X-Git-Url: http://matita.cs.unibo.it/gitweb/?a=blobdiff_plain;ds=sidebyside;f=helm%2Fhttp_getter%2Fhttp_getter.ml;h=40f38dc4eec38598258f23e0fbb4a3f817689af0;hb=f54edaccaf3d56499632ce349276be5c2e711a6c;hp=44f24dee9247fe6a3ab3a3f1765b6569cc9464b1;hpb=32bc9336091a84376f93b06325c08dab58124a0d;p=helm.git diff --git a/helm/http_getter/http_getter.ml b/helm/http_getter/http_getter.ml index 44f24dee9..40f38dc4e 100644 --- a/helm/http_getter/http_getter.ml +++ b/helm/http_getter/http_getter.ml @@ -1,5 +1,7 @@ (* - * Copyright (C) 2003, HELM Team. + * Copyright (C) 2003: + * Stefano Zacchiroli + * for the HELM Team http://helm.cs.unibo.it/ * * This file is part of HELM, an Hypertextual, Electronic * Library of Mathematics, developed at the Computer Science @@ -21,17 +23,23 @@ * MA 02111-1307, USA. * * For details, see the HELM World-Wide-Web page, - * http://cs.unibo.it/helm/. + * http://helm.cs.unibo.it/ *) -(* TODO optimization: precompile regexps *) - open Http_getter_common;; open Http_getter_misc;; open Http_getter_types;; open Http_getter_debugger;; open Printf;; + (* constants *) + +let common_headers = [ + "Cache-Control", "no-cache"; + "Pragma", "no-cache"; + "Expires", "0" +] + (* HTTP queries argument parsing *) let parse_enc (req: Http_types.request) = @@ -42,7 +50,7 @@ let parse_enc (req: Http_types.request) = | s -> raise (Http_getter_bad_request ("Invalid format: " ^ s))) with Http_types.Param_not_found _ -> Enc_normal ;; -let parse_patch_dtd (req: Http_types.request) = +let parse_patch (req: Http_types.request) = match req#param "patch_dtd" with | s when String.lowercase s = "yes" -> true | s when String.lowercase s = "no" -> false @@ -54,35 +62,51 @@ let parse_output_format (req: Http_types.request) = | s when String.lowercase s = "xml" -> Fmt_xml | s -> raise (Http_getter_bad_request ("Invalid /ls format: " ^ s)) ;; -let parse_ls_uri (req: Http_types.request) = - let baseuri = req#param "baseuri" in - let subs = - Pcre.extract ~pat:"^(\\w+):(.*)$" (Pcre.replace ~pat:"/+$" baseuri) - in - match (subs.(1), subs.(2)) with - | "cic", uri -> Cic uri - | "theory", uri -> Theory uri - | _ -> raise (Http_getter_bad_request ("Invalid /ls baseuri: " ^ baseuri)) +let parse_ls_uri = + let parse_ls_RE = Pcre.regexp "^(\\w+):(.*)$" in + let trailing_slash_RE = Pcre.regexp "/+$" in + fun (req: Http_types.request) -> + let baseuri = req#param "baseuri" in + let subs = + Pcre.extract ~rex:parse_ls_RE + (Pcre.replace ~rex:trailing_slash_RE baseuri) + in + match (subs.(1), subs.(2)) with + | "cic", uri -> Cic uri + | "theory", uri -> Theory uri + | _ -> raise (Http_getter_bad_request ("Invalid /ls baseuri: " ^ baseuri)) ;; (* global maps, shared by all threads *) -let xml_map = new Http_getter_map.map Http_getter_env.xml_dbm in +let cic_map = new Http_getter_map.map Http_getter_env.cic_dbm in +let nuprl_map = new Http_getter_map.map Http_getter_env.nuprl_dbm in let rdf_map = new Http_getter_map.map Http_getter_env.rdf_dbm in let xsl_map = new Http_getter_map.map Http_getter_env.xsl_dbm in -let save_maps () = xml_map#close; rdf_map#close; xsl_map#close in +let save_maps () = + cic_map#close; nuprl_map#close; rdf_map#close; xsl_map#close in let map_of_uri = function - | uri when is_xml_uri uri -> xml_map + | uri when is_cic_uri uri -> cic_map + | uri when is_nuprl_uri uri -> nuprl_map | uri when is_rdf_uri uri -> rdf_map | uri when is_xsl_uri uri -> xsl_map | uri -> raise (Http_getter_unresolvable_URI uri) in -let resolve uri = (map_of_uri uri)#resolve uri in +let resolve uri = + try + (map_of_uri uri)#resolve uri + with Http_getter_map.Key_not_found _ -> + raise (Http_getter_unresolvable_URI uri) +in let register uri = (map_of_uri uri )#add uri in let return_all_foo_uris map doctype filter outchan = + (** return all URIs contained in 'map' which satisfy predicate 'filter'; URIs + are written in an XMLish format ('doctype' is the XML doctype) onto 'outchan' + *) Http_daemon.send_basic_headers ~code:200 outchan; Http_daemon.send_header "Content-Type" "text/xml" outchan; + Http_daemon.send_headers common_headers outchan; Http_daemon.send_CRLF outchan; output_string outchan @@ -102,7 +126,7 @@ let return_all_foo_uris map doctype filter outchan = output_string outchan (sprintf "\t\n" uri)); output_string outchan (sprintf "\n" doctype) in -let return_all_xml_uris = return_all_foo_uris xml_map "alluris" in +let return_all_xml_uris = return_all_foo_uris cic_map "alluris" in let return_all_rdf_uris = return_all_foo_uris rdf_map "allrdfuris" in let return_ls = let (++) (oldann, oldtypes, oldbody) (newann, newtypes, newbody) = @@ -150,7 +174,7 @@ let return_ls = in Hashtbl.replace objs basepart (oldflags ++ newflags) in - xml_map#iter (* BLEARGH Dbm module lacks support for fold-like functions *) + cic_map#iter (* BLEARGH Dbm module lacks support for fold-like functions *) (fun key _ -> match key with | uri when Pcre.pmatch ~rex:dir_RE uri -> (* directory hit *) @@ -176,7 +200,8 @@ let return_ls = objs "") in Http_daemon.respond - ~headers:["Content-Type", "text/plain"] ~body outchan + ~headers:(("Content-Type", "text/plain") :: common_headers) + ~body outchan | Fmt_xml -> let body = sprintf @@ -209,26 +234,31 @@ let return_ls = objs "")) in Http_daemon.respond - ~headers:["Content-Type", "text/xml"] ~body outchan + ~headers:(("Content-Type", "text/xml") :: common_headers) + ~body outchan in -let (index_line_sep_RE, index_sep_RE) = - (Pcre.regexp "[ \t]+", Pcre.regexp "\n+") +let (index_line_sep_RE, index_sep_RE, trailing_types_RE, + heading_cic_RE, heading_theory_RE, + heading_rdf_cic_RE, heading_rdf_theory_RE) = + (Pcre.regexp "[ \t]+", Pcre.regexp "\n+", Pcre.regexp "\\.types$", + Pcre.regexp "^cic:", Pcre.regexp "^theory:", + Pcre.regexp "^helm:rdf.*//cic:", Pcre.regexp "^helm:rdf.*//theory:") in let update_from_server logmsg server_url = (* use global maps *) debug_print ("Updating information from " ^ server_url); let xml_url_of_uri = function (* TODO missing sanity checks on server_url, e.g. it can contains $1 *) - | uri when (Pcre.pmatch ~pat:"^cic:" uri) -> - Pcre.replace ~pat:"^cic:" ~templ:server_url uri - | uri when (Pcre.pmatch ~pat:"^theory:" uri) -> - Pcre.replace ~pat:"^theory:" ~templ:server_url uri + | uri when (Pcre.pmatch ~rex:heading_cic_RE uri) -> + Pcre.replace ~rex:heading_cic_RE ~templ:server_url uri + | uri when (Pcre.pmatch ~rex:heading_theory_RE uri) -> + Pcre.replace ~rex:heading_theory_RE ~templ:server_url uri | uri -> raise (Http_getter_invalid_URI uri) in let rdf_url_of_uri = function (* TODO as above *) - | uri when (Pcre.pmatch ~pat:"^helm:rdf.*//cic:" uri) -> - Pcre.replace ~pat:"^helm:rdf.*//cic:" ~templ:server_url uri - | uri when (Pcre.pmatch ~pat:"^helm:rdf.*//theory:" uri) -> - Pcre.replace ~pat:"^helm:rdf.*//theory:" ~templ:server_url uri + | uri when (Pcre.pmatch ~rex:heading_rdf_cic_RE uri) -> + Pcre.replace ~rex:heading_rdf_cic_RE ~templ:server_url uri + | uri when (Pcre.pmatch ~rex:heading_rdf_theory_RE uri) -> + Pcre.replace ~rex:heading_rdf_theory_RE ~templ:server_url uri | uri -> raise (Http_getter_invalid_URI uri) in let log = ref (logmsg ^ "Processing server: " ^ server_url ^ "
\n") in @@ -247,8 +277,12 @@ let update_from_server logmsg server_url = (* use global maps *) (fun l -> try (match Pcre.split ~rex:index_line_sep_RE l with - | [uri; "gz"] -> xml_map#add uri ((xml_url_of_uri uri) ^ ".xml.gz") - | [uri] -> xml_map#add uri ((xml_url_of_uri uri) ^ ".xml") + | [uri; "gz"] -> + assert (is_cic_uri uri || is_nuprl_uri uri) ; + (map_of_uri uri)#add uri ((xml_url_of_uri uri) ^ ".xml.gz") + | [uri] -> + assert (is_cic_uri uri || is_nuprl_uri uri) ; + (map_of_uri uri)#add uri ((xml_url_of_uri uri) ^ ".xml") | _ -> log := !log ^ "Ignoring invalid line: " ^ l ^ "
\n") with Http_getter_invalid_URI uri -> log := !log ^ "Ignoring invalid XML URI: " ^ uri ^ "
\n") @@ -287,56 +321,57 @@ let callback (req: Http_types.request) outchan = debug_print ("Connection from " ^ req#clientAddr); debug_print ("Received request: " ^ req#path); (match req#path with - | "/help" -> return_html_msg Http_getter_const.usage_string outchan + | "/help" -> return_html_raw Http_getter_const.usage_string outchan | "/getxml" | "/getxslt" | "/getdtd" | "/resolve" | "/register" -> (let uri = req#param "uri" in (* common parameter *) match req#path with | "/getxml" -> let enc = parse_enc req in - let patch_dtd = - try parse_patch_dtd req with Http_types.Param_not_found _ -> true + let patch = + try parse_patch req with Http_types.Param_not_found _ -> true in Http_getter_cache.respond_xml - ~url:(resolve uri) ~uri ~enc ~patch_dtd outchan + ~url:(resolve uri) ~uri ~enc ~patch outchan | "/getxslt" -> - let patch_dtd = - try parse_patch_dtd req with Http_types.Param_not_found _ -> true + let patch = + try parse_patch req with Http_types.Param_not_found _ -> true in - Http_getter_cache.respond_xsl ~url:(resolve uri) ~patch_dtd outchan + Http_getter_cache.respond_xsl ~url:(resolve uri) ~patch outchan | "/getdtd" -> - let patch_dtd = - try parse_patch_dtd req with Http_types.Param_not_found _ -> true + let patch = + try parse_patch req with Http_types.Param_not_found _ -> true in Http_getter_cache.respond_dtd - ~patch_dtd ~url:(Http_getter_env.dtd_dir ^ "/" ^ uri) outchan + ~patch ~url:(Http_getter_env.dtd_dir ^ "/" ^ uri) outchan | "/resolve" -> (try - return_xml_msg + return_xml_raw (sprintf "\n" (resolve uri)) outchan with Http_getter_unresolvable_URI uri -> - return_xml_msg "\n" outchan) + return_xml_raw "\n" outchan) | "/register" -> let url = req#param "url" in register uri url; return_html_msg "Register done" outchan | _ -> assert false) | "/update" -> - (xml_map#clear; rdf_map#clear; xsl_map#clear; + (Http_getter_env.reload (); (* reload servers list from servers file *) + cic_map#clear; nuprl_map#clear; rdf_map#clear; xsl_map#clear; let log = List.fold_left update_from_server "" (* initial logmsg: empty *) (* reverse order: 1st server is the most important one *) - (List.rev Http_getter_env.servers) + (List.rev !Http_getter_env.servers) in - xml_map#sync; rdf_map#sync; xsl_map#sync; + cic_map#sync; nuprl_map#sync; rdf_map#sync; xsl_map#sync; return_html_msg log outchan) | "/getalluris" -> return_all_xml_uris (fun uri -> - (Pcre.pmatch ~pat:"^cic:" uri) && - not (Pcre.pmatch ~pat:"\\.types$" uri)) + (Pcre.pmatch ~rex:heading_cic_RE uri) && + not (Pcre.pmatch ~rex:trailing_types_RE uri)) outchan | "/getallrdfuris" -> (let classs = req#param "class" in