-rw-r--r-- | lib/util.cc | 39 |
1 files changed, 33 insertions, 6 deletions
diff --git a/lib/util.cc b/lib/util.cc index 29e6738..a6e08e2 100644 --- a/lib/util.cc +++ b/lib/util.cc | |||
@@ -1,77 +1,76 @@ | |||
1 | #include <errno.h> | 1 | #include <errno.h> |
2 | #include <cassert> | 2 | #include <cassert> |
3 | #include <cctype> | 3 | #include <cctype> |
4 | #include <cstring> | 4 | #include <cstring> |
5 | #include <vector> | 5 | #include <vector> |
6 | #include <string> | 6 | #include <string> |
7 | #include <stack> | 7 | #include <stack> |
8 | #include <algorithm> | 8 | #include <algorithm> |
9 | #include <openssl/bio.h> | 9 | #include <openssl/bio.h> |
10 | #include <openssl/evp.h> | 10 | #include <openssl/evp.h> |
11 | #include <openssl/sha.h> | 11 | #include <openssl/sha.h> |
12 | #include <openssl/hmac.h> | 12 | #include <openssl/hmac.h> |
13 | #include <curl/curl.h> | ||
14 | #include <opkele/util.h> | 13 | #include <opkele/util.h> |
15 | #include <opkele/exception.h> | 14 | #include <opkele/exception.h> |
16 | #include <opkele/data.h> | 15 | #include <opkele/data.h> |
17 | #include <opkele/debug.h> | 16 | #include <opkele/debug.h> |
18 | 17 | ||
19 | #include <config.h> | 18 | #include <config.h> |
20 | #ifdef HAVE_DEMANGLE | 19 | #ifdef HAVE_DEMANGLE |
21 | # include <cxxabi.h> | 20 | # include <cxxabi.h> |
22 | #endif | 21 | #endif |
23 | 22 | ||
24 | namespace opkele { | 23 | namespace opkele { |
25 | using namespace std; | 24 | using namespace std; |
26 | 25 | ||
27 | namespace util { | 26 | namespace util { |
28 | 27 | ||
29 | /* | 28 | /* |
30 | * base64 | 29 | * base64 |
31 | */ | 30 | */ |
32 | string encode_base64(const void *data,size_t length) { | 31 | string encode_base64(const void *data,size_t length) { |
33 | BIO *b64 = 0, *bmem = 0; | 32 | BIO *b64 = 0, *bmem = 0; |
34 | try { | 33 | try { |
35 | b64 = BIO_new(BIO_f_base64()); | 34 | b64 = BIO_new(BIO_f_base64()); |
36 | if(!b64) | 35 | if(!b64) |
37 | throw exception_openssl(OPKELE_CP_ "failed to BIO_new() base64 encoder"); | 36 | throw exception_openssl(OPKELE_CP_ "failed to BIO_new() base64 encoder"); |
38 | BIO_set_flags(b64,BIO_FLAGS_BASE64_NO_NL); | 37 | BIO_set_flags(b64,BIO_FLAGS_BASE64_NO_NL); |
39 | bmem = BIO_new(BIO_s_mem()); | 38 | bmem = BIO_new(BIO_s_mem()); |
40 | BIO_set_flags(b64,BIO_CLOSE); | 39 | BIO_set_flags(b64,BIO_CLOSE); |
41 | if(!bmem) | 40 | if(!bmem) |
42 | throw exception_openssl(OPKELE_CP_ "failed to BIO_new() memory buffer"); | 41 | throw exception_openssl(OPKELE_CP_ "failed to BIO_new() memory buffer"); |
43 | BIO_push(b64,bmem); | 42 | BIO_push(b64,bmem); |
44 | if(((size_t)BIO_write(b64,data,length))!=length) | 43 | if(((size_t)BIO_write(b64,data,length))!=length) |
45 | throw exception_openssl(OPKELE_CP_ "failed to BIO_write()"); | 44 | throw exception_openssl(OPKELE_CP_ "failed to BIO_write()"); |
46 | if(BIO_flush(b64)!=1) | 45 | if(BIO_flush(b64)!=1) |
47 | throw exception_openssl(OPKELE_CP_ "failed to BIO_flush()"); | 46 | throw exception_openssl(OPKELE_CP_ "failed to BIO_flush()"); |
48 | char *rvd; | 47 | char *rvd; |
49 | long rvl = BIO_get_mem_data(bmem,&rvd); | 48 | long rvl = BIO_get_mem_data(bmem,&rvd); |
50 | string rv(rvd,rvl); | 49 | string rv(rvd,rvl); |
51 | BIO_free_all(b64); | 50 | BIO_free_all(b64); |
52 | return rv; | 51 | return rv; |
53 | }catch(...) { | 52 | }catch(...) { |
54 | if(b64) BIO_free_all(b64); | 53 | if(b64) BIO_free_all(b64); |
55 | throw; | 54 | throw; |
56 | } | 55 | } |
57 | } | 56 | } |
58 | 57 | ||
59 | void decode_base64(const string& data,vector<unsigned char>& rv) { | 58 | void decode_base64(const string& data,vector<unsigned char>& rv) { |
60 | BIO *b64 = 0, *bmem = 0; | 59 | BIO *b64 = 0, *bmem = 0; |
61 | rv.clear(); | 60 | rv.clear(); |
62 | try { | 61 | try { |
63 | bmem = BIO_new_mem_buf((void*)data.data(),data.size()); | 62 | bmem = BIO_new_mem_buf((void*)data.data(),data.size()); |
64 | if(!bmem) | 63 | if(!bmem) |
65 | throw exception_openssl(OPKELE_CP_ "failed to BIO_new_mem_buf()"); | 64 | throw exception_openssl(OPKELE_CP_ "failed to BIO_new_mem_buf()"); |
66 | b64 = BIO_new(BIO_f_base64()); | 65 | b64 = BIO_new(BIO_f_base64()); |
67 | if(!b64) | 66 | if(!b64) |
68 | throw exception_openssl(OPKELE_CP_ "failed to BIO_new() base64 decoder"); | 67 | throw exception_openssl(OPKELE_CP_ "failed to BIO_new() base64 decoder"); |
69 | BIO_set_flags(b64,BIO_FLAGS_BASE64_NO_NL); | 68 | BIO_set_flags(b64,BIO_FLAGS_BASE64_NO_NL); |
70 | BIO_push(b64,bmem); | 69 | BIO_push(b64,bmem); |
71 | unsigned char tmp[512]; | 70 | unsigned char tmp[512]; |
72 | size_t rb = 0; | 71 | size_t rb = 0; |
73 | while((rb=BIO_read(b64,tmp,sizeof(tmp)))>0) | 72 | while((rb=BIO_read(b64,tmp,sizeof(tmp)))>0) |
74 | rv.insert(rv.end(),tmp,&tmp[rb]); | 73 | rv.insert(rv.end(),tmp,&tmp[rb]); |
75 | BIO_free_all(b64); | 74 | BIO_free_all(b64); |
76 | }catch(...) { | 75 | }catch(...) { |
77 | if(b64) BIO_free_all(b64); | 76 | if(b64) BIO_free_all(b64); |
@@ -95,134 +94,162 @@ namespace opkele { | |||
95 | BIGNUM *dec_to_bignum(const string& dec) { | 94 | BIGNUM *dec_to_bignum(const string& dec) { |
96 | BIGNUM *rv = 0; | 95 | BIGNUM *rv = 0; |
97 | if(!BN_dec2bn(&rv,dec.c_str())) | 96 | if(!BN_dec2bn(&rv,dec.c_str())) |
98 | throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()"); | 97 | throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()"); |
99 | return rv; | 98 | return rv; |
100 | } | 99 | } |
101 | 100 | ||
102 | string bignum_to_base64(const BIGNUM *bn) { | 101 | string bignum_to_base64(const BIGNUM *bn) { |
103 | vector<unsigned char> bin(BN_num_bytes(bn)+1); | 102 | vector<unsigned char> bin(BN_num_bytes(bn)+1); |
104 | unsigned char *binptr = &(bin.front())+1; | 103 | unsigned char *binptr = &(bin.front())+1; |
105 | int l = BN_bn2bin(bn,binptr); | 104 | int l = BN_bn2bin(bn,binptr); |
106 | if(l && (*binptr)&0x80){ | 105 | if(l && (*binptr)&0x80){ |
107 | (*(--binptr)) = 0; ++l; | 106 | (*(--binptr)) = 0; ++l; |
108 | } | 107 | } |
109 | return encode_base64(binptr,l); | 108 | return encode_base64(binptr,l); |
110 | } | 109 | } |
111 | 110 | ||
112 | /* | 111 | /* |
113 | * w3c times | 112 | * w3c times |
114 | */ | 113 | */ |
115 | 114 | ||
116 | string time_to_w3c(time_t t) { | 115 | string time_to_w3c(time_t t) { |
117 | struct tm tm_t; | 116 | struct tm tm_t; |
118 | if(!gmtime_r(&t,&tm_t)) | 117 | if(!gmtime_r(&t,&tm_t)) |
119 | throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()"); | 118 | throw failed_conversion(OPKELE_CP_ "failed to BN_dec2bn()"); |
120 | char rv[25]; | 119 | char rv[25]; |
121 | 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)) |
122 | throw failed_conversion(OPKELE_CP_ "failed to strftime()"); | 121 | throw failed_conversion(OPKELE_CP_ "failed to strftime()"); |
123 | return rv; | 122 | return rv; |
124 | } | 123 | } |
125 | 124 | ||
126 | time_t w3c_to_time(const string& w) { | 125 | time_t w3c_to_time(const string& w) { |
127 | int fraction; | 126 | int fraction; |
128 | struct tm tm_t; | 127 | struct tm tm_t; |
129 | memset(&tm_t,0,sizeof(tm_t)); | 128 | memset(&tm_t,0,sizeof(tm_t)); |
130 | if( ( | 129 | if( ( |
131 | sscanf( | 130 | sscanf( |
132 | w.c_str(), | 131 | w.c_str(), |
133 | "%04d-%02d-%02dT%02d:%02d:%02dZ", | 132 | "%04d-%02d-%02dT%02d:%02d:%02dZ", |
134 | &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday, | 133 | &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday, |
135 | &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec | 134 | &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec |
136 | ) != 6 | 135 | ) != 6 |
137 | ) && ( | 136 | ) && ( |
138 | sscanf( | 137 | sscanf( |
139 | w.c_str(), | 138 | w.c_str(), |
140 | "%04d-%02d-%02dT%02d:%02d:%02d.%03dZ", | 139 | "%04d-%02d-%02dT%02d:%02d:%02d.%03dZ", |
141 | &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday, | 140 | &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday, |
142 | &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec, | 141 | &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec, |
143 | &fraction | 142 | &fraction |
144 | ) != 7 | 143 | ) != 7 |
145 | ) ) | 144 | ) ) |
146 | throw failed_conversion(OPKELE_CP_ "failed to sscanf()"); | 145 | throw failed_conversion(OPKELE_CP_ "failed to sscanf()"); |
147 | tm_t.tm_mon--; | 146 | tm_t.tm_mon--; |
148 | tm_t.tm_year-=1900; | 147 | tm_t.tm_year-=1900; |
149 | time_t rv = mktime(&tm_t); | 148 | time_t rv = mktime(&tm_t); |
150 | if(rv==(time_t)-1) | 149 | if(rv==(time_t)-1) |
151 | throw failed_conversion(OPKELE_CP_ "failed to mktime()"); | 150 | throw failed_conversion(OPKELE_CP_ "failed to mktime()"); |
152 | return rv-timezone; | 151 | return rv-timezone; |
153 | } | 152 | } |
154 | 153 | ||
155 | /* | 154 | /* |
156 | * | 155 | * |
157 | */ | 156 | */ |
158 | 157 | ||
158 | static inline bool isrfc3986unreserved(int c) { | ||
159 | if(c<'-') return false; | ||
160 | if(c<='.') return true; | ||
161 | if(c<'0') return false; if(c<='9') return true; | ||
162 | if(c<'A') return false; if(c<='Z') return true; | ||
163 | if(c<'_') return false; | ||
164 | if(c=='_') return true; | ||
165 | if(c<'a') return false; if(c<='z') return true; | ||
166 | if(c=='~') return true; | ||
167 | return false; | ||
168 | } | ||
169 | |||
170 | struct __url_encoder : public unary_function<char,void> { | ||
171 | public: | ||
172 | string& rv; | ||
173 | |||
174 | __url_encoder(string& r) : rv(r) { } | ||
175 | |||
176 | result_type operator()(argument_type c) { | ||
177 | if(isrfc3986unreserved(c)) | ||
178 | rv += c; | ||
179 | else{ | ||
180 | char tmp[4]; | ||
181 | snprintf(tmp,sizeof(tmp),"%%%02X", | ||
182 | (c&0xff)); | ||
183 | rv += tmp; | ||
184 | } | ||
185 | } | ||
186 | }; | ||
187 | |||
159 | string url_encode(const string& str) { | 188 | string url_encode(const string& str) { |
160 | char * t = curl_escape(str.c_str(),str.length()); | 189 | string rv; |
161 | if(!t) | 190 | for_each(str.begin(),str.end(), |
162 | throw failed_conversion(OPKELE_CP_ "failed to curl_escape()"); | 191 | __url_encoder(rv)); |
163 | string rv(t); | ||
164 | curl_free(t); | ||
165 | return rv; | 192 | return rv; |
166 | } | 193 | } |
167 | 194 | ||
168 | string attr_escape(const string& str) { | 195 | string attr_escape(const string& str) { |
169 | static const char *unsafechars = "<>&\n\"'"; | 196 | static const char *unsafechars = "<>&\n\"'"; |
170 | string rv; | 197 | string rv; |
171 | string::size_type p=0; | 198 | string::size_type p=0; |
172 | while(true) { | 199 | while(true) { |
173 | string::size_type us = str.find_first_of(unsafechars,p); | 200 | string::size_type us = str.find_first_of(unsafechars,p); |
174 | if(us==string::npos) { | 201 | if(us==string::npos) { |
175 | if(p!=str.length()) | 202 | if(p!=str.length()) |
176 | rv.append(str,p,str.length()-p); | 203 | rv.append(str,p,str.length()-p); |
177 | return rv; | 204 | return rv; |
178 | } | 205 | } |
179 | rv.append(str,p,us-p); | 206 | rv.append(str,p,us-p); |
180 | rv += "&#"; | 207 | rv += "&#"; |
181 | rv += long_to_string((long)str[us]); | 208 | rv += long_to_string((long)str[us]); |
182 | rv += ';'; | 209 | rv += ';'; |
183 | p = us+1; | 210 | p = us+1; |
184 | } | 211 | } |
185 | } | 212 | } |
186 | 213 | ||
187 | string long_to_string(long l) { | 214 | string long_to_string(long l) { |
188 | char rv[32]; | 215 | char rv[32]; |
189 | int r=snprintf(rv,sizeof(rv),"%ld",l); | 216 | int r=snprintf(rv,sizeof(rv),"%ld",l); |
190 | if(r<0 || r>=(int)sizeof(rv)) | 217 | if(r<0 || r>=(int)sizeof(rv)) |
191 | throw failed_conversion(OPKELE_CP_ "failed to snprintf()"); | 218 | throw failed_conversion(OPKELE_CP_ "failed to snprintf()"); |
192 | return rv; | 219 | return rv; |
193 | } | 220 | } |
194 | 221 | ||
195 | long string_to_long(const string& s) { | 222 | long string_to_long(const string& s) { |
196 | char *endptr = 0; | 223 | char *endptr = 0; |
197 | long rv = strtol(s.c_str(),&endptr,10); | 224 | long rv = strtol(s.c_str(),&endptr,10); |
198 | if((!endptr) || endptr==s.c_str()) | 225 | if((!endptr) || endptr==s.c_str()) |
199 | throw failed_conversion(OPKELE_CP_ "failed to strtol()"); | 226 | throw failed_conversion(OPKELE_CP_ "failed to strtol()"); |
200 | return rv; | 227 | return rv; |
201 | } | 228 | } |
202 | 229 | ||
203 | /* | 230 | /* |
204 | * Normalize URL according to the rules, described in rfc 3986, section 6 | 231 | * Normalize URL according to the rules, described in rfc 3986, section 6 |
205 | * | 232 | * |
206 | * - uppercase hex triplets (e.g. %ab -> %AB) | 233 | * - uppercase hex triplets (e.g. %ab -> %AB) |
207 | * - lowercase scheme and host | 234 | * - lowercase scheme and host |
208 | * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3, | 235 | * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3, |
209 | * that is - [:alpha:][:digit:]._~- | 236 | * that is - [:alpha:][:digit:]._~- |
210 | * - remove dot segments | 237 | * - remove dot segments |
211 | * - remove empty and default ports | 238 | * - remove empty and default ports |
212 | * - if there's no path component, add '/' | 239 | * - if there's no path component, add '/' |
213 | */ | 240 | */ |
214 | string rfc_3986_normalize_uri(const string& uri) { | 241 | string rfc_3986_normalize_uri(const string& uri) { |
215 | string rv; | 242 | string rv; |
216 | string::size_type ns = uri.find_first_not_of(data::_whitespace_chars); | 243 | string::size_type ns = uri.find_first_not_of(data::_whitespace_chars); |
217 | if(ns==string::npos) | 244 | if(ns==string::npos) |
218 | throw bad_input(OPKELE_CP_ "Can't normalize empty URI"); | 245 | throw bad_input(OPKELE_CP_ "Can't normalize empty URI"); |
219 | string::size_type colon = uri.find(':',ns); | 246 | string::size_type colon = uri.find(':',ns); |
220 | if(colon==string::npos) | 247 | if(colon==string::npos) |
221 | throw bad_input(OPKELE_CP_ "No scheme specified in URI"); | 248 | throw bad_input(OPKELE_CP_ "No scheme specified in URI"); |
222 | transform( | 249 | transform( |
223 | uri.begin()+ns, uri.begin()+colon+1, | 250 | uri.begin()+ns, uri.begin()+colon+1, |
224 | back_inserter(rv), ::tolower ); | 251 | back_inserter(rv), ::tolower ); |
225 | bool s; | 252 | bool s; |
226 | string::size_type ul = uri.find_last_not_of(data::_whitespace_chars)+1; | 253 | string::size_type ul = uri.find_last_not_of(data::_whitespace_chars)+1; |
227 | if(ul <= (colon+3)) | 254 | if(ul <= (colon+3)) |
228 | throw bad_input(OPKELE_CP_ "Unexpected end of URI being normalized encountered"); | 255 | throw bad_input(OPKELE_CP_ "Unexpected end of URI being normalized encountered"); |