Add some special cases for parser errors

* unexpected token """ -> unexpected double-quote mark
* unexpected quoted string "foo" -> unexpected single-quoted string
  "foo" / unexpected double-quoted string "foo"
* unexpected illegal character "_" -> unexpected character 0xNN
  (where _ is almost certainly a control character, and NN is the
   hexadecimal value of the byte)

Author: IMSoP

Zend/zend_language_parser.y

@@ -1454,6 +1454,15 @@ static YYSIZE_T zend_yytnamerr(char *yyres, const char *yystr)
 			return sizeof("token \"\\\"")-1;
 		}
 
+		/* Avoid unreadable """ */
+		/* "'" would theoretically be just as bad, but is never currently parsed as a separate token */
+		if (strcmp(toktype, "'\"'") == 0) {
+			if (yyres) {
+				yystpcpy(yyres, "double-quote mark");
+			}
+			return sizeof("double-quote mark");
+		}
+
 		/* Strip off the outer quote marks */
 		if (toktype_len >= 2 && *toktype == '"') {
 			toktype++;
@@ -1475,12 +1484,34 @@ static YYSIZE_T zend_yytnamerr(char *yyres, const char *yystr)
 		tokcontent = LANG_SCNG(yy_text);
 		tokcontent_len = LANG_SCNG(yy_leng);
 
+		/* For T_BAD_CHARACTER, the content probably won't be a printable char */
+		/* Also, "unexpected invalid character" sounds a bit redundant */
+		if (tokcontent_len == 1 && strcmp(yystr, "\"invalid character\"") == 0) {
+			if (yyres) {
+				snprintf(buffer, sizeof(buffer), "character 0x%02hhX", *tokcontent);
+				yystpcpy(yyres, buffer);
+			}
+			return sizeof("character 0x00")-1;
+		}
+
 		/* Truncate at line end to avoid messing up log formats */
 		tokcontent_end = memchr(tokcontent, '\n', LANG_SCNG(yy_leng));
 		if (tokcontent_end != NULL) {
 			tokcontent_len = (tokcontent_end - tokcontent);
 		}
 
+		/* Try to be helpful about what kind of string was found, before stripping the quotes */
+		if (tokcontent_len > 0 && strcmp(yystr, "\"quoted string\"") == 0) {
+			if (*tokcontent=='"') {
+				toktype="double-quoted string";
+				toktype_len=sizeof("double-quoted string")-1;
+			}
+			else if (*tokcontent=='\'') {
+				toktype="single-quoted string";
+				toktype_len=sizeof("single-quoted string")-1;
+			}
+		}
+
 		/* For quoted strings, strip off another layer of quotes to avoid putting quotes inside quotes */
 		if (tokcontent_len > 0 && (*tokcontent=='\'' || *tokcontent=='"'))  {
 			tokcontent++;