summaryrefslogtreecommitdiffabout
path: root/lib/util.cc
Unidiff
Diffstat (limited to 'lib/util.cc') (more/less context) (show whitespace changes)
-rw-r--r--lib/util.cc2
1 files changed, 1 insertions, 1 deletions
diff --git a/lib/util.cc b/lib/util.cc
index a46ba2a..249eeed 100644
--- a/lib/util.cc
+++ b/lib/util.cc
@@ -21,392 +21,392 @@
21#endif 21#endif
22 22
23namespace opkele { 23namespace opkele {
24 using namespace std; 24 using namespace std;
25 25
26 namespace util { 26 namespace util {
27 27
28 /* 28 /*
29 * base64 29 * base64
30 */ 30 */
31 string encode_base64(const void *data,size_t length) { 31 string encode_base64(const void *data,size_t length) {
32 BIO *b64 = 0, *bmem = 0; 32 BIO *b64 = 0, *bmem = 0;
33 try { 33 try {
34 b64 = BIO_new(BIO_f_base64()); 34 b64 = BIO_new(BIO_f_base64());
35 if(!b64) 35 if(!b64)
36 throw exception_openssl(OPKELE_CP_ "failed to BIO_new() base64 encoder"); 36 throw exception_openssl(OPKELE_CP_ "failed to BIO_new() base64 encoder");
37 BIO_set_flags(b64,BIO_FLAGS_BASE64_NO_NL); 37 BIO_set_flags(b64,BIO_FLAGS_BASE64_NO_NL);
38 bmem = BIO_new(BIO_s_mem()); 38 bmem = BIO_new(BIO_s_mem());
39 BIO_set_flags(b64,BIO_CLOSE); 39 BIO_set_flags(b64,BIO_CLOSE);
40 if(!bmem) 40 if(!bmem)
41 throw exception_openssl(OPKELE_CP_ "failed to BIO_new() memory buffer"); 41 throw exception_openssl(OPKELE_CP_ "failed to BIO_new() memory buffer");
42 BIO_push(b64,bmem); 42 BIO_push(b64,bmem);
43 if(((size_t)BIO_write(b64,data,length))!=length) 43 if(((size_t)BIO_write(b64,data,length))!=length)
44 throw exception_openssl(OPKELE_CP_ "failed to BIO_write()"); 44 throw exception_openssl(OPKELE_CP_ "failed to BIO_write()");
45 if(BIO_flush(b64)!=1) 45 if(BIO_flush(b64)!=1)
46 throw exception_openssl(OPKELE_CP_ "failed to BIO_flush()"); 46 throw exception_openssl(OPKELE_CP_ "failed to BIO_flush()");
47 char *rvd; 47 char *rvd;
48 long rvl = BIO_get_mem_data(bmem,&rvd); 48 long rvl = BIO_get_mem_data(bmem,&rvd);
49 string rv(rvd,rvl); 49 string rv(rvd,rvl);
50 BIO_free_all(b64); 50 BIO_free_all(b64);
51 return rv; 51 return rv;
52 }catch(...) { 52 }catch(...) {
53 if(b64) BIO_free_all(b64); 53 if(b64) BIO_free_all(b64);
54 throw; 54 throw;
55 } 55 }
56 } 56 }
57 57
58 void decode_base64(const string& data,vector<unsigned char>& rv) { 58 void decode_base64(const string& data,vector<unsigned char>& rv) {
59 BIO *b64 = 0, *bmem = 0; 59 BIO *b64 = 0, *bmem = 0;
60 rv.clear(); 60 rv.clear();
61 try { 61 try {
62 bmem = BIO_new_mem_buf((void*)data.data(),data.size()); 62 bmem = BIO_new_mem_buf((void*)data.data(),data.size());
63 if(!bmem) 63 if(!bmem)
64 throw exception_openssl(OPKELE_CP_ "failed to BIO_new_mem_buf()"); 64 throw exception_openssl(OPKELE_CP_ "failed to BIO_new_mem_buf()");
65 b64 = BIO_new(BIO_f_base64()); 65 b64 = BIO_new(BIO_f_base64());
66 if(!b64) 66 if(!b64)
67 throw exception_openssl(OPKELE_CP_ "failed to BIO_new() base64 decoder"); 67 throw exception_openssl(OPKELE_CP_ "failed to BIO_new() base64 decoder");
68 BIO_set_flags(b64,BIO_FLAGS_BASE64_NO_NL); 68 BIO_set_flags(b64,BIO_FLAGS_BASE64_NO_NL);
69 BIO_push(b64,bmem); 69 BIO_push(b64,bmem);
70 unsigned char tmp[512]; 70 unsigned char tmp[512];
71 size_t rb = 0; 71 size_t rb = 0;
72 while((rb=BIO_read(b64,tmp,sizeof(tmp)))>0) 72 while((rb=BIO_read(b64,tmp,sizeof(tmp)))>0)
73 rv.insert(rv.end(),tmp,&tmp[rb]); 73 rv.insert(rv.end(),tmp,&tmp[rb]);
74 BIO_free_all(b64); 74 BIO_free_all(b64);
75 }catch(...) { 75 }catch(...) {
76 if(b64) BIO_free_all(b64); 76 if(b64) BIO_free_all(b64);
77 throw; 77 throw;
78 } 78 }
79 } 79 }
80 80
81 /* 81 /*
82 * big numerics 82 * big numerics
83 */ 83 */
84 84
85 BIGNUM *base64_to_bignum(const string& b64) { 85 BIGNUM *base64_to_bignum(const string& b64) {
86 vector<unsigned char> bin; 86 vector<unsigned char> bin;
87 decode_base64(b64,bin); 87 decode_base64(b64,bin);
88 BIGNUM *rv = BN_bin2bn(&(bin.front()),bin.size(),0); 88 BIGNUM *rv = BN_bin2bn(&(bin.front()),bin.size(),0);
89 if(!rv) 89 if(!rv)
90 throw failed_conversion(OPKELE_CP_ "failed to BN_bin2bn()"); 90 throw failed_conversion(OPKELE_CP_ "failed to BN_bin2bn()");
91 return rv; 91 return rv;
92 } 92 }
93 93
94 BIGNUM *dec_to_bignum(const string& dec) { 94 BIGNUM *dec_to_bignum(const string& dec) {
95 BIGNUM *rv = 0; 95 BIGNUM *rv = 0;
96 if(!BN_dec2bn(&rv,dec.c_str())) 96 if(!BN_dec2bn(&rv,dec.c_str()))
97 throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()"); 97 throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()");
98 return rv; 98 return rv;
99 } 99 }
100 100
101 string bignum_to_base64(const BIGNUM *bn) { 101 string bignum_to_base64(const BIGNUM *bn) {
102 vector<unsigned char> bin(BN_num_bytes(bn)+1); 102 vector<unsigned char> bin(BN_num_bytes(bn)+1);
103 unsigned char *binptr = &(bin.front())+1; 103 unsigned char *binptr = &(bin.front())+1;
104 int l = BN_bn2bin(bn,binptr); 104 int l = BN_bn2bin(bn,binptr);
105 if(l && (*binptr)&0x80){ 105 if(l && (*binptr)&0x80){
106 (*(--binptr)) = 0; ++l; 106 (*(--binptr)) = 0; ++l;
107 } 107 }
108 return encode_base64(binptr,l); 108 return encode_base64(binptr,l);
109 } 109 }
110 110
111 /* 111 /*
112 * w3c times 112 * w3c times
113 */ 113 */
114 114
115 string time_to_w3c(time_t t) { 115 string time_to_w3c(time_t t) {
116 struct tm tm_t; 116 struct tm tm_t;
117 if(!gmtime_r(&t,&tm_t)) 117 if(!gmtime_r(&t,&tm_t))
118 throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()"); 118 throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()");
119 char rv[25]; 119 char rv[25];
120 if(!strftime(rv,sizeof(rv)-1,"%Y-%m-%dT%H:%M:%SZ",&tm_t)) 120 if(!strftime(rv,sizeof(rv)-1,"%Y-%m-%dT%H:%M:%SZ",&tm_t))
121 throw failed_conversion(OPKELE_CP_ "failed to strftime()"); 121 throw failed_conversion(OPKELE_CP_ "failed to strftime()");
122 return rv; 122 return rv;
123 } 123 }
124 124
125#ifndef HAVE_TIMEGM 125#ifndef HAVE_TIMEGM
126 static time_t timegm(struct tm *t) { 126 static time_t timegm(struct tm *t) {
127 char *tz = getenv("TZ"); 127 char *tz = getenv("TZ");
128 setenv("TZ","",1); tzset(); 128 setenv("TZ","",1); tzset();
129 time_t rv = mktime(t); 129 time_t rv = mktime(t);
130 if(tz) 130 if(tz)
131 setenv("TZ",tz,1); 131 setenv("TZ",tz,1);
132 else 132 else
133 unsetenv("TZ"); 133 unsetenv("TZ");
134 tzset(); 134 tzset();
135 return rv; 135 return rv;
136 } 136 }
137 #define timegm opkele::util::timegm 137 #define timegm opkele::util::timegm
138#endif /* HAVE_TIMEGM */ 138#endif /* HAVE_TIMEGM */
139 139
140 time_t w3c_to_time(const string& w) { 140 time_t w3c_to_time(const string& w) {
141 int fraction; 141 int fraction;
142 struct tm tm_t; 142 struct tm tm_t;
143 memset(&tm_t,0,sizeof(tm_t)); 143 memset(&tm_t,0,sizeof(tm_t));
144 if( ( 144 if( (
145 sscanf( 145 sscanf(
146 w.c_str(), 146 w.c_str(),
147 "%04d-%02d-%02dT%02d:%02d:%02dZ", 147 "%04d-%02d-%02dT%02d:%02d:%02dZ",
148 &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday, 148 &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday,
149 &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec 149 &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec
150 ) != 6 150 ) != 6
151 ) && ( 151 ) && (
152 sscanf( 152 sscanf(
153 w.c_str(), 153 w.c_str(),
154 "%04d-%02d-%02dT%02d:%02d:%02d.%03dZ", 154 "%04d-%02d-%02dT%02d:%02d:%02d.%03dZ",
155 &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday, 155 &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday,
156 &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec, 156 &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec,
157 &fraction 157 &fraction
158 ) != 7 158 ) != 7
159 ) ) 159 ) )
160 throw failed_conversion(OPKELE_CP_ "failed to sscanf()"); 160 throw failed_conversion(OPKELE_CP_ "failed to sscanf()");
161 tm_t.tm_mon--; 161 tm_t.tm_mon--;
162 tm_t.tm_year-=1900; 162 tm_t.tm_year-=1900;
163 time_t rv = timegm(&tm_t); 163 time_t rv = timegm(&tm_t);
164 if(rv==(time_t)-1) 164 if(rv==(time_t)-1)
165 throw failed_conversion(OPKELE_CP_ "failed to gmtime()"); 165 throw failed_conversion(OPKELE_CP_ "failed to gmtime()");
166 return rv; 166 return rv;
167 } 167 }
168 168
169 /* 169 /*
170 * 170 *
171 */ 171 */
172 172
173 static inline bool isrfc3986unreserved(int c) { 173 static inline bool isrfc3986unreserved(int c) {
174 if(c<'-') return false; 174 if(c<'-') return false;
175 if(c<='.') return true; 175 if(c<='.') return true;
176 if(c<'0') return false; if(c<='9') return true; 176 if(c<'0') return false; if(c<='9') return true;
177 if(c<'A') return false; if(c<='Z') return true; 177 if(c<'A') return false; if(c<='Z') return true;
178 if(c<'_') return false; 178 if(c<'_') return false;
179 if(c=='_') return true; 179 if(c=='_') return true;
180 if(c<'a') return false; if(c<='z') return true; 180 if(c<'a') return false; if(c<='z') return true;
181 if(c=='~') return true; 181 if(c=='~') return true;
182 return false; 182 return false;
183 } 183 }
184 184
185 struct __url_encoder : public unary_function<char,void> { 185 struct __url_encoder : public unary_function<char,void> {
186 public: 186 public:
187 string& rv; 187 string& rv;
188 188
189 __url_encoder(string& r) : rv(r) { } 189 __url_encoder(string& r) : rv(r) { }
190 190
191 result_type operator()(argument_type c) { 191 result_type operator()(argument_type c) {
192 if(isrfc3986unreserved(c)) 192 if(isrfc3986unreserved(c))
193 rv += c; 193 rv += c;
194 else{ 194 else{
195 char tmp[4]; 195 char tmp[4];
196 snprintf(tmp,sizeof(tmp),"%%%02X", 196 snprintf(tmp,sizeof(tmp),"%%%02X",
197 (c&0xff)); 197 (c&0xff));
198 rv += tmp; 198 rv += tmp;
199 } 199 }
200 } 200 }
201 }; 201 };
202 202
203 string url_encode(const string& str) { 203 string url_encode(const string& str) {
204 string rv; 204 string rv;
205 for_each(str.begin(),str.end(), 205 for_each(str.begin(),str.end(),
206 __url_encoder(rv)); 206 __url_encoder(rv));
207 return rv; 207 return rv;
208 } 208 }
209 209
210 string url_decode(const string& str) { 210 string url_decode(const string& str) {
211 string rv; 211 string rv;
212 back_insert_iterator<string> ii(rv); 212 back_insert_iterator<string> ii(rv);
213 char tmp[3]; tmp[2] = 0;
213 for(string::const_iterator i=str.begin(),ie=str.end(); 214 for(string::const_iterator i=str.begin(),ie=str.end();
214 i!=ie;++i) { 215 i!=ie;++i) {
215 switch(*i) { 216 switch(*i) {
216 case '+': 217 case '+':
217 *(ii++) = ' '; break; 218 *(ii++) = ' '; break;
218 case '%': 219 case '%':
219 ++i; 220 ++i;
220 static char tmp[3] = {0,0,0};
221 if(i==ie) 221 if(i==ie)
222 throw failed_conversion(OPKELE_CP_ "trailing percent in the url-encoded string"); 222 throw failed_conversion(OPKELE_CP_ "trailing percent in the url-encoded string");
223 tmp[0] = *(i++); 223 tmp[0] = *(i++);
224 if(i==ie) 224 if(i==ie)
225 throw failed_conversion(OPKELE_CP_ "not enough hexadecimals after the percent sign in url-encoded string"); 225 throw failed_conversion(OPKELE_CP_ "not enough hexadecimals after the percent sign in url-encoded string");
226 tmp[1] = *i; 226 tmp[1] = *i;
227 if(!(isxdigit(tmp[0]) && isxdigit(tmp[1]))) 227 if(!(isxdigit(tmp[0]) && isxdigit(tmp[1])))
228 throw failed_conversion(OPKELE_CP_ "non-hex follows percent in url-encoded string"); 228 throw failed_conversion(OPKELE_CP_ "non-hex follows percent in url-encoded string");
229 *(ii++) = (char)strtol(tmp,0,16); 229 *(ii++) = (char)strtol(tmp,0,16);
230 break; 230 break;
231 default: 231 default:
232 *(ii++) = *i; break; 232 *(ii++) = *i; break;
233 } 233 }
234 } 234 }
235 return rv; 235 return rv;
236 } 236 }
237 237
238 string attr_escape(const string& str) { 238 string attr_escape(const string& str) {
239 static const char *unsafechars = "<>&\n\"'"; 239 static const char *unsafechars = "<>&\n\"'";
240 string rv; 240 string rv;
241 string::size_type p=0; 241 string::size_type p=0;
242 while(true) { 242 while(true) {
243 string::size_type us = str.find_first_of(unsafechars,p); 243 string::size_type us = str.find_first_of(unsafechars,p);
244 if(us==string::npos) { 244 if(us==string::npos) {
245 if(p!=str.length()) 245 if(p!=str.length())
246 rv.append(str,p,str.length()-p); 246 rv.append(str,p,str.length()-p);
247 return rv; 247 return rv;
248 } 248 }
249 rv.append(str,p,us-p); 249 rv.append(str,p,us-p);
250 rv += "&#"; 250 rv += "&#";
251 rv += long_to_string((long)str[us]); 251 rv += long_to_string((long)str[us]);
252 rv += ';'; 252 rv += ';';
253 p = us+1; 253 p = us+1;
254 } 254 }
255 } 255 }
256 256
257 string long_to_string(long l) { 257 string long_to_string(long l) {
258 char rv[32]; 258 char rv[32];
259 int r=snprintf(rv,sizeof(rv),"%ld",l); 259 int r=snprintf(rv,sizeof(rv),"%ld",l);
260 if(r<0 || r>=(int)sizeof(rv)) 260 if(r<0 || r>=(int)sizeof(rv))
261 throw failed_conversion(OPKELE_CP_ "failed to snprintf()"); 261 throw failed_conversion(OPKELE_CP_ "failed to snprintf()");
262 return rv; 262 return rv;
263 } 263 }
264 264
265 long string_to_long(const string& s) { 265 long string_to_long(const string& s) {
266 char *endptr = 0; 266 char *endptr = 0;
267 long rv = strtol(s.c_str(),&endptr,10); 267 long rv = strtol(s.c_str(),&endptr,10);
268 if((!endptr) || endptr==s.c_str()) 268 if((!endptr) || endptr==s.c_str())
269 throw failed_conversion(OPKELE_CP_ "failed to strtol()"); 269 throw failed_conversion(OPKELE_CP_ "failed to strtol()");
270 return rv; 270 return rv;
271 } 271 }
272 272
273 /* 273 /*
274 * Normalize URL according to the rules, described in rfc 3986, section 6 274 * Normalize URL according to the rules, described in rfc 3986, section 6
275 * 275 *
276 * - uppercase hex triplets (e.g. %ab -> %AB) 276 * - uppercase hex triplets (e.g. %ab -> %AB)
277 * - lowercase scheme and host 277 * - lowercase scheme and host
278 * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3, 278 * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3,
279 * that is - [:alpha:][:digit:]._~- 279 * that is - [:alpha:][:digit:]._~-
280 * - remove dot segments 280 * - remove dot segments
281 * - remove empty and default ports 281 * - remove empty and default ports
282 * - if there's no path component, add '/' 282 * - if there's no path component, add '/'
283 */ 283 */
284 string rfc_3986_normalize_uri(const string& uri) { 284 string rfc_3986_normalize_uri(const string& uri) {
285 string rv; 285 string rv;
286 string::size_type ns = uri.find_first_not_of(data::_whitespace_chars); 286 string::size_type ns = uri.find_first_not_of(data::_whitespace_chars);
287 if(ns==string::npos) 287 if(ns==string::npos)
288 throw bad_input(OPKELE_CP_ "Can't normalize empty URI"); 288 throw bad_input(OPKELE_CP_ "Can't normalize empty URI");
289 string::size_type colon = uri.find(':',ns); 289 string::size_type colon = uri.find(':',ns);
290 if(colon==string::npos) 290 if(colon==string::npos)
291 throw bad_input(OPKELE_CP_ "No scheme specified in URI"); 291 throw bad_input(OPKELE_CP_ "No scheme specified in URI");
292 transform( 292 transform(
293 uri.begin()+ns, uri.begin()+colon+1, 293 uri.begin()+ns, uri.begin()+colon+1,
294 back_inserter(rv), ::tolower ); 294 back_inserter(rv), ::tolower );
295 bool s; 295 bool s;
296 string::size_type ul = uri.find_last_not_of(data::_whitespace_chars)+1; 296 string::size_type ul = uri.find_last_not_of(data::_whitespace_chars)+1;
297 if(ul <= (colon+3)) 297 if(ul <= (colon+3))
298 throw bad_input(OPKELE_CP_ "Unexpected end of URI being normalized encountered"); 298 throw bad_input(OPKELE_CP_ "Unexpected end of URI being normalized encountered");
299 if(uri[colon+1]!='/' || uri[colon+2]!='/') 299 if(uri[colon+1]!='/' || uri[colon+2]!='/')
300 throw bad_input(OPKELE_CP_ "Unexpected input in URI being normalized after scheme component"); 300 throw bad_input(OPKELE_CP_ "Unexpected input in URI being normalized after scheme component");
301 if(rv=="http:") 301 if(rv=="http:")
302 s = false; 302 s = false;
303 else if(rv=="https:") 303 else if(rv=="https:")
304 s = true; 304 s = true;
305 else{ 305 else{
306 /* TODO: support more schemes. e.g. xri. How do we normalize 306 /* TODO: support more schemes. e.g. xri. How do we normalize
307 * xri? 307 * xri?
308 */ 308 */
309 rv.append(uri,colon+1,ul-colon-1); 309 rv.append(uri,colon+1,ul-colon-1);
310 return rv; 310 return rv;
311 } 311 }
312 rv += "//"; 312 rv += "//";
313 string::size_type interesting = uri.find_first_of(":/#?",colon+3); 313 string::size_type interesting = uri.find_first_of(":/#?",colon+3);
314 if(interesting==string::npos) { 314 if(interesting==string::npos) {
315 transform( 315 transform(
316 uri.begin()+colon+3,uri.begin()+ul, 316 uri.begin()+colon+3,uri.begin()+ul,
317 back_inserter(rv), ::tolower ); 317 back_inserter(rv), ::tolower );
318 rv += '/'; return rv; 318 rv += '/'; return rv;
319 } 319 }
320 transform( 320 transform(
321 uri.begin()+colon+3,uri.begin()+interesting, 321 uri.begin()+colon+3,uri.begin()+interesting,
322 back_inserter(rv), ::tolower ); 322 back_inserter(rv), ::tolower );
323 bool qf = false; 323 bool qf = false;
324 char ic = uri[interesting]; 324 char ic = uri[interesting];
325 if(ic==':') { 325 if(ic==':') {
326 string::size_type ni = uri.find_first_of("/#?%",interesting+1); 326 string::size_type ni = uri.find_first_of("/#?%",interesting+1);
327 const char *nptr = uri.data()+interesting+1; 327 const char *nptr = uri.data()+interesting+1;
328 char *eptr = 0; 328 char *eptr = 0;
329 long port = strtol(nptr,&eptr,10); 329 long port = strtol(nptr,&eptr,10);
330 if( (port>0) && (port<65535) && port!=(s?443:80) ) { 330 if( (port>0) && (port<65535) && port!=(s?443:80) ) {
331 char tmp[8]; 331 char tmp[8];
332 snprintf(tmp,sizeof(tmp),":%ld",port); 332 snprintf(tmp,sizeof(tmp),":%ld",port);
333 rv += tmp; 333 rv += tmp;
334 } 334 }
335 if(ni==string::npos) { 335 if(ni==string::npos) {
336 rv += '/'; return rv; 336 rv += '/'; return rv;
337 } 337 }
338 interesting = ni; 338 interesting = ni;
339 }else if(ic!='/') { 339 }else if(ic!='/') {
340 rv += '/'; rv += ic; 340 rv += '/'; rv += ic;
341 qf = true; 341 qf = true;
342 ++interesting; 342 ++interesting;
343 } 343 }
344 string::size_type n = interesting; 344 string::size_type n = interesting;
345 char tmp[3] = { 0,0,0 }; 345 char tmp[3] = { 0,0,0 };
346 stack<string::size_type> psegs; psegs.push(rv.length()); 346 stack<string::size_type> psegs; psegs.push(rv.length());
347 string pseg; 347 string pseg;
348 for(;n<ul;) { 348 for(;n<ul;) {
349 string::size_type unsafe = uri.find_first_of(qf?"%":"%/?#",n); 349 string::size_type unsafe = uri.find_first_of(qf?"%":"%/?#",n);
350 if(unsafe==string::npos) { 350 if(unsafe==string::npos) {
351 pseg.append(uri,n,ul-n-1); n = ul-1; 351 pseg.append(uri,n,ul-n-1); n = ul-1;
352 }else{ 352 }else{
353 pseg.append(uri,n,unsafe-n); 353 pseg.append(uri,n,unsafe-n);
354 n = unsafe; 354 n = unsafe;
355 } 355 }
356 char c = uri[n++]; 356 char c = uri[n++];
357 if(c=='%') { 357 if(c=='%') {
358 if((n+1)>=ul) 358 if((n+1)>=ul)
359 throw bad_input(OPKELE_CP_ "Unexpected end of URI encountered while parsing percent-encoded character"); 359 throw bad_input(OPKELE_CP_ "Unexpected end of URI encountered while parsing percent-encoded character");
360 tmp[0] = uri[n++]; 360 tmp[0] = uri[n++];
361 tmp[1] = uri[n++]; 361 tmp[1] = uri[n++];
362 if(!( isxdigit(tmp[0]) && isxdigit(tmp[1]) )) 362 if(!( isxdigit(tmp[0]) && isxdigit(tmp[1]) ))
363 throw bad_input(OPKELE_CP_ "Invalid percent-encoded character in URI being normalized"); 363 throw bad_input(OPKELE_CP_ "Invalid percent-encoded character in URI being normalized");
364 int cc = strtol(tmp,0,16); 364 int cc = strtol(tmp,0,16);
365 if( isalpha(cc) || isdigit(cc) || strchr("._~-",cc) ) 365 if( isalpha(cc) || isdigit(cc) || strchr("._~-",cc) )
366 pseg += (char)cc; 366 pseg += (char)cc;
367 else{ 367 else{
368 pseg += '%'; 368 pseg += '%';
369 pseg += (char)toupper(tmp[0]); pseg += (char)toupper(tmp[1]); 369 pseg += (char)toupper(tmp[0]); pseg += (char)toupper(tmp[1]);
370 } 370 }
371 }else if(qf) { 371 }else if(qf) {
372 rv += pseg; rv += c; 372 rv += pseg; rv += c;
373 pseg.clear(); 373 pseg.clear();
374 }else if(n>=ul || strchr("?/#",c)) { 374 }else if(n>=ul || strchr("?/#",c)) {
375 if(pseg.empty() || pseg==".") { 375 if(pseg.empty() || pseg==".") {
376 }else if(pseg=="..") { 376 }else if(pseg=="..") {
377 if(psegs.size()>1) { 377 if(psegs.size()>1) {
378 rv.resize(psegs.top()); psegs.pop(); 378 rv.resize(psegs.top()); psegs.pop();
379 } 379 }
380 }else{ 380 }else{
381 psegs.push(rv.length()); 381 psegs.push(rv.length());
382 if(c!='/') { 382 if(c!='/') {
383 pseg += c; 383 pseg += c;
384 qf = true; 384 qf = true;
385 } 385 }
386 rv += '/'; rv += pseg; 386 rv += '/'; rv += pseg;
387 } 387 }
388 if(c=='/' && (n>=ul || strchr("?#",uri[n])) ) { 388 if(c=='/' && (n>=ul || strchr("?#",uri[n])) ) {
389 rv += '/'; 389 rv += '/';
390 if(n<ul) 390 if(n<ul)
391 qf = true; 391 qf = true;
392 }else if(strchr("?#",c)) { 392 }else if(strchr("?#",c)) {
393 if(psegs.size()==1 && psegs.top()==rv.length()) 393 if(psegs.size()==1 && psegs.top()==rv.length())
394 rv += '/'; 394 rv += '/';
395 if(pseg.empty()) 395 if(pseg.empty())
396 rv += c; 396 rv += c;
397 qf = true; 397 qf = true;
398 } 398 }
399 pseg.clear(); 399 pseg.clear();
400 }else{ 400 }else{
401 pseg += c; 401 pseg += c;
402 } 402 }
403 } 403 }
404 if(!pseg.empty()) { 404 if(!pseg.empty()) {
405 if(!qf) rv += '/'; 405 if(!qf) rv += '/';
406 rv += pseg; 406 rv += pseg;
407 } 407 }
408 return rv; 408 return rv;
409 } 409 }
410 410
411 string& strip_uri_fragment_part(string& u) { 411 string& strip_uri_fragment_part(string& u) {
412 string::size_type q = u.find('?'), f = u.find('#'); 412 string::size_type q = u.find('?'), f = u.find('#');