X-Git-Url: http://matita.cs.unibo.it/gitweb/?a=blobdiff_plain;f=helm%2Focaml%2Fgetter%2Fhttp_getter_common.ml;h=d56cf6909dcb6867bd576c60bddea372f414aea4;hb=97c2d258a5c524eb5c4b85208899d80751a2c82f;hp=62a52c792f1713b9f152881421b889a1d76f46dd;hpb=7dfa9f509a84cdb3ff9577e359d1f0f17b349c97;p=helm.git diff --git a/helm/ocaml/getter/http_getter_common.ml b/helm/ocaml/getter/http_getter_common.ml index 62a52c792..d56cf6909 100644 --- a/helm/ocaml/getter/http_getter_common.ml +++ b/helm/ocaml/getter/http_getter_common.ml @@ -31,8 +31,8 @@ open Printf;; let string_of_ls_flag = function No -> "NO" | Yes -> "YES" | Ann -> "ANN" let string_of_encoding = function - | Enc_normal -> "Normal" - | Enc_gzipped -> "GZipped" + | `Normal -> "Normal" + | `Gzipped -> "GZipped" let is_cic_obj_uri uri = Pcre.pmatch ~pat:"^cic:" uri let is_theory_uri uri = Pcre.pmatch ~pat:"^theory:" uri @@ -52,75 +52,115 @@ let rec uri_of_string = function in Rdf_uri (prefix, rest) | _ -> raise (Invalid_URI uri)) - | uri when is_cic_uri uri -> Cic_uri (Cic (Pcre.replace ~pat:"^cic:" uri)) + | uri when is_cic_obj_uri uri -> Cic_uri (Cic (Pcre.replace ~pat:"^cic:" uri)) | uri when is_nuprl_uri uri -> Nuprl_uri (Pcre.replace ~pat:"^nuprl:" uri) | uri when is_theory_uri uri -> Cic_uri (Theory (Pcre.replace ~pat:"^theory:" uri)) | uri -> raise (Invalid_URI uri) -let patch_xml line = - Pcre.replace - ~pat:(sprintf "DOCTYPE (.*) SYSTEM\\s+\"%s/" - (Lazy.force Http_getter_env.dtd_base_url)) - ~templ:(sprintf "DOCTYPE $1 SYSTEM \"%s/getdtd?uri=" - (Lazy.force Http_getter_env.my_own_url)) - line -let patch_xsl line = - let mk_patch_fun tag line = - Pcre.replace - ~pat:(sprintf "%s\\s+href=\"" tag) - ~templ:(sprintf "%s href=\"%s/getxslt?uri=" - tag (Lazy.force Http_getter_env.my_own_url)) - line +let patch_xsl ?(via_http = true) () = + fun line -> + let mk_patch_fun tag line = + Pcre.replace + ~pat:(sprintf "%s\\s+href=\"" tag) + ~templ:(sprintf "%s href=\"%s/getxslt?uri=" + tag (Lazy.force Http_getter_env.my_own_url)) + line + in + let (patch_import, patch_include) = + (mk_patch_fun "xsl:import", mk_patch_fun "xsl:include") + in + patch_include (patch_import line) + +let patch_system kind ?(via_http = true) () = + let rex = + Pcre.regexp (sprintf "%s (.*) SYSTEM\\s+\"((%s)/)?" kind + (String.concat "|" (Lazy.force Http_getter_env.dtd_base_urls))) in - let (patch_import, patch_include) = - (mk_patch_fun "xsl:import", mk_patch_fun "xsl:include") + let templ = + if via_http then + sprintf "%s $1 SYSTEM \"%s/getdtd?uri=" kind + (Lazy.force Http_getter_env.my_own_url) + else + sprintf "%s $1 SYSTEM \"file://%s/" kind + (Lazy.force Http_getter_env.dtd_dir) in - patch_include (patch_import line) -let patch_dtd line = - Pcre.replace - ~pat:(sprintf "ENTITY (.*) SYSTEM\\s+\"(%s/)?" - (Lazy.force Http_getter_env.dtd_base_url)) - ~templ:(sprintf "ENTITY $1 SYSTEM \"%s/getdtd?uri=" - (Lazy.force Http_getter_env.my_own_url)) - line + fun line -> Pcre.replace ~rex ~templ line + +let patch_entity = patch_system "ENTITY" +let patch_doctype = patch_system "DOCTYPE" + +let patch_xmlbase = + let rex = Pcre.regexp "^(\\s*<\\w[^ ]*)(\\s|>)" in + fun xmlbases baseurl baseuri s -> + let s' = + Pcre.replace ~rex + ~templ:(sprintf "$1 xml:base=\"%s\" helm:base=\"%s\"$2" baseurl baseuri) + s + in + if s <> s' then xmlbases := None; + s' + +let patch_dtd = patch_entity +let patch_xml ?via_http ?xmlbases () = + let xmlbases = ref xmlbases in + fun line -> + match !xmlbases with + | None -> patch_doctype ?via_http () (patch_entity ?via_http () line) + | Some (xmlbaseuri, xmlbaseurl) -> + patch_xmlbase xmlbases xmlbaseurl xmlbaseuri + (patch_doctype ?via_http () (patch_entity ?via_http () line)) let return_file - ~fname ?contype ?contenc ?(patch_fun = fun x -> x) ?(gunzip = false) outchan - = - let headers = - match (contype, contenc) with - | (Some t, Some e) -> ["Content-Encoding", e; "Content-Type", t] - | (Some t, None) -> ["Content-Type" , t] - | (None, Some e) -> ["Content-Encoding", e] - | (None, None) -> [] - in - Http_daemon.send_basic_headers ~code:200 outchan; - Http_daemon.send_headers headers outchan; - Http_daemon.send_CRLF outchan; - if gunzip then begin (* gunzip needed, uncompress file, apply patch_fun to - it, compress the result and sent it to client *) - let (tmp1, tmp2) = - (Http_getter_misc.tempfile (), Http_getter_misc.tempfile ()) + ~fname ?contype ?contenc ?patch_fun ?(gunzip = false) ?(via_http = true) + ~enc outchan += + if via_http then begin + let headers = + match (contype, contenc) with + | (Some t, Some e) -> ["Content-Encoding", e; "Content-Type", t] + | (Some t, None) -> ["Content-Type" , t] + | (None, Some e) -> ["Content-Encoding", e] + | (None, None) -> [] in - try - Http_getter_misc.gunzip ~keep:true ~output:tmp1 fname;(* gunzip to tmp1 *) - let new_file = open_out tmp2 in - Http_getter_misc.iter_file (* tmp2 = patch(tmp1) *) - (fun line -> output_string new_file (patch_fun line ^ "\n")) - tmp1; - close_out new_file; - Http_getter_misc.gzip ~output:tmp1 tmp2; (* tmp1 = gzip(tmp2); rm tmp2 *) - Http_getter_misc.iter_file (* send tmp1 to client as is*) - (fun line -> output_string outchan (line ^ "\n")) - tmp1; - Sys.remove tmp1 (* rm tmp1 *) - with e -> - Sys.remove tmp1; - raise e - end else (* no need to gunzip, apply patch_fun directly to file *) - Http_getter_misc.iter_file - (fun line -> output_string outchan (patch_fun line ^ "\n")) - fname + Http_daemon.send_basic_headers ~code:(`Code 200) outchan; + Http_daemon.send_headers headers outchan; + Http_daemon.send_CRLF outchan + end; + match gunzip, patch_fun with + | true, Some patch_fun -> + Http_getter_logger.log ~level:2 + "Patch required, uncompress/compress cycle needed :-("; + (* gunzip needed, uncompress file, apply patch_fun to it, compress the + * result and sent it to client *) + let (tmp1, tmp2) = + (Http_getter_misc.tempfile (), Http_getter_misc.tempfile ()) + in + (try + Http_getter_misc.gunzip ~keep:true ~output:tmp1 fname; (* gunzip tmp1 *) + let new_file = open_out tmp2 in + Http_getter_misc.iter_file (* tmp2 = patch(tmp1) *) + (fun line -> + output_string new_file (patch_fun line ^ "\n"); + flush outchan) + tmp1; + close_out new_file; + Http_getter_misc.gzip ~output:tmp1 tmp2;(* tmp1 = gzip(tmp2); rm tmp2 *) + Http_getter_misc.iter_file (* send tmp1 to client as is*) + (fun line -> output_string outchan (line ^ "\n"); flush outchan) + tmp1; + Sys.remove tmp1 (* rm tmp1 *) + with e -> + Sys.remove tmp1; + raise e) + | false, Some patch_fun -> + (match enc with + | `Normal -> + Http_getter_misc.iter_file + (fun line -> output_string outchan (patch_fun (line ^ "\n"))) + fname + | `Gzipped -> assert false) + (* dangerous case, if this happens it needs to be investigated *) + | _, None -> Http_getter_misc.iter_file_data (output_string outchan) fname ;;