1 (* COCANWIKI - a wiki written in Objective CAML.
2 * Written by Richard W.M. Jones <rich@merjis.com>.
3 * Copyright (C) 2004 Merjis Ltd.
4 * $Id: page.ml,v 1.51 2006/08/03 13:52:58 rich Exp $
6 * This program is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License as published by
8 * the Free Software Foundation; either version 2 of the License, or
9 * (at your option) any later version.
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU General Public License for more details.
16 * You should have received a copy of the GNU General Public License
17 * along with this program; see the file COPYING. If not, write to
18 * the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19 * Boston, MA 02111-1307, USA.
31 open Cocanwiki_template
34 open Cocanwiki_server_settings
36 open Cocanwiki_extensions
37 open Cocanwiki_strings
39 type fp_status = FPOK of int32 * string * string * Calendar.t * bool
40 | FPInternalRedirect of string
41 | FPExternalRedirect of string
44 (* Referer strings which help us decide if the user came from
45 * a search engine and highlight terms in the page appropriately.
47 let search_engines = [
48 Pcre.regexp "^http://.*google\\.", [ "q"; "as_q"; "as_epq"; "as_oq" ];
49 Pcre.regexp "^http://.*yahoo\\.", [ "p" ];
50 Pcre.regexp "^http://.*msn\\.", [ "q"; "MT" ]
52 let split_words = Pcre.regexp "\\W+"
54 let split_qs_re = Pcre.regexp "\\?"
56 let xhtml_re = Pcre.regexp "<.*?>|[^<>]+"
58 let run r (q : cgi) dbh hostid
59 ({ edit_anon = edit_anon; view_anon = view_anon } as host)
61 let page = q#param "page" in
62 let page = if page = "" then "index" else page in
64 (* The main "page" template is split in two to improve the speed of
65 * delivery of the page. The very first part ("page_header.html")
66 * contains the page <head>, crucially including all the links to the
67 * stylesheets. We send this first and flush it out to the client so
68 * that the client can begin requesting stylesheets, background images
69 * and so on. After this we compose the main page ("page.html") and
73 let template_page_header =
74 get_template ~page dbh hostid "page_header.html" in
75 let template_page = get_template ~page dbh hostid "page.html" in
77 (* This is the simpler template for 404 pages. *)
78 let template_404_header = get_template dbh hostid "page_404_header.html" in
79 let template_404 = get_template dbh hostid "page_404.html" in
81 (* Host-specific fields. *)
84 "select feedback_email is not null,
85 mailing_list, navigation
86 from hosts where id = $hostid" in
87 let has_feedback_email, mailing_list, navigation =
89 | [Some has_feedback_email, mailing_list, navigation] ->
90 has_feedback_email, mailing_list, navigation
91 | _ -> assert false in
93 (* User permissions. *)
94 let can_edit = can_edit host user in
95 let can_manage_users = can_manage_users host user in
97 (* Do we have a stats page set up? *)
98 let has_stats = server_settings_stats_page dbh <> None in
100 (* Given the referer string, return the list of search terms. If none
101 * can be found, then throws Not_found.
103 let search_terms_from_referer referer =
105 List.find (fun (rex, _) -> Pcre.pmatch ~rex referer) search_engines in
107 match Pcre.split ~rex:split_qs_re ~max:2 referer with
108 | [url] | [url;""] -> url, ""
109 | [url;qs] -> url, qs
110 | _ -> assert false in
111 let args = Cgi_args.parse qs in
113 List.find (fun argname -> List.mem_assoc argname args) argnames in
114 let search_string = List.assoc argname args in
115 Pcre.split ~rex:split_words search_string
118 (* Given a full page of XHTML, highlight search terms found in the
119 * <body> part of the page.
121 let highlight_search_terms xhtml search_terms span_class =
122 (* Split the original XHTML into strings and tags. For example if
123 * the original string is: "This is some <b>bold</b> text.<br/>", then
124 * after this step we will have the following list:
125 * [ "This is some "; "<b>"; "bold"; "</b>"; " text."; "<br/>" ]
127 let xhtml = Pcre.extract_all ~rex:xhtml_re xhtml in
128 let xhtml = Array.to_list xhtml in
129 let xhtml = List.map (fun matches -> matches.(0)) xhtml in
131 (* Find the <body> ... </body> tags. We only want to apply
132 * highlighting to tags within this area.
134 let rec list_split f acc = function
135 | [] -> List.rev acc, []
136 | ((x :: _) as xs) when f x -> List.rev acc, xs
138 let acc = x :: acc in
142 list_split (fun str -> String.starts_with str "<body") [] xhtml in
144 list_split ((=) "</body>") [] body in
145 (* NB: Hopefully, xhtml = head @ body @ tail. *)
147 (* The search terms are a list of simple words. Turn into a big
148 * regular expression, because we want to substitute for each. We
149 * end up with a regexp like '(word1|word2|word3)'.
152 Pcre.regexp ~flags:[`CASELESS]
153 ("(" ^ String.concat "|" search_terms ^ ")") in
155 (* Do the substitution, but only on text, not elements! *)
158 "<span class=\"" ^ span_class ^ "\">" ^ text ^ "</span>"
161 if String.length str > 0 && str.[0] != '<' then
162 Pcre.substitute ~rex ~subst str
166 (* Join the XHTML fragments back together again. *)
167 String.concat "" (List.concat [ head ; body ; tail ])
170 (* Check the templates table for extensions. *)
171 let get_extension url =
175 PGSQL(dbh) "select extension from templates
176 where $url ~ url_regexp
180 Some (List.assoc name !extensions)
182 Not_found | ExtList.List.Empty_list | Failure "hd" -> None
185 (* This code generates ordinary pages. *)
186 let make_page title description pageid last_modified_date has_page_css
187 version page page' extension =
188 let t = template_page in
189 let th = template_page_header in
191 th#set "title" title;
192 t#set "last_modified_date" last_modified_date;
194 (match description with
195 None -> th#conditional "has_description" false
196 | Some description ->
197 th#conditional "has_description" true;
198 th#set "description" description);
200 if page <> page' then (* redirection *) (
203 t#set "original_page" page; (* XXX title - get it from database *)
204 t#conditional "redirected" true
208 t#conditional "redirected" false
211 th#conditional "has_page_css" has_page_css;
213 (* Are we showing an old version of the page? If so, warn. *)
216 t#conditional "is_old_version" false;
217 th#conditional "is_old_version" false
219 t#conditional "is_old_version" true;
220 th#conditional "is_old_version" true;
221 t#set "old_version" (Int32.to_string pageid);
222 th#set "old_version" (Int32.to_string pageid));
224 (* Just before we show the header, call any registered pre-page
225 * handlers. They might want to send cookies.
227 List.iter (fun handler ->
228 handler r q dbh hostid page') !pre_page_handlers;
230 (* At this point, we can print out the header and flush it back to
231 * the user, allowing the browser to start fetching stylesheets
232 * and background images while we compose the page.
235 ignore (print_string r th#to_string);
236 ignore (Request.rflush r);
238 t#conditional "has_feedback_email" has_feedback_email;
239 t#conditional "mailing_list" mailing_list;
240 t#conditional "navigation" navigation;
242 t#conditional "can_edit" can_edit;
243 t#conditional "can_manage_users" can_manage_users;
244 t#conditional "has_stats" has_stats;
246 (* Pull out the sections in this page. *)
251 let rows = PGSQL(dbh)
252 "select ordering, sectionname, content, divname, jsgo
253 from contents where pageid = $pageid order by ordering" in
256 (fun (ordering, sectionname, content, divname, jsgo) ->
257 let divname, has_divname =
260 | Some divname -> divname, true in
264 | Some jsgo -> jsgo, true in
265 let sectionname, has_sectionname =
266 match sectionname with
268 | Some sectionname -> sectionname, true in
269 let linkname = linkname_of_sectionname sectionname in
270 [ "ordering", Template.VarString (Int32.to_string ordering);
272 Template.VarConditional has_sectionname;
273 "sectionname", Template.VarString sectionname;
274 "linkname", Template.VarString linkname;
277 (Wikilib.xhtml_of_content r dbh hostid content);
278 "has_divname", Template.VarConditional has_divname;
279 "divname", Template.VarString divname;
280 "has_jsgo", Template.VarConditional has_jsgo;
281 "jsgo", Template.VarString jsgo ]) rows in
283 (* Call an extension to generate the first section in this page? *)
288 let content = extension r dbh hostid page' in
290 "ordering", Template.VarString "0";
291 "has_sectionname", Template.VarConditional false;
292 "linkname", Template.VarString "";
293 "content", Template.VarString content;
294 "has_divname", Template.VarConditional true;
295 "divname", Template.VarString "form_div";
296 "has_jsgo", Template.VarConditional false;
297 "jsgo", Template.VarString "";
299 section :: sections in
301 t#table "sections" sections;
306 t#conditional "user_logged_in" false
307 | User (_, username, _, _) ->
308 t#conditional "user_logged_in" true;
309 t#set "username" username);
311 (* Can anonymous users create accounts? If not them we don't
312 * want to offer to create accounts for them.
314 t#conditional "create_account_anon" host.create_account_anon;
316 (* If logged in, we want to update the recently_visited table. *)
317 if pageid <> None then (
319 | User (userid, _, _, _) ->
322 "delete from recently_visited
323 where hostid = $hostid and userid = $userid
326 "insert into recently_visited (hostid, userid, url)
327 values ($hostid, $userid, $page')";
331 (* Exceptions here are non-fatal. Just print them. *)
332 prerr_endline "exception updating recently_visited:";
333 prerr_endline (Printexc.to_string exn);
334 PGOCaml.rollback dbh;
336 PGOCaml.begin_work dbh;
340 (* Navigation links. *)
342 let max_links = 18 in (* Show no more links than this. *)
344 (* What links here. *)
345 let wlh = what_links_here dbh hostid page' in
346 let wlh = List.take max_links wlh in
347 let wlh_urls = List.map fst wlh in (* Just the URLs ... *)
351 | User (userid, _, _, _) ->
352 (* Recently visited URLs, but don't repeat any from the 'what
353 * links here' section, and don't link to self.
355 let not_urls = page' :: wlh_urls in
356 let limit = Int32.of_int (max_links - List.length wlh_urls) in
359 "select rv.url, p.title, rv.visit_time
360 from recently_visited rv, pages p
361 where rv.hostid = $hostid and rv.userid = $userid
362 and rv.url not in $@not_urls
363 and rv.hostid = p.hostid and rv.url = p.url
367 fun (url, title, _) -> url, title
372 let f (page, title) = [ "page", Template.VarString page;
373 "title", Template.VarString title ] in
374 let table = List.map f wlh in
375 t#table "what_links_here" table;
376 t#conditional "has_what_links_here" (wlh <> []);
378 let table = List.map f rv in
379 t#table "recently_visited" table;
380 t#conditional "has_recently_visited" (rv <> []);
382 (* If both lists are empty (ie. an empty navigation box would
383 * appear), then disable navigation altogether.
385 if wlh = [] && rv = [] then t#conditional "navigation" false
388 (* If we are coming from a search engine then we want to highlight
389 * search terms throughout the whole page ...
392 let referer = Table.get (Request.headers_in r) "Referer" in
393 let search_terms = search_terms_from_referer referer in
395 (* Highlight the search terms. *)
396 let xhtml = t#to_string in
397 let xhtml = highlight_search_terms xhtml search_terms "search_term" in
399 (* Deliver the page. *)
400 ignore (print_string r xhtml)
403 (* No referer / no search terms / not a search engine referer. *)
404 ignore (print_string r t#to_string)
407 (* This code generates 404 pages. *)
409 Request.set_status r 404; (* Return a 404 error code. *)
411 let th = template_404_header in
417 ('a'..'z' | 'A'..'Z' | '0'..'9') as c -> c
420 th#set "search_terms" search_terms;
422 (* Flush out the header while we start the search. *)
424 ignore (print_string r th#to_string);
425 ignore (Request.rflush r);
427 let t = template_404 in
428 t#set "query" search_terms;
429 t#set "canonical_hostname" host.canonical_hostname;
431 (* This is a simplified version of the code in search.ml. *)
433 (* Get the keywords from the query string. *)
434 let keywords = Pcre.split ~rex:split_words search_terms in
436 List.filter (fun s -> not (string_is_whitespace s)) keywords in
437 let keywords = List.map String.lowercase keywords in
439 (* Turn the keywords into a tsearch2 ts_query string. *)
440 let tsquery = String.concat "&" keywords in
442 (* Search the titles first. *)
445 "select url, title, last_modified_date,
446 (lower (title) = lower ($search_terms)) as exact
448 where hostid = $hostid
451 and title_description_fti @@ to_tsquery ('default', $tsquery)
452 order by exact desc, last_modified_date desc, title" in
456 | (Some url, title, last_modified, _) ->
457 url, title, last_modified
458 | _ -> assert false) rows in
460 let have_titles = titles <> [] in
461 t#conditional "have_titles" have_titles;
463 (* Search the contents. *)
466 "select c.id, p.url, p.title, p.last_modified_date
467 from contents c, pages p
468 where c.pageid = p.id
469 and p.hostid = $hostid
471 and p.redirect is null
472 and c.content_fti @@ to_tsquery ('default', $tsquery)
473 order by p.last_modified_date desc, p.title
478 | (contentid, Some url, title, last_modified) ->
479 contentid, url, title, last_modified
480 | _ -> assert false) rows in
482 let have_contents = contents <> [] in
483 t#conditional "have_contents" have_contents;
485 (* Pull out the actual text which matched so we can generate a summary.
486 * XXX tsearch2 can actually do better than this by emboldening
487 * the text which maps.
490 if contents = [] then []
494 List.map (fun (contentid, _,_,_) -> contentid) contents in
496 "select id, sectionname, content from contents
497 where id in $@contentids" in
498 List.map (fun (id, sectionname, content) ->
499 id, (sectionname, content)) rows
502 (* Generate the final tables. *)
504 List.map (fun (url, title, last_modified) ->
505 let last_modified = printable_date last_modified in
506 [ "url", Template.VarString url;
507 "title", Template.VarString title;
508 "last_modified", Template.VarString last_modified ]
510 t#table "titles" table;
514 (fun (contentid, url, title, last_modified) ->
515 let sectionname, content = List.assoc contentid content_map in
516 let have_sectionname, sectionname =
517 match sectionname with
519 | Some sectionname -> true, sectionname in
522 (Wikilib.text_of_xhtml
523 (Wikilib.xhtml_of_content r dbh hostid content)) in
524 let linkname = linkname_of_sectionname sectionname in
525 let last_modified = printable_date last_modified in
526 [ "url", Template.VarString url;
527 "title", Template.VarString title;
528 "have_sectionname", Template.VarConditional have_sectionname;
529 "sectionname", Template.VarString sectionname;
530 "linkname", Template.VarString linkname;
531 "content", Template.VarString content;
532 "last_modified", Template.VarString last_modified ]
534 t#table "contents" table;
536 (* Do we have any results? *)
537 let have_results = have_titles || have_contents in
539 t#conditional "have_results" have_results;
541 (* Deliver the rest of the page. *)
542 ignore (print_string r t#to_string)
545 (* Fetch a page by name. This function can give three answers:
546 * (1) Page fetched OK (fetches some details of the page).
547 * (2) Page is a redirect (fetches the name of the redirect page).
548 * (3) Page not found in database, could be template or 404 error.
550 let fetch_page page version allow_redirect =
553 if allow_redirect then (
554 let rows = PGSQL(dbh)
555 "select url, redirect, id, title, description,
556 last_modified_date, css is not null
558 where hostid = $hostid and lower (url) = lower ($page)" in
560 | [Some page', _, _, _, _, _, _]
561 when page <> page' -> (* different case *)
562 FPExternalRedirect page'
563 | [ _, None, id, title, description,
564 last_modified_date, has_page_css ] ->
565 let has_page_css = Option.get has_page_css in
566 FPOK (id, title, description, last_modified_date,
568 | [_, Some redirect, _, _, _, _, _] ->
569 FPInternalRedirect redirect
572 ) else (* redirects not allowed ... *) (
573 let rows = PGSQL(dbh)
574 "select id, title, description, last_modified_date,
576 from pages where hostid = $hostid and url = $page" in
578 | [ id, title, description,
579 last_modified_date, has_page_css ] ->
580 let has_page_css = Option.get has_page_css in
581 FPOK (id, title, description, last_modified_date,
587 let rows = PGSQL(dbh)
588 "select id, title, description, last_modified_date,
591 where hostid = $hostid and id = $version and
592 (url = $page or url_deleted = $page)" in
594 | [ id, title, description,
595 last_modified_date, has_page_css ] ->
596 let has_page_css = Option.get has_page_css in
597 FPOK (id, title, description, last_modified_date,
603 (* Here we deal with the complex business of redirects and versions. *)
604 (* Only allow the no_redirect and version syntax for editors. *)
605 let allow_redirect, version =
607 not (q#param_true "no_redirect"),
608 try Some (Int32.of_string (q#param "version")) with Not_found -> None
612 let rec loop page' i =
613 if i > max_redirect then (
614 error ~title:"Too many redirections" ~back_button:true
616 ("Too many redirects between pages. This may happen because " ^
617 "of a cycle of redirections.");
620 match fetch_page page' version allow_redirect with
621 | FPOK (pageid, title, description, last_modified_date, has_page_css)->
622 (* Check if the page is also a template. *)
623 let extension = get_extension page' in
624 make_page title (Some description) (Some pageid)
625 (printable_date last_modified_date) has_page_css
626 version page page' extension
627 | FPInternalRedirect page' ->
629 | FPExternalRedirect page' ->
630 (* This normally happens when a user has requested an uppercase
631 * page name. We redirect to the true (lowercase) version.
633 q#redirect ("http://" ^ host.hostname ^ "/" ^ page')
635 (* Might be a templated page with no content in it. *)
636 let extension = get_extension page' in
637 (match extension with
638 | (Some _) as extension ->
640 make_page title None None
641 "Now" false None page page'
649 register_script ~restrict:[CanView] run