-rw-r--r-- | lib/util.cc | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/lib/util.cc b/lib/util.cc index ee75d29..a9b9bed 100644 --- a/lib/util.cc +++ b/lib/util.cc | |||
@@ -180,136 +180,137 @@ namespace opkele { | |||
180 | * that is - [:alpha:][:digit:]._~- | 180 | * that is - [:alpha:][:digit:]._~- |
181 | * - remove dot segments | 181 | * - remove dot segments |
182 | * - remove empty and default ports | 182 | * - remove empty and default ports |
183 | * - if there's no path component, add '/' | 183 | * - if there's no path component, add '/' |
184 | */ | 184 | */ |
185 | string rfc_3986_normalize_uri(const string& uri) { | 185 | string rfc_3986_normalize_uri(const string& uri) { |
186 | static const char *whitespace = " \t\r\n"; | 186 | static const char *whitespace = " \t\r\n"; |
187 | string rv; | 187 | string rv; |
188 | string::size_type ns = uri.find_first_not_of(whitespace); | 188 | string::size_type ns = uri.find_first_not_of(whitespace); |
189 | if(ns==string::npos) | 189 | if(ns==string::npos) |
190 | throw bad_input(OPKELE_CP_ "Can't normalize empty URI"); | 190 | throw bad_input(OPKELE_CP_ "Can't normalize empty URI"); |
191 | string::size_type colon = uri.find(':',ns); | 191 | string::size_type colon = uri.find(':',ns); |
192 | if(colon==string::npos) | 192 | if(colon==string::npos) |
193 | throw bad_input(OPKELE_CP_ "No scheme specified in URI"); | 193 | throw bad_input(OPKELE_CP_ "No scheme specified in URI"); |
194 | transform( | 194 | transform( |
195 | uri.begin()+ns, uri.begin()+colon+1, | 195 | uri.begin()+ns, uri.begin()+colon+1, |
196 | back_inserter(rv), ::tolower ); | 196 | back_inserter(rv), ::tolower ); |
197 | bool s; | 197 | bool s; |
198 | string::size_type ul = uri.find_last_not_of(whitespace)+1; | 198 | string::size_type ul = uri.find_last_not_of(whitespace)+1; |
199 | if(ul <= (colon+3)) | 199 | if(ul <= (colon+3)) |
200 | 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"); |
201 | if(uri[colon+1]!='/' || uri[colon+2]!='/') | 201 | if(uri[colon+1]!='/' || uri[colon+2]!='/') |
202 | 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"); |
203 | if(rv=="http:") | 203 | if(rv=="http:") |
204 | s = false; | 204 | s = false; |
205 | else if(rv=="https:") | 205 | else if(rv=="https:") |
206 | s = true; | 206 | s = true; |
207 | else{ | 207 | else{ |
208 | /* TODO: support more schemes. | 208 | /* TODO: support more schemes. |
209 | * e.g. xri. How do we normalize | 209 | * e.g. xri. How do we normalize |
210 | * xri? | 210 | * xri? |
211 | */ | 211 | */ |
212 | rv.append(uri,colon+1,ul-colon-1); | 212 | rv.append(uri,colon+1,ul-colon-1); |
213 | return rv; | 213 | return rv; |
214 | } | 214 | } |
215 | rv += "//"; | 215 | rv += "//"; |
216 | string::size_type interesting = uri.find_first_of(":/#?",colon+3); | 216 | string::size_type interesting = uri.find_first_of(":/#?",colon+3); |
217 | if(interesting==string::npos) { | 217 | if(interesting==string::npos) { |
218 | transform( | 218 | transform( |
219 | uri.begin()+colon+3,uri.begin()+ul, | 219 | uri.begin()+colon+3,uri.begin()+ul, |
220 | back_inserter(rv), ::tolower ); | 220 | back_inserter(rv), ::tolower ); |
221 | rv += '/'; return rv; | 221 | rv += '/'; return rv; |
222 | } | 222 | } |
223 | transform( | 223 | transform( |
224 | uri.begin()+colon+3,uri.begin()+interesting, | 224 | uri.begin()+colon+3,uri.begin()+interesting, |
225 | back_inserter(rv), ::tolower ); | 225 | back_inserter(rv), ::tolower ); |
226 | bool qf = false; | 226 | bool qf = false; |
227 | char ic = uri[interesting]; | 227 | char ic = uri[interesting]; |
228 | if(ic==':') { | 228 | if(ic==':') { |
229 | string::size_type ni = uri.find_first_of("/#?%",interesting+1); | 229 | string::size_type ni = uri.find_first_of("/#?%",interesting+1); |
230 | const char *nptr = uri.data()+interesting+1; | 230 | const char *nptr = uri.data()+interesting+1; |
231 | char *eptr = 0; | 231 | char *eptr = 0; |
232 | long port = strtol(nptr,&eptr,10); | 232 | long port = strtol(nptr,&eptr,10); |
233 | if( (port>0) && (port<65535) && port!=(s?443:80) ) { | 233 | if( (port>0) && (port<65535) && port!=(s?443:80) ) { |
234 | char tmp[8]; | 234 | char tmp[8]; |
235 | snprintf(tmp,sizeof(tmp),":%ld",port); | 235 | snprintf(tmp,sizeof(tmp),":%ld",port); |
236 | rv += tmp; | 236 | rv += tmp; |
237 | } | 237 | } |
238 | if(ni==string::npos) { | 238 | if(ni==string::npos) { |
239 | rv += '/'; return rv; | 239 | rv += '/'; return rv; |
240 | } | 240 | } |
241 | interesting = ni; | 241 | interesting = ni; |
242 | }else if(ic!='/') { | 242 | }else if(ic!='/') { |
243 | rv += '/'; rv += ic; | 243 | rv += '/'; rv += ic; |
244 | qf = true; | 244 | qf = true; |
245 | ++interesting; | 245 | ++interesting; |
246 | } | 246 | } |
247 | string::size_type n = interesting; | 247 | string::size_type n = interesting; |
248 | char tmp[3] = { 0,0,0 }; | 248 | char tmp[3] = { 0,0,0 }; |
249 | stack<string::size_type> psegs; psegs.push(rv.length()); | 249 | stack<string::size_type> psegs; psegs.push(rv.length()); |
250 | string pseg; | 250 | string pseg; |
251 | for(;n<ul;) { | 251 | for(;n<ul;) { |
252 | string::size_type unsafe = uri.find_first_of(qf?"%":"%/?#",n); | 252 | string::size_type unsafe = uri.find_first_of(qf?"%":"%/?#",n); |
253 | if(unsafe==string::npos) { | 253 | if(unsafe==string::npos) { |
254 | pseg.append(uri,n,ul-n-1); n = ul-1; | 254 | pseg.append(uri,n,ul-n-1); n = ul-1; |
255 | }else{ | 255 | }else{ |
256 | pseg.append(uri,n,unsafe-n); | 256 | pseg.append(uri,n,unsafe-n); |
257 | n = unsafe; | 257 | n = unsafe; |
258 | } | 258 | } |
259 | char c = uri[n++]; | 259 | char c = uri[n++]; |
260 | if(c=='%') { | 260 | if(c=='%') { |
261 | if((n+1)>=ul) | 261 | if((n+1)>=ul) |
262 | throw bad_input(OPKELE_CP_ "Unexpected end of URI encountered while parsing percent-encoded character"); | 262 | throw bad_input(OPKELE_CP_ "Unexpected end of URI encountered while parsing percent-encoded character"); |
263 | tmp[0] = uri[n++]; | 263 | tmp[0] = uri[n++]; |
264 | tmp[1] = uri[n++]; | 264 | tmp[1] = uri[n++]; |
265 | if(!( isxdigit(tmp[0]) && isxdigit(tmp[1]) )) | 265 | if(!( isxdigit(tmp[0]) && isxdigit(tmp[1]) )) |
266 | throw bad_input(OPKELE_CP_ "Invalid percent-encoded character in URI being normalized"); | 266 | throw bad_input(OPKELE_CP_ "Invalid percent-encoded character in URI being normalized"); |
267 | int cc = strtol(tmp,0,16); | 267 | int cc = strtol(tmp,0,16); |
268 | if( isalpha(cc) || isdigit(cc) || strchr("._~-",cc) ) | 268 | if( isalpha(cc) || isdigit(cc) || strchr("._~-",cc) ) |
269 | pseg += cc; | 269 | pseg += cc; |
270 | else{ | 270 | else{ |
271 | pseg += '%'; | 271 | pseg += '%'; |
272 | pseg += toupper(tmp[0]); pseg += toupper(tmp[1]); | 272 | pseg += toupper(tmp[0]); pseg += toupper(tmp[1]); |
273 | } | 273 | } |
274 | }else if(qf) { | 274 | }else if(qf) { |
275 | rv += pseg; rv += c; | 275 | rv += pseg; rv += c; |
276 | pseg.clear(); | 276 | pseg.clear(); |
277 | }else if(n>=ul || strchr("?/#",c)) { | 277 | }else if(n>=ul || strchr("?/#",c)) { |
278 | if(pseg.empty() || pseg==".") { | 278 | if(pseg.empty() || pseg==".") { |
279 | }else if(pseg=="..") { | 279 | }else if(pseg=="..") { |
280 | if(psegs.size()>1) { | 280 | if(psegs.size()>1) { |
281 | rv.resize(psegs.top()); psegs.pop(); | 281 | rv.resize(psegs.top()); psegs.pop(); |
282 | } | 282 | } |
283 | }else{ | 283 | }else{ |
284 | psegs.push(rv.length()); | 284 | psegs.push(rv.length()); |
285 | if(c!='/') { | 285 | if(c!='/') { |
286 | pseg += c; | 286 | pseg += c; |
287 | qf = true; | 287 | qf = true; |
288 | } | 288 | } |
289 | rv += '/'; rv += pseg; | 289 | rv += '/'; rv += pseg; |
290 | } | 290 | } |
291 | if(c=='/' && (n>=ul || strchr("?#",uri[n])) ) { | 291 | if(c=='/' && (n>=ul || strchr("?#",uri[n])) ) { |
292 | rv += '/'; | 292 | rv += '/'; |
293 | if(n<ul) | 293 | if(n<ul) |
294 | qf = true; | 294 | qf = true; |
295 | }else if(strchr("?#",c)) { | 295 | }else if(strchr("?#",c)) { |
296 | if(psegs.size()==1 && psegs.top()==rv.length()) | 296 | if(psegs.size()==1 && psegs.top()==rv.length()) |
297 | rv += '/'; | 297 | rv += '/'; |
298 | if(pseg.empty()) | 298 | if(pseg.empty()) |
299 | rv += c; | 299 | rv += c; |
300 | qf = true; | 300 | qf = true; |
301 | } | 301 | } |
302 | pseg.clear(); | 302 | pseg.clear(); |
303 | }else{ | 303 | }else{ |
304 | pseg += c; | 304 | pseg += c; |
305 | } | 305 | } |
306 | } | 306 | } |
307 | if(!pseg.empty()) { | 307 | if(!pseg.empty()) { |
308 | rv += '/'; rv += pseg; | 308 | if(!qf) rv += '/'; |
309 | rv += pseg; | ||
309 | } | 310 | } |
310 | return rv; | 311 | return rv; |
311 | } | 312 | } |
312 | 313 | ||
313 | } | 314 | } |
314 | 315 | ||
315 | } | 316 | } |