/*
* Parse succesive blocks of XML data, generating events for the
* handlers/callbacks as we go. State is maintained in the
* simple_xml_parser object.
* If the top level XML document ends before the last character,
* the "read" parameter indicates how much input was consumed.
*/
hcerr_t xml_parse(xml_parser *formal_parser, char s[], hc_long_t size, hc_long_t *read){
simple_xml_parser *parser = (simple_xml_parser *) formal_parser;
int i = 0;
if (DEBUG == TRUE){
print_state(parser->state, parser->depth);
printf ("in parser with " LL_FORMAT " %s\n", size, s);
fflush(stdout);
}
while (i < size){
switch(parser->state){
case OUTSIDE_ELEMENT:
if (is_white_space(s[i])){
/*skip_white_space */
break;
}
if (s[i] == '<'){
parser->start_tag = TRUE;
change_state(&parser->state, EXPECTING_OPEN_OR_CLOSE_TAG, parser->depth);
}
else {
HC_ERR_LOG(("Expected '<', read %c at %d %s\n", s[i], i, s));
return HCERR_XML_EXPECTED_LT;
}
break;
case DOCUMENT_ELEMENT:
/* discard document element */
if (s[i] != '>'){
if (DEBUG == TRUE)
printf("discarding %c", s[i]);
break;
}
else{
parser->state = OUTSIDE_ELEMENT;
break;
}
case EXPECTING_OPEN_OR_CLOSE_TAG:
if (is_white_space(s[i])){
/*skip_white_space */
break;
}
if (s[i] == '/'){
if (DEBUG)
printf("parser->start_tag = FALSE\n");
parser->start_tag = FALSE;
break;
}
case EXPECTING_TAG:
if (is_name_first_char(s[i]) == TRUE){
change_state(&parser->state, SCANNING_TAG, parser->depth);
require_ok(token_append(&parser->buffer_list, s[i]));
break;
}
/* Discard document element */
else if (s[i] == '?' && parser->depth == 0){
parser->state = DOCUMENT_ELEMENT;
break;
}
else{
HC_ERR_LOG(("Invalid first character for element name : %c %d %s\n", s[i], i, s));
return HCERR_XML_INVALID_ELEMENT_TAG;
}
// FALLTHRU INTENTIONAL???
/* Start tag is terminated by whitespace, /, or >
End tag is terminated by whitespace or >
*/
case SCANNING_TAG:
/* Still reading token */
if (is_name_char(s[i]) == TRUE){
require_ok(token_append(&parser->buffer_list, s[i]));
break;
}
else if (is_white_space(s[i]) == TRUE) {
parser->current_tag = token_finish(&parser->buffer_list);
if (parser->start_tag == TRUE){
/*printf("Start element: %s\n", parser->current_tag);*/
change_state(&parser->state, SCANNING_ATTRIBUTES, parser->depth);
break;
}
else{
change_state(&parser->state, SCANNING_CLOSE_TAG, parser->depth);
break;
}
}
else if (s[i] == '>') {
if (DEBUG == TRUE)
printf("parser->depth: %d\n", parser->depth);
require_ok(close_tag(&i, parser));
if (DEBUG == TRUE)
printf("parser->depth: %d\n", parser->depth);
if (parser->depth == 0){
*read = i + 1;
return HCERR_OK;
}
}
/* <element/> */
else if (s[i] == '/' && parser->start_tag == TRUE) {
if (DEBUG == TRUE){
printf("Start element: %s\n", parser->current_tag);
printf("End element: %s\n", parser->current_tag);
}
change_state(&parser->state, EXPECTING_RIGHT_BRACKET, parser->depth);
break;
}
else {
HC_ERR_LOG(("Invalid character '%c' in tag. %i %s\n", s[i], i, s));
return HCERR_XML_INVALID_ELEMENT_TAG;
}
break;
case EXPECTING_RIGHT_BRACKET:
if (s[i] != '>') {
HC_ERR_LOG(("Unexpected character %c after close element. %d %s", s[i], i, s));
return HCERR_XML_MALFORMED_START_ELEMENT;
}
if (parser->depth == 0){
*read = i + 1;
return HCERR_OK;
}
change_state(&parser->state, OUTSIDE_ELEMENT, parser->depth);
break;
case SCANNING_CLOSE_TAG:
if (is_white_space(s[i])) {
break;
}
if (DEBUG == TRUE)
fprintf(stdout, "End element: %s\n", parser->current_tag);
if (s[i] != '>') {
HC_ERR_LOG(("Unexpected character %c after close element. %d %s", s[i], i, s));
return HCERR_XML_MALFORMED_END_ELEMENT;
}
require_ok((*parser->end_element_callback)(parser->current_tag, parser->data));
parser->depth--;
if (parser->depth == 0){
*read = i + 1;
return HCERR_OK;
}
change_state(&parser->state, OUTSIDE_ELEMENT, parser->depth);
break;
/* Expected tokens:
* attribute_name
* '/'
* >
*/
case SCANNING_ATTRIBUTES:
if (is_white_space(s[i])){
/*skip_white_space */
break;
}
if (is_name_first_char(s[i]) == TRUE) {
change_state(&parser->state, SCANNING_ATTRIBUTE_NAME, parser->depth);
require_ok(token_append(&parser->buffer_list, s[i]));
}
else if (s[i] == '/' && parser->start_tag == TRUE) {
if (DEBUG == TRUE){
int j = 0;
printf("SA Start element: %s\n", parser->current_tag);
fprintf(stdout, "Start element: %s %d\n", parser->current_tag, parser->current_attribute);
for (j = 0; j < parser->current_attribute; j++){
printf(" %s=\"%s\"", *(parser->attribute_names + j), *(parser->attribute_values + j));
}
fprintf(stdout, "End element: %s\n", parser->current_tag);
fflush(stdout);
}
require_ok((*parser->start_element_callback)(parser->current_tag,
parser->data,
parser->attribute_names,
parser->attribute_values,
parser->current_attribute));
require_ok((*parser->end_element_callback)(parser->current_tag, parser->data));
parser->current_attribute = 0;
change_state(&parser->state, EXPECTING_RIGHT_BRACKET, parser->depth);
}
else if (s[i] == '>') {
if (DEBUG == TRUE){
int j = 0;
fprintf(stdout, "Start element event: %s %d\n", parser->current_tag, parser->current_attribute);
for (j = 0; j < parser->current_attribute; j++){
printf(" %s=\"%s\"", *(parser->attribute_names + j), *(parser->attribute_values + j));
}
}
require_ok((*parser->start_element_callback)(parser->current_tag,
parser->data,
parser->attribute_names,
parser->attribute_values,
parser->current_attribute));
parser->current_attribute = 0;
parser->depth++;
change_state(&parser->state, OUTSIDE_ELEMENT, parser->depth);
}
else{
HC_ERR_LOG(("Unexpected character %c after close element. %d %s", s[i], i, s));
return HCERR_XML_MALFORMED_START_ELEMENT;
}
break;
case SCANNING_ATTRIBUTE_NAME:
if (s[i] == '='){
if (parser->current_attribute == parser->attribute_arrays_size){
require_ok(grow_attribute_arrays(parser));
}
parser->attribute_names[parser->current_attribute] = token_finish(&parser->buffer_list);
change_state(&parser->state, SCANNING_START_ATTRIBUTE_VALUE, parser->depth);
}
else if (is_name_char(s[i]) == TRUE) {
require_ok(token_append(&parser->buffer_list, s[i]));
}
else{
HC_ERR_LOG(("Illegal char %c in attribute name. %i <<%s>>\n", s[i], i, s));
return HCERR_XML_BAD_ATTRIBUTE_NAME;
}
break;
case SCANNING_START_ATTRIBUTE_VALUE:
if (is_white_space(s[i])){
break;
}
else if (s[i] != '"'){
HC_ERR_LOG(("Attribute value does not begin with quote: '%c'. %i %s\n", s[i], i, s));
return HCERR_XML_BAD_ATTRIBUTE_NAME;
}
change_state(&parser->state, SCANNING_ATTRIBUTE_VALUE, parser->depth);
break;
case SCANNING_ATTRIBUTE_VALUE:
if (s[i] == '\\') {
if (parser->backslash == TRUE){
parser->backslash = FALSE;
}
else{
parser->backslash = TRUE;
}
}
else if (s[i] == '"' && parser->backslash == FALSE) {
parser->attribute_values[parser->current_attribute++] = token_finish(&parser->buffer_list);
change_state(&parser->state, SCANNING_ATTRIBUTES, parser->depth);
break;
}
require_ok(token_append(&parser->buffer_list, s[i]));
break;
}
i++;
}
return HCERR_OK;
}
int tokenize(struct token_list* tk_list, char* file_buffer)
{
enum Status status;
line_num;
size_t token_begin, token_end;
token_begin = 0, token_end = 0;
status = STATUS_INVALID;
str_toupper(file_buffer);
/*
* Careful: it seems an error to let "i <= len",
* but we need one more execution to flush the last token into token list.
*/
size_t line_num = 1;
for (size_t i = 0; ; ++i) {
struct token_node* tok_node;
switch (status) {
case STATUS_LETTER:
if (!IS_LETTER(file_buffer[i]) && !IS_DIGIT(file_buffer[i])) {
token_end = i;
tok_node = create_token(TOKEN_LABEL, file_buffer + token_begin, token_end - token_begin);
tok_node->type = letter_type(tok_node->liter, tok_node->len);
token_append(tk_list, tok_node);
token_begin = i;
status = next_status(status, file_buffer[i]);
}
break;
case STATUS_PRAGMA:
if (!IS_LETTER(file_buffer[i])) {
int type;
token_end = i;
type = pragma_type(file_buffer + token_begin, token_end - token_begin);
if (type < 0) {
error("invalid pragma ad line %d\n", line_num);
return -4;
}
tok_node = create_token(type, file_buffer + token_begin, token_end - token_begin);
token_append(tk_list, tok_node);
token_begin = i;
status = next_status(status, file_buffer[i]);
}
break;
case STATUS_PUNCTUATION:
token_end = i;
tok_node = create_token(file_buffer[token_begin], file_buffer + token_begin, token_end - token_begin);
token_append(tk_list, tok_node);
token_begin = i;
status = next_status(status, file_buffer[i]);
break;
case STATUS_NUMBER:
if (!IS_NUMBER(file_buffer[i])) {
token_end = i;
if (!check_number(file_buffer + token_begin, token_end - token_begin)) {
error("invalid number format at line %d\n", line_num);
return -2;
}
tok_node = create_token(TOKEN_NUMBER, file_buffer + token_begin, token_end - token_begin);
tok_node->data = parse_number(tok_node->liter);
token_append(tk_list, tok_node);
token_begin = i;
status = next_status(status, file_buffer[i]);
}
break;
case STATUS_BLANK:
if (!IS_BLANK(file_buffer[i])) {
token_begin = i;
status = next_status(status, file_buffer[i]);
}
break;
case STATUS_COMMENTS:
//once status is in comments, it will always be in comments
if ('\n' == file_buffer[i]) {
token_begin = i;
status = next_status(status, file_buffer[i]);
}
break;
case STATUS_INVALID:
token_begin = i;
status = next_status(status, file_buffer[i]);
if (STATUS_INVALID == status && 0 != file_buffer[i]) {
error("invalid format at line %d\n", line_num);
return -3;
}
break;
}
if (0 == file_buffer[i])
break;
else if ('\n' == file_buffer[i])
++line_num;
}
return 0;
}