- Improved parse error messages

NEWS

@@ -91,6 +91,7 @@ PHP                                                                        NEWS
   . Changed $GLOBALS into a JIT autoglobal, so it's initialized only if used.
     (this may affect opcode caches!)
   . Improved performance of @ (silence) operator.
+  . Improved parse error messages. (Felipe)
   . Simplified string offset reading. $str[1][0] is now a legal construct.
   . Added caches to eliminate repeatable run-time bindings of functions,
     classes, constants, methods and properties.

Zend/tests/bug48693.phpt

@@ -20,7 +20,7 @@ var_dump(
 --EXPECTF--
 Parse error: syntax error, unexpected '}' in %s(%d) : runtime-created function on line 1
 
-Parse error: syntax error, unexpected $end in %s(%d) : runtime-created function on line 1
+Parse error: syntax error, unexpected end of file in %s(%d) : runtime-created function on line 1
 bool(false)
 int(2)
 bool(false)

Zend/tests/function_arguments_001.phpt

@@ -5,5 +5,5 @@ Argument parsing error #001
 function foo($arg1 string) {}
 ?>
 --EXPECTF--
-Parse error: syntax error, unexpected T_STRING, expecting ')' in %sfunction_arguments_001.php on line %d
+Parse error: syntax error, unexpected 'string' (T_STRING), expecting ')' in %sfunction_arguments_001.php on line %d
 

Zend/zend_globals.h

@@ -101,6 +101,7 @@ struct _zend_compiler_globals {
 
 	HashTable *auto_globals;
 
+	zend_bool parse_error;
 	zend_bool in_compilation;
 	zend_bool short_tags;
 	zend_bool asp_tags;

Zend/zend_language_parser.y

@@ -35,6 +35,9 @@
 #include "zend_API.h"
 #include "zend_constants.h"
 
+#define YYSIZE_T size_t
+#define yytnamerr zend_yytnamerr
+static YYSIZE_T zend_yytnamerr(char*, const char*);
 
 #define YYERROR_VERBOSE
 #define YYSTYPE znode
@@ -49,107 +52,160 @@
 %pure_parser
 %expect 2
 
+%token END 0 "end of file"
 %left T_INCLUDE T_INCLUDE_ONCE T_EVAL T_REQUIRE T_REQUIRE_ONCE
+%token T_INCLUDE      "include (T_INCLUDE)"
+%token T_INCLUDE_ONCE "include_once (T_INCLUDE_ONCE)"
+%token T_EVAL         "eval (T_EVAL)"
+%token T_REQUIRE      "require (T_REQUIRE)"
+%token T_REQUIRE_ONCE "require_once (T_REQUIRE_ONCE)"
 %left ','
 %left T_LOGICAL_OR
+%token T_LOGICAL_OR   "or (T_LOGICAL_OR)"
 %left T_LOGICAL_XOR
+%token T_LOGICAL_XOR  "xor (T_LOGICAL_XOR)"
 %left T_LOGICAL_AND
+%token T_LOGICAL_AND  "and (T_LOGICAL_AND)"
 %right T_PRINT
+%token T_PRINT        "print (T_PRINT)"
 %left '=' T_PLUS_EQUAL T_MINUS_EQUAL T_MUL_EQUAL T_DIV_EQUAL T_CONCAT_EQUAL T_MOD_EQUAL T_AND_EQUAL T_OR_EQUAL T_XOR_EQUAL T_SL_EQUAL T_SR_EQUAL
+%token T_PLUS_EQUAL   "+= (T_PLUS_EQUAL)"
+%token T_MINUS_EQUAL  "-= (T_MINUS_EQUAL)"
+%token T_MUL_EQUAL    "*= (T_MUL_EQUAL)"
+%token T_DIV_EQUAL    "/= (T_DIV_EQUAL)"
+%token T_CONCAT_EQUAL ".= (T_CONCAT_EQUAL)"
+%token T_MOD_EQUAL    "%= (T_MOD_EQUAL)"
+%token T_AND_EQUAL    "&= (T_AND_EQUAL)"
+%token T_OR_EQUAL     "|= (T_OR_EQUAL)"
+%token T_XOR_EQUAL    "^= (T_XOR_EQUAL)"
+%token T_SL_EQUAL     "<<= (T_SL_EQUAL)"
+%token T_SR_EQUAL     ">>= (T_SR_EQUAL)"
 %left '?' ':'
 %left T_BOOLEAN_OR
+%token T_BOOLEAN_OR   "|| (T_BOOLEAN_OR)"
 %left T_BOOLEAN_AND 
+%token T_BOOLEAN_AND  "&& (T_BOOLEAN_AND)"
 %left '|'
 %left '^'
 %left '&'
 %nonassoc T_IS_EQUAL T_IS_NOT_EQUAL T_IS_IDENTICAL T_IS_NOT_IDENTICAL
+%token T_IS_EQUAL     "== (T_IS_EQUAL)"
+%token T_IS_NOT_EQUAL "!= (T_IS_NOT_EQUAL)"
+%token T_IS_IDENTICAL "=== (T_IS_IDENTICAL)"
+%token T_IS_NOT_IDENTICAL "!== (T_IS_NOT_IDENTICAL)"
 %nonassoc '<' T_IS_SMALLER_OR_EQUAL '>' T_IS_GREATER_OR_EQUAL
+%token T_IS_SMALLER_OR_EQUAL "<= (T_IS_SMALLER_OR_EQUAL)"
+%token T_IS_GREATER_OR_EQUAL ">= (T_IS_GREATER_OR_EQUAL)"
 %left T_SL T_SR
+%token T_SL "<< (T_SL)"
+%token T_SR ">> (T_SR)"
 %left '+' '-' '.'
 %left '*' '/' '%'
 %right '!'
 %nonassoc T_INSTANCEOF
+%token T_INSTANCEOF  "instanceof (T_INSTANCEOF)"
 %right '~' T_INC T_DEC T_INT_CAST T_DOUBLE_CAST T_STRING_CAST T_ARRAY_CAST T_OBJECT_CAST T_BOOL_CAST T_UNSET_CAST '@'
+%token T_INC "++ (T_INC)"
+%token T_DEC "-- (T_DEC)"
+%token T_INT_CAST    "(int) (T_INT_CAST)"
+%token T_SCALAR_CAST "(scalar) (T_SCALAR_CAST)"
+%token T_DOUBLE_CAST "(double) (T_DOUBLE_CAST)"
+%token T_STRING_CAST "(string) (T_STRING_CAST)"
+%token T_ARRAY_CAST  "(array) (T_ARRAY_CAST)"
+%token T_OBJECT_CAST "(object) (T_OBJECT_CAST)"
+%token T_BOOL_CAST   "(bool) (T_BOOL_CAST)"
+%token T_UNSET_CAST  "(unset) (T_UNSET_CAST)"
 %right '['
 %nonassoc T_NEW T_CLONE
-%token T_EXIT
+%token T_NEW       "new (T_NEW)"
-%token T_IF
+%token T_CLONE     "clone (T_CLONE)"
+%token T_EXIT      "exit (T_EXIT)"
+%token T_IF        "if (T_IF)"
 %left T_ELSEIF
+%token T_ELSEIF    "elseif (T_ELSEIF)"
 %left T_ELSE 
+%token T_ELSE      "else (T_ELSE)"
 %left T_ENDIF 
-%token T_LNUMBER
+%token T_ENDIF     "endif (T_ENDIF)"
-%token T_DNUMBER
+%token T_LNUMBER   "integer number (T_LNUMBER)"
-%token T_STRING
+%token T_DNUMBER   "floating-point number (T_DNUMBER)"
-%token T_STRING_VARNAME
+%token T_STRING    "identifier (T_STRING)"
-%token T_VARIABLE
+%token T_STRING_VARNAME "variable name (T_STRING_VARNAME)"
-%token T_NUM_STRING
+%token T_VARIABLE  "variable (T_VARIABLE)"
+%token T_NUM_STRING "number (T_NUM_STRING)"
 %token T_INLINE_HTML
 %token T_CHARACTER
 %token T_BAD_CHARACTER
-%token T_ENCAPSED_AND_WHITESPACE
+%token T_ENCAPSED_AND_WHITESPACE  "quoted-string and whitespace (T_ENCAPSED_AND_WHITESPACE)"
-%token T_CONSTANT_ENCAPSED_STRING
+%token T_CONSTANT_ENCAPSED_STRING "quoted-string (T_CONSTANT_ENCAPSED_STRING)"
-%token T_ECHO
+%token T_ECHO       "echo (T_ECHO)"
-%token T_DO
+%token T_DO         "do (T_DO)"
-%token T_WHILE
+%token T_WHILE      "while (T_WHILE)"
-%token T_ENDWHILE
+%token T_ENDWHILE   "endwhile (T_ENDWHILE)"
-%token T_FOR
+%token T_FOR        "for (T_FOR)"
-%token T_ENDFOR
+%token T_ENDFOR     "endfor (T_ENDFOR)"
-%token T_FOREACH
+%token T_FOREACH    "foreach (T_FOREACH)"
-%token T_ENDFOREACH
+%token T_ENDFOREACH "endforeach (T_ENDFOREACH)"
-%token T_DECLARE
+%token T_DECLARE    "declare (T_DECLARE)"
-%token T_ENDDECLARE
+%token T_ENDDECLARE "enddeclare (T_ENDDECLARE)"
-%token T_AS
+%token T_AS         "as (T_AS)"
-%token T_SWITCH
+%token T_SWITCH     "switch (T_SWITCH)"
-%token T_ENDSWITCH
+%token T_ENDSWITCH  "endswitch (T_ENDSWITCH)"
-%token T_CASE
+%token T_CASE       "case (T_CASE)"
-%token T_DEFAULT
+%token T_DEFAULT    "default (T_DEFAULT)"
-%token T_BREAK
+%token T_BREAK      "break (T_BREAK)"
-%token T_CONTINUE
+%token T_CONTINUE   "continue (T_CONTINUE)"
-%token T_GOTO
+%token T_GOTO       "goto (T_GOTO)"
-%token T_FUNCTION
+%token T_FUNCTION   "function (T_FUNCTION)"
-%token T_CONST
+%token T_CONST      "const (T_CONST)"
-%token T_RETURN
+%token T_RETURN     "return (T_RETURN)"
-%token T_TRY
+%token T_TRY        "try (T_TRY)"
-%token T_CATCH
+%token T_CATCH      "catch (T_CATCH)"
-%token T_THROW
+%token T_THROW      "throw (T_THROW)"
-%token T_USE
+%token T_USE        "use (T_USE)"
-%token T_INSTEADOF
+%token T_INSTEADOF  "insteadof (T_INSTEADOF)"
-%token T_GLOBAL
+%token T_GLOBAL     "global (T_GLOBAL)"
 %right T_STATIC T_ABSTRACT T_FINAL T_PRIVATE T_PROTECTED T_PUBLIC
-%token T_VAR
+%token T_STATIC     "static (T_STATIC)"
-%token T_UNSET
+%token T_ABSTRACT   "abstract (T_ABSTRACT)"
-%token T_ISSET
+%token T_FINAL      "final (T_FINAL)"
-%token T_EMPTY
+%token T_PRIVATE    "private (T_PRIVATE)"
-%token T_HALT_COMPILER
+%token T_PROTECTED  "protected (T_PROTECTED)"
-%token T_CLASS
+%token T_PUBLIC     "public (T_PUBLIC)"
-%token T_TRAIT
+%token T_VAR        "var (T_VAR)"
-%token T_INTERFACE
+%token T_UNSET      "unset (T_UNSET)"
-%token T_EXTENDS
+%token T_ISSET      "isset (T_ISSET)"
-%token T_IMPLEMENTS
+%token T_EMPTY      "empty (T_EMPTY)"
-%token T_OBJECT_OPERATOR
+%token T_HALT_COMPILER "__halt_compiler (T_HALT_COMPILER)"
-%token T_DOUBLE_ARROW
+%token T_CLASS      "class (T_CLASS)"
-%token T_LIST
+%token T_TRAIT      "trait (T_TRAIT)"
-%token T_ARRAY
+%token T_INTERFACE  "interface (T_INTERFACE)"
-%token T_CLASS_C
+%token T_EXTENDS    "extends (T_EXTENDS)"
-%token T_METHOD_C
+%token T_IMPLEMENTS "implements (T_IMPLEMENTS)"
-%token T_FUNC_C
+%token T_OBJECT_OPERATOR "-> (T_OBJECT_OPERATOR)"
-%token T_LINE
+%token T_DOUBLE_ARROW    "=> (T_DOUBLE_ARROW)"
-%token T_FILE
+%token T_LIST            "list (T_LIST)"
-%token T_COMMENT
+%token T_ARRAY           "array (T_ARRAY)"
-%token T_DOC_COMMENT
+%token T_CLASS_C         "__CLASS__ (T_CLASS_C)"
-%token T_OPEN_TAG
+%token T_METHOD_C        "__METHOD__ (T_METHOD_C)"
-%token T_OPEN_TAG_WITH_ECHO
+%token T_FUNC_C          "__FUNCTION__ (T_FUNC_C)"
-%token T_CLOSE_TAG
+%token T_LINE            "__LINE__ (T_LINE)"
-%token T_WHITESPACE
+%token T_FILE            "__FILE__ (T_FILE)"
-%token T_START_HEREDOC
+%token T_COMMENT         "comment (T_COMMENT)"
-%token T_END_HEREDOC
+%token T_DOC_COMMENT     "doc comment (T_DOC_COMMENT)"
-%token T_DOLLAR_OPEN_CURLY_BRACES
+%token T_OPEN_TAG        "open tag (T_OPEN_TAG)"
-%token T_CURLY_OPEN
+%token T_OPEN_TAG_WITH_ECHO "open tag with echo (T_OPEN_TAG_WITH_ECHO)"
-%token T_PAAMAYIM_NEKUDOTAYIM
+%token T_CLOSE_TAG       "close tag (T_CLOSE_TAG)"
-%token T_NAMESPACE
+%token T_WHITESPACE      "whitespace (T_WHITESPACE)"
-%token T_NS_C
+%token T_START_HEREDOC   "heredoc start (T_START_HEREDOC)"
-%token T_DIR
+%token T_END_HEREDOC     "heredoc end (T_END_HEREDOC)"
-%token T_NS_SEPARATOR
+%token T_DOLLAR_OPEN_CURLY_BRACES "${ (T_DOLLAR_OPEN_CURLY_BRACES)"
+%token T_CURLY_OPEN      "{$ (T_CURLY_OPEN)"
+%token T_PAAMAYIM_NEKUDOTAYIM ":: (T_PAAMAYIM_NEKUDOTAYIM)"
+%token T_NAMESPACE       "namespace (T_NAMESPACE)"
+%token T_NS_C            "__NAMESPACE__ (T_NS_C)"
+%token T_DIR             "__DIR__ (T_DIR)"
+%token T_NS_SEPARATOR    "\\ (T_NS_SEPARATOR)"
 
 %% /* Rules */
 
@@ -1079,6 +1135,70 @@ class_constant:
 
 %%
 
+/* Copy to YYRES the contents of YYSTR after stripping away unnecessary
+   quotes and backslashes, so that it's suitable for yyerror.  The
+   heuristic is that double-quoting is unnecessary unless the string
+   contains an apostrophe, a comma, or backslash (other than
+   backslash-backslash).  YYSTR is taken from yytname.  If YYRES is
+   null, do not copy; instead, return the length of what the result
+   would have been.  */
+static YYSIZE_T zend_yytnamerr(char *yyres, const char *yystr)
+{
+	if (!yyres) {
+		return yystrlen(yystr);
+	}
+	{
+		TSRMLS_FETCH();
+		if (CG(parse_error) == 0) {
+			char buffer[120], *end, *str, *tok1 = NULL, *tok2 = NULL;
+			unsigned int len = 0, toklen = 0, yystr_len;
+			
+			CG(parse_error) = 1;
+
+			if (LANG_SCNG(yy_text)[0] == 0 &&
+				LANG_SCNG(yy_leng) == 1 &&
+				memcmp(yystr, ZEND_STRL("\"end of file\"")) == 0) {
+				return yystpcpy(yyres, "end of file") - yyres;
+			}
+			
+			str = LANG_SCNG(yy_text);
+			end = memchr(str, '\n', LANG_SCNG(yy_leng));
+			yystr_len = yystrlen(yystr);
+			
+			if ((tok1 = memchr(yystr, '(', yystr_len)) != NULL
+				&& (tok2 = zend_memrchr(yystr, ')', yystr_len)) != NULL) {
+				toklen = (tok2 - tok1) + 1;
+			} else {
+				tok1 = tok2 = NULL;
+				toklen = 0;
+			}
+			
+			if (end == NULL) {
+				len = LANG_SCNG(yy_leng) > 30 ? 30 : LANG_SCNG(yy_leng);
+			} else {
+				len = (end - str) > 30 ? 30 : (end - str);
+			}
+			if (toklen) {
+				snprintf(buffer, sizeof(buffer), "'%.*s' %.*s", len, str, toklen, tok1);
+			} else {
+				snprintf(buffer, sizeof(buffer), "'%.*s'", len, str);
+			}
+			return yystpcpy(yyres, buffer) - yyres;
+		}		
+	}	
+	if (*yystr == '"') {
+		YYSIZE_T yyn = 0;
+		const char *yyp = yystr;
+
+		for (; *++yyp != '"'; ++yyn) {
+			yyres[yyn] = *yyp;
+		}
+		yyres[yyn] = '\0';
+		return yyn;
+	}
+	return yystpcpy(yyres, yystr) - yyres;
+}
+
 /*
  * Local variables:
  * tab-width: 4

Zend/zend_language_scanner.l

@@ -174,6 +174,7 @@ static void yy_scan_buffer(char *str, unsigned int len TSRMLS_DC)
 
 void startup_scanner(TSRMLS_D)
 {
+	CG(parse_error) = 0;
 	CG(heredoc) = NULL;
 	CG(heredoc_len) = 0;
 	CG(doc_comment) = NULL;
@@ -187,6 +188,7 @@ void shutdown_scanner(TSRMLS_D)
 		efree(CG(heredoc));
 		CG(heredoc_len)=0;
 	}
+	CG(parse_error) = 0;
 	zend_stack_destroy(&SCNG(state_stack));
 	RESET_DOC_COMMENT();
 }

Zend/zend_language_scanner_defs.h

@@ -1,4 +1,4 @@
-/* Generated by re2c 0.13.5 on Thu Jun  2 17:34:19 2011 */
+/* Generated by re2c 0.13.5 on Thu Jun 23 19:47:38 2011 */
 #line 3 "Zend/zend_language_scanner_defs.h"
 
 enum YYCONDTYPE {

main/main.c

@@ -990,7 +990,7 @@ static void php_error_cb(int type, const char *error_filename, const uint error_
 				char *append_string = INI_STR("error_append_string");
 
 				if (PG(html_errors)) {
-					if (type == E_ERROR) {
+					if (type == E_ERROR || type == E_PARSE) {
 						size_t len;
 						char *buf = php_escape_html_entities(buffer, buffer_len, &len, 0, ENT_COMPAT, NULL TSRMLS_CC);
 						php_printf("%s<br />\n<b>%s</b>:  %s in <b>%s</b> on line <b>%d</b><br />\n%s", STR_PRINT(prepend_string), error_type_str, buf, error_filename, error_lineno, STR_PRINT(append_string));
@@ -1061,7 +1061,9 @@ static void php_error_cb(int type, const char *error_filename, const uint error_
 					sapi_header_op(SAPI_HEADER_REPLACE, &ctr TSRMLS_CC);
 				}
 				/* the parser would return 1 (failure), we can bail out nicely */
-				if (type != E_PARSE) {
+				if (type == E_PARSE) {
+					CG(parse_error) = 0;
+				} else {
 					/* restore memory limit */
 					zend_set_memory_limit(PG(memory_limit));
 					efree(buffer);