XHP comment syntax

XML comment syntax for XHP is one of the top wishlist items within the UIE group (+ some other people).

Currently, the only way to have comments inside XHP blocks is to use expressions and block comments, e.g.

  <div>{/* comment */}</div>

After this change, you can use the more familiar syntax. This syntax is limited to XHP child contexts:

  // Only single-line, doc, block comments work outside XHP
  $xhp =
    <div>
      <!--XML comments only work within XHP children contexts-->
    </div>;

Based on http://www.w3.org/TR/REC-xml/#sec-comments - except we ignore the rule about double hyphens ('--') within a comment because we don't care about SGML.

hphp/compiler/parser/hphp.tab.cpp

@@ -365,7 +365,7 @@ static void xhp_attribute(Parser *_p, Token &out, Token &type, Token &label,
                           Token &def, Token &req) {
   /**
    * The basic builtin types "bool", "int", "double", and "string" all map to
-   * T_STRING in the parser, and the parser uses always uses type code 5 for
+   * T_STRING in the parser, and the parser always uses type code 5 for
    * T_STRING. However, XHP uses different type codes for these basic builtin
    * types, so we need to fix up the type code here to make XHP happy.
    */

hphp/test/quick/xhp_comment.php

@@ -0,0 +1,42 @@
+<?php
+
+class :node {
+  private $children;
+  public function __construct($_, $children) {
+    $this->children = $children;
+  }
+  public function __toString() {
+    $o = '';
+    foreach ($this->children as $child) {
+      $o .= (string)$child;
+    }
+    $o .= count($this->children);
+    return $o;
+  }
+}
+
+$n1 =
+  <node>
+    <!--hey-->
+  </node>;
+
+$n2 =
+  <node>
+    {'a'}
+    <!--hey-->
+    <!--
+      multiline comment
+      with multiple lines
+    -->
+    {'b'}
+    <node>
+      {'c'}
+      <!--nested comment-->
+    </node>
+    <!--comment-with -- extra--hyphens----->
+  </node>;
+
+echo
+  (string)$n1.
+  (string)$n2.
+  "\n";

hphp/test/quick/xhp_comment.php.expect

@@ -0,0 +1 @@
+0abc13

hphp/util/parser/hphp.ll

@@ -172,6 +172,7 @@ static int getNextTokenType(int t) {
 %x ST_XHP_END_SINGLETON_TAG
 %x ST_XHP_END_CLOSE_TAG
 %x ST_XHP_CHILD
+%x ST_XHP_COMMENT
 
 %option stack
 
@@ -811,6 +812,20 @@ BACKQUOTE_CHARS     ("{"*([^$`\\{]|("\\"{ANY_CHAR}))|{BACKQUOTE_LITERAL_DOLLAR})
         yymore();
 }
 
+<ST_XHP_COMMENT>[^-]+ {
+        yymore();
+}
+
+<ST_XHP_COMMENT>"-->" {
+        STEPPOS;
+        yy_pop_state(yyscanner);
+        return T_COMMENT;
+}
+
+<ST_XHP_COMMENT>"-" {
+        yymore();
+}
+
 <ST_IN_SCRIPTING>"?>"{NEWLINE}? {
         if (_scanner->isHackMode()) {
           _scanner->error("Hack mode: ?> not allowed");
@@ -955,6 +970,11 @@ BACKQUOTE_CHARS     ("{"*([^$`\\{]|("\\"{ANY_CHAR}))|{BACKQUOTE_LITERAL_DOLLAR})
   return T_XHP_TAG_GT;
 }
 
+<ST_XHP_CHILD>"<!--" {
+  yy_push_state(ST_XHP_COMMENT, yyscanner);
+  yymore();
+}
+
 <ST_XHP_CHILD>[^{<]+ {
   SETTOKEN;
   return T_XHP_TEXT;

hphp/util/parser/hphp.y

@@ -286,7 +286,7 @@ static void xhp_attribute(Parser *_p, Token &out, Token &type, Token &label,
                           Token &def, Token &req) {
   /**
    * The basic builtin types "bool", "int", "double", and "string" all map to
-   * T_STRING in the parser, and the parser uses always uses type code 5 for
+   * T_STRING in the parser, and the parser always uses type code 5 for
    * T_STRING. However, XHP uses different type codes for these basic builtin
    * types, so we need to fix up the type code here to make XHP happy.
    */