-rw-r--r-- | lib/util.cc | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/lib/util.cc b/lib/util.cc index ac70938..416e2cc 100644 --- a/lib/util.cc +++ b/lib/util.cc | |||
@@ -166,99 +166,99 @@ namespace opkele { | |||
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 | string rv; | 176 | string rv; |
177 | string::size_type colon = uri.find(':'); | 177 | string::size_type colon = uri.find(':'); |
178 | if(colon==string::npos) | 178 | if(colon==string::npos) |
179 | throw bad_input(OPKELE_CP_ "No scheme specified in URI"); | 179 | throw bad_input(OPKELE_CP_ "No scheme specified in URI"); |
180 | transform( | 180 | transform( |
181 | uri.begin(), uri.begin()+colon+1, | 181 | uri.begin(), uri.begin()+colon+1, |
182 | back_inserter(rv), ::tolower ); | 182 | back_inserter(rv), ::tolower ); |
183 | bool s; | 183 | bool s; |
184 | if(rv=="http:") | 184 | if(rv=="http:") |
185 | s = false; | 185 | s = false; |
186 | else if(rv=="https:") | 186 | else if(rv=="https:") |
187 | s = true; | 187 | s = true; |
188 | else | 188 | else |
189 | throw not_implemented(OPKELE_CP_ "Only http(s) URIs can be normalized here"); | 189 | throw not_implemented(OPKELE_CP_ "Only http(s) URIs can be normalized here"); |
190 | string::size_type ul = uri.length(); | 190 | string::size_type ul = uri.length(); |
191 | if(ul <= (colon+3)) | 191 | if(ul <= (colon+3)) |
192 | throw bad_input(OPKELE_CP_ "Unexpected end of URI being normalized encountered"); | 192 | throw bad_input(OPKELE_CP_ "Unexpected end of URI being normalized encountered"); |
193 | if(uri[colon+1]!='/' || uri[colon+2]!='/') | 193 | if(uri[colon+1]!='/' || uri[colon+2]!='/') |
194 | throw bad_input(OPKELE_CP_ "Unexpected input in URI being normalized after scheme component"); | 194 | throw bad_input(OPKELE_CP_ "Unexpected input in URI being normalized after scheme component"); |
195 | rv += "//"; | 195 | rv += "//"; |
196 | string::size_type interesting = uri.find_first_of(":/#?",colon+3); | 196 | string::size_type interesting = uri.find_first_of(":/#?",colon+3); |
197 | if(interesting==string::npos) { | 197 | if(interesting==string::npos) { |
198 | transform( | 198 | transform( |
199 | uri.begin()+colon+3,uri.end(), | 199 | uri.begin()+colon+3,uri.end(), |
200 | back_inserter(rv), ::tolower ); | 200 | back_inserter(rv), ::tolower ); |
201 | rv += '/'; return rv; | 201 | rv += '/'; return rv; |
202 | } | 202 | } |
203 | transform( | 203 | transform( |
204 | uri.begin()+colon+3,uri.begin()+interesting, | 204 | uri.begin()+colon+3,uri.begin()+interesting, |
205 | back_inserter(rv), ::tolower ); | 205 | back_inserter(rv), ::tolower ); |
206 | bool qf = false; | 206 | bool qf = false; |
207 | char ic = uri[interesting]; | 207 | char ic = uri[interesting]; |
208 | if(ic==':') { | 208 | if(ic==':') { |
209 | string::size_type ni = uri.find_first_of("/#?%",interesting+1); | 209 | string::size_type ni = uri.find_first_of("/#?%",interesting+1); |
210 | const char *nptr = uri.data()+interesting+1; | 210 | const char *nptr = uri.data()+interesting+1; |
211 | char *eptr = 0; | 211 | char *eptr = 0; |
212 | long port = strtol(nptr,&eptr,10); | 212 | long port = strtol(nptr,&eptr,10); |
213 | if( (port>0) && (port<65535) && port!=(s?443:80) ) { | 213 | if( (port>0) && (port<65535) && port!=(s?443:80) ) { |
214 | char tmp[6]; | 214 | char tmp[8]; |
215 | snprintf(tmp,sizeof(tmp),"%ld",port); | 215 | snprintf(tmp,sizeof(tmp),":%ld",port); |
216 | rv += ':'; rv += tmp; | 216 | rv += tmp; |
217 | } | 217 | } |
218 | if(ni==string::npos) { | 218 | if(ni==string::npos) { |
219 | rv += '/'; return rv; | 219 | rv += '/'; return rv; |
220 | } | 220 | } |
221 | interesting = ni; | 221 | interesting = ni; |
222 | }else if(ic!='/') { | 222 | }else if(ic!='/') { |
223 | rv += '/'; rv += ic; | 223 | rv += '/'; rv += ic; |
224 | qf = true; | 224 | qf = true; |
225 | ++interesting; | 225 | ++interesting; |
226 | } | 226 | } |
227 | string::size_type n = interesting; | 227 | string::size_type n = interesting; |
228 | char tmp[3] = { 0,0,0 }; | 228 | char tmp[3] = { 0,0,0 }; |
229 | stack<string::size_type> psegs; psegs.push(rv.length()); | 229 | stack<string::size_type> psegs; psegs.push(rv.length()); |
230 | string pseg; | 230 | string pseg; |
231 | for(;n<ul;) { | 231 | for(;n<ul;) { |
232 | string::size_type unsafe = uri.find_first_of(qf?"%":"%/?#",n); | 232 | string::size_type unsafe = uri.find_first_of(qf?"%":"%/?#",n); |
233 | if(unsafe==string::npos) { | 233 | if(unsafe==string::npos) { |
234 | pseg.append(uri,n,ul-n-1); n = ul-1; | 234 | pseg.append(uri,n,ul-n-1); n = ul-1; |
235 | }else{ | 235 | }else{ |
236 | pseg.append(uri,n,unsafe-n); | 236 | pseg.append(uri,n,unsafe-n); |
237 | n = unsafe; | 237 | n = unsafe; |
238 | } | 238 | } |
239 | char c = uri[n++]; | 239 | char c = uri[n++]; |
240 | if(c=='%') { | 240 | if(c=='%') { |
241 | if((n+1)>=ul) | 241 | if((n+1)>=ul) |
242 | throw bad_input(OPKELE_CP_ "Unexpected end of URI encountered while parsing percent-encoded character"); | 242 | throw bad_input(OPKELE_CP_ "Unexpected end of URI encountered while parsing percent-encoded character"); |
243 | tmp[0] = uri[n++]; | 243 | tmp[0] = uri[n++]; |
244 | tmp[1] = uri[n++]; | 244 | tmp[1] = uri[n++]; |
245 | if(!( isxdigit(tmp[0]) && isxdigit(tmp[1]) )) | 245 | if(!( isxdigit(tmp[0]) && isxdigit(tmp[1]) )) |
246 | throw bad_input(OPKELE_CP_ "Invalid percent-encoded character in URI being normalized"); | 246 | throw bad_input(OPKELE_CP_ "Invalid percent-encoded character in URI being normalized"); |
247 | int cc = strtol(tmp,0,16); | 247 | int cc = strtol(tmp,0,16); |
248 | if( isalpha(cc) || isdigit(cc) || strchr("._~-",cc) ) | 248 | if( isalpha(cc) || isdigit(cc) || strchr("._~-",cc) ) |
249 | pseg += cc; | 249 | pseg += cc; |
250 | else{ | 250 | else{ |
251 | pseg += '%'; | 251 | pseg += '%'; |
252 | pseg += toupper(tmp[0]); pseg += toupper(tmp[1]); | 252 | pseg += toupper(tmp[0]); pseg += toupper(tmp[1]); |
253 | } | 253 | } |
254 | }else if(qf) { | 254 | }else if(qf) { |
255 | rv += pseg; rv += c; | 255 | rv += pseg; rv += c; |
256 | pseg.clear(); | 256 | pseg.clear(); |
257 | }else if(n>=ul || strchr("?/#",c)) { | 257 | }else if(n>=ul || strchr("?/#",c)) { |
258 | if(pseg.empty() || pseg==".") { | 258 | if(pseg.empty() || pseg==".") { |
259 | }else if(pseg=="..") { | 259 | }else if(pseg=="..") { |
260 | if(psegs.size()>1) { | 260 | if(psegs.size()>1) { |
261 | rv.resize(psegs.top()); psegs.pop(); | 261 | rv.resize(psegs.top()); psegs.pop(); |
262 | } | 262 | } |
263 | }else{ | 263 | }else{ |
264 | psegs.push(rv.length()); | 264 | psegs.push(rv.length()); |