2 * Copyright (C) 2003-2004:
3 * Stefano Zacchiroli <zack@cs.unibo.it>
4 * for the HELM Team http://helm.cs.unibo.it/
6 * This file is part of HELM, an Hypertextual, Electronic
7 * Library of Mathematics, developed at the Computer Science
8 * Department, University of Bologna, Italy.
10 * HELM is free software; you can redistribute it and/or
11 * modify it under the terms of the GNU General Public License
12 * as published by the Free Software Foundation; either version 2
13 * of the License, or (at your option) any later version.
15 * HELM is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with HELM; if not, write to the Free Software
22 * Foundation, Inc., 59 Temple Place - Suite 330, Boston,
25 * For details, see the HELM World-Wide-Web page,
26 * http://helm.cs.unibo.it/
29 (* TODO cache expires control!!! *)
31 if two large proof (already in cache) are requested at the same time by two
32 clients, uwobo (java implementation, not yet tested with the OCaml one)
33 starts looping sending output to one of the client *)
35 open Http_getter_common
37 open Http_getter_types
40 (* expose ThreadSafe.threadSafe methods *)
43 inherit ThreadSafe.threadSafe
44 method virtual doCritical : 'a. 'a lazy_t -> 'a
45 method virtual doReader : 'a. 'a lazy_t -> 'a
46 method virtual doWriter : 'a. 'a lazy_t -> 'a
49 let threadSafe = new threadSafe
51 let finally cleanup f =
53 let res = Lazy.force f in
58 raise (Http_getter_types.Cache_failure (Printexc.to_string e))
60 let resource_type_of_url = function
61 | url when Pcre.pmatch ~pat:"\\.xml\\.gz$" url -> `Gzipped
62 | url when Pcre.pmatch ~pat:"\\.xml$" url -> `Normal
63 | url -> raise (Invalid_URL url)
65 let extension_of_resource_type = function
67 | `Gzipped -> "xml.gz"
69 (* basename = resource name without trailing ".gz", if any *)
70 let is_in_cache basename =
72 (match Lazy.force Http_getter_env.cache_mode with
74 | `Gzipped -> basename ^ ".gz")
77 ?(via_http = true) ?(enc = `Normal) ?(patch = true) ~url ~uri outchan
79 let resource_type = resource_type_of_url url in
80 let extension = extension_of_resource_type resource_type in
82 match uri_of_string uri with (* parse uri *)
83 | Cic_uri (Cic baseuri) | Cic_uri (Theory baseuri) ->
84 (* assumption: baseuri starts with "/" *)
85 sprintf "%s%s.%s" (Lazy.force Http_getter_env.cic_dir) baseuri extension
86 | Nuprl_uri baseuri ->
87 (* assumption: baseuri starts with "/" *)
88 sprintf "%s%s.%s" (Lazy.force Http_getter_env.nuprl_dir) baseuri
90 | Rdf_uri (prefix, ((Cic baseuri) as qbaseuri))
91 | Rdf_uri (prefix, ((Theory baseuri) as qbaseuri)) ->
93 (Pcre.replace ~pat:"/" ~templ:"_"
94 (Pcre.replace ~pat:"_" ~templ:"__"
98 | Theory _ -> "//theory:"))))
101 (Lazy.force Http_getter_env.rdf_dir) escaped_prefix baseuri extension
104 if patch then Http_getter_common.patch_xml ~via_http () else (fun x -> x)
106 let basename = Pcre.replace ~pat:"\\.gz$" downloadname in
107 let contype = "text/xml" in
108 (* Fill cache if needed and return a short circuit file.
109 Short circuit is needed in situations like:
110 resource type = normal, cache type = gzipped, required encoding = normal
111 we would like to avoid normal -> gzipped -> normal conversions. To avoid
112 this tmp_short_circuit is used to remember the name of the intermediate
115 threadSafe#doWriter (lazy(
116 if not (is_in_cache basename) then begin (* cache MISS *)
117 Http_getter_logger.log ~level:2 "Cache MISS :-(";
118 mkdir ~parents:true (Filename.dirname downloadname);
119 match (resource_type, Lazy.force Http_getter_env.cache_mode) with
120 | `Normal, `Normal | `Gzipped, `Gzipped ->
121 wget ~output:downloadname url;
123 | `Normal, `Gzipped -> (* resource normal, cache gzipped *)
124 let tmp = tempfile () in
126 if enc = `Normal then (* user wants normal: don't delete it! *)
127 (Some (tmp, enc), (fun () -> ()))
129 (None, (fun () -> Sys.remove tmp))
131 finally cleanup (lazy (
132 wget ~output:tmp url;
133 gzip ~output:(basename ^ ".gz") ~keep:true tmp; (* fill cache *)
136 | `Gzipped, `Normal -> (* resource gzipped, cache normal *)
137 let tmp = tempfile () in
139 if enc = `Gzipped then (* user wants .gz: don't delete it! *)
140 (Some (tmp, enc), (fun () -> ()))
142 (None, (fun () -> Sys.remove tmp))
144 finally cleanup (lazy (
145 wget ~output:tmp url;
146 gunzip ~output:basename ~keep:true tmp; (* fill cache *)
150 Http_getter_logger.log ~level:2 "Cache HIT :-)";
154 let tmp_short_circuit = fill_cache () in
155 threadSafe#doReader (lazy(
156 assert (is_in_cache basename);
157 match (enc, Lazy.force Http_getter_env.cache_mode) with
158 | `Normal, `Normal | `Gzipped, `Gzipped ->
159 (* resource in cache is already in the required format *)
162 Http_getter_logger.log ~level:2
163 "No format mangling required (encoding = normal)";
164 return_file ~via_http ~fname:basename ~contype ~patch_fun outchan
166 Http_getter_logger.log ~level:2
167 "No format mangling required (encoding = gzipped)";
169 ~via_http ~fname:(basename ^ ".gz") ~contype ~contenc:"x-gzip"
170 ~patch_fun ~gunzip:true
172 | `Normal, `Gzipped | `Gzipped, `Normal ->
173 (match tmp_short_circuit with
174 | None -> (* no short circuit possible, use cache *)
175 Http_getter_logger.log ~level:2
176 "No short circuit available, use cache";
177 let tmp = tempfile () in
178 finally (fun () -> Sys.remove tmp) (lazy (
181 (* required format is normal, cached version is gzipped *)
182 gunzip (* gunzip to tmp *)
183 ~output:tmp ~keep:true (basename ^ ".gz");
184 return_file ~via_http ~fname:tmp ~contype ~patch_fun outchan;
186 (* required format is gzipped, cached version is normal *)
187 gzip ~output:tmp ~keep:true basename; (* gzip to tmp *)
189 ~via_http ~fname:tmp ~contype ~contenc:"x-gzip"
190 ~patch_fun ~gunzip:true
193 | Some (fname, `Normal) -> (* short circuit available, use it! *)
194 Http_getter_logger.log ~level:2
195 "Using short circuit (encoding = normal)";
196 finally (fun () -> Sys.remove fname) (lazy (
197 return_file ~via_http ~fname ~contype ~patch_fun outchan
199 | Some (fname, `Gzipped) -> (* short circuit available, use it! *)
200 Http_getter_logger.log ~level:2
201 "Using short circuit (encoding = gzipped)";
202 finally (fun () -> Sys.remove fname) (lazy (
203 return_file ~via_http ~fname ~contype ~contenc:"x-gzip" ~patch_fun
208 (* TODO enc is not yet supported *)
210 ?(via_http = true) ?(enc = `Normal) ?(patch = true) ~url outchan
213 if patch then Http_getter_common.patch_xsl ~via_http () else (fun x -> x)
215 let fname = tempfile () in
216 finally (fun () -> Sys.remove fname) (lazy (
217 wget ~output:fname url;
218 return_file ~via_http ~fname ~contype:"text/xml" ~patch_fun outchan
221 (* TODO enc is not yet supported *)
223 ?(via_http = true) ?(enc = `Normal) ?(patch = true) ~url outchan
226 if patch then Http_getter_common.patch_dtd ~via_http () else (fun x -> x)
228 if Sys.file_exists url then
229 (* TODO check this: old getter here used text/xml *)
230 return_file ~via_http ~fname:url ~contype:"text/plain" ~patch_fun outchan
232 raise (Dtd_not_found url)
235 let module E = Http_getter_env in
237 (function dir -> ignore (Unix.system ("rm -rf " ^ dir ^ "/*")))
238 [ Lazy.force E.cic_dir; Lazy.force E.nuprl_dir; Lazy.force E.rdf_dir ]