Mercurial > irccd
annotate plugins/links/links.c @ 992:462e12e434fb
cmake: add CMake configuration
| author | David Demelier <markand@malikania.fr> |
|---|---|
| date | Wed, 10 Feb 2021 21:31:32 +0100 |
| parents | daaf92c097e6 |
| children | 936cbd66b4b8 |
| rev | line source |
|---|---|
| 987 | 1 /* |
| 2 * links.c -- links plugin | |
| 3 * | |
| 4 * Copyright (c) 2013-2021 David Demelier <markand@malikania.fr> | |
| 5 * | |
| 6 * Permission to use, copy, modify, and/or distribute this software for any | |
| 7 * purpose with or without fee is hereby granted, provided that the above | |
| 8 * copyright notice and this permission notice appear in all copies. | |
| 9 * | |
| 10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES | |
| 11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF | |
| 12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR | |
| 13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES | |
| 14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN | |
| 15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF | |
| 16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. | |
| 17 */ | |
| 18 | |
| 19 #include <assert.h> | |
| 20 #include <ctype.h> | |
| 21 #include <err.h> | |
| 22 #include <errno.h> | |
| 23 #include <regex.h> | |
| 24 #include <stdio.h> | |
| 25 #include <stdlib.h> | |
| 26 #include <string.h> | |
| 27 | |
| 28 #include <curl/curl.h> | |
| 29 | |
|
992
462e12e434fb
cmake: add CMake configuration
David Demelier <markand@malikania.fr>
parents:
990
diff
changeset
|
30 #include <irccd/compat.h> |
| 987 | 31 #include <irccd/irccd.h> |
| 32 #include <irccd/limits.h> | |
| 33 #include <irccd/server.h> | |
| 34 #include <irccd/subst.h> | |
| 35 #include <irccd/util.h> | |
| 36 | |
| 37 /* | |
| 38 * Since most websites are now bloated, we need a very large page size to | |
| 39 * analyse. Use 5MB for now. | |
| 40 */ | |
| 41 #define PAGE_MAX 5242880 | |
| 42 | |
| 43 struct req { | |
| 44 pthread_t thr; | |
| 45 CURL *curl; | |
| 46 struct irc_server *server; | |
| 47 int status; | |
| 48 char *link; | |
| 49 char *chan; | |
| 50 char *nickname; | |
| 51 char *origin; | |
| 52 FILE *fp; | |
| 53 char buf[]; | |
| 54 }; | |
| 55 | |
| 56 enum { | |
| 57 TPL_INFO | |
| 58 }; | |
| 59 | |
|
990
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
60 static unsigned long timeout = 30; |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
61 |
| 987 | 62 static char templates[][512] = { |
| 63 [TPL_INFO] = "#{nickname}, voici le lien: #{title}" | |
| 64 }; | |
| 65 | |
| 66 static size_t | |
| 67 callback(char *ptr, size_t size, size_t nmemb, struct req *req) | |
| 68 { | |
| 69 fprintf(req->fp, "%.*s", (int)(size * nmemb), ptr); | |
| 70 | |
| 71 if (feof(req->fp) || ferror(req->fp)) | |
| 72 return 0; | |
| 73 | |
| 74 return size * nmemb; | |
| 75 } | |
| 76 | |
| 77 static const char * | |
| 78 parse(struct req *req) | |
| 79 { | |
| 80 regex_t regex; | |
| 81 regmatch_t match[2]; | |
| 82 char *ret = NULL; | |
| 83 | |
| 84 if (regcomp(®ex, "<title>([^<]+)<\\/title>", REG_EXTENDED | REG_ICASE) != 0) | |
| 85 return NULL; | |
| 86 | |
| 87 if (regexec(®ex, req->buf, 2, match, 0) == 0) { | |
| 88 ret = &req->buf[match[1].rm_so]; | |
| 89 ret[match[1].rm_eo - match[1].rm_so] = '\0'; | |
| 90 } | |
| 91 | |
| 92 regfree(®ex); | |
| 93 | |
| 94 return ret; | |
| 95 } | |
| 96 | |
| 97 static const char * | |
| 98 fmt(const struct req *req, const char *title) | |
| 99 { | |
| 100 static char line[IRC_MESSAGE_LEN]; | |
| 101 struct irc_subst subst = { | |
| 102 .time = time(NULL), | |
| 103 .flags = IRC_SUBST_DATE | IRC_SUBST_KEYWORDS | IRC_SUBST_IRC_ATTRS, | |
| 104 .keywords = (struct irc_subst_keyword []) { | |
| 105 { "channel", req->chan }, | |
| 106 { "nickname", req->nickname }, | |
| 107 { "origin", req->origin }, | |
| 108 { "server", req->server->name }, | |
| 109 { "title", title } | |
| 110 }, | |
| 111 .keywordsz = 5 | |
| 112 }; | |
| 113 | |
| 114 irc_subst(line, sizeof (line), templates[TPL_INFO], &subst); | |
| 115 | |
| 116 return line; | |
| 117 } | |
| 118 | |
| 119 static void | |
| 120 req_finish(struct req *); | |
| 121 | |
| 122 static void | |
| 123 complete(struct req *req) | |
| 124 { | |
| 125 const char *title; | |
| 126 | |
| 127 if (req->status && (title = parse(req))) | |
| 128 irc_server_message(req->server, req->chan, fmt(req, title)); | |
| 129 | |
| 130 pthread_join(req->thr, NULL); | |
| 131 req_finish(req); | |
| 132 } | |
| 133 | |
| 134 /* | |
| 135 * This function is running in a separate thread. | |
| 136 */ | |
| 137 static void * | |
| 138 routine(struct req *req) | |
| 139 { | |
| 140 typedef void (*func_t)(void *); | |
| 141 | |
| 142 if (curl_easy_perform(req->curl) == 0) | |
| 143 req->status = 1; | |
| 144 | |
| 145 irc_bot_post((func_t)complete, req); | |
| 146 | |
| 147 return NULL; | |
| 148 } | |
| 149 | |
| 150 static void | |
| 151 req_finish(struct req *req) | |
| 152 { | |
| 153 assert(req); | |
| 154 | |
| 155 if (req->server) | |
| 156 irc_server_decref(req->server); | |
| 157 if (req->curl) | |
| 158 curl_easy_cleanup(req->curl); | |
| 159 if (req->fp) | |
| 160 fclose(req->fp); | |
| 161 | |
| 162 free(req->chan); | |
| 163 free(req->nickname); | |
| 164 free(req->origin); | |
| 165 free(req); | |
| 166 } | |
| 167 | |
| 168 static struct req * | |
| 169 req_new(struct irc_server *server, const char *origin, const char *channel, char *link) | |
| 170 { | |
| 171 assert(link); | |
| 172 | |
| 173 struct req *req; | |
| 174 struct irc_server_user user; | |
| 175 | |
| 176 if (!(req = calloc(1, sizeof (*req) + PAGE_MAX + 1))) { | |
| 177 free(link); | |
| 178 return NULL; | |
| 179 } | |
| 180 if (!(req->curl = curl_easy_init())) | |
| 181 goto enomem; | |
| 182 if (!(req->fp = fmemopen(req->buf, PAGE_MAX, "w"))) | |
| 183 goto enomem; | |
| 184 | |
| 185 irc_server_incref(server); | |
| 186 irc_server_split(origin, &user); | |
| 187 req->link = link; | |
| 188 req->server = server; | |
| 189 req->chan = strdup(channel); | |
| 190 req->nickname = strdup(user.nickname); | |
| 191 req->origin = strdup(origin); | |
| 192 | |
| 193 curl_easy_setopt(req->curl, CURLOPT_URL, link); | |
| 194 curl_easy_setopt(req->curl, CURLOPT_FOLLOWLOCATION, 1L); | |
| 195 curl_easy_setopt(req->curl, CURLOPT_WRITEFUNCTION, callback); | |
| 196 curl_easy_setopt(req->curl, CURLOPT_WRITEDATA, req); | |
|
990
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
197 curl_easy_setopt(req->curl, CURLOPT_TIMEOUT, timeout); |
| 987 | 198 |
| 199 return req; | |
| 200 | |
| 201 enomem: | |
| 202 errno = ENOMEM; | |
| 203 req_finish(req); | |
| 204 | |
| 205 return NULL; | |
| 206 } | |
| 207 | |
| 208 static void | |
| 209 req_start(struct req *req) | |
| 210 { | |
| 211 typedef void *(*func_t)(void *); | |
| 212 | |
| 213 if (pthread_create(&req->thr, NULL, (func_t)routine, req) != 0) | |
| 214 req_finish(req); | |
| 215 } | |
| 216 | |
| 217 void | |
| 218 links_event(const struct irc_event *ev) | |
| 219 { | |
| 220 struct req *req; | |
| 221 char *loc, *link, *end; | |
| 222 | |
| 223 if (ev->type != IRC_EVENT_MESSAGE) | |
| 224 return; | |
| 225 | |
| 226 /* Parse link. */ | |
| 227 if (!(loc = strstr(ev->message.message, "http://")) && | |
| 228 !(loc = strstr(ev->message.message, "https://"))) | |
| 229 return; | |
| 230 | |
| 231 /* Compute end to allocate only what's needed. */ | |
| 232 for (end = loc; *end && !isspace(*end); ) | |
| 233 ++end; | |
| 234 | |
| 235 link = irc_util_strndup(loc, end - loc); | |
| 236 | |
| 237 /* If the function fails, link is free'd anyway. */ | |
| 238 if (!(req = req_new(ev->server, ev->message.origin, ev->message.channel, link))) | |
| 239 return; | |
| 240 | |
| 241 req_start(req); | |
| 242 } | |
| 243 | |
| 244 void | |
| 245 links_set_template(const char *key, const char *value) | |
| 246 { | |
| 247 if (strcmp(key, "info") == 0) | |
| 248 strlcpy(templates[TPL_INFO], value, sizeof (templates[TPL_INFO])); | |
| 249 } | |
| 250 | |
| 251 const char * | |
| 252 links_get_template(const char *key) | |
| 253 { | |
| 254 if (strcmp(key, "info") == 0) | |
| 255 return templates[TPL_INFO]; | |
| 256 | |
| 257 return NULL; | |
| 258 } | |
| 259 | |
| 260 const char ** | |
| 261 links_get_templates(void) | |
| 262 { | |
| 263 static const char *keys[] = { "info", NULL }; | |
| 264 | |
| 265 return keys; | |
| 266 } | |
|
988
2e02fbb8b32b
irccd: load metadata from native plugins
David Demelier <markand@malikania.fr>
parents:
987
diff
changeset
|
267 |
|
990
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
268 void |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
269 links_set_option(const char *key, const char *value) |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
270 { |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
271 if (strcmp(key, "timeout") == 0) |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
272 timeout = atol(value); |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
273 } |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
274 |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
275 const char * |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
276 links_get_option(const char *key) |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
277 { |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
278 static char out[32]; |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
279 |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
280 if (strcmp(key, "timeout") == 0) { |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
281 snprintf(out, sizeof (out), "%lu", timeout); |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
282 return out; |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
283 } |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
284 |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
285 return NULL; |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
286 } |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
287 |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
288 const char ** |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
289 links_get_options(void) |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
290 { |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
291 static const char *keys[] = { "timeout", NULL }; |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
292 |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
293 return keys; |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
294 } |
|
daaf92c097e6
plugin links: add timeout support
David Demelier <markand@malikania.fr>
parents:
988
diff
changeset
|
295 |
|
988
2e02fbb8b32b
irccd: load metadata from native plugins
David Demelier <markand@malikania.fr>
parents:
987
diff
changeset
|
296 const char *links_description = "Parse links from messages"; |
|
2e02fbb8b32b
irccd: load metadata from native plugins
David Demelier <markand@malikania.fr>
parents:
987
diff
changeset
|
297 const char *links_version = IRCCD_VERSION; |
|
2e02fbb8b32b
irccd: load metadata from native plugins
David Demelier <markand@malikania.fr>
parents:
987
diff
changeset
|
298 const char *links_license = "ISC"; |
|
2e02fbb8b32b
irccd: load metadata from native plugins
David Demelier <markand@malikania.fr>
parents:
987
diff
changeset
|
299 const char *links_author = "David Demelier <markand@malikania.fr>"; |