summaryrefslogtreecommitdiffabout
path: root/lib/util.cc
Unidiff
Diffstat (limited to 'lib/util.cc') (more/less context) (ignore whitespace changes)
-rw-r--r--lib/util.cc16
1 files changed, 12 insertions, 4 deletions
diff --git a/lib/util.cc b/lib/util.cc
index ac70938..69d37b5 100644
--- a/lib/util.cc
+++ b/lib/util.cc
@@ -48,248 +48,256 @@ namespace opkele {
48 48
49 void decode_base64(const string& data,vector<unsigned char>& rv) { 49 void decode_base64(const string& data,vector<unsigned char>& rv) {
50 BIO *b64 = 0, *bmem = 0; 50 BIO *b64 = 0, *bmem = 0;
51 rv.clear(); 51 rv.clear();
52 try { 52 try {
53 bmem = BIO_new_mem_buf((void*)data.data(),data.size()); 53 bmem = BIO_new_mem_buf((void*)data.data(),data.size());
54 if(!bmem) 54 if(!bmem)
55 throw exception_openssl(OPKELE_CP_ "failed to BIO_new_mem_buf()"); 55 throw exception_openssl(OPKELE_CP_ "failed to BIO_new_mem_buf()");
56 b64 = BIO_new(BIO_f_base64()); 56 b64 = BIO_new(BIO_f_base64());
57 if(!b64) 57 if(!b64)
58 throw exception_openssl(OPKELE_CP_ "failed to BIO_new() base64 decoder"); 58 throw exception_openssl(OPKELE_CP_ "failed to BIO_new() base64 decoder");
59 BIO_set_flags(b64,BIO_FLAGS_BASE64_NO_NL); 59 BIO_set_flags(b64,BIO_FLAGS_BASE64_NO_NL);
60 BIO_push(b64,bmem); 60 BIO_push(b64,bmem);
61 unsigned char tmp[512]; 61 unsigned char tmp[512];
62 size_t rb = 0; 62 size_t rb = 0;
63 while((rb=BIO_read(b64,tmp,sizeof(tmp)))>0) 63 while((rb=BIO_read(b64,tmp,sizeof(tmp)))>0)
64 rv.insert(rv.end(),tmp,&tmp[rb]); 64 rv.insert(rv.end(),tmp,&tmp[rb]);
65 BIO_free_all(b64); 65 BIO_free_all(b64);
66 }catch(...) { 66 }catch(...) {
67 if(b64) BIO_free_all(b64); 67 if(b64) BIO_free_all(b64);
68 throw; 68 throw;
69 } 69 }
70 } 70 }
71 71
72 /* 72 /*
73 * big numerics 73 * big numerics
74 */ 74 */
75 75
76 BIGNUM *base64_to_bignum(const string& b64) { 76 BIGNUM *base64_to_bignum(const string& b64) {
77 vector<unsigned char> bin; 77 vector<unsigned char> bin;
78 decode_base64(b64,bin); 78 decode_base64(b64,bin);
79 BIGNUM *rv = BN_bin2bn(&(bin.front()),bin.size(),0); 79 BIGNUM *rv = BN_bin2bn(&(bin.front()),bin.size(),0);
80 if(!rv) 80 if(!rv)
81 throw failed_conversion(OPKELE_CP_ "failed to BN_bin2bn()"); 81 throw failed_conversion(OPKELE_CP_ "failed to BN_bin2bn()");
82 return rv; 82 return rv;
83 } 83 }
84 84
85 BIGNUM *dec_to_bignum(const string& dec) { 85 BIGNUM *dec_to_bignum(const string& dec) {
86 BIGNUM *rv = 0; 86 BIGNUM *rv = 0;
87 if(!BN_dec2bn(&rv,dec.c_str())) 87 if(!BN_dec2bn(&rv,dec.c_str()))
88 throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()"); 88 throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()");
89 return rv; 89 return rv;
90 } 90 }
91 91
92 string bignum_to_base64(const BIGNUM *bn) { 92 string bignum_to_base64(const BIGNUM *bn) {
93 vector<unsigned char> bin(BN_num_bytes(bn)+1); 93 vector<unsigned char> bin(BN_num_bytes(bn)+1);
94 unsigned char *binptr = &(bin.front())+1; 94 unsigned char *binptr = &(bin.front())+1;
95 int l = BN_bn2bin(bn,binptr); 95 int l = BN_bn2bin(bn,binptr);
96 if(l && (*binptr)&0x80){ 96 if(l && (*binptr)&0x80){
97 (*(--binptr)) = 0; ++l; 97 (*(--binptr)) = 0; ++l;
98 } 98 }
99 return encode_base64(binptr,l); 99 return encode_base64(binptr,l);
100 } 100 }
101 101
102 /* 102 /*
103 * w3c times 103 * w3c times
104 */ 104 */
105 105
106 string time_to_w3c(time_t t) { 106 string time_to_w3c(time_t t) {
107 struct tm tm_t; 107 struct tm tm_t;
108 if(!gmtime_r(&t,&tm_t)) 108 if(!gmtime_r(&t,&tm_t))
109 throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()"); 109 throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()");
110 char rv[25]; 110 char rv[25];
111 if(!strftime(rv,sizeof(rv)-1,"%Y-%m-%dT%H:%M:%SZ",&tm_t)) 111 if(!strftime(rv,sizeof(rv)-1,"%Y-%m-%dT%H:%M:%SZ",&tm_t))
112 throw failed_conversion(OPKELE_CP_ "failed to strftime()"); 112 throw failed_conversion(OPKELE_CP_ "failed to strftime()");
113 return rv; 113 return rv;
114 } 114 }
115 115
116 time_t w3c_to_time(const string& w) { 116 time_t w3c_to_time(const string& w) {
117 struct tm tm_t; 117 struct tm tm_t;
118 memset(&tm_t,0,sizeof(tm_t)); 118 memset(&tm_t,0,sizeof(tm_t));
119 if( 119 if(
120 sscanf( 120 sscanf(
121 w.c_str(), 121 w.c_str(),
122 "%04d-%02d-%02dT%02d:%02d:%02dZ", 122 "%04d-%02d-%02dT%02d:%02d:%02dZ",
123 &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday, 123 &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday,
124 &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec 124 &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec
125 ) != 6 ) 125 ) != 6 )
126 throw failed_conversion(OPKELE_CP_ "failed to sscanf()"); 126 throw failed_conversion(OPKELE_CP_ "failed to sscanf()");
127 tm_t.tm_mon--; 127 tm_t.tm_mon--;
128 tm_t.tm_year-=1900; 128 tm_t.tm_year-=1900;
129 time_t rv = mktime(&tm_t); 129 time_t rv = mktime(&tm_t);
130 if(rv==(time_t)-1) 130 if(rv==(time_t)-1)
131 throw failed_conversion(OPKELE_CP_ "failed to mktime()"); 131 throw failed_conversion(OPKELE_CP_ "failed to mktime()");
132 return rv; 132 return rv;
133 } 133 }
134 134
135 /* 135 /*
136 * 136 *
137 */ 137 */
138 138
139 string url_encode(const string& str) { 139 string url_encode(const string& str) {
140 char * t = curl_escape(str.c_str(),str.length()); 140 char * t = curl_escape(str.c_str(),str.length());
141 if(!t) 141 if(!t)
142 throw failed_conversion(OPKELE_CP_ "failed to curl_escape()"); 142 throw failed_conversion(OPKELE_CP_ "failed to curl_escape()");
143 string rv(t); 143 string rv(t);
144 curl_free(t); 144 curl_free(t);
145 return rv; 145 return rv;
146 } 146 }
147 147
148 string long_to_string(long l) { 148 string long_to_string(long l) {
149 char rv[32]; 149 char rv[32];
150 int r=snprintf(rv,sizeof(rv),"%ld",l); 150 int r=snprintf(rv,sizeof(rv),"%ld",l);
151 if(r<0 || r>=(int)sizeof(rv)) 151 if(r<0 || r>=(int)sizeof(rv))
152 throw failed_conversion(OPKELE_CP_ "failed to snprintf()"); 152 throw failed_conversion(OPKELE_CP_ "failed to snprintf()");
153 return rv; 153 return rv;
154 } 154 }
155 155
156 long string_to_long(const string& s) { 156 long string_to_long(const string& s) {
157 char *endptr = 0; 157 char *endptr = 0;
158 long rv = strtol(s.c_str(),&endptr,10); 158 long rv = strtol(s.c_str(),&endptr,10);
159 if((!endptr) || endptr==s.c_str()) 159 if((!endptr) || endptr==s.c_str())
160 throw failed_conversion(OPKELE_CP_ "failed to strtol()"); 160 throw failed_conversion(OPKELE_CP_ "failed to strtol()");
161 return rv; 161 return rv;
162 } 162 }
163 163
164 /* 164 /*
165 * Normalize URL according to the rules, described in rfc 3986, section 6 165 * Normalize URL according to the rules, described in rfc 3986, section 6
166 * 166 *
167 * - uppercase hext triplets (e.g. %ab -> %AB) 167 * - uppercase hext triplets (e.g. %ab -> %AB)
168 * - lowercase scheme and host 168 * - lowercase scheme and host
169 * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3, 169 * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3,
170 * that is - [:alpha:][:digit:]._~- 170 * that is - [:alpha:][:digit:]._~-
171 * - remove dot segments 171 * - remove dot segments
172 * - remove empty and default ports 172 * - remove empty and default ports
173 * - if there's no path component, add '/' 173 * - if there's no path component, add '/'
174 */ 174 */
175 string rfc_3986_normalize_uri(const string& uri) { 175 string rfc_3986_normalize_uri(const string& uri) {
176 static const char *whitespace = " \t\r\n";
176 string rv; 177 string rv;
177 string::size_type colon = uri.find(':'); 178 string::size_type ns = uri.find_first_not_of(whitespace);
179 if(ns==string::npos)
180 throw bad_input(OPKELE_CP_ "Can't normalize empty URI");
181 string::size_type colon = uri.find(':',ns);
178 if(colon==string::npos) 182 if(colon==string::npos)
179 throw bad_input(OPKELE_CP_ "No scheme specified in URI"); 183 throw bad_input(OPKELE_CP_ "No scheme specified in URI");
180 transform( 184 transform(
181 uri.begin(), uri.begin()+colon+1, 185 uri.begin()+ns, uri.begin()+colon+1,
182 back_inserter(rv), ::tolower ); 186 back_inserter(rv), ::tolower );
183 bool s; 187 bool s;
184 if(rv=="http:") 188 if(rv=="http:")
185 s = false; 189 s = false;
186 else if(rv=="https:") 190 else if(rv=="https:")
187 s = true; 191 s = true;
192#ifndef NDEBUG
193 else if(rv=="file:")
194 s = false;
195#endif /* XXX: or try to make tests work some other way */
188 else 196 else
189 throw not_implemented(OPKELE_CP_ "Only http(s) URIs can be normalized here"); 197 throw not_implemented(OPKELE_CP_ "Only http(s) URIs can be normalized here");
190 string::size_type ul = uri.length(); 198 string::size_type ul = uri.find_last_not_of(whitespace)+1;
191 if(ul <= (colon+3)) 199 if(ul <= (colon+3))
192 throw bad_input(OPKELE_CP_ "Unexpected end of URI being normalized encountered"); 200 throw bad_input(OPKELE_CP_ "Unexpected end of URI being normalized encountered");
193 if(uri[colon+1]!='/' || uri[colon+2]!='/') 201 if(uri[colon+1]!='/' || uri[colon+2]!='/')
194 throw bad_input(OPKELE_CP_ "Unexpected input in URI being normalized after scheme component"); 202 throw bad_input(OPKELE_CP_ "Unexpected input in URI being normalized after scheme component");
195 rv += "//"; 203 rv += "//";
196 string::size_type interesting = uri.find_first_of(":/#?",colon+3); 204 string::size_type interesting = uri.find_first_of(":/#?",colon+3);
197 if(interesting==string::npos) { 205 if(interesting==string::npos) {
198 transform( 206 transform(
199 uri.begin()+colon+3,uri.end(), 207 uri.begin()+colon+3,uri.begin()+ul,
200 back_inserter(rv), ::tolower ); 208 back_inserter(rv), ::tolower );
201 rv += '/'; return rv; 209 rv += '/'; return rv;
202 } 210 }
203 transform( 211 transform(
204 uri.begin()+colon+3,uri.begin()+interesting, 212 uri.begin()+colon+3,uri.begin()+interesting,
205 back_inserter(rv), ::tolower ); 213 back_inserter(rv), ::tolower );
206 bool qf = false; 214 bool qf = false;
207 char ic = uri[interesting]; 215 char ic = uri[interesting];
208 if(ic==':') { 216 if(ic==':') {
209 string::size_type ni = uri.find_first_of("/#?%",interesting+1); 217 string::size_type ni = uri.find_first_of("/#?%",interesting+1);
210 const char *nptr = uri.data()+interesting+1; 218 const char *nptr = uri.data()+interesting+1;
211 char *eptr = 0; 219 char *eptr = 0;
212 long port = strtol(nptr,&eptr,10); 220 long port = strtol(nptr,&eptr,10);
213 if( (port>0) && (port<65535) && port!=(s?443:80) ) { 221 if( (port>0) && (port<65535) && port!=(s?443:80) ) {
214 char tmp[6]; 222 char tmp[6];
215 snprintf(tmp,sizeof(tmp),"%ld",port); 223 snprintf(tmp,sizeof(tmp),"%ld",port);
216 rv += ':'; rv += tmp; 224 rv += ':'; rv += tmp;
217 } 225 }
218 if(ni==string::npos) { 226 if(ni==string::npos) {
219 rv += '/'; return rv; 227 rv += '/'; return rv;
220 } 228 }
221 interesting = ni; 229 interesting = ni;
222 }else if(ic!='/') { 230 }else if(ic!='/') {
223 rv += '/'; rv += ic; 231 rv += '/'; rv += ic;
224 qf = true; 232 qf = true;
225 ++interesting; 233 ++interesting;
226 } 234 }
227 string::size_type n = interesting; 235 string::size_type n = interesting;
228 char tmp[3] = { 0,0,0 }; 236 char tmp[3] = { 0,0,0 };
229 stack<string::size_type> psegs; psegs.push(rv.length()); 237 stack<string::size_type> psegs; psegs.push(rv.length());
230 string pseg; 238 string pseg;
231 for(;n<ul;) { 239 for(;n<ul;) {
232 string::size_type unsafe = uri.find_first_of(qf?"%":"%/?#",n); 240 string::size_type unsafe = uri.find_first_of(qf?"%":"%/?#",n);
233 if(unsafe==string::npos) { 241 if(unsafe==string::npos) {
234 pseg.append(uri,n,ul-n-1); n = ul-1; 242 pseg.append(uri,n,ul-n-1); n = ul-1;
235 }else{ 243 }else{
236 pseg.append(uri,n,unsafe-n); 244 pseg.append(uri,n,unsafe-n);
237 n = unsafe; 245 n = unsafe;
238 } 246 }
239 char c = uri[n++]; 247 char c = uri[n++];
240 if(c=='%') { 248 if(c=='%') {
241 if((n+1)>=ul) 249 if((n+1)>=ul)
242 throw bad_input(OPKELE_CP_ "Unexpected end of URI encountered while parsing percent-encoded character"); 250 throw bad_input(OPKELE_CP_ "Unexpected end of URI encountered while parsing percent-encoded character");
243 tmp[0] = uri[n++]; 251 tmp[0] = uri[n++];
244 tmp[1] = uri[n++]; 252 tmp[1] = uri[n++];
245 if(!( isxdigit(tmp[0]) && isxdigit(tmp[1]) )) 253 if(!( isxdigit(tmp[0]) && isxdigit(tmp[1]) ))
246 throw bad_input(OPKELE_CP_ "Invalid percent-encoded character in URI being normalized"); 254 throw bad_input(OPKELE_CP_ "Invalid percent-encoded character in URI being normalized");
247 int cc = strtol(tmp,0,16); 255 int cc = strtol(tmp,0,16);
248 if( isalpha(cc) || isdigit(cc) || strchr("._~-",cc) ) 256 if( isalpha(cc) || isdigit(cc) || strchr("._~-",cc) )
249 pseg += cc; 257 pseg += cc;
250 else{ 258 else{
251 pseg += '%'; 259 pseg += '%';
252 pseg += toupper(tmp[0]); pseg += toupper(tmp[1]); 260 pseg += toupper(tmp[0]); pseg += toupper(tmp[1]);
253 } 261 }
254 }else if(qf) { 262 }else if(qf) {
255 rv += pseg; rv += c; 263 rv += pseg; rv += c;
256 pseg.clear(); 264 pseg.clear();
257 }else if(n>=ul || strchr("?/#",c)) { 265 }else if(n>=ul || strchr("?/#",c)) {
258 if(pseg.empty() || pseg==".") { 266 if(pseg.empty() || pseg==".") {
259 }else if(pseg=="..") { 267 }else if(pseg=="..") {
260 if(psegs.size()>1) { 268 if(psegs.size()>1) {
261 rv.resize(psegs.top()); psegs.pop(); 269 rv.resize(psegs.top()); psegs.pop();
262 } 270 }
263 }else{ 271 }else{
264 psegs.push(rv.length()); 272 psegs.push(rv.length());
265 if(c!='/') { 273 if(c!='/') {
266 pseg += c; 274 pseg += c;
267 qf = true; 275 qf = true;
268 } 276 }
269 rv += '/'; rv += pseg; 277 rv += '/'; rv += pseg;
270 } 278 }
271 if(c=='/' && (n>=ul || strchr("?#",uri[n])) ) { 279 if(c=='/' && (n>=ul || strchr("?#",uri[n])) ) {
272 rv += '/'; 280 rv += '/';
273 if(n<ul) 281 if(n<ul)
274 qf = true; 282 qf = true;
275 }else if(strchr("?#",c)) { 283 }else if(strchr("?#",c)) {
276 if(psegs.size()==1 && psegs.top()==rv.length()) 284 if(psegs.size()==1 && psegs.top()==rv.length())
277 rv += '/'; 285 rv += '/';
278 if(pseg.empty()) 286 if(pseg.empty())
279 rv += c; 287 rv += c;
280 qf = true; 288 qf = true;
281 } 289 }
282 pseg.clear(); 290 pseg.clear();
283 }else{ 291 }else{
284 pseg += c; 292 pseg += c;
285 } 293 }
286 } 294 }
287 if(!pseg.empty()) { 295 if(!pseg.empty()) {
288 rv += '/'; rv += pseg; 296 rv += '/'; rv += pseg;
289 } 297 }
290 return rv; 298 return rv;
291 } 299 }
292 300
293 } 301 }
294 302
295} 303}