// Content-Type
bool HttpMime::parseContentType(const char *field, size_t fieldLen) {
static const char s_contentType[] = "content-type";
static const size_t s_contentTypeLen = strlen(s_contentType);
if (fieldLen == s_contentTypeLen && strncasecmp(field, s_contentType, fieldLen) == 0) {
const char *value = NULL;
size_t valueLen = 0;
if (getValue(&value, &valueLen)) {
m_contentTypePos = value;
m_contentType = getContentTypePrivate(value, valueLen);
}
return true;
}
return false;
}
// returns false on bad mime
bool HttpMime::parse ( char *mime , long mimeLen , Url *url ) {
// reset locUrl to 0
m_locUrl.reset();
// return if we have no valid complete mime
if ( mimeLen == 0 ) return false;
// status is on first line
m_status = -1;
// skip HTTP/x.x till we hit a space
char *p = mime;
char *pend = mime + mimeLen;
while ( p < pend && !is_wspace_a(*p) ) p++;
// then skip over spaces
while ( p < pend && is_wspace_a(*p) ) p++;
// return false on a problem
if ( p == pend ) return false;
// then read in the http status
m_status = atol2 ( p , pend - p );
// if no Content-Type: mime field was provided, assume html
m_contentType = CT_HTML;
// assume default charset
m_charset = NULL;
m_charsetLen = 0;
// set contentLen, lastModifiedDate, m_cookie
p = mime;
while ( p < pend ) {
// compute the length of the string starting at p and ending
// at a \n or \r
long len = 0;
while ( &p[len] < pend && p[len]!='\n' && p[len]!='\r' ) len++;
// . if we could not find a \n or \r there was an error
// . MIMEs must always end in \n or \r
if ( &p[len] >= pend ) return false;
// . stick a NULL at the end of the line
// . overwrites \n or \r TEMPORARILY
char c = p [ len ];
p [ len ] = '\0';
// parse out some meaningful data
if ( strncasecmp ( p , "Content-Length:" ,15) == 0 ) {
m_contentLengthPos = p + 15;
m_contentLen = atol( m_contentLengthPos);
}
else if ( strncasecmp ( p , "Last-Modified:" ,14) == 0 ) {
m_lastModifiedDate=atotime(p+14);
// do not let them exceed current time for purposes
// of sorting by date using datedb (see Msg16.cpp)
time_t now = time(NULL);
if (m_lastModifiedDate > now) m_lastModifiedDate = now;
}
else if ( strncasecmp ( p , "Content-Type:" ,13) == 0 )
m_contentType = getContentTypePrivate ( p + 13 );
else if ( strncasecmp ( p , "Set-Cookie: " ,11) == 0 ) {
m_cookie = p + 11;
m_cookieLen = gbstrlen ( p + 11 );
}
else if ( strncasecmp ( p , "Location:" , 9) == 0 ) {
// point to it
char *tt = p + 9;
// skip if space
if ( *tt == ' ' ) tt++;
if ( *tt == ' ' ) tt++;
// at least set this for Msg13.cpp to use
m_locationField = tt;
m_locationFieldLen = gbstrlen(tt);
// . we don't add the "www." because of slashdot.com
// . we skip initial spaces in this Url::set() routine
if(url)
m_locUrl.set ( url, p + 9, len - 9,
false/*addWWW?*/);
}
else if ( strncasecmp ( p , "Content-Encoding:", 17) == 0 ) {
//only support gzip now, it doesn't seem like servers
//implement the other types much
m_contentEncodingPos = p+17;
if(strstr(m_contentEncodingPos, "gzip")) {
m_contentEncoding = ET_GZIP;
}
else if(strstr(m_contentEncodingPos, "deflate")) {
//zlib's compression
m_contentEncoding = ET_DEFLATE;
}
}
//else if ( strncasecmp ( p, "Cookie:", 7) == 0 )
// log (LOG_INFO, "mime: Got Cookie = %s", (p+7));
// re-insert the character that we replaced with a '\0'
p [ len ] = c;
// go to next line
p += len;
// skip over the cruft at the end of this line
while ( p < pend && ( *p=='\r' || *p=='\n' ) ) p++;
}
return true;
}