_string })->to_array, ['<li>Ax1</li>'],
+      'selector "' . $selector =~ s/\t/\\t/r . '"';
+  }
+};
+
 done_testing();
