summaryrefslogtreecommitdiffabout
path: root/lib/util.cc
Unidiff
Diffstat (limited to 'lib/util.cc') (more/less context) (ignore whitespace changes)
-rw-r--r--lib/util.cc41
1 files changed, 38 insertions, 3 deletions
diff --git a/lib/util.cc b/lib/util.cc
index bb8a2e8..29e6738 100644
--- a/lib/util.cc
+++ b/lib/util.cc
@@ -4,24 +4,25 @@
4#include <cstring> 4#include <cstring>
5#include <vector> 5#include <vector>
6#include <string> 6#include <string>
7#include <stack> 7#include <stack>
8#include <algorithm> 8#include <algorithm>
9#include <openssl/bio.h> 9#include <openssl/bio.h>
10#include <openssl/evp.h> 10#include <openssl/evp.h>
11#include <openssl/sha.h> 11#include <openssl/sha.h>
12#include <openssl/hmac.h> 12#include <openssl/hmac.h>
13#include <curl/curl.h> 13#include <curl/curl.h>
14#include <opkele/util.h> 14#include <opkele/util.h>
15#include <opkele/exception.h> 15#include <opkele/exception.h>
16#include <opkele/data.h>
16#include <opkele/debug.h> 17#include <opkele/debug.h>
17 18
18#include <config.h> 19#include <config.h>
19#ifdef HAVE_DEMANGLE 20#ifdef HAVE_DEMANGLE
20# include <cxxabi.h> 21# include <cxxabi.h>
21#endif 22#endif
22 23
23namespace opkele { 24namespace opkele {
24 using namespace std; 25 using namespace std;
25 26
26 namespace util { 27 namespace util {
27 28
@@ -202,37 +203,36 @@ namespace opkele {
202 /* 203 /*
203 * Normalize URL according to the rules, described in rfc 3986, section 6 204 * Normalize URL according to the rules, described in rfc 3986, section 6
204 * 205 *
205 * - uppercase hex triplets (e.g. %ab -> %AB) 206 * - uppercase hex triplets (e.g. %ab -> %AB)
206 * - lowercase scheme and host 207 * - lowercase scheme and host
207 * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3, 208 * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3,
208 * that is - [:alpha:][:digit:]._~- 209 * that is - [:alpha:][:digit:]._~-
209 * - remove dot segments 210 * - remove dot segments
210 * - remove empty and default ports 211 * - remove empty and default ports
211 * - if there's no path component, add '/' 212 * - if there's no path component, add '/'
212 */ 213 */
213 string rfc_3986_normalize_uri(const string& uri) { 214 string rfc_3986_normalize_uri(const string& uri) {
214 static const char *whitespace = " \t\r\n";
215 string rv; 215 string rv;
216 string::size_type ns = uri.find_first_not_of(whitespace); 216 string::size_type ns = uri.find_first_not_of(data::_whitespace_chars);
217 if(ns==string::npos) 217 if(ns==string::npos)
218 throw bad_input(OPKELE_CP_ "Can't normalize empty URI"); 218 throw bad_input(OPKELE_CP_ "Can't normalize empty URI");
219 string::size_type colon = uri.find(':',ns); 219 string::size_type colon = uri.find(':',ns);
220 if(colon==string::npos) 220 if(colon==string::npos)
221 throw bad_input(OPKELE_CP_ "No scheme specified in URI"); 221 throw bad_input(OPKELE_CP_ "No scheme specified in URI");
222 transform( 222 transform(
223 uri.begin()+ns, uri.begin()+colon+1, 223 uri.begin()+ns, uri.begin()+colon+1,
224 back_inserter(rv), ::tolower ); 224 back_inserter(rv), ::tolower );
225 bool s; 225 bool s;
226 string::size_type ul = uri.find_last_not_of(whitespace)+1; 226 string::size_type ul = uri.find_last_not_of(data::_whitespace_chars)+1;
227 if(ul <= (colon+3)) 227 if(ul <= (colon+3))
228 throw bad_input(OPKELE_CP_ "Unexpected end of URI being normalized encountered"); 228 throw bad_input(OPKELE_CP_ "Unexpected end of URI being normalized encountered");
229 if(uri[colon+1]!='/' || uri[colon+2]!='/') 229 if(uri[colon+1]!='/' || uri[colon+2]!='/')
230 throw bad_input(OPKELE_CP_ "Unexpected input in URI being normalized after scheme component"); 230 throw bad_input(OPKELE_CP_ "Unexpected input in URI being normalized after scheme component");
231 if(rv=="http:") 231 if(rv=="http:")
232 s = false; 232 s = false;
233 else if(rv=="https:") 233 else if(rv=="https:")
234 s = true; 234 s = true;
235 else{ 235 else{
236 /* TODO: support more schemes. e.g. xri. How do we normalize 236 /* TODO: support more schemes. e.g. xri. How do we normalize
237 * xri? 237 * xri?
238 */ 238 */
@@ -421,15 +421,50 @@ namespace opkele {
421 evpmd = EVP_sha1(); 421 evpmd = EVP_sha1();
422 else 422 else
423 throw unsupported(OPKELE_CP_ "unknown association type"); 423 throw unsupported(OPKELE_CP_ "unknown association type");
424 unsigned int md_len = 0; 424 unsigned int md_len = 0;
425 unsigned char md[SHA256_DIGEST_LENGTH]; 425 unsigned char md[SHA256_DIGEST_LENGTH];
426 HMAC(evpmd, 426 HMAC(evpmd,
427 &(secret.front()),secret.size(), 427 &(secret.front()),secret.size(),
428 (const unsigned char*)kv.data(),kv.length(), 428 (const unsigned char*)kv.data(),kv.length(),
429 md,&md_len); 429 md,&md_len);
430 return encode_base64(md,md_len); 430 return encode_base64(md,md_len);
431 } 431 }
432 432
433 string normalize_identifier(const string& usi,bool strip_fragment) {
434 if(usi.empty())
435 return usi;
436 string rv;
437 string::size_type fsc = usi.find_first_not_of(data::_whitespace_chars);
438 if(fsc==string::npos)
439 return rv;
440 string::size_type lsc = usi.find_last_not_of(data::_whitespace_chars);
441 assert(lsc!=string::npos);
442 if(!strncasecmp(usi.c_str()+fsc,"xri://",sizeof("xri://")-1))
443 fsc += sizeof("xri://")-1;
444 if( (fsc+1) >= lsc )
445 return rv;
446 rv.assign(usi,fsc,lsc-fsc+1);
447 if(strchr(data::_iname_leaders,rv[0])) {
448 /* TODO: further normalize xri identity, fold case or
449 * whatever... */
450 }else{
451 if(rv.find("://")==string::npos)
452 rv.insert(0,"http://");
453 if(strip_fragment) {
454 string::size_type fp = rv.find('#');
455 if(fp!=string::npos) {
456 string::size_type qp = rv.find('?');
457 if(qp==string::npos || qp<fp)
458 rv.erase(fp);
459 else if(qp>fp)
460 rv.erase(fp,qp-fp);
461 }
462 }
463 rv = rfc_3986_normalize_uri(rv);
464 }
465 return rv;
466 }
467
433 } 468 }
434 469
435} 470}