static unsigned char *cmark_clean_autolink(cmark_chunk *url, int is_email)
{
cmark_strbuf buf = GH_BUF_INIT;
cmark_chunk_trim(url);
if (url->len == 0)
return NULL;
if (is_email)
cmark_strbuf_puts(&buf, "mailto:");
houdini_unescape_html_f(&buf, url->data, url->len);
return cmark_strbuf_detach(&buf);
}
// Clean a URL: remove surrounding whitespace, and remove \ that escape
// punctuation.
cmark_chunk cmark_clean_url(cmark_mem *mem, cmark_chunk *url) {
cmark_strbuf buf = CMARK_BUF_INIT(mem);
cmark_chunk_trim(url);
if (url->len == 0) {
cmark_chunk result = CMARK_CHUNK_EMPTY;
return result;
}
houdini_unescape_html_f(&buf, url->data, url->len);
cmark_strbuf_unescape(&buf);
return cmark_chunk_buf_detach(&buf);
}
static cmark_chunk cmark_clean_autolink(cmark_mem *mem, cmark_chunk *url,
int is_email) {
cmark_strbuf buf = CMARK_BUF_INIT(mem);
cmark_chunk_trim(url);
if (url->len == 0) {
cmark_chunk result = CMARK_CHUNK_EMPTY;
return result;
}
if (is_email)
cmark_strbuf_puts(&buf, "mailto:");
houdini_unescape_html_f(&buf, url->data, url->len);
return cmark_chunk_buf_detach(&buf);
}
// Clean a URL: remove surrounding whitespace and surrounding <>,
// and remove \ that escape punctuation.
unsigned char *cmark_clean_url(cmark_chunk *url)
{
cmark_strbuf buf = GH_BUF_INIT;
cmark_chunk_trim(url);
if (url->len == 0)
return NULL;
if (url->data[0] == '<' && url->data[url->len - 1] == '>') {
houdini_unescape_html_f(&buf, url->data + 1, url->len - 2);
} else {
houdini_unescape_html_f(&buf, url->data, url->len);
}
cmark_strbuf_unescape(&buf);
return cmark_strbuf_detach(&buf);
}
// Parse a link label. Returns 1 if successful.
// Note: unescaped brackets are not allowed in labels.
// The label begins with `[` and ends with the first `]` character
// encountered. Backticks in labels do not start code spans.
static int link_label(subject* subj, cmark_chunk *raw_label)
{
bufsize_t startpos = subj->pos;
int length = 0;
unsigned char c;
// advance past [
if (peek_char(subj) == '[') {
advance(subj);
} else {
return 0;
}
while ((c = peek_char(subj)) && c != '[' && c != ']') {
if (c == '\\') {
advance(subj);
length++;
if (cmark_ispunct(peek_char(subj))) {
advance(subj);
length++;
}
} else {
advance(subj);
length++;
}
if (length > MAX_LINK_LABEL_LENGTH) {
goto noMatch;
}
}
if (c == ']') { // match found
*raw_label = cmark_chunk_dup(&subj->input, startpos + 1, subj->pos - (startpos + 1));
cmark_chunk_trim(raw_label);
advance(subj); // advance past ]
return 1;
}
noMatch:
subj->pos = startpos; // rewind
return 0;
}
// Clean a URL: remove surrounding whitespace and surrounding <>,
// and remove \ that escape punctuation.
cmark_chunk cmark_clean_url(cmark_chunk *url)
{
cmark_strbuf buf = GH_BUF_INIT;
cmark_chunk_trim(url);
if (url->len == 0) {
cmark_chunk result = CMARK_CHUNK_EMPTY;
return result;
}
if (url->data[0] == '<' && url->data[url->len - 1] == '>') {
houdini_unescape_html_f(&buf, url->data + 1, url->len - 2);
} else {
houdini_unescape_html_f(&buf, url->data, url->len);
}
cmark_strbuf_unescape(&buf);
return cmark_chunk_buf_detach(&buf);
}
