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 open Http_getter_debugger;;
32 let trailing_dot_gz_RE = Pcre.regexp "\\.gz$" (* for g{,un}zip *)
33 let url_RE = Pcre.regexp "^([\\w.]+)(:(\\d+))?(/.*)?$"
34 let http_scheme_RE = Pcre.regexp ~flags:[`CASELESS] "^http://"
35 let file_scheme_RE = Pcre.regexp ~flags:[`CASELESS] "^file://"
36 let dir_sep_RE = Pcre.regexp "/"
37 let heading_slash_RE = Pcre.regexp "^/"
39 let bufsiz = 16384 (* for file system I/O *)
40 let tcp_bufsiz = 4096 (* for TCP I/O *)
42 let fold_file f init fname =
43 let inchan = open_in fname in
44 let rec fold_lines' value =
46 let line = input_line inchan in
47 fold_lines' (f value line)
48 with End_of_file -> value
50 let res = (try fold_lines' init with e -> (close_in inchan; raise e)) in
53 let iter_file f = fold_file (fun _ line -> f line) ()
55 let hashtbl_sorted_fold f tbl init =
57 List.sort compare (Hashtbl.fold (fun key _ keys -> key::keys) tbl [])
59 List.fold_left (fun acc k -> f k (Hashtbl.find tbl k) acc) init sorted_keys
62 let (ic, oc) = (open_in src, open_out dst) in
63 let buf = String.create bufsiz in
66 let bytes = input ic buf 0 bufsiz in
67 if bytes = 0 then raise End_of_file else output oc buf 0 bytes
69 with End_of_file -> ());
70 close_in ic; close_out oc
75 Pcre.extract ~rex:url_RE (Pcre.replace ~rex:http_scheme_RE url)
78 (if subs.(2) = "" then 80 else int_of_string subs.(3)),
79 (if subs.(4) = "" then "/" else subs.(4)))
82 (sprintf "Can't parse url: %s (exception: %s)"
83 url (Printexc.to_string exc))
84 let init_socket addr port =
85 let inet_addr = (Unix.gethostbyname addr).Unix.h_addr_list.(0) in
86 let sockaddr = Unix.ADDR_INET (inet_addr, port) in
87 let suck = Unix.socket Unix.PF_INET Unix.SOCK_STREAM 0 in
88 Unix.connect suck sockaddr;
89 let outchan = Unix.out_channel_of_descr suck in
90 let inchan = Unix.in_channel_of_descr suck in
92 let http_get_iter_buf ~callback url =
93 let (address, port, path) = parse_url url in
94 let buf = String.create tcp_bufsiz in
95 let (inchan, outchan) = init_socket address port in
96 output_string outchan (sprintf "GET %s\r\n" path);
100 match input inchan buf 0 tcp_bufsiz with
101 | 0 -> raise End_of_file
102 | bytes when bytes = tcp_bufsiz -> (* buffer full, no need to slice it *)
104 | bytes when bytes < tcp_bufsiz -> (* buffer not full, slice it *)
105 callback (String.sub buf 0 bytes)
106 | _ -> (* ( bytes < 0 ) || ( bytes > tcp_bufsiz ) *)
109 with End_of_file -> ());
110 close_in inchan (* close also outchan, same fd *)
112 let wget ?output url =
114 (sprintf "wgetting %s (output: %s)" url
115 (match output with None -> "default" | Some f -> f));
117 | url when Pcre.pmatch ~rex:file_scheme_RE url -> (* file:// *)
118 (let src_fname = Pcre.replace ~rex:file_scheme_RE url in
120 | Some dst_fname -> cp src_fname dst_fname
122 let dst_fname = Filename.basename src_fname in
123 if src_fname <> dst_fname then
124 cp src_fname dst_fname
125 else (* src and dst are the same: do nothing *)
127 | url when Pcre.pmatch ~rex:http_scheme_RE url -> (* http:// *)
129 open_out (match output with Some f -> f | None -> Filename.basename url)
131 http_get_iter_buf ~callback:(fun data -> output_string oc data) url;
133 | scheme -> (* unsupported scheme *)
134 failwith ("Http_getter_misc.wget: unsupported scheme: " ^ scheme)
136 let gzip ?(keep = false) fname =
137 debug_print (sprintf "gzipping %s (keep: %b)" fname keep);
138 let (ic, oc) = (open_in fname, Gzip.open_out (fname ^ ".gz")) in
139 let buf = String.create bufsiz in
142 let bytes = input ic buf 0 bufsiz in
143 if bytes = 0 then raise End_of_file else Gzip.output oc buf 0 bytes
145 with End_of_file -> ());
146 close_in ic; Gzip.close_out oc;
147 if not keep then Sys.remove fname
149 let gunzip ?(keep = false) fname =
150 debug_print (sprintf "gunzipping %s (keep: %b)" fname keep);
151 let basename = Pcre.replace ~rex:trailing_dot_gz_RE fname in
152 assert (basename <> fname);
153 let (ic, oc) = (Gzip.open_in fname, open_out basename) in
154 let buf = String.create bufsiz in
157 let bytes = Gzip.input ic buf 0 bufsiz in
158 if bytes = 0 then raise End_of_file else output oc buf 0 bytes
160 with End_of_file -> ());
161 Gzip.close_in ic; close_out oc;
162 if not keep then Sys.remove fname
164 let tempfile () = Filename.temp_file "http_getter_" ""
166 exception Mkdir_failure of string * string;; (* dirname, failure reason *)
169 let mkdir ?(parents = false) dirname =
172 let split = Pcre.split ~rex:dir_sep_RE dirname in
173 if Pcre.pmatch ~rex:heading_slash_RE dirname then
182 sprintf "%s%s%s" pre (match pre with "/" | "" -> "" | _ -> "/") dir
185 (match (Unix.stat next_dir).Unix.st_kind with
186 | Unix.S_DIR -> () (* dir component already exists, go on! *)
187 | _ -> (* dir component already exists but isn't a dir, abort! *)
189 (Mkdir_failure (dirname,
190 sprintf "'%s' already exists but is not a dir" next_dir)))
191 with Unix.Unix_error (Unix.ENOENT, "stat", _) ->
192 (* dir component doesn't exists, create it and go on! *)
193 Unix.mkdir next_dir dir_perm);
197 if parents then mkdirhier () else Unix.mkdir dirname dir_perm
199 let string_of_proc_status = function
200 | Unix.WEXITED code -> sprintf "[Exited: %d]" code
201 | Unix.WSIGNALED sg -> sprintf "[Killed: %d]" sg
202 | Unix.WSTOPPED sg -> sprintf "[Stopped: %d]" sg
205 if Pcre.pmatch ~rex:file_scheme_RE url then begin
206 (* file:// URL. Read data from file system *)
207 let fname = Pcre.replace ~rex:file_scheme_RE url in
209 let size = (Unix.stat fname).Unix.st_size in
210 let buf = String.create size in
211 let ic = open_in fname in
212 really_input ic buf 0 size;
215 with Unix.Unix_error (Unix.ENOENT, "stat", _) -> None
216 end else (* other URL, pass it to netclient *)
218 Some (Http_client.Convenience.http_get url)
219 with Http_client.Http_error (code, _) -> None