psyced/world/net/xml/parse.c-retro

66 lines
2.7 KiB
Plaintext

// $Id: parse.c,v 1.10 2005/03/14 10:23:28 lynx Exp $ // vim:syntax=lpc
//
// the actual RSS parser
// why the file is called "parse.c" i don't know
// it certainly wouldn't be a good idea to have several parsers in one file
// so you may as well rename it into rss.c? TODO
//
#include <net.h>
inherit NET_PATH "xml/common";
xmlparse(str) {
// DOM-style XML parser for parsing RSS and RDF
// see CHANGESTODO for discussion on how to make it compliant
mapping dom;
string namespace;
string tag, lasttag, data, params;
int pos, close;
pos = 0;
close = -1;
namespace = "";
dom = ([ ]);
while(pos = strstr(str, "<", pos) + 1){
// D2(D("looping xmlparser...\n");)
data = xmlunquote(str[close + 1..pos - 2]);
close = strstr(str, ">", pos);
sscanf(str[pos..close - 1], "%s%t%s", tag, params) || tag = str[pos..close - 1];
if (strlen(tag) && (tag[0] == '!' || tag[0] == '?' || tag[0..2] == "rdf" || tag[0..2] == "rss")) {
if(tag[0..2] == "rdf" || tag[0..2] == "rss")
dom["type"] = tag[0..2];
} else if (strlen(tag) && tag[0] == '/') {
// D2(D("closing " + tag + "\n");)
// closing tag
if (tag[1..] != lasttag){
// D2(D("warning: XML may be malformed\n");)
;
} else {
// handle data
if (data != "\n") {
if (stringp(dom[namespace]) )
dom[namespace] = ({ dom[namespace], data });
else if (pointerp(dom[namespace]))
dom[namespace] += ({ data });
else
dom[namespace] = data;
}
namespace = namespace[..<strlen(lasttag) + 2];
lasttag = explode(namespace, "_")[<1];
}
} else {
// open tag
if ((params && params[<1] == '/') || tag[<1] == '/') {
// better than before, but not really
// correct to simply skip it
continue;
}
// D2(D("opening " + tag + "\n");)
namespace += "_" + tag;
lasttag = tag;
}
pos = close;
}
// P2(("DOM: %O\n", dom))
return dom;
}