-rw-r--r-- | lib/discovery.cc | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/lib/discovery.cc b/lib/discovery.cc index b4ed3b6..5585e12 100644 --- a/lib/discovery.cc +++ b/lib/discovery.cc | |||
@@ -1,251 +1,251 @@ | |||
1 | #include <list> | 1 | #include <list> |
2 | #include <opkele/curl.h> | 2 | #include <opkele/curl.h> |
3 | #include <opkele/expat.h> | 3 | #include <opkele/expat.h> |
4 | #include <opkele/uris.h> | 4 | #include <opkele/uris.h> |
5 | #include <opkele/discovery.h> | 5 | #include <opkele/discovery.h> |
6 | #include <opkele/exception.h> | 6 | #include <opkele/exception.h> |
7 | #include <opkele/util.h> | 7 | #include <opkele/util.h> |
8 | #include <opkele/data.h> | 8 | #include <opkele/data.h> |
9 | #include <opkele/debug.h> | 9 | #include <opkele/debug.h> |
10 | 10 | ||
11 | #include "config.h" | 11 | #include "config.h" |
12 | 12 | ||
13 | #include <opkele/tidy.h> | 13 | #include <opkele/tidy.h> |
14 | 14 | ||
15 | #define XRDS_HEADER "X-XRDS-Location" | 15 | #define XRDS_HEADER "X-XRDS-Location" |
16 | #define CT_HEADER "Content-Type" | 16 | #define CT_HEADER "Content-Type" |
17 | 17 | ||
18 | namespace opkele { | 18 | namespace opkele { |
19 | using std::list; | 19 | using std::list; |
20 | using xrd::XRD_t; | 20 | using xrd::XRD_t; |
21 | using xrd::service_t; | 21 | using xrd::service_t; |
22 | 22 | ||
23 | /* TODO: the whole discovery thing needs cleanup and optimization due to | 23 | /* TODO: the whole discovery thing needs cleanup and optimization due to |
24 | * many changes of concept. */ | 24 | * many changes of concept. */ |
25 | 25 | ||
26 | static const size_t max_html = 16384; | 26 | static const size_t max_html = 16384; |
27 | 27 | ||
28 | static const struct service_type_t { | 28 | static const struct service_type_t { |
29 | const char *uri; | 29 | const char *uri; |
30 | const char *forceid; | 30 | const char *forceid; |
31 | } op_service_types[] = { | 31 | } op_service_types[] = { |
32 | { STURI_OPENID20_OP, IDURI_SELECT20 }, | 32 | { STURI_OPENID20_OP, IDURI_SELECT20 }, |
33 | { STURI_OPENID20, 0 }, | 33 | { STURI_OPENID20, 0 }, |
34 | { STURI_OPENID11, 0 }, | 34 | { STURI_OPENID11, 0 }, |
35 | { STURI_OPENID10, 0 } | 35 | { STURI_OPENID10, 0 } |
36 | }; | 36 | }; |
37 | enum { | 37 | enum { |
38 | st_index_1 = 2, st_index_2 = 1 | 38 | st_index_1 = 2, st_index_2 = 1 |
39 | }; | 39 | }; |
40 | 40 | ||
41 | 41 | ||
42 | static inline bool is_qelement(const XML_Char *n,const char *qen) { | 42 | static inline bool is_qelement(const XML_Char *n,const char *qen) { |
43 | return !strcasecmp(n,qen); | 43 | return !strcasecmp(n,qen); |
44 | } | 44 | } |
45 | static inline bool is_element(const XML_Char *n,const char *en) { | 45 | static inline bool is_element(const XML_Char *n,const char *en) { |
46 | if(!strcasecmp(n,en)) return true; | 46 | if(!strcasecmp(n,en)) return true; |
47 | int nl = strlen(n), enl = strlen(en); | 47 | int nl = strlen(n), enl = strlen(en); |
48 | if( (nl>=(enl+1)) && n[nl-enl-1]=='\t' | 48 | if( (nl>=(enl+1)) && n[nl-enl-1]=='\t' |
49 | && !strcasecmp(&n[nl-enl],en) ) | 49 | && !strcasecmp(&n[nl-enl],en) ) |
50 | return true; | 50 | return true; |
51 | return false; | 51 | return false; |
52 | } | 52 | } |
53 | 53 | ||
54 | static long element_priority(const XML_Char **a) { | 54 | static long element_priority(const XML_Char **a) { |
55 | for(;*a;++a) | 55 | for(;*a;++a) |
56 | if(!strcasecmp(*(a++),"priority")) { | 56 | if(!strcasecmp(*(a++),"priority")) { |
57 | long rv; | 57 | long rv; |
58 | return (sscanf(*a,"%ld",&rv)==1)?rv:-1; | 58 | return (sscanf(*a,"%ld",&rv)==1)?rv:-1; |
59 | } | 59 | } |
60 | return -1; | 60 | return -1; |
61 | } | 61 | } |
62 | /* TODO: ideally all attributes should be | 62 | /* TODO: ideally all attributes should be |
63 | * retrieved in one run */ | 63 | * retrieved in one run */ |
64 | static const char *element_attr(const XML_Char **a, const char *at) { | 64 | static const char *element_attr(const XML_Char **a, const char *at) { |
65 | for(;*a;++a) | 65 | for(;*a;++a) |
66 | if(!strcasecmp(*(a++),at)) { | 66 | if(!strcasecmp(*(a++),at)) { |
67 | return *a; | 67 | return *a; |
68 | } | 68 | } |
69 | return 0; | 69 | return 0; |
70 | } | 70 | } |
71 | 71 | ||
72 | class idigger_t : public util::curl_t, public util::expat_t { | 72 | class idigger_t : public util::curl_t, public util::expat_t { |
73 | public: | 73 | public: |
74 | string xri_proxy; | 74 | string xri_proxy; |
75 | 75 | ||
76 | enum { | 76 | enum { |
77 | xmode_html = 1, xmode_xrd = 2, xmode_cid = 4, | 77 | xmode_html = 1, xmode_xrd = 2, xmode_cid = 4, |
78 | xmode_noredirs = 8 | 78 | xmode_noredirs = 8 |
79 | }; | 79 | }; |
80 | int xmode; | 80 | int xmode; |
81 | 81 | ||
82 | string xrds_location; | 82 | string xrds_location; |
83 | string http_content_type; | 83 | string http_content_type; |
84 | service_t html_openid1; | 84 | service_t html_openid1; |
85 | service_t html_openid2; | 85 | service_t html_openid2; |
86 | string cdata_buf; | 86 | string cdata_buf; |
87 | long status_code; | 87 | long status_code; |
88 | string status_string; | 88 | string status_string; |
89 | 89 | ||
90 | typedef list<string> pt_stack_t; | 90 | typedef list<string> pt_stack_t; |
91 | pt_stack_t pt_stack; | 91 | pt_stack_t pt_stack; |
92 | int skipping; | 92 | int skipping; |
93 | bool parser_choked; | 93 | bool parser_choked; |
94 | string save_html; | 94 | string save_html; |
95 | 95 | ||
96 | XRD_t *xrd; | 96 | XRD_t *xrd; |
97 | service_t *xrd_service; | 97 | service_t *xrd_service; |
98 | string* cdata; | 98 | string* cdata; |
99 | 99 | ||
100 | idigger_t() | 100 | idigger_t() |
101 | : util::curl_t(easy_init()), | 101 | : util::curl_t(easy_init()), |
102 | util::expat_t(0), | 102 | util::expat_t(0), |
103 | xri_proxy(XRI_PROXY_URL) { | 103 | xri_proxy(XRI_PROXY_URL) { |
104 | CURLcode r; | 104 | CURLcode r; |
105 | (r=misc_sets()) | 105 | (r=misc_sets()) |
106 | || (r=set_write()) | 106 | || (r=set_write()) |
107 | || (r=set_header()) | 107 | || (r=set_header()) |
108 | ; | 108 | ; |
109 | if(r) | 109 | if(r) |
110 | throw exception_curl(OPKELE_CP_ "failed to set curly options",r); | 110 | throw exception_curl(OPKELE_CP_ "failed to set curly options",r); |
111 | } | 111 | } |
112 | ~idigger_t() throw() { } | 112 | ~idigger_t() throw() { } |
113 | 113 | ||
114 | void yadiscover(endpoint_discovery_iterator oi,const string& yurl,const char **types,bool redirs) { | 114 | void yadiscover(endpoint_discovery_iterator oi,const string& yurl,const char **types,bool redirs) { |
115 | idiscovery_t idis; | 115 | idiscovery_t idis; |
116 | idis.xri_identity = false; | 116 | idis.xri_identity = false; |
117 | discover_at(idis,yurl,xmode_html|xmode_xrd|(redirs?0:xmode_noredirs)); | 117 | discover_at(idis,yurl,xmode_html|xmode_xrd|(redirs?0:xmode_noredirs)); |
118 | if(!xrds_location.empty()) { | 118 | if(!xrds_location.empty()) { |
119 | idis.clear(); | 119 | idis.clear(); |
120 | discover_at(idis,xrds_location,xmode_xrd); | 120 | discover_at(idis,xrds_location,xmode_xrd); |
121 | } | 121 | } |
122 | idis.normalized_id = idis.canonicalized_id = yurl; | 122 | idis.normalized_id = idis.canonicalized_id = yurl; |
123 | service_type_t st; | 123 | service_type_t st = { 0, 0 }; |
124 | for(st.uri=*types;*types;st.uri=*(++types)) | 124 | for(st.uri=*types;*types;st.uri=*(++types)) |
125 | queue_endpoints(oi,idis,&st); | 125 | queue_endpoints(oi,idis,&st); |
126 | } | 126 | } |
127 | 127 | ||
128 | string discover(endpoint_discovery_iterator& oi,const string& identity) { | 128 | string discover(endpoint_discovery_iterator& oi,const string& identity) { |
129 | string rv; | 129 | string rv; |
130 | idiscovery_t idis; | 130 | idiscovery_t idis; |
131 | string::size_type fsc = identity.find_first_not_of(data::_whitespace_chars); | 131 | string::size_type fsc = identity.find_first_not_of(data::_whitespace_chars); |
132 | if(fsc==string::npos) | 132 | if(fsc==string::npos) |
133 | throw bad_input(OPKELE_CP_ "whitespace-only identity"); | 133 | throw bad_input(OPKELE_CP_ "whitespace-only identity"); |
134 | string::size_type lsc = identity.find_last_not_of(data::_whitespace_chars); | 134 | string::size_type lsc = identity.find_last_not_of(data::_whitespace_chars); |
135 | assert(lsc!=string::npos); | 135 | assert(lsc!=string::npos); |
136 | if(!strncasecmp(identity.c_str()+fsc,"xri://",sizeof("xri://")-1)) | 136 | if(!strncasecmp(identity.c_str()+fsc,"xri://",sizeof("xri://")-1)) |
137 | fsc += sizeof("xri://")-1; | 137 | fsc += sizeof("xri://")-1; |
138 | if((fsc+1)>=lsc) | 138 | if((fsc+1)>=lsc) |
139 | throw bad_input(OPKELE_CP_ "not a character of importance in identity"); | 139 | throw bad_input(OPKELE_CP_ "not a character of importance in identity"); |
140 | string id(identity,fsc,lsc-fsc+1); | 140 | string id(identity,fsc,lsc-fsc+1); |
141 | idis.clear(); | 141 | idis.clear(); |
142 | if(strchr(data::_iname_leaders,id[0])) { | 142 | if(strchr(data::_iname_leaders,id[0])) { |
143 | /* TODO: further normalize xri identity? Like folding case | 143 | /* TODO: further normalize xri identity? Like folding case |
144 | * or whatever... */ | 144 | * or whatever... */ |
145 | rv = id; | 145 | rv = id; |
146 | set<string> cids; | 146 | set<string> cids; |
147 | for(const struct service_type_t *st=op_service_types; | 147 | for(const struct service_type_t *st=op_service_types; |
148 | st<&op_service_types[sizeof(op_service_types)/sizeof(*op_service_types)];++st) { | 148 | st<&op_service_types[sizeof(op_service_types)/sizeof(*op_service_types)];++st) { |
149 | idis.clear(); | 149 | idis.clear(); |
150 | discover_at( idis, | 150 | discover_at( idis, |
151 | xri_proxy + util::url_encode(id)+ | 151 | xri_proxy + util::url_encode(id)+ |
152 | "?_xrd_t="+util::url_encode(st->uri)+ | 152 | "?_xrd_t="+util::url_encode(st->uri)+ |
153 | "&_xrd_r=application/xrd%2Bxml" | 153 | "&_xrd_r=application/xrd%2Bxml" |
154 | ";sep=true;refs=true", | 154 | ";sep=true;refs=true", |
155 | xmode_xrd ); | 155 | xmode_xrd ); |
156 | if(status_code==241) continue; | 156 | if(status_code==241) continue; |
157 | if(status_code!=100) | 157 | if(status_code!=100) |
158 | throw failed_xri_resolution(OPKELE_CP_ | 158 | throw failed_xri_resolution(OPKELE_CP_ |
159 | "XRI resolution failed with '"+status_string+"' message" | 159 | "XRI resolution failed with '"+status_string+"' message" |
160 | ", while looking for SEP with type '"+st->uri+"'", status_code); | 160 | ", while looking for SEP with type '"+st->uri+"'", status_code); |
161 | if(idis.xrd.canonical_ids.empty()) | 161 | if(idis.xrd.canonical_ids.empty()) |
162 | throw opkele::failed_discovery(OPKELE_CP_ "No CanonicalID for XRI identity found"); | 162 | throw opkele::failed_discovery(OPKELE_CP_ "No CanonicalID for XRI identity found"); |
163 | string cid = idis.xrd.canonical_ids.begin()->second; | 163 | string cid = idis.xrd.canonical_ids.begin()->second; |
164 | if(cids.find(cid)==cids.end()) { | 164 | if(cids.find(cid)==cids.end()) { |
165 | cids.insert(cid); | 165 | cids.insert(cid); |
166 | idis.clear(); | 166 | idis.clear(); |
167 | discover_at( idis, | 167 | discover_at( idis, |
168 | xri_proxy + util::url_encode(id)+ | 168 | xri_proxy + util::url_encode(id)+ |
169 | "?_xrd_t="+util::url_encode(st->uri)+ | 169 | "?_xrd_t="+util::url_encode(st->uri)+ |
170 | "&_xrd_r=application/xrd%2Bxml" | 170 | "&_xrd_r=application/xrd%2Bxml" |
171 | ";sep=true;refs=true", | 171 | ";sep=true;refs=true", |
172 | xmode_xrd ); | 172 | xmode_xrd ); |
173 | if(status_code==241) continue; | 173 | if(status_code==241) continue; |
174 | if(status_code!=100) | 174 | if(status_code!=100) |
175 | throw failed_xri_resolution(OPKELE_CP_ | 175 | throw failed_xri_resolution(OPKELE_CP_ |
176 | "XRI resolution failed with '"+status_string+"' message" | 176 | "XRI resolution failed with '"+status_string+"' message" |
177 | ", while looking for SEP with type '"+st->uri+"'" | 177 | ", while looking for SEP with type '"+st->uri+"'" |
178 | " on canonical id", status_code); | 178 | " on canonical id", status_code); |
179 | } | 179 | } |
180 | idis.canonicalized_id = cid; | 180 | idis.canonicalized_id = cid; |
181 | idis.normalized_id = rv; idis.xri_identity = true; | 181 | idis.normalized_id = rv; idis.xri_identity = true; |
182 | queue_endpoints(oi,idis,st); | 182 | queue_endpoints(oi,idis,st); |
183 | } | 183 | } |
184 | }else{ | 184 | }else{ |
185 | idis.xri_identity = false; | 185 | idis.xri_identity = false; |
186 | if(id.find("://")==string::npos) | 186 | if(id.find("://")==string::npos) |
187 | id.insert(0,"http://"); | 187 | id.insert(0,"http://"); |
188 | string::size_type fp = id.find('#'); | 188 | string::size_type fp = id.find('#'); |
189 | if(fp!=string::npos) { | 189 | if(fp!=string::npos) { |
190 | string::size_type qp = id.find('?'); | 190 | string::size_type qp = id.find('?'); |
191 | if(qp==string::npos || qp<fp) | 191 | if(qp==string::npos || qp<fp) |
192 | id.erase(fp); | 192 | id.erase(fp); |
193 | else if(qp>fp) | 193 | else if(qp>fp) |
194 | id.erase(fp,qp-fp); | 194 | id.erase(fp,qp-fp); |
195 | } | 195 | } |
196 | rv = idis.normalized_id = util::rfc_3986_normalize_uri(id); | 196 | rv = idis.normalized_id = util::rfc_3986_normalize_uri(id); |
197 | discover_at(idis,id,xmode_html|xmode_xrd); | 197 | discover_at(idis,id,xmode_html|xmode_xrd); |
198 | const char * eu = 0; | 198 | const char * eu = 0; |
199 | CURLcode r = easy_getinfo(CURLINFO_EFFECTIVE_URL,&eu); | 199 | CURLcode r = easy_getinfo(CURLINFO_EFFECTIVE_URL,&eu); |
200 | if(r) | 200 | if(r) |
201 | throw exception_curl(OPKELE_CP_ "failed to get CURLINFO_EFFECTIVE_URL",r); | 201 | throw exception_curl(OPKELE_CP_ "failed to get CURLINFO_EFFECTIVE_URL",r); |
202 | string cid = util::strip_uri_fragment_part( idis.canonicalized_id = util::rfc_3986_normalize_uri(eu) ); | 202 | string cid = util::strip_uri_fragment_part( idis.canonicalized_id = util::rfc_3986_normalize_uri(eu) ); |
203 | if(xrds_location.empty()) { | 203 | if(xrds_location.empty()) { |
204 | if(idis.xrd.empty()) | 204 | if(idis.xrd.empty()) |
205 | html2xrd(oi,idis); | 205 | html2xrd(oi,idis); |
206 | else{ | 206 | else{ |
207 | for(const service_type_t *st=op_service_types; | 207 | for(const service_type_t *st=op_service_types; |
208 | st<&op_service_types[sizeof(op_service_types)/sizeof(*op_service_types)];++st) | 208 | st<&op_service_types[sizeof(op_service_types)/sizeof(*op_service_types)];++st) |
209 | queue_endpoints(oi,idis,st); | 209 | queue_endpoints(oi,idis,st); |
210 | } | 210 | } |
211 | }else{ | 211 | }else{ |
212 | idis.clear(); | 212 | idis.clear(); |
213 | idis.canonicalized_id = cid; | 213 | idis.canonicalized_id = cid; |
214 | discover_at(idis,xrds_location,xmode_xrd); | 214 | discover_at(idis,xrds_location,xmode_xrd); |
215 | if(idis.xrd.empty()) | 215 | if(idis.xrd.empty()) |
216 | html2xrd(oi,idis); | 216 | html2xrd(oi,idis); |
217 | else{ | 217 | else{ |
218 | for(const service_type_t *st=op_service_types; | 218 | for(const service_type_t *st=op_service_types; |
219 | st<&op_service_types[sizeof(op_service_types)/sizeof(*op_service_types)];++st) | 219 | st<&op_service_types[sizeof(op_service_types)/sizeof(*op_service_types)];++st) |
220 | queue_endpoints(oi,idis,st); | 220 | queue_endpoints(oi,idis,st); |
221 | } | 221 | } |
222 | } | 222 | } |
223 | } | 223 | } |
224 | return rv; | 224 | return rv; |
225 | } | 225 | } |
226 | 226 | ||
227 | void discover_at(idiscovery_t& idis,const string& url,int xm) { | 227 | void discover_at(idiscovery_t& idis,const string& url,int xm) { |
228 | CURLcode r = easy_setopt(CURLOPT_MAXREDIRS, (xm&xmode_noredirs)?0:5); | 228 | CURLcode r = easy_setopt(CURLOPT_MAXREDIRS, (xm&xmode_noredirs)?0:5); |
229 | if(r) | 229 | if(r) |
230 | throw exception_curl(OPKELE_CP_ "failed to set curly maxredirs option"); | 230 | throw exception_curl(OPKELE_CP_ "failed to set curly maxredirs option"); |
231 | if( (r=easy_setopt(CURLOPT_URL,url.c_str())) ) | 231 | if( (r=easy_setopt(CURLOPT_URL,url.c_str())) ) |
232 | throw exception_curl(OPKELE_CP_ "failed to set curly urlie",r); | 232 | throw exception_curl(OPKELE_CP_ "failed to set curly urlie",r); |
233 | 233 | ||
234 | http_content_type.clear(); | 234 | http_content_type.clear(); |
235 | xmode = xm; | 235 | xmode = xm; |
236 | prepare_to_parse(); | 236 | prepare_to_parse(); |
237 | if(xmode&xmode_html) { | 237 | if(xmode&xmode_html) { |
238 | xrds_location.clear(); | 238 | xrds_location.clear(); |
239 | save_html.clear(); | 239 | save_html.clear(); |
240 | save_html.reserve(max_html); | 240 | save_html.reserve(max_html); |
241 | } | 241 | } |
242 | xrd = &idis.xrd; | 242 | xrd = &idis.xrd; |
243 | 243 | ||
244 | r = easy_perform(); | 244 | r = easy_perform(); |
245 | if(r && r!=CURLE_WRITE_ERROR) | 245 | if(r && r!=CURLE_WRITE_ERROR) |
246 | throw exception_curl(OPKELE_CP_ "failed to perform curly request",r); | 246 | throw exception_curl(OPKELE_CP_ "failed to perform curly request",r); |
247 | 247 | ||
248 | if(!parser_choked) { | 248 | if(!parser_choked) { |
249 | parse(0,0,true); | 249 | parse(0,0,true); |
250 | }else if(xmode&xmode_html){ | 250 | }else if(xmode&xmode_html){ |
251 | /* TODO: do not bother if we've seen xml */ | 251 | /* TODO: do not bother if we've seen xml */ |