static int xml_parse_internal(XMLState *s, const char *buf_start, int buf_len,
EditBuffer *b, int offset_start)
{
int ch, offset, offset0, text_offset_start, ret, offset_end;
const char *buf_end, *buf;
buf = buf_start;
buf_end = buf + buf_len;
offset = offset_start;
offset_end = offset_start + buf_len;
offset0 = 0; /* not used */
text_offset_start = 0; /* not used */
for (;;) {
if (buf) {
if (buf >= buf_end)
break;
ch = charset_decode(&s->charset_state, &buf);
} else {
if (offset >= offset_end)
break;
offset0 = offset;
ch = eb_nextc(b, offset, &offset);
}
/* increment line number to signal errors */
if (ch == '\n') {
/* well, should add counter, but we test abort here */
if (s->abort_func(s->abort_opaque))
return -1;
s->line_num++;
}
switch (s->state) {
case XML_STATE_TAG:
if (ch == '>') {
strbuf_addch(&s->str, '\0');
ret = parse_tag(s, (char *)s->str.buf);
switch (ret) {
default:
case XML_STATE_TEXT:
xml_text:
strbuf_reset(&s->str);
s->state = XML_STATE_TEXT;
text_offset_start = offset;
break;
case XML_STATE_PRETAG:
strbuf_reset(&s->str);
s->state = XML_STATE_PRETAG;
text_offset_start = offset;
break;
}
} else {
strbuf_addch(&s->str, ch);
/* test comment */
if (s->str.size == 3 &&
s->str.buf[0] == '!' &&
s->str.buf[1] == '-' &&
s->str.buf[2] == '-') {
s->state = XML_STATE_COMMENT;
}
}
break;
case XML_STATE_TEXT:
if (ch == '<') {
/* XXX: not strictly correct with comments : should
not flush if comment */
if (buf) {
strbuf_addch(&s->str, '\0');
flush_text(s, (char *)s->str.buf);
strbuf_reset(&s->str);
} else {
flush_text_buffer(s, text_offset_start, offset0);
}
s->state = XML_STATE_TAG;
} else {
if (buf) {
/* evaluate entities */
if (ch == '&') {
buf--;
ch = parse_entity(&buf);
}
strbuf_addch(&s->str, ch);
}
}
break;
case XML_STATE_COMMENT:
if (ch == '-')
s->state = XML_STATE_COMMENT1;
break;
case XML_STATE_COMMENT1:
if (ch == '-')
s->state = XML_STATE_COMMENT2;
else
s->state = XML_STATE_COMMENT;
break;
case XML_STATE_COMMENT2:
if (ch == '>') {
goto xml_text;
} else if (ch != '-') {
s->state = XML_STATE_COMMENT;
}
break;
case XML_STATE_PRETAG:
{
int len, taglen;
strbuf_addch(&s->str, ch);
taglen = s->pretaglen + 2;
len = s->str.size - taglen;
if (len >= 0 &&
s->str.buf[len] == '<' &&
s->str.buf[len + 1] == '/' &&
!xml_tagcmp((char *)s->str.buf + len + 2, s->pretag)) {
s->str.buf[len] = '\0';
if (!xml_tagcmp(s->pretag, "style")) {
if (s->style_sheet) {
CSSParseState b1, *b = &b1;
b->ptr = (char *)s->str.buf;
b->line_num = s->line_num; /* XXX: incorrect */
b->filename = s->filename;
b->ignore_case = s->ignore_case;
css_parse_style_sheet(s->style_sheet, b);
}
} else if (!xml_tagcmp(s->pretag, "script")) {
/* XXX: handle script */
} else {
/* just add the content */
if (buf) {
flush_text(s, (char *)s->str.buf);
} else {
/* XXX: would be incorrect if non ascii chars */
flush_text_buffer(s, text_offset_start, offset - taglen);
}
strbuf_reset(&s->str);
if (s->box)
s->box = s->box->parent;
}
s->state = XML_STATE_WAIT_EOT;
}
}
break;
case XML_STATE_WAIT_EOT:
/* wait end of tag */
if (ch == '>')
goto xml_text;
break;
}
}
return buf - buf_start;
}
/*
* Transition from string context to text context.
*/
static int
sync_text_state(gx_device_pdf *pdev)
{
pdf_text_state_t *pts = pdev->text->text_state;
stream *s = pdev->strm;
int code;
if (pts->buffer.count_chars == 0)
return 0; /* nothing to output */
if (pts->continue_line)
return flush_text_buffer(pdev);
/* Bring text state parameters up to date. */
if (pts->out.character_spacing != pts->in.character_spacing) {
pprintg1(s, "%g Tc\n", pts->in.character_spacing);
pts->out.character_spacing = pts->in.character_spacing;
}
if (pts->out.pdfont != pts->in.pdfont || pts->out.size != pts->in.size) {
pdf_font_resource_t *pdfont = pts->in.pdfont;
code = pdf_assign_font_object_id(pdev, pdfont);
if (code < 0)
return code;
pprints1(s, "/%s ", pdfont->rname);
pprintg1(s, "%g Tf\n", pts->in.size);
pts->out.pdfont = pdfont;
pts->out.size = pts->in.size;
/*
* In PDF, the only place to specify WMode is in the CMap
* (a.k.a. Encoding) of a Type 0 font.
*/
pts->wmode =
(pdfont->FontType == ft_composite ?
pdfont->u.type0.WMode : 0);
code = pdf_used_charproc_resources(pdev, pdfont);
if (code < 0)
return code;
}
if (memcmp(&pts->in.matrix, &pts->out.matrix, sizeof(pts->in.matrix)) ||
((pts->start.x != pts->out_pos.x || pts->start.y != pts->out_pos.y) &&
(pts->buffer.count_chars != 0 || pts->buffer.count_moves != 0))) {
/* pdf_set_text_matrix sets out.matrix = in.matrix */
code = pdf_set_text_matrix(pdev);
if (code < 0)
return code;
}
if (pts->out.render_mode != pts->in.render_mode) {
pprintg1(s, "%g Tr\n", pts->in.render_mode);
pts->out.render_mode = pts->in.render_mode;
}
if (pts->out.word_spacing != pts->in.word_spacing) {
if (memchr(pts->buffer.chars, 32, pts->buffer.count_chars)) {
pprintg1(s, "%g Tw\n", pts->in.word_spacing);
pts->out.word_spacing = pts->in.word_spacing;
}
}
return flush_text_buffer(pdev);
}
