(*
- * Copyright (C) 2000, HELM Team.
+ * Copyright (C) 2003, HELM Team.
*
* This file is part of HELM, an Hypertextual, Electronic
* Library of Mathematics, developed at the Computer Science
* http://cs.unibo.it/helm/.
*)
-(* TODO optimization: precompile almost all regexp *)
-
open Http_getter_common;;
+open Http_getter_misc;;
open Http_getter_types;;
+open Http_getter_debugger;;
open Printf;;
- (* debugging settings *)
-let debug = true;;
-let debug_print s = if debug then prerr_endline ("[HTTP-Getter] " ^ s);;
-let http_debug = false;;
-Http_common.debug := http_debug;;
+ (* constants *)
-let http_get url =
-(* debug_print ("Downloading URL: " ^ url); *)
- try
- Some (Http_client.Convenience.http_get url)
- with Http_client.Http_error (code, _) -> None
-(*
- (debug_print
- (sprintf "Failed to download %s, HTTP response was %d" url code);
- None)
-*)
-;;
+let common_headers = [
+ "Cache-Control", "no-cache";
+ "Pragma", "no-cache";
+ "Expires", "0"
+]
+
+ (* HTTP queries argument parsing *)
let parse_enc (req: Http_types.request) =
try
| 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 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 resolve uri = (map_of_uri uri)#resolve 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
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
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, 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 ^ "<br />\n") in
List.iter
(fun l ->
try
- (match Pcre.split ~pat:"[ \\t]+" l with
+ (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")
| _ -> log := !log ^ "Ignoring invalid line: " ^ l ^ "<br />\n")
with Http_getter_invalid_URI uri ->
log := !log ^ "Ignoring invalid XML URI: " ^ uri ^ "<br />\n")
- (Pcre.split ~pat:"\n+" xml_index)) (* xml_index lines *)
+ (Pcre.split ~rex:index_sep_RE xml_index)) (* xml_index lines *)
| None -> ());
(match rdf_index with
| Some rdf_index ->
List.iter
(fun l ->
try
- (match Pcre.split ~pat:"[ \\t]+" l with
+ (match Pcre.split ~rex:index_line_sep_RE l with
| [uri; "gz"] -> rdf_map#add uri ((rdf_url_of_uri uri) ^ ".xml.gz")
| [uri] -> rdf_map#add uri ((rdf_url_of_uri uri) ^ ".xml")
| _ -> log := !log ^ "Ignoring invalid line: " ^ l ^ "<br />\n")
with Http_getter_invalid_URI uri ->
log := !log ^ "Ignoring invalid RDF URI: " ^ uri ^ "<br />\n")
- (Pcre.split ~pat:"\n+" rdf_index)) (* rdf_index lines *)
+ (Pcre.split ~rex:index_sep_RE rdf_index)) (* rdf_index lines *)
| None -> ());
(match xsl_index with
| Some xsl_index ->
(log := !log ^ "Updating XSLT db ...<br />\n";
List.iter
(fun l -> xsl_map#add l (server_url ^ "/" ^ l))
- (Pcre.split ~pat:"\n+" xsl_index);
+ (Pcre.split ~rex:index_sep_RE xsl_index);
log := !log ^ "All done!<br />\n")
| None -> ());
+ debug_print "done with this server";
!log
in
(* thread action *)
+
let callback (req: Http_types.request) outchan =
try
debug_print ("Connection from " ^ req#clientAddr);
| "/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
Http_daemon.respond ~body:Http_getter_const.empty_xml outchan
| invalid_request ->
Http_daemon.respond_error ~status:(`Client_error `Bad_request) outchan);
- debug_print "Done!"
+ debug_print "Done!\n"
with
| Http_types.Param_not_found attr_name ->
return_400 (sprintf "Parameter '%s' is missing" attr_name) outchan
| Http_getter_bad_request msg -> return_html_error msg outchan
| Http_getter_internal_error msg -> return_html_internal_error msg outchan
+ | Shell.Subprocess_error l ->
+ return_html_internal_error
+ (String.concat "<br />\n"
+ (List.map
+ (fun (cmd, code) ->
+ sprintf "Command '%s' returned %s"
+ cmd (string_of_proc_status code))
+ l))
+ outchan
| exc ->
return_html_error
("Uncaught exception: " ^ (Printexc.to_string exc))
outchan
in
- (* daemon initialization *)
-Http_getter_env.dump_env ();
-flush stdout;
-Unix.putenv "http_proxy" "";
-Sys.catch_break true;
-try
- Http_daemon.start'
- ~timeout:None ~port:Http_getter_env.port ~mode:`Thread callback
-with Sys.Break ->
- save_maps ()
+ (* daemon initialization *)
+
+let main () =
+ Http_getter_env.dump_env ();
+ Unix.putenv "http_proxy" "";
+ at_exit save_maps;
+ Sys.catch_break true;
+ try
+ Http_daemon.start'
+ ~timeout:(Some 600) ~port:Http_getter_env.port ~mode:`Thread callback
+ with Sys.Break -> () (* 'save_maps' already registered with 'at_exit' *)
+in
+
+main ()