open Printf
+let file_scheme_prefix = "file://"
+
let trailing_dot_gz_RE = Pcre.regexp "\\.gz$" (* for g{,un}zip *)
let url_RE = Pcre.regexp "^([\\w.-]+)(:(\\d+))?(/.*)?$"
let http_scheme_RE = Pcre.regexp ~flags:[`CASELESS] "^http://"
-let file_scheme_RE = Pcre.regexp ~flags:[`CASELESS] "^file://"
+let file_scheme_RE = Pcre.regexp ~flags:[`CASELESS] ("^" ^ file_scheme_prefix)
let dir_sep_RE = Pcre.regexp "/"
let heading_slash_RE = Pcre.regexp "^/"
+let local_url =
+ let rex = Pcre.regexp ("^(" ^ file_scheme_prefix ^ ")(.*)(.gz)$") in
+ fun s ->
+ try
+ Some ((Pcre.extract ~rex s).(2))
+ with Not_found -> None
+
let bufsiz = 16384 (* for file system I/O *)
let tcp_bufsiz = 4096 (* for TCP I/O *)
let iter_file f = fold_file (fun line _ -> f line) ()
+let iter_buf_size = 10240
+
+let iter_file_data f fname =
+ let ic = open_in fname in
+ let buf = String.create iter_buf_size in
+ try
+ while true do
+ let bytes = input ic buf 0 iter_buf_size in
+ if bytes = 0 then raise End_of_file;
+ f (String.sub buf 0 bytes)
+ done
+ with End_of_file -> close_in ic
+
let hashtbl_sorted_fold f tbl init =
let sorted_keys =
List.sort compare (Hashtbl.fold (fun key _ keys -> key::keys) tbl [])
let sorted_keys =
List.sort compare (Hashtbl.fold (fun key _ keys -> key::keys) tbl [])
in
- List.iter (fun k -> f k (Hashtbl.find tbl k)) sorted_keys
+ List.iter (fun k -> f k (Hashtbl.find tbl k)) sorted_keys
let cp src dst =
- let (ic, oc) = (open_in src, open_out dst) in
- let buf = String.create bufsiz in
- (try
- while true do
- let bytes = input ic buf 0 bufsiz in
- if bytes = 0 then raise End_of_file else output oc buf 0 bytes
- done
- with End_of_file -> ());
- close_in ic; close_out oc
+ try
+ let ic = open_in src in
+ try
+ let oc = open_out dst in
+ let buf = String.create bufsiz in
+ (try
+ while true do
+ let bytes = input ic buf 0 bufsiz in
+ if bytes = 0 then raise End_of_file else output oc buf 0 bytes
+ done
+ with
+ End_of_file -> ()
+ );
+ close_in ic; close_out oc
+ with
+ Sys_error s ->
+ Http_getter_logger.log s;
+ close_in ic
+ | e ->
+ Http_getter_logger.log (Printexc.to_string e);
+ close_in ic;
+ raise e
+ with
+ Sys_error s ->
+ Http_getter_logger.log s
+ | e ->
+ Http_getter_logger.log (Printexc.to_string e);
+ raise e
let wget ?output url =
Http_getter_logger.log
(let oc =
open_out (match output with Some f -> f | None -> Filename.basename url)
in
- Http_client.http_get_iter (fun data -> output_string oc data) url;
+ Http_user_agent.get_iter (fun data -> output_string oc data) url;
close_out oc)
| scheme -> (* unsupported scheme *)
failwith ("Http_getter_misc.wget: unsupported scheme: " ^ scheme)
in
Http_getter_logger.log ~level:3
(sprintf "gunzipping %s (keep: %b, output: %s)" fname keep output);
- let (ic, oc) = (Gzip.open_in fname, open_out output) in
- let buf = String.create bufsiz in
- (try
- while true do
- let bytes = Gzip.input ic buf 0 bufsiz in
- if bytes = 0 then raise End_of_file else Pervasives.output oc buf 0 bytes
- done
- with End_of_file -> ());
- Gzip.close_in ic; close_out oc;
+ (* Open the zipped file manually since Gzip.open_in may
+ * leak the descriptor if it raises an exception *)
+ let zic = open_in fname in
+ begin
+ try
+ let ic = Gzip.open_in_chan zic in
+ let oc = open_out output in
+ let buf = String.create bufsiz in
+ (try
+ while true do
+ let bytes = Gzip.input ic buf 0 bufsiz in
+ if bytes = 0 then raise End_of_file else Pervasives.output oc buf 0 bytes
+ done
+ with End_of_file -> ());
+ close_out oc;
+ Gzip.close_in ic
+ with
+ e -> close_in zic ; raise e
+ end ;
if not keep then Sys.remove fname
;;
let size = (Unix.stat fname).Unix.st_size in
let buf = String.create size in
let ic = open_in fname in
- really_input ic buf 0 size;
+ really_input ic buf 0 size ;
close_in ic;
Some buf
with Unix.Unix_error (Unix.ENOENT, "stat", _) -> None
- end else (* other URL, pass it to Http_client *)
+ end else (* other URL, pass it to Http_user_agent *)
try
- Some (Http_client.http_get url)
+ Some (Http_user_agent.get url)
with e ->
Http_getter_logger.log (sprintf
- "Warning: Http_client failed on url %s with exception: %s"
+ "Warning: Http_user_agent failed on url %s with exception: %s"
url (Printexc.to_string e));
None
fun line ->
Pcre.pmatch ~rex:blank_line_RE line
+let normalize_dir s = (* append "/" if missing *)
+ let len = String.length s in
+ try
+ if s.[len - 1] = '/' then s
+ else s ^ "/"
+ with Invalid_argument _ -> (* string is empty *) "/"
+
+let strip_trailing_slash s =
+ try
+ let len = String.length s in
+ if s.[len - 1] = '/' then String.sub s 0 (len - 1)
+ else s
+ with Invalid_argument _ -> s
+
+let strip_suffix ~suffix s =
+ try
+ let s_len = String.length s in
+ let suffix_len = String.length suffix in
+ let suffix_sub = String.sub s (s_len - suffix_len) suffix_len in
+ if suffix_sub <> suffix then raise (Invalid_argument "");
+ String.sub s 0 (s_len - suffix_len)
+ with Invalid_argument _ ->
+ raise (Invalid_argument "Http_getter_misc.strip_suffix")
+
+let rec list_uniq = function
+ | [] -> []
+ | h::[] -> [h]
+ | h1::h2::tl when h1 = h2 -> list_uniq (h2 :: tl)
+ | h1::tl (* when h1 <> h2 *) -> h1 :: list_uniq tl
+
+let extension s =
+ try
+ let idx = String.rindex s '.' in
+ String.sub s idx (String.length s - idx)
+ with Not_found -> ""
+
+let temp_file_of_uri uri =
+ let flat_string s s' c =
+ let cs = String.copy s in
+ for i = 0 to (String.length s) - 1 do
+ if String.contains s' s.[i] then cs.[i] <- c
+ done;
+ cs
+ in
+ let user = try Unix.getlogin () with _ -> "" in
+ Filename.open_temp_file (user ^ flat_string uri ".-=:;!?/&" '_') ""
+
+let backtick cmd =
+ let ic = Unix.open_process_in cmd in
+ let res = input_line ic in
+ ignore (Unix.close_process_in ic);
+ res
+