void yf::HttpRewrite1::Rep::rewrite_response(mp::odr &o, Z_HTTP_Response *hres)
{
const char *ctype = z_HTTP_header_lookup(hres->headers, "Content-Type");
if (ctype && hres->content_buf)
{
std::string text(hres->content_buf, hres->content_len);
std::list<Rule>::const_iterator it;
int number_of_replaces = 0;
for (it = rules.begin(); it != rules.end(); it++)
{
if (strcmp(ctype, it->content_type.c_str()) == 0)
{
boost::regex::flag_type b_mode = boost::regex::perl;
if (it->mode.find_first_of('i') != std::string::npos)
b_mode |= boost::regex::icase;
boost::regex e(it->pattern, b_mode);
boost::match_flag_type match_mode = boost::format_first_only;
if (it->mode.find_first_of('g') != std::string::npos)
match_mode = boost::format_all;
text = regex_replace(text, e, it->replacement, match_mode);
number_of_replaces++;
}
}
if (number_of_replaces > 0)
{
hres->content_buf = odr_strdup(o, text.c_str());
hres->content_len = strlen(hres->content_buf);
}
}
}
int yaz_srw_check_content_type(Z_HTTP_Response *hres)
{
const char *content_type = z_HTTP_header_lookup(hres->headers,
"Content-Type");
if (content_type)
{
if (!yaz_strcmp_del("text/xml", content_type, "; "))
return 1;
if (!yaz_strcmp_del("application/xml", content_type, "; "))
return 1;
}
return 0;
}
int yaz_encode_http_request(ODR o, Z_HTTP_Request *hr)
{
Z_HTTP_Header *h;
int top0 = o->top;
odr_write(o, (unsigned char *) hr->method,
strlen(hr->method));
odr_write(o, (unsigned char *) " ", 1);
odr_write(o, (unsigned char *) hr->path,
strlen(hr->path));
odr_write(o, (unsigned char *) " HTTP/", 6);
odr_write(o, (unsigned char *) hr->version,
strlen(hr->version));
odr_write(o, (unsigned char *) "\r\n", 2);
if (hr->content_len &&
!z_HTTP_header_lookup(hr->headers,
"Content-Length"))
{
char lstr[60];
sprintf(lstr, "Content-Length: %d\r\n",
hr->content_len);
odr_write(o, (unsigned char *) lstr, strlen(lstr));
}
for (h = hr->headers; h; h = h->next)
{
odr_write(o, (unsigned char *) h->name, strlen(h->name));
odr_write(o, (unsigned char *) ": ", 2);
odr_write(o, (unsigned char *) h->value, strlen(h->value));
odr_write(o, (unsigned char *) "\r\n", 2);
}
odr_write(o, (unsigned char *) "\r\n", 2);
if (hr->content_buf)
odr_write(o, (unsigned char *)
hr->content_buf,
hr->content_len);
if (o->direction == ODR_PRINT)
{
odr_printf(o, "-- HTTP request:\n%.*s\n", o->top - top0,
o->buf + top0);
odr_printf(o, "-- \n");
}
return 1;
}
int yaz_encode_http_response(ODR o, Z_HTTP_Response *hr)
{
char sbuf[80];
Z_HTTP_Header *h;
int top0 = o->top;
sprintf(sbuf, "HTTP/%s %d %s\r\n", hr->version,
hr->code,
z_HTTP_errmsg(hr->code));
odr_write(o, (unsigned char *) sbuf, strlen(sbuf));
/* apply Content-Length if not already applied */
if (!z_HTTP_header_lookup(hr->headers,
"Content-Length"))
{
char lstr[60];
sprintf(lstr, "Content-Length: %d\r\n",
hr->content_len);
odr_write(o, (unsigned char *) lstr, strlen(lstr));
}
for (h = hr->headers; h; h = h->next)
{
odr_write(o, (unsigned char *) h->name, strlen(h->name));
odr_write(o, (unsigned char *) ": ", 2);
odr_write(o, (unsigned char *) h->value, strlen(h->value));
odr_write(o, (unsigned char *) "\r\n", 2);
}
odr_write(o, (unsigned char *) "\r\n", 2);
if (hr->content_buf)
odr_write(o, (unsigned char *)
hr->content_buf,
hr->content_len);
if (o->direction == ODR_PRINT)
{
odr_printf(o, "-- HTTP response:\n%.*s\n", o->top - top0,
o->buf + top0);
odr_printf(o, "-- \n");
}
return 1;
}
/**
* Look for authentication tokens in HTTP Basic parameters or in x-username/x-password
* parameters. Added by SH.
*/
static void yaz_srw_decodeauth(Z_SRW_PDU *sr, Z_HTTP_Request *hreq,
char *username, char *password, ODR decode)
{
const char *basic = z_HTTP_header_lookup(hreq->headers, "Authorization");
if (username)
sr->username = username;
if (password)
sr->password = password;
if (basic)
{
int len;
char out[256];
char ubuf[256] = "", pbuf[256] = "", *p;
if (strncmp(basic, "Basic ", 6))
return;
basic += 6;
len = strlen(basic);
if (!len || len > 256)
return;
yaz_base64decode(basic, out);
/* Format of out should be username:password at this point */
strcpy(ubuf, out);
if ((p = strchr(ubuf, ':')))
{
*(p++) = '\0';
if (*p)
strcpy(pbuf, p);
}
if (*ubuf)
sr->username = odr_strdup(decode, ubuf);
if (*pbuf)
sr->password = odr_strdup(decode, pbuf);
}
}
/**
http://www.loc.gov/z3950/agency/zing/srw/service.html
*/
int yaz_sru_decode(Z_HTTP_Request *hreq, Z_SRW_PDU **srw_pdu,
Z_SOAP **soap_package, ODR decode, char **charset,
Z_SRW_diagnostic **diag, int *num_diag)
{
#if YAZ_HAVE_XML2
static Z_SOAP_Handler soap_handlers[2] = {
{YAZ_XMLNS_SRU_v1_1, 0, (Z_SOAP_fun) yaz_srw_codec},
{0, 0, 0}
};
#endif
const char *content_type = z_HTTP_header_lookup(hreq->headers,
"Content-Type");
/*
SRU GET: ignore content type.
SRU POST: we support "application/x-www-form-urlencoded";
not "multipart/form-data" .
*/
if (!strcmp(hreq->method, "GET")
||
(!strcmp(hreq->method, "POST") && content_type &&
!yaz_strcmp_del("application/x-www-form-urlencoded",
content_type, "; ")))
{
char *db = "Default";
const char *p0 = hreq->path, *p1;
#if YAZ_HAVE_XML2
const char *operation = 0;
char *version = 0;
char *query = 0;
char *queryType = "cql";
char *username = 0;
char *password = 0;
char *sortKeys = 0;
char *stylesheet = 0;
char *scanClause = 0;
char *recordXPath = 0;
char *recordSchema = 0;
char *recordXMLEscaping = 0;
char *recordPacking = 0;
char *maximumRecords = 0;
char *startRecord = 0;
char *maximumTerms = 0;
char *responsePosition = 0;
const char *facetLimit = 0;
const char *facetStart = 0;
const char *facetSort = 0;
Z_SRW_extra_arg *extra_args = 0;
#endif
char **uri_name;
char **uri_val;
grab_charset(decode, content_type, charset);
if (charset && *charset == 0 && !strcmp(hreq->method, "GET"))
*charset = "UTF-8";
if (*p0 == '/')
p0++;
p1 = strchr(p0, '?');
if (!p1)
p1 = p0 + strlen(p0);
if (p1 != p0)
db = yaz_decode_sru_dbpath_odr(decode, p0, p1 - p0);
if (!strcmp(hreq->method, "POST"))
p1 = hreq->content_buf;
yaz_uri_to_array(p1, decode, &uri_name, &uri_val);
#if YAZ_HAVE_XML2
if (uri_name)
{
int i;
for (i = 0; uri_name[i]; i++)
{
char *n = uri_name[i];
char *v = uri_val[i];
if (!strcmp(n, "query"))
query = v;
else if (!strcmp(n, "x-pquery"))
{
query = v;
queryType = "pqf";
}
else if (!strcmp(n, "queryType"))
queryType = v;
else if (!strcmp(n, "x-username"))
username = v;
else if (!strcmp(n, "x-password"))
password = v;
else if (!strcmp(n, "operation"))
operation = v;
else if (!strcmp(n, "stylesheet"))
stylesheet = v;
else if (!strcmp(n, "sortKeys"))
sortKeys = v;
else if (!strcmp(n, "recordXPath"))
recordXPath = v;
else if (!strcmp(n, "recordSchema"))
recordSchema = v;
else if (!strcmp(n, "recordPacking"))
recordPacking = v;
else if (!strcmp(n, "recordXMLEscaping"))
recordXMLEscaping = v;
else if (!strcmp(n, "version"))
version = v;
else if (!strcmp(n, "scanClause"))
scanClause = v;
else if (!strcmp(n, "x-pScanClause"))
{
scanClause = v;
queryType = "pqf";
}
else if (!strcmp(n, "maximumRecords"))
maximumRecords = v;
else if (!strcmp(n, "startRecord"))
startRecord = v;
else if (!strcmp(n, "maximumTerms"))
maximumTerms = v;
else if (!strcmp(n, "responsePosition"))
responsePosition = v;
else if (!strcmp(n, "facetLimit"))
facetLimit = v;
else if (!strcmp(n, "facetStart"))
facetStart = v;
else if (!strcmp(n, "facetSort"))
facetSort = v;
else if (!strcmp(n, "extraRequestData"))
; /* ignoring extraRequestData */
else if (n[0] == 'x' && n[1] == '-')
{
append_extra_arg(decode, &extra_args, n, v);
}
else
{
if (*num_diag < 10)
yaz_add_srw_diagnostic(decode, diag, num_diag,
YAZ_SRW_UNSUPP_PARAMETER, n);
}
}
}
if (!operation)
{
if (query)
operation = "searchRetrieve";
else if (scanClause)
operation = "scan";
else
operation = "explain";
}
version = yaz_negotiate_sru_version(version);
if (!version)
{ /* negotiation failed. */
yaz_add_srw_diagnostic(decode, diag, num_diag,
YAZ_SRW_UNSUPP_VERSION, "2.0");
version = "2.0";
}
if (!operation)
{
if (uri_name)
yaz_add_srw_diagnostic(
decode, diag, num_diag,
YAZ_SRW_MANDATORY_PARAMETER_NOT_SUPPLIED, "operation");
operation = "explain";
}
if (strcmp(version, "2.0"))
{
if (recordXMLEscaping)
{
yaz_add_srw_diagnostic(decode, diag, num_diag,
YAZ_SRW_UNSUPP_PARAMETER,
"recordXMLEscaping");
}
recordXMLEscaping = recordPacking;
recordPacking = "packed";
}
if (!recordXMLEscaping)
recordXMLEscaping = "xml";
if (!strcmp(operation, "searchRetrieve"))
{
Z_SRW_PDU *sr = yaz_srw_get(decode, Z_SRW_searchRetrieve_request);
sr->srw_version = version;
sr->extra_args = extra_args;
*srw_pdu = sr;
yaz_srw_decodeauth(sr, hreq, username, password, decode);
sr->u.request->queryType = queryType;
sr->u.request->query = query;
if (!query)
yaz_add_srw_diagnostic(
decode, diag, num_diag,
YAZ_SRW_MANDATORY_PARAMETER_NOT_SUPPLIED, "query");
if (sortKeys)
{
sr->u.request->sort_type = Z_SRW_sort_type_sort;
sr->u.request->sort.sortKeys = sortKeys;
}
sr->u.request->recordXPath = recordXPath;
sr->u.request->recordSchema = recordSchema;
sr->u.request->recordPacking = recordXMLEscaping;
sr->u.request->packing = recordPacking;
sr->u.request->stylesheet = stylesheet;
yaz_sru_facet_request(decode , &sr->u.request->facetList,
&facetLimit, &facetStart, &facetSort);
yaz_sru_decode_integer(decode, "maximumRecords", maximumRecords,
&sr->u.request->maximumRecords,
diag, num_diag, 0);
yaz_sru_decode_integer(decode, "startRecord", startRecord,
&sr->u.request->startRecord,
diag, num_diag, 1);
sr->u.request->database = db;
(*soap_package) = (Z_SOAP *)
odr_malloc(decode, sizeof(**soap_package));
(*soap_package)->which = Z_SOAP_generic;
(*soap_package)->u.generic = (Z_SOAP_Generic *)
odr_malloc(decode, sizeof(*(*soap_package)->u.generic));
(*soap_package)->u.generic->p = sr;
(*soap_package)->u.generic->ns = soap_handlers[0].ns;
(*soap_package)->u.generic->no = 0;
(*soap_package)->ns = "SRU";
return 0;
}
else if (!strcmp(operation, "explain"))
{
/* Transfer SRU explain parameters to common struct */
/* http://www.loc.gov/z3950/agency/zing/srw/explain.html */
Z_SRW_PDU *sr = yaz_srw_get(decode, Z_SRW_explain_request);
sr->srw_version = version;
sr->extra_args = extra_args;
yaz_srw_decodeauth(sr, hreq, username, password, decode);
*srw_pdu = sr;
sr->u.explain_request->recordPacking = recordXMLEscaping;
sr->u.explain_request->packing = recordPacking;
sr->u.explain_request->database = db;
sr->u.explain_request->stylesheet = stylesheet;
(*soap_package) = (Z_SOAP *)
odr_malloc(decode, sizeof(**soap_package));
(*soap_package)->which = Z_SOAP_generic;
(*soap_package)->u.generic = (Z_SOAP_Generic *)
odr_malloc(decode, sizeof(*(*soap_package)->u.generic));
(*soap_package)->u.generic->p = sr;
(*soap_package)->u.generic->ns = soap_handlers[0].ns;
(*soap_package)->u.generic->no = 0;
(*soap_package)->ns = "SRU";
return 0;
}
else if (!strcmp(operation, "scan"))
{
/* Transfer SRU scan parameters to common struct */
/* http://www.loc.gov/z3950/agency/zing/srw/scan.html */
Z_SRW_PDU *sr = yaz_srw_get(decode, Z_SRW_scan_request);
sr->srw_version = version;
sr->extra_args = extra_args;
*srw_pdu = sr;
yaz_srw_decodeauth(sr, hreq, username, password, decode);
sr->u.scan_request->queryType = queryType;
sr->u.scan_request->scanClause = scanClause;
if (!scanClause)
yaz_add_srw_diagnostic(
decode, diag, num_diag,
YAZ_SRW_MANDATORY_PARAMETER_NOT_SUPPLIED, "scanClause");
sr->u.scan_request->database = db;
yaz_sru_decode_integer(decode, "maximumTerms",
maximumTerms,
&sr->u.scan_request->maximumTerms,
diag, num_diag, 0);
yaz_sru_decode_integer(decode, "responsePosition",
responsePosition,
&sr->u.scan_request->responsePosition,
diag, num_diag, 0);
sr->u.scan_request->stylesheet = stylesheet;
(*soap_package) = (Z_SOAP *)
odr_malloc(decode, sizeof(**soap_package));
(*soap_package)->which = Z_SOAP_generic;
(*soap_package)->u.generic = (Z_SOAP_Generic *)
odr_malloc(decode, sizeof(*(*soap_package)->u.generic));
(*soap_package)->u.generic->p = sr;
(*soap_package)->u.generic->ns = soap_handlers[0].ns;
(*soap_package)->u.generic->no = 0;
(*soap_package)->ns = "SRU";
return 0;
}
else
{
/* unsupported operation ... */
/* Act as if we received a explain request and throw diagnostic. */
Z_SRW_PDU *sr = yaz_srw_get(decode, Z_SRW_explain_request);
sr->srw_version = version;
*srw_pdu = sr;
sr->u.explain_request->recordPacking = recordPacking;
sr->u.explain_request->database = db;
sr->u.explain_request->stylesheet = stylesheet;
(*soap_package) = (Z_SOAP *)
odr_malloc(decode, sizeof(**soap_package));
(*soap_package)->which = Z_SOAP_generic;
(*soap_package)->u.generic = (Z_SOAP_Generic *)
odr_malloc(decode, sizeof(*(*soap_package)->u.generic));
(*soap_package)->u.generic->p = sr;
(*soap_package)->u.generic->ns = soap_handlers[0].ns;
(*soap_package)->u.generic->no = 0;
(*soap_package)->ns = "SRU";
yaz_add_srw_diagnostic(decode, diag, num_diag,
YAZ_SRW_UNSUPP_OPERATION, operation);
return 0;
}
#else
return 1;
#endif
}
return 2;
}
int yaz_srw_decode(Z_HTTP_Request *hreq, Z_SRW_PDU **srw_pdu,
Z_SOAP **soap_package, ODR decode, char **charset)
{
if (!strcmp(hreq->method, "POST"))
{
const char *content_type = z_HTTP_header_lookup(hreq->headers,
"Content-Type");
if (content_type &&
(!yaz_strcmp_del("text/xml", content_type, "; ") ||
!yaz_strcmp_del("application/soap+xml", content_type, "; ") ||
!yaz_strcmp_del("text/plain", content_type, "; ")))
{
char *db = "Default";
const char *p0 = hreq->path, *p1;
int ret = -1;
static Z_SOAP_Handler soap_handlers[5] = {
#if YAZ_HAVE_XML2
{ YAZ_XMLNS_SRU_v1_1, 0, (Z_SOAP_fun) yaz_srw_codec },
{ YAZ_XMLNS_SRU_v1_0, 0, (Z_SOAP_fun) yaz_srw_codec },
{ YAZ_XMLNS_UPDATE_v0_9, 0, (Z_SOAP_fun) yaz_ucp_codec },
{ YAZ_XMLNS_SRU_v2_mask, 0, (Z_SOAP_fun) yaz_srw_codec },
#endif
{0, 0, 0}
};
if (*p0 == '/')
p0++;
p1 = strchr(p0, '?');
if (!p1)
p1 = p0 + strlen(p0);
if (p1 != p0)
db = yaz_decode_sru_dbpath_odr(decode, p0, p1 - p0);
ret = z_soap_codec(decode, soap_package,
&hreq->content_buf, &hreq->content_len,
soap_handlers);
if (ret == 0 && (*soap_package)->which == Z_SOAP_generic)
{
*srw_pdu = (Z_SRW_PDU*) (*soap_package)->u.generic->p;
yaz_srw_decodeauth(*srw_pdu, hreq, 0, 0, decode);
/* last entry in handlers - SRU 2.0 - is turned into
offset 0.. due to other pieces relying on it */
if ((*soap_package)->u.generic->no == 3)
(*soap_package)->u.generic->no = 0;
if ((*srw_pdu)->which == Z_SRW_searchRetrieve_request &&
(*srw_pdu)->u.request->database == 0)
(*srw_pdu)->u.request->database = db;
if ((*srw_pdu)->which == Z_SRW_explain_request &&
(*srw_pdu)->u.explain_request->database == 0)
(*srw_pdu)->u.explain_request->database = db;
if ((*srw_pdu)->which == Z_SRW_scan_request &&
(*srw_pdu)->u.scan_request->database == 0)
(*srw_pdu)->u.scan_request->database = db;
if ((*srw_pdu)->which == Z_SRW_update_request &&
(*srw_pdu)->u.update_request->database == 0)
(*srw_pdu)->u.update_request->database = db;
return 0;
}
return 1;
}
}
return 2;
}
Z_HTTP_Response *yaz_url_exec(yaz_url_t p, const char *uri,
const char *method,
Z_HTTP_Header *user_headers,
const char *buf, size_t len)
{
Z_HTTP_Response *res = 0;
int number_of_redirects = 0;
yaz_cookies_reset(p->cookies);
wrbuf_rewind(p->w_error);
while (1)
{
void *add;
COMSTACK conn = 0;
int code;
const char *location = 0;
char *http_user = 0;
char *http_pass = 0;
char *uri_lean = 0;
Z_GDU *gdu;
extract_user_pass(p->odr_out->mem, uri, &uri_lean,
&http_user, &http_pass);
gdu = z_get_HTTP_Request_uri(p->odr_out, uri_lean, 0, p->proxy ? 1 : 0);
gdu->u.HTTP_Request->method = odr_strdup(p->odr_out, method);
yaz_cookies_request(p->cookies, p->odr_out, gdu->u.HTTP_Request);
for ( ; user_headers; user_headers = user_headers->next)
{
/* prefer new Host over user-supplied Host */
if (!strcmp(user_headers->name, "Host"))
;
/* prefer user-supplied User-Agent over YAZ' own */
else if (!strcmp(user_headers->name, "User-Agent"))
z_HTTP_header_set(p->odr_out, &gdu->u.HTTP_Request->headers,
user_headers->name, user_headers->value);
else
z_HTTP_header_add(p->odr_out, &gdu->u.HTTP_Request->headers,
user_headers->name, user_headers->value);
}
if (http_user && http_pass)
z_HTTP_header_add_basic_auth(p->odr_out,
&gdu->u.HTTP_Request->headers,
http_user, http_pass);
res = 0;
if (buf && len)
{
gdu->u.HTTP_Request->content_buf = (char *) buf;
gdu->u.HTTP_Request->content_len = len;
}
if (!z_GDU(p->odr_out, &gdu, 0, 0))
{
wrbuf_printf(p->w_error, "Can not encode HTTP request for URL %s",
uri);
log_warn(p);
return 0;
}
conn = cs_create_host_proxy(uri_lean, 1, &add, p->proxy);
if (!conn)
{
wrbuf_printf(p->w_error, "Can not resolve URL %s", uri);
log_warn(p);
}
else if (cs_connect(conn, add) < 0)
{
wrbuf_printf(p->w_error, "Can not connect to URL %s", uri);
log_warn(p);
}
else
{
int len;
char *buf = odr_getbuf(p->odr_out, &len, 0);
if (p->verbose)
fwrite(buf, 1, len, stdout);
if (cs_put(conn, buf, len) < 0)
{
wrbuf_printf(p->w_error, "cs_put fail for URL %s", uri);
log_warn(p);
}
else
{
char *netbuffer = 0;
int netlen = 0;
int cs_res = cs_get(conn, &netbuffer, &netlen);
if (cs_res <= 0)
{
wrbuf_printf(p->w_error, "cs_get failed for URL %s", uri);
log_warn(p);
}
else
{
Z_GDU *gdu;
if (p->verbose)
fwrite(netbuffer, 1, cs_res, stdout);
odr_setbuf(p->odr_in, netbuffer, cs_res, 0);
if (!z_GDU(p->odr_in, &gdu, 0, 0)
|| gdu->which != Z_GDU_HTTP_Response)
{
wrbuf_printf(p->w_error, "HTTP decoding fail for "
"URL %s", uri);
log_warn(p);
}
else
{
res = gdu->u.HTTP_Response;
}
}
xfree(netbuffer);
}
}
if (conn)
cs_close(conn);
if (!res)
break;
code = res->code;
location = z_HTTP_header_lookup(res->headers, "Location");
if (++number_of_redirects <= p->max_redirects &&
location && (code == 301 || code == 302 || code == 307))
{
int host_change = 0;
const char *nlocation = yaz_check_location(p->odr_in, uri,
location, &host_change);
odr_reset(p->odr_out);
uri = odr_strdup(p->odr_out, nlocation);
}
else
break;
yaz_cookies_response(p->cookies, res);
odr_reset(p->odr_in);
}
return res;
}