// [24] ws ::= #x9 | #xA | #xD | #x20 | comment
static inline bool
read_ws(SerdReader* reader)
{
const uint8_t c = peek_byte(reader);
switch (c) {
case 0x9: case 0xA: case 0xD: case 0x20:
eat_byte_safe(reader, c);
return true;
case '#':
read_comment(reader);
return true;
default:
return false;
}
}
static int runtest(char *filename, char *ironout)
{
struct input *input = input_open(filename);
char current_line[MAXLINELEN];
char cmd[128];
char *line;
if (!input)
return 1;
while ((line = input_line(input))) {
int result = -1;
strcpy(current_line, line);
nthtoken(cmd, line, " \n", 2);
if (!strcmp(cmd, "comment") || !strcmp(cmd, "#") || !*cmd)
result = read_comment(input);
if (!strcmp(cmd, "write") || !strcmp(cmd, ">"))
result = write_file(input);
if (!strcmp(cmd, "read") || !strcmp(cmd, "<"))
result = read_file(input);
if (!strcmp(cmd, "ironout"))
result = exec_ironout(input, ironout);
if (result == -1) {
printf("unknown cmd: %s\n", cmd);
return 1;
}
if (result > 0) {
char *testname = filename;
if (strchr(testname, '/'))
testname = strrchr(testname, '/') + 1;
printf("%s:%d %s",
testname, input->lineno, current_line);
return 1;
}
}
input_free(input);
return 0;
}
VALUE
ox_parse(char *xml, ParseCallbacks pcb, char **endp, int trace, Effort effort) {
struct _PInfo pi;
int body_read = 0;
if (0 == xml) {
raise_error("Invalid arg, xml string can not be null", xml, 0);
}
if (DEBUG <= trace) {
printf("Parsing xml:\n%s\n", xml);
}
/* initialize parse info */
pi.str = xml;
pi.s = xml;
pi.h = 0;
pi.pcb = pcb;
pi.obj = Qnil;
pi.circ_array = 0;
pi.encoding = 0;
pi.trace = trace;
pi.effort = effort;
while (1) {
next_non_white(&pi); // skip white space
if ('\0' == *pi.s) {
break;
}
if (body_read && 0 != endp) {
*endp = pi.s;
break;
}
if ('<' != *pi.s) { // all top level entities start with <
raise_error("invalid format, expected <", pi.str, pi.s);
}
pi.s++; // past <
switch (*pi.s) {
case '?': // prolog
pi.s++;
read_instruction(&pi);
break;
case '!': /* comment or doctype */
pi.s++;
if ('\0' == *pi.s) {
raise_error("invalid format, DOCTYPE or comment not terminated", pi.str, pi.s);
} else if ('-' == *pi.s) {
pi.s++; // skip -
if ('-' != *pi.s) {
raise_error("invalid format, bad comment format", pi.str, pi.s);
} else {
pi.s++; // skip second -
read_comment(&pi);
}
} else if (0 == strncmp("DOCTYPE", pi.s, 7)) {
pi.s += 7;
read_doctype(&pi);
} else {
raise_error("invalid format, DOCTYPE or comment expected", pi.str, pi.s);
}
break;
case '\0':
raise_error("invalid format, document not terminated", pi.str, pi.s);
default:
read_element(&pi);
body_read = 1;
break;
}
}
return pi.obj;
}
/* 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;
}
}
}
}
}
macro_fn getmacro(int c) {
if (c == '"') {
return read_string;
} else if (c == '(') {
return read_list;
} else if (c == ')') {
// lambda's are awesome!
// but! note that this is only ok as long as the capture list is empty
// TODO: and i should probably make sure this doesn't incur any
// run time overhead as apposed to using pure functions
return [] (std::istream &) -> std::shared_ptr<Object> {
throw "Unmatched delimiter: )";
};
} else if (c == '[') {
return read_vector;
} else if (c == ']') {
return [] (std::istream &) -> std::shared_ptr<Object> {
throw "Unmatched delimiter: ]";
};
} else if (c == '{') {
return read_map;
} else if (c == '}') {
return [] (std::istream &) -> std::shared_ptr<Object> {
throw "Unmatched delimiter: }";
};
} else if (c == '\'') {
return [] (std::istream &in) -> std::shared_ptr<Object> {
return wrap_read(in, Symbol::create("quote"));
};
} else if (c == '@') {
return [] (std::istream &in) -> std::shared_ptr<Object> {
return wrap_read(in, Symbol::create("clojure.core", "deref"));
};
} else if (c == ';') {
return read_comment;
} else if (c == '\\') {
return read_character;
} else if (c == '^') {
return read_meta;
} else if (c == '`') {
return [] (std::istream &in) -> std::shared_ptr<Object> {
throw "TODO: implement syntax quote reader";
};
} else if (c == '~') {
// unquote reader
return [] (std::istream &in) -> std::shared_ptr<Object> {
int ch = in.get();
if (in.eof()) {
throw "EOF while reading character";
}
if (ch == '@') {
return wrap_read(in, UNQUOTE_SPLICING);
} else {
in.unget();
return wrap_read(in, UNQUOTE);
}
};
} else if (c == '%') {
return [] (std::istream &in) -> std::shared_ptr<Object> {
throw "TODO: implement arg reader";
};
} else if (c == '#') {
// dispatch macro reader (lambda)
// TODO: is this over-using lambdas? what is the overhead of a lambda
// over a pure function, esp if the function is being called via a ptr
return [] (std::istream &in) -> std::shared_ptr<Object> {
int c = in.get();
if (in.eof()) {
throw "EOF while reading character";
}
if (c == '{') {
return read_set(in);
} else if (c == '^') {
return read_meta(in);
} else if (c == '\'') {
return wrap_read(in, Symbol::create("var"));
} else if (c == '"') {
return read_regex(in);
} else if (c == '(') {
throw "TODO: implement function reader";
} else if (c == '=') {
throw "TODO: implement eval reader";
} else if (c == '!') {
return read_comment(in);
} else if (c == '<') {
throw "Unreadable form";
} else if (c == '_') {
read(in, true, Object::nil, true);
return NOOP;
} else {
// try ctor reader
// TODO: implement ctor reader
}
throw "No dispatch macro for: " + std::string{ (char)c };
};
} else {
return 0;
}
}
struct generic_section_config *
parse_param(char const *path,
FILE *f,
const struct config_section_info *params,
int quiet_flag,
int _ncond_var,
cfg_cond_var_t *_cond_vars,
int *p_cond_count)
{
struct generic_section_config *cfg = NULL;
struct generic_section_config **psect = &cfg, *sect = NULL;
const struct config_section_info *cur_info = NULL;
char sectname[32];
char varname[32];
char varvalue[1024];
int c, sindex;
parsecfg_state.ncond_var = _ncond_var;
parsecfg_state.cond_vars = _cond_vars;
parsecfg_state.cond_stack = 0;
parsecfg_state.output_enabled = 1;
parsecfg_state.lineno = 1;
if (p_cond_count) *p_cond_count = 0;
/* found the global section description */
for (sindex = 0; params[sindex].name; sindex++) {
if (!strcmp(params[sindex].name, "global")) {
cur_info = ¶ms[sindex];
break;
}
}
/*
if (!cur_info) {
fprintf(stderr, "Cannot find description of section [global]\n");
goto cleanup;
}
*/
if (!f && !(f = fopen(path, "r"))) {
fprintf(stderr, "Cannot open configuration file %s\n", path);
goto cleanup;
}
if (cur_info) {
cfg = (struct generic_section_config*) xcalloc(1, cur_info->size);
if (cur_info->init_func) cur_info->init_func(cfg);
cfg->next = NULL;
psect = &cfg->next;
}
while (1) {
c = read_first_char(f);
if (c == EOF || c == '[') break;
if (c == '#' || c== '%' || c == ';') {
read_comment(f);
continue;
}
if (c == '@') {
if (handle_conditional(f) < 0) goto cleanup;
if (p_cond_count) (*p_cond_count)++;
continue;
}
if (!parsecfg_state.output_enabled) {
read_comment(f);
continue;
}
if (read_variable(f, varname, sizeof(varname),
varvalue, sizeof(varvalue)) < 0) goto cleanup;
if (!quiet_flag) {
printf("%d: Value: %s = %s\n", parsecfg_state.lineno - 1, varname, varvalue);
}
if (!cur_info) {
fprintf(stderr, "Cannot find description of section [global]\n");
goto cleanup;
}
if (copy_param(cfg, cur_info, varname, varvalue) < 0) goto cleanup;
}
while (c != EOF) {
if (read_section_name(f, sectname, sizeof(sectname)) < 0) goto cleanup;
if (!quiet_flag) {
printf("%d: New section %s\n", parsecfg_state.lineno - 1, sectname);
}
if (!strcmp(sectname, "global")) {
fprintf(stderr, "Section global cannot be specified explicitly\n");
goto cleanup;
}
for (sindex = 0; params[sindex].name; sindex++) {
if (!strcmp(params[sindex].name, sectname)) {
cur_info = ¶ms[sindex];
break;
}
}
if (!cur_info) {
fprintf(stderr, "Cannot find description of section [%s]\n", sectname);
goto cleanup;
}
if (cur_info->pcounter) (*cur_info->pcounter)++;
sect = (struct generic_section_config*) xcalloc(1, cur_info->size);
strcpy(sect->name, sectname);
if (cur_info->init_func) cur_info->init_func(sect);
sect->next = NULL;
*psect = sect;
psect = §->next;
while (1) {
c = read_first_char(f);
if (c == EOF || c == '[') break;
if (c == '#' || c == '%' || c == ';') {
read_comment(f);
continue;
}
if (c == '@') {
if (handle_conditional(f) < 0) goto cleanup;
if (p_cond_count) (*p_cond_count)++;
continue;
}
if (!parsecfg_state.output_enabled) {
read_comment(f);
continue;
}
if (read_variable(f, varname, sizeof(varname),
varvalue, sizeof(varvalue)) < 0) goto cleanup;
if (!quiet_flag) {
printf("%d: Value: %s = %s\n", parsecfg_state.lineno - 1, varname, varvalue);
}
if (copy_param(sect, cur_info, varname, varvalue) < 0) goto cleanup;
}
}
if (parsecfg_state.cond_stack) {
fprintf(stderr, "%d: unclosed conditional compilation\n", parsecfg_state.lineno);
goto cleanup;
}
fflush(stdout);
if (f) fclose(f);
return cfg;
cleanup:
xfree(cfg);
if (f) fclose(f);
return NULL;
}
/********************************************************
* next_token -- read the next token in an input stream *
* *
* Returns *
* next token *
********************************************************/
enum TOKEN_TYPE next_token(void)
{
if (in_comment)
return (read_comment());
while (is_char_type(in_cur_char(), C_WHITE)) {
in_read_char();
}
if (in_cur_char() == EOF)
return (T_EOF);
switch (get_char_type(in_cur_char())) {
case C_NEWLINE:
in_read_char();
return (T_NEWLINE);
case C_ALPHA:
while (is_char_type(in_cur_char(), C_ALPHA_NUMERIC))
in_read_char();
return (T_ID);
case C_DIGIT:
in_read_char();
if ((in_cur_char() == 'X') || (in_cur_char() == 'x')) {
in_read_char();
while (is_char_type(in_cur_char(), C_HEX_DIGIT))
in_read_char();
return (T_NUMBER);
}
while (is_char_type(in_cur_char(), C_DIGIT))
in_read_char();
return (T_NUMBER);
case C_SLASH:
/* Check for '/', '*' characters */
if (in_next_char() == '*') {
return (read_comment());
}
/* Fall through */
case C_OPERATOR:
in_read_char();
return (T_OPERATOR);
case C_L_PAREN:
in_read_char();
return (T_L_PAREN);
case C_R_PAREN:
in_read_char();
return (T_R_PAREN);
case C_L_CURLY:
in_read_char();
return (T_L_CURLY);
case C_R_CURLY:
in_read_char();
return (T_R_CURLY);
case C_DOUBLE:
while (1) {
in_read_char();
/* Check for end of string */
if (in_cur_char() == '"')
break;
/* Escape character, then skip the next character */
if (in_cur_char() == '\\')
in_read_char();
}
in_read_char();
return (T_STRING);
case C_SINGLE:
while (1) {
in_read_char();
/* Check for end of character */
if (in_cur_char() == '\'')
break;
/* Escape character, then skip the next character */
if (in_cur_char() == '\\')
in_read_char();
}
in_read_char();
return (T_STRING);
default:
fprintf(stderr, "Internal error: Very strange character\n");
abort();
}
fprintf(stderr, "Internal error: We should never get here\n");
abort();
return (T_EOF); /* Should never get here either */
/* But we put in the return to avoid a compiler */
/* warning. */
}
/********************************************************
* next_token -- read the next token in an input stream *
* *
* Parameters *
* in_file -- file to read *
* *
* Returns *
* next token *
********************************************************/
TOKEN_TYPE token::next_token(input_file& in_file)
{
if (need_to_read_one)
in_file.read_char();
need_to_read_one = 0;
if (in_comment)
return (read_comment(in_file));
while (char_type.is(in_file.cur_char, char_type::C_WHITE)) {
in_file.read_char();
}
if (in_file.cur_char == EOF)
return (T_EOF);
switch (char_type.type(in_file.cur_char)) {
case char_type::C_NEWLINE:
in_file.read_char();
return (T_NEWLINE);
case char_type::C_ALPHA:
while (char_type.is(in_file.cur_char,
char_type::C_ALPHA_NUMERIC))
in_file.read_char();
return (T_ID);
case char_type::C_DIGIT:
in_file.read_char();
if ((in_file.cur_char == 'X') || (in_file.cur_char == 'x')) {
in_file.read_char();
while (char_type.is(in_file.cur_char,
char_type::C_HEX_DIGIT)) {
in_file.read_char();
}
return (T_NUMBER);
}
while (char_type.is(in_file.cur_char, char_type::C_DIGIT))
in_file.read_char();
return (T_NUMBER);
case char_type::C_SLASH:
// Check for /* characters
if (in_file.next_char == '*') {
return (read_comment(in_file));
}
// Now check for double slash comments
if (in_file.next_char == '/') {
while (true) {
// Comment starting with // and ending with EOF is legal
if (in_file.cur_char == EOF)
return (T_COMMENT);
if (in_file.cur_char == '\n')
return (T_COMMENT);
in_file.read_char();
}
}
// Fall through
case char_type::C_OPERATOR:
in_file.read_char();
return (T_OPERATOR);
case char_type::C_L_PAREN:
in_file.read_char();
return (T_L_PAREN);
case char_type::C_R_PAREN:
in_file.read_char();
return (T_R_PAREN);
case char_type::C_L_CURLY:
in_file.read_char();
return (T_L_CURLY);
case char_type::C_R_CURLY:
in_file.read_char();
return (T_R_CURLY);
case char_type::C_DOUBLE:
while (true) {
in_file.read_char();
// Check for end of string
if (in_file.cur_char == '"')
break;
// Escape character, then skip the next character
if (in_file.cur_char == '\\')
in_file.read_char();
}
in_file.read_char();
return (T_STRING);
case char_type::C_SINGLE:
while (true) {
in_file.read_char();
// Check for end of character
if (in_file.cur_char == '\'')
break;
// Escape character, then skip the next character
if (in_file.cur_char == '\\')
in_file.read_char();
}
in_file.read_char();
return (T_STRING);
default:
assert("Internal error: Very strange character" != 0);
}
assert("Internal error: We should never get here" != 0);
return (T_EOF); // Should never get here either
}
void
AsxTokenizer::move_next ()
{
char c;
while ((c = read_char ()) != -1) {
if (in_data) {
bool entirely_whitespace;
current_token->set_type (TOKEN_DATA);
current_token->set_value (read_data (c, entirely_whitespace));
current_token->set_entirely_whitespace (entirely_whitespace);
return;
}
if (c == OP_OPEN_ELEMENT) {
if (is_comment_open (c)) {
read_comment ();
continue;
}
current_token->set_type (TOKEN_OPEN_ELEMENT);
current_token->set_value (NULL);
in_data = false;
return;
}
if (c == OP_CLOSE_ELEMENT) {
current_token->set_type (TOKEN_CLOSE_ELEMENT);
current_token->set_value (NULL);
in_data = true;
return;
}
if (c == OP_ASSIGNMENT) {
current_token->set_type (TOKEN_ASSIGNMENT);
current_token->set_value (NULL);
return;
}
if (c == OP_SLASH) {
current_token->set_type (TOKEN_SLASH);
current_token->set_value (NULL);
return;
}
if (g_ascii_isspace (c)) {
current_token->set_type (TOKEN_WHITESPACE);
current_token->set_value (NULL);
return;
}
if (is_quote_char (c)) {
current_token->set_type (TOKEN_QUOTED_STRING);
current_token->set_value (read_quoted_string (c));
return;
}
if (is_name_char (c)) {
current_token->set_type (TOKEN_NAME);
current_token->set_value (read_name (c));
return;
}
}
current_token->set_type (TOKEN_EOF);
current_token->set_value (NULL);
}
bool LoadPPM(const char *pathname, unsigned char *&pixels,
int &width, int &height, int &maxval)
{
FILE *fp;
if ( ( fp = fopen( pathname, "rb" ) ) == NULL )
{
printf( "fail to open file %s!\n", pathname );
return false;
}
//read header
read_comment(fp);
char magic[16];
fscanf(fp, "%s", magic);
if (strcmp(magic, "P6")!=0)
{
printf("LoadPPM : invalid magic number in %s!\n", pathname);
return false;
}
fgetc(fp);
read_comment(fp);
fscanf(fp, "%d", &width);
if (width <= 0)
{
printf("LoadPPM : invalid width in %s!\n", pathname);
return false;
}
fgetc(fp);
read_comment(fp);
fscanf(fp, "%d", &height);
if (height <= 0)
{
printf("LoadPPM : invalid height in %s!\n", pathname);
return false;
}
fgetc(fp);
read_comment(fp);
fscanf(fp, "%d", &maxval);
if (maxval < 1 || maxval > 65535)
{
printf("LoadPPM : invalid maxval in %s\n", pathname);
return false;
}
fgetc(fp);
//read pixels
int pixDepth = ( maxval < 256 ? 1 : 2 );
int numPixels = width * height * 3;
if (pixels == NULL)
{
pixels = new unsigned char[pixDepth*numPixels];
}
size_t n = fread(pixels, sizeof(unsigned char)*pixDepth, numPixels, fp);
fclose(fp);
if (n == numPixels)
{
//PPM is Most Significant Byte First for 16 bit
//convert to Little-Endian for PC
if (pixDepth == 2) SwapBytes(pixels, numPixels, pixDepth);
return true;
}
else
{
printf("LoadPGM : fail to read pixels in %s!\n", pathname);
return false;
}
}
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);
}
}
}
}
/*************************************************
* read_token -- read token in input stream
* argument
* in_file -- input file
* return
* next token
*************************************************/
TOKEN_TYPE token::next_token(input_file& in_file) {
if (need_to_read_one) {
in_file.read_char();
}
need_to_read_one = 0;
if (in_comment) {
return (read_comment(in_file));
}
while (char_type.is(in_file.cur_char, char_type::C_WHITE)) {
in_file.read_char();
}
if (in_file.cur_char == EOF) {
return T_EOF;
}
switch (char_type.type(in_file.cur_char)) {
case char_type::C_NEWLINE:
in_file.read_char();
return T_NEWLINE;
case char_type::C_ALPHA:
while (char_type.is(in_file.cur_char, char_type::C_ALPHA_NUMERIC)) {
in_file.read_char();
}
return T_ID;
case char_type::C_DIGIT:
in_file.read_char();
if ((in_file.cur_char == 'X') || (in_file.cur_char == 'x')) {
in_file.read_char();
while (char_type.is(in_file.cur_char, char_type::C_HEX_DIGIT)) {
in_file.read_char();
}
return T_NUMBER;
}
while (char_type.is(in_file.cur_char, char_type::C_DIGIT)) {
in_file.read_char();
}
return T_NUMBER;
case char_type::C_SLASH:
// check '/*' comment
if (in_file.next_char == '*') {
return (read_comment(in_file));
}
// check '//' comment
if (in_file.next_char == '/') {
while (true) {
if (in_file.cur_char == EOF) {
return T_COMMENT;
}
if (in_file.cur_char == '\n') {
return T_COMMENT;
}
in_file.read_char();
}
}
// fall through
case char_type::C_OPERATOR:
in_file.read_char();
return T_OPERATOR;
case char_type::C_L_PAREN:
in_file.read_char();
return T_L_PAREN;
case char_type::C_R_PAREN:
in_file.read_char();
return T_R_PAREN;
case char_type::C_L_CURLY:
in_file.read_char();
return T_L_CURLY;
case char_type::C_R_CURLY:
in_file.read_char();
return T_R_CURLY;
case char_type::C_DOUBLE:
while (true) {
in_file.read_char();
if (in_file.cur_char == '"') {
break;
}
if (in_file.cur_char == '\\') {
in_file.read_char();
}
}
in_file.read_char();
return T_STRING;
case char_type::C_SINGLE:
while (true) {
in_file.read_char();
if (in_file.cur_char == '\'') {
break;
}
if (in_file.cur_char == '\\') {
in_file.read_char();
}
}
in_file.read_char();
return T_STRING;
default:
assert("Internal error: Very strange character" != 0);
}
assert("Internal error: We should never get here" != 0);
return T_EOF;
}
