lib/libxml2
1
0
Fork 0
mirror of https://gitlab.gnome.org/GNOME/libxml2.git synced 2025-07-29 11:41:22 +03:00
Code Activity

html: Optimize htmlParseHTMLName

Browse Source
This commit is contained in:
Nick Wellnhofer
2024-09-12 04:01:38 +02:00
parent c34d0ae9cc
commit 440bd64c69
1 changed files with 70 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

84
HTMLparser.c
View File

@ -2391,28 +2391,84 @@ htmlNewDoc(const xmlChar *URI, const xmlChar *ExternalID) {
static const xmlChar * static const xmlChar *
htmlParseHTMLName(htmlParserCtxtPtr ctxt, int attr) { htmlParseHTMLName(htmlParserCtxtPtr ctxt, int attr) {
const xmlChar *ret;
int nbchar = 0;
int c, l;
int stop = attr ? '=' : 0;
xmlChar buf[HTML_PARSER_BUFFER_SIZE]; xmlChar buf[HTML_PARSER_BUFFER_SIZE];
const xmlChar *ret, *in;
size_t avail;
int eof = PARSER_PROGRESSIVE(ctxt);
int nbchar = 0;
int stop = attr ? '=' : ' ';
c = CUR_CHAR(l); in = ctxt->input->cur;
while ((c != 0) && (c != '/') && (c != '>') && avail = ctxt->input->end - in;
((nbchar == 0) || (c != stop)) &&
(!IS_WS_HTML(c))) { while (1) {
if (nbchar + l <= HTML_PARSER_BUFFER_SIZE) { int c, size;
if ((c >= 'A') && (c <= 'Z')) {
buf[nbchar++] = c + 0x20; if ((!eof) && (avail < 32)) {
size_t oldAvail = avail;
ctxt->input->cur = in;
SHRINK;
xmlParserGrow(ctxt);
in = ctxt->input->cur;
avail = ctxt->input->end - in;
if (oldAvail == avail)
eof = 1;
}
if (avail == 0)
break;
c = *in;
size = 1;
if ((nbchar != 0) &&
((c == '/') || (c == '>') || (c == stop) ||
(IS_WS_HTML(c))))
break;
if (c == 0) {
if (nbchar + 3 <= HTML_PARSER_BUFFER_SIZE) {
buf[nbchar++] = 0xEF;
buf[nbchar++] = 0xBF;
buf[nbchar++] = 0xBD;
}
} else if (c < 0x80) {
if (nbchar < HTML_PARSER_BUFFER_SIZE) {
if ((c >= 'A') && (c <= 'Z'))
c += 0x20;
buf[nbchar++] = c;
}
} else {
size = htmlValidateUtf8(ctxt, in, avail);
if (size > 0) {
if (nbchar + size <= HTML_PARSER_BUFFER_SIZE) {
memcpy(buf + nbchar, in, size);
nbchar += size;
}
} else { } else {
COPY_BUF(buf, nbchar, c); size = 1;
if (nbchar + 3 <= HTML_PARSER_BUFFER_SIZE) {
buf[nbchar++] = 0xEF;
buf[nbchar++] = 0xBF;
buf[nbchar++] = 0xBD;
}
} }
} }
NEXTL(l); in += size;
c = CUR_CHAR(l); avail -= size;
} }
ctxt->input->cur = in;
SHRINK;
ret = xmlDictLookup(ctxt->dict, buf, nbchar); ret = xmlDictLookup(ctxt->dict, buf, nbchar);
if (ret == NULL) if (ret == NULL)
htmlErrMemory(ctxt); htmlErrMemory(ctxt);