-rw-r--r-- | lib/util.cc | 51 |
1 files changed, 36 insertions, 15 deletions
diff --git a/lib/util.cc b/lib/util.cc index 416e2cc..a9b9bed 100644 --- a/lib/util.cc +++ b/lib/util.cc | |||
@@ -111,28 +111,38 @@ namespace opkele { | |||
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 | int fraction; | ||
117 | struct tm tm_t; | 118 | struct tm tm_t; |
118 | memset(&tm_t,0,sizeof(tm_t)); | 119 | memset(&tm_t,0,sizeof(tm_t)); |
119 | if( | 120 | if( ( |
120 | sscanf( | 121 | sscanf( |
121 | w.c_str(), | 122 | w.c_str(), |
122 | "%04d-%02d-%02dT%02d:%02d:%02dZ", | 123 | "%04d-%02d-%02dT%02d:%02d:%02dZ", |
123 | &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday, | 124 | &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 | 125 | &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec |
125 | ) != 6 ) | 126 | ) != 6 |
127 | ) && ( | ||
128 | sscanf( | ||
129 | w.c_str(), | ||
130 | "%04d-%02d-%02dT%02d:%02d:%02d.%03dZ", | ||
131 | &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday, | ||
132 | &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec, | ||
133 | &fraction | ||
134 | ) != 7 | ||
135 | ) ) | ||
126 | throw failed_conversion(OPKELE_CP_ "failed to sscanf()"); | 136 | throw failed_conversion(OPKELE_CP_ "failed to sscanf()"); |
127 | tm_t.tm_mon--; | 137 | tm_t.tm_mon--; |
128 | tm_t.tm_year-=1900; | 138 | tm_t.tm_year-=1900; |
129 | time_t rv = mktime(&tm_t); | 139 | time_t rv = mktime(&tm_t); |
130 | if(rv==(time_t)-1) | 140 | if(rv==(time_t)-1) |
131 | throw failed_conversion(OPKELE_CP_ "failed to mktime()"); | 141 | throw failed_conversion(OPKELE_CP_ "failed to mktime()"); |
132 | return rv; | 142 | return rv-timezone; |
133 | } | 143 | } |
134 | 144 | ||
135 | /* | 145 | /* |
136 | * | 146 | * |
137 | */ | 147 | */ |
138 | 148 | ||
@@ -161,45 +171,55 @@ namespace opkele { | |||
161 | return rv; | 171 | return rv; |
162 | } | 172 | } |
163 | 173 | ||
164 | /* | 174 | /* |
165 | * Normalize URL according to the rules, described in rfc 3986, section 6 | 175 | * Normalize URL according to the rules, described in rfc 3986, section 6 |
166 | * | 176 | * |
167 | * - uppercase hext triplets (e.g. %ab -> %AB) | 177 | * - uppercase hex triplets (e.g. %ab -> %AB) |
168 | * - lowercase scheme and host | 178 | * - lowercase scheme and host |
169 | * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3, | 179 | * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3, |
170 | * that is - [:alpha:][:digit:]._~- | 180 | * that is - [:alpha:][:digit:]._~- |
171 | * - remove dot segments | 181 | * - remove dot segments |
172 | * - remove empty and default ports | 182 | * - remove empty and default ports |
173 | * - if there's no path component, add '/' | 183 | * - if there's no path component, add '/' |
174 | */ | 184 | */ |
175 | 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"; | ||
176 | string rv; | 187 | string rv; |
177 | string::size_type colon = uri.find(':'); | 188 | string::size_type ns = uri.find_first_not_of(whitespace); |
189 | if(ns==string::npos) | ||
190 | throw bad_input(OPKELE_CP_ "Can't normalize empty URI"); | ||
191 | string::size_type colon = uri.find(':',ns); | ||
178 | if(colon==string::npos) | 192 | if(colon==string::npos) |
179 | throw bad_input(OPKELE_CP_ "No scheme specified in URI"); | 193 | throw bad_input(OPKELE_CP_ "No scheme specified in URI"); |
180 | transform( | 194 | transform( |
181 | uri.begin(), uri.begin()+colon+1, | 195 | uri.begin()+ns, uri.begin()+colon+1, |
182 | back_inserter(rv), ::tolower ); | 196 | back_inserter(rv), ::tolower ); |
183 | bool s; | 197 | bool s; |
184 | if(rv=="http:") | 198 | string::size_type ul = uri.find_last_not_of(whitespace)+1; |
185 | s = false; | ||
186 | else if(rv=="https:") | ||
187 | s = true; | ||
188 | else | ||
189 | throw not_implemented(OPKELE_CP_ "Only http(s) URIs can be normalized here"); | ||
190 | string::size_type ul = uri.length(); | ||
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"); |
203 | if(rv=="http:") | ||
204 | s = false; | ||
205 | else if(rv=="https:") | ||
206 | s = true; | ||
207 | else{ | ||
208 | /* TODO: support more schemes. | ||
209 | * e.g. xri. How do we normalize | ||
210 | * xri? | ||
211 | */ | ||
212 | rv.append(uri,colon+1,ul-colon-1); | ||
213 | return rv; | ||
214 | } | ||
195 | rv += "//"; | 215 | rv += "//"; |
196 | string::size_type interesting = uri.find_first_of(":/#?",colon+3); | 216 | string::size_type interesting = uri.find_first_of(":/#?",colon+3); |
197 | if(interesting==string::npos) { | 217 | if(interesting==string::npos) { |
198 | transform( | 218 | transform( |
199 | uri.begin()+colon+3,uri.end(), | 219 | uri.begin()+colon+3,uri.begin()+ul, |
200 | back_inserter(rv), ::tolower ); | 220 | back_inserter(rv), ::tolower ); |
201 | rv += '/'; return rv; | 221 | rv += '/'; return rv; |
202 | } | 222 | } |
203 | transform( | 223 | transform( |
204 | uri.begin()+colon+3,uri.begin()+interesting, | 224 | uri.begin()+colon+3,uri.begin()+interesting, |
205 | back_inserter(rv), ::tolower ); | 225 | back_inserter(rv), ::tolower ); |
@@ -282,13 +302,14 @@ namespace opkele { | |||
282 | pseg.clear(); | 302 | pseg.clear(); |
283 | }else{ | 303 | }else{ |
284 | pseg += c; | 304 | pseg += c; |
285 | } | 305 | } |
286 | } | 306 | } |
287 | if(!pseg.empty()) { | 307 | if(!pseg.empty()) { |
288 | rv += '/'; rv += pseg; | 308 | if(!qf) rv += '/'; |
309 | rv += pseg; | ||
289 | } | 310 | } |
290 | return rv; | 311 | return rv; |
291 | } | 312 | } |
292 | 313 | ||
293 | } | 314 | } |
294 | 315 | ||