int
main(int argc, char *argv[])
{
FILE *in;
char *prog, *opath;
prog = argv[1];
opath = 0;
in = stdin;
argc--;
argv++;
while (argc > 0) {
if (argv[0][0] == '-') {
switch (argv[0][1]) {
case 'o':
argc--;
argv++;
opath = argv[0];
break;
case 'x':
argc--;
argv++;
if ((in = fopen(argv[0], "r")) == 0)
fprintf(stderr,
"%s: unable to open composition exclusion file %s\n",
prog, argv[0]);
else {
read_compexdata(in);
fclose(in);
in = 0;
}
break;
default:
usage(prog);
}
} else {
if (in != stdin && in != NULL)
fclose(in);
if ((in = fopen(argv[0], "r")) == 0)
fprintf(stderr, "%s: unable to open ctype file %s\n",
prog, argv[0]);
else {
read_cdata(in);
fclose(in);
in = 0;
}
}
argc--;
argv++;
}
if (opath == 0)
opath = ".";
write_cdata(opath);
return 0;
}
/* Entered after the '<' and the first character after that. Returns status
* code.
*/
static void
read_element(PInfo pi) {
struct _Attr attrs[MAX_ATTRS];
Attr ap = attrs;
char *name;
char *ename;
char *end;
char c;
long elen;
int hasChildren = 0;
int done = 0;
ename = read_name_token(pi);
end = pi->s;
elen = end - ename;
next_non_white(pi);
c = *pi->s;
*end = '\0';
if ('/' == c) {
/* empty element, no attributes and no children */
pi->s++;
if ('>' != *pi->s) {
//printf("*** '%s' ***\n", pi->s);
raise_error("invalid format, element not closed", pi->str, pi->s);
}
pi->s++; /* past > */
ap->name = 0;
pi->pcb->add_element(pi, ename, attrs, hasChildren);
pi->pcb->end_element(pi, ename);
return;
}
/* read attribute names until the close (/ or >) is reached */
while (!done) {
if ('\0' == c) {
next_non_white(pi);
c = *pi->s;
}
switch (c) {
case '\0':
raise_error("invalid format, document not terminated", pi->str, pi->s);
case '/':
// Element with just attributes.
pi->s++;
if ('>' != *pi->s) {
raise_error("invalid format, element not closed", pi->str, pi->s);
}
pi->s++;
ap->name = 0;
pi->pcb->add_element(pi, ename, attrs, hasChildren);
pi->pcb->end_element(pi, ename);
return;
case '>':
// has either children or a value
pi->s++;
hasChildren = 1;
done = 1;
ap->name = 0;
pi->pcb->add_element(pi, ename, attrs, hasChildren);
break;
default:
// Attribute name so it's an element and the attribute will be
// added to it.
ap->name = read_name_token(pi);
end = pi->s;
next_non_white(pi);
if ('=' != *pi->s++) {
raise_error("invalid format, no attribute value", pi->str, pi->s);
}
*end = '\0'; // terminate name
// read value
next_non_white(pi);
ap->value = read_quoted_value(pi);
if (0 != strchr(ap->value, '&')) {
if (0 != collapse_special((char*)ap->value)) {
raise_error("invalid format, special character does not end with a semicolon", pi->str, pi->s);
}
}
ap++;
if (MAX_ATTRS <= (ap - attrs)) {
raise_error("too many attributes", pi->str, pi->s);
}
break;
}
c = '\0';
}
if (hasChildren) {
char *start;
done = 0;
// read children
while (!done) {
start = pi->s;
next_non_white(pi);
c = *pi->s++;
if ('\0' == c) {
raise_error("invalid format, document not terminated", pi->str, pi->s);
}
if ('<' == c) {
switch (*pi->s) {
case '!': /* better be a comment or CDATA */
pi->s++;
if ('-' == *pi->s && '-' == *(pi->s + 1)) {
pi->s += 2;
read_comment(pi);
} else if (0 == strncmp("[CDATA[", pi->s, 7)) {
pi->s += 7;
read_cdata(pi);
} else {
raise_error("invalid format, invalid comment or CDATA format", pi->str, pi->s);
}
break;
case '/':
pi->s++;
name = read_name_token(pi);
end = pi->s;
next_non_white(pi);
c = *pi->s;
*end = '\0';
if (0 != strcmp(name, ename)) {
raise_error("invalid format, elements overlap", pi->str, pi->s);
}
if ('>' != c) {
raise_error("invalid format, element not closed", pi->str, pi->s);
}
pi->s++;
pi->pcb->end_element(pi, ename);
return;
case '\0':
raise_error("invalid format, document not terminated", pi->str, pi->s);
default:
// a child element
read_element(pi);
break;
}
} else { // read as TEXT
pi->s = start;
//pi->s--;
read_text(pi);
//read_reduced_text(pi);
// to exit read_text with no errors the next character must be <
if ('/' == *(pi->s + 1) &&
0 == strncmp(ename, pi->s + 2, elen) &&
'>' == *(pi->s + elen + 2)) {
// close tag after text so treat as a value
pi->s += elen + 3;
pi->pcb->end_element(pi, ename);
return;
}
}
}
}
}
static void
parse(SaxDrive dr) {
char c = skipBOM(dr);
int state = START_STATE;
while ('\0' != c) {
buf_protect(&dr->buf);
if (is_white(c) && '\0' == (c = buf_next_non_white(&dr->buf))) {
break;
}
if ('<' == c) {
c = buf_get(&dr->buf);
switch (c) {
case '?': /* instructions (xml or otherwise) */
c = read_instruction(dr);
break;
case '!': /* comment or doctype */
buf_protect(&dr->buf);
c = buf_get(&dr->buf);
if ('\0' == c) {
ox_sax_drive_error(dr, NO_TERM "DOCTYPE or comment not terminated");
goto DONE;
} else if ('-' == c) {
c = buf_get(&dr->buf); /* skip first - and get next character */
if ('-' != c) {
ox_sax_drive_error(dr, INVALID_FORMAT "bad comment format, expected <!--");
} else {
c = buf_get(&dr->buf); /* skip second - */
}
c = read_comment(dr);
} else {
int i;
int spaced = 0;
int line = dr->buf.line;
int col = dr->buf.col;
if (is_white(c)) {
spaced = 1;
c = buf_next_non_white(&dr->buf);
}
dr->buf.str = dr->buf.tail - 1;
for (i = 7; 0 < i; i--) {
c = buf_get(&dr->buf);
}
if (0 == strncmp("DOCTYPE", dr->buf.str, 7)) {
if (spaced) {
ox_sax_drive_error_at(dr, WRONG_CHAR "<!DOCTYPE can not included spaces", line, col);
}
if (START_STATE != state) {
ox_sax_drive_error(dr, OUT_OF_ORDER "DOCTYPE can not come after an element");
}
c = read_doctype(dr);
} else if (0 == strncasecmp("DOCTYPE", dr->buf.str, 7)) {
ox_sax_drive_error(dr, CASE_ERROR "expected DOCTYPE all in caps");
if (START_STATE != state) {
ox_sax_drive_error(dr, OUT_OF_ORDER "DOCTYPE can not come after an element");
}
c = read_doctype(dr);
} else if (0 == strncmp("[CDATA[", dr->buf.str, 7)) {
if (spaced) {
ox_sax_drive_error_at(dr, WRONG_CHAR "<![CDATA[ can not included spaces", line, col);
}
c = read_cdata(dr);
} else if (0 == strncasecmp("[CDATA[", dr->buf.str, 7)) {
ox_sax_drive_error(dr, CASE_ERROR "expected CDATA all in caps");
c = read_cdata(dr);
} else {
ox_sax_drive_error_at(dr, WRONG_CHAR "DOCTYPE, CDATA, or comment expected", line, col);
c = read_name_token(dr);
if ('>' == c) {
c = buf_get(&dr->buf);
}
}
}
break;
case '/': /* element end */
c = read_element_end(dr);
if (0 == stack_peek(&dr->stack)) {
state = AFTER_STATE;
}
break;
case '\0':
goto DONE;
default:
buf_backup(&dr->buf);
if (AFTER_STATE == state) {
ox_sax_drive_error(dr, OUT_OF_ORDER "multiple top level elements");
}
state = BODY_STATE;
c = read_element_start(dr);
if (0 == stack_peek(&dr->stack)) {
state = AFTER_STATE;
}
break;
}
} else {
buf_reset(&dr->buf);
c = read_text(dr);
}
}
DONE:
if (dr->stack.head < dr->stack.tail) {
char msg[256];
Nv sp;
if (dr->has.line) {
rb_ivar_set(dr->handler, ox_at_line_id, LONG2NUM(dr->buf.line));
}
if (dr->has.column) {
rb_ivar_set(dr->handler, ox_at_column_id, LONG2NUM(dr->buf.col));
}
for (sp = dr->stack.tail - 1; dr->stack.head <= sp; sp--) {
snprintf(msg, sizeof(msg) - 1, "%selement '%s' not closed", EL_MISMATCH, sp->name);
ox_sax_drive_error_at(dr, msg, dr->buf.line, dr->buf.col);
if (dr->has.end_element) {
VALUE args[1];
args[0] = sp->val;
rb_funcall2(dr->handler, ox_end_element_id, 1, args);
}
}
}
}