Make the SGML scanner (optionally) keep track of line numbers

A new line is either \n or \f. The main logic for counting lines is in
skip_sgml{,_chars,_space}. For the general case where line numbers are not
wanted the code tries to avoid the extra checks for newlines.

This will be useful for reporting errors when loading the XBEL file.

src/dom/scanner.c

@@ -154,7 +154,7 @@ init_dom_scanner_info(struct dom_scanner_info *scanner_info)
 
 void
 init_dom_scanner(struct dom_scanner *scanner, struct dom_scanner_info *scanner_info,
-		 struct dom_string *string, int state)
+		 struct dom_string *string, int state, int count_lines)
 {
 	if (!scanner_info->initialized) {
 		init_dom_scanner_info(scanner_info);
@@ -169,5 +169,7 @@ init_dom_scanner(struct dom_scanner *scanner, struct dom_scanner_info *scanner_i
 	scanner->current = scanner->table;
 	scanner->info = scanner_info;
 	scanner->state = state;
+	scanner->count_lines = !!count_lines;
+	scanner->lineno = scanner->count_lines;
 	scanner->info->scan(scanner);
 }

src/dom/scanner.h

@@ -92,7 +92,7 @@ struct dom_scanner_info {
 
 /* Initializes the scanner. */
 void init_dom_scanner(struct dom_scanner *scanner, struct dom_scanner_info *scanner_info,
-		      struct dom_string *string, int state);
+		      struct dom_string *string, int state, int count_lines);
 
 /* The number of tokens in the scanners token table:
  * At best it should be big enough to contain properties with space separated
@@ -123,6 +123,9 @@ struct dom_scanner {
 	int line;
 #endif
 
+	unsigned int count_lines:1;
+	unsigned int lineno;
+
 	/* Some state indicator only meaningful to the scanner internals */
 	int state;
 

src/dom/select.c

@@ -391,7 +391,7 @@ parse_dom_select(struct dom_select *select, struct dom_stack *stack,
 	struct dom_scanner scanner;
 	struct dom_select_node sel;
 
-	init_dom_scanner(&scanner, &dom_css_scanner_info, string, 0);
+	init_dom_scanner(&scanner, &dom_css_scanner_info, string, 0, 0);
 
 	memset(&sel, 0, sizeof(sel));
 

src/dom/sgml/parser.c

@@ -296,7 +296,8 @@ parse_sgml_plain(struct dom_stack *stack, struct dom_scanner *scanner)
 				struct dom_scanner attr_scanner;
 
 				init_dom_scanner(&attr_scanner, &sgml_scanner_info,
-						 &token->string, SGML_STATE_ELEMENT);
+						 &token->string, SGML_STATE_ELEMENT,
+						 scanner->count_lines);
 
 				if (dom_scanner_has_tokens(&attr_scanner))
 					parse_sgml_attributes(stack, &attr_scanner);
@@ -361,7 +362,7 @@ sgml_parsing_push(struct dom_stack *stack, struct dom_node *node, void *data)
 	parsing->depth = parser->stack.depth;
 	get_dom_stack_top(&parser->stack)->immutable = 1;
 	init_dom_scanner(&parsing->scanner, &sgml_scanner_info, &node->string,
-			 SGML_STATE_TEXT);
+			 SGML_STATE_TEXT, 0);
 }
 
 static void

src/dom/sgml/scanner.c

@@ -36,7 +36,7 @@ static struct dom_scan_table_info sgml_scan_table_info[] = {
 	DOM_SCAN_TABLE_STRING("-_:.",	 SGML_CHAR_IDENT | SGML_CHAR_ENTITY),
 	DOM_SCAN_TABLE_STRING("#",	 SGML_CHAR_ENTITY),
 	DOM_SCAN_TABLE_STRING(" \f\n\r\t\v", SGML_CHAR_WHITESPACE),
-	DOM_SCAN_TABLE_STRING("\f\n\r",	 SGML_CHAR_NEWLINE),
+	DOM_SCAN_TABLE_STRING("\f\n",	 SGML_CHAR_NEWLINE),
 	DOM_SCAN_TABLE_STRING("<&",	 SGML_CHAR_NOT_TEXT),
 	DOM_SCAN_TABLE_STRING("<=>",	 SGML_CHAR_NOT_ATTRIBUTE),
 
@@ -74,6 +74,7 @@ struct dom_scanner_info sgml_scanner_info = {
 #define	is_sgml_ident(c)	check_sgml_table(c, SGML_CHAR_IDENT)
 #define	is_sgml_entity(c)	check_sgml_table(c, SGML_CHAR_ENTITY)
 #define	is_sgml_space(c)	check_sgml_table(c, SGML_CHAR_WHITESPACE)
+#define	is_sgml_newline(c)	check_sgml_table(c, SGML_CHAR_NEWLINE)
 #define	is_sgml_text(c)		!check_sgml_table(c, SGML_CHAR_NOT_TEXT)
 #define	is_sgml_token_start(c)	check_sgml_table(c, SGML_CHAR_TOKEN_START)
 #define	is_sgml_attribute(c)	!check_sgml_table(c, SGML_CHAR_NOT_ATTRIBUTE | SGML_CHAR_WHITESPACE)
@@ -83,7 +84,16 @@ skip_sgml_space(struct dom_scanner *scanner, unsigned char **string)
 {
 	unsigned char *pos = *string;
 
-	scan_sgml(scanner, pos, SGML_CHAR_WHITESPACE);
+	if (!scanner->count_lines) {
+		scan_sgml(scanner, pos, SGML_CHAR_WHITESPACE);
+	} else {
+		while (pos < scanner->end && is_sgml_space(*pos)) {
+			if (is_sgml_newline(*pos))
+				scanner->lineno++;
+			pos++;
+		}
+	}
+
 	*string = pos;
 }
 
@@ -157,11 +167,26 @@ static inline unsigned char *
 skip_sgml_chars(struct dom_scanner *scanner, unsigned char *string,
 		unsigned char skipto)
 {
+	int newlines;
+
 	assert(string >= scanner->position && string <= scanner->end);
 
-	for (; string < scanner->end; string++) {
-		if (*string == skipto)
+	if (!scanner->count_lines) {
+		size_t length = scanner->end - string;
+
+		return memchr(string, skipto, length);
+	}
+
+	for (newlines = 0; string < scanner->end; string++) {
+		if (is_sgml_newline(*string))
+			newlines++;
+		if (*string == skipto) {
+			/* Only count newlines if we actually find the
+			 * requested char. Else callers are assumed to discard
+			 * the scanning. */
+			scanner->lineno += newlines;
 			return string;
+		}
 	}
 
 	return NULL;
@@ -189,6 +214,9 @@ skip_sgml(struct dom_scanner *scanner, unsigned char **string, unsigned char ski
 
 			end = skip_sgml_chars(scanner, pos + 1, *pos);
 			if (end) pos = end;
+
+		} else if (scanner->count_lines && is_sgml_newline(*pos)) {
+			scanner->lineno++;
 		}
 	}