--- /dev/null
+
+
+let make_iso enc =
+ let s = ref "" in
+ for i = 0 to 255 do
+ let u = try Netconversion.makechar (enc :> Netconversion.encoding) i
+ with Not_found -> "" in
+ s := !s ^ u
+ done;
+ !s
+;;
+
+let make_ucs2 start stop =
+ let s = String.create ((stop - start) * 2) in
+ for i = 0 to stop-start-1 do
+ let k = 2 * i in
+ let c = i + start in
+ s.[k] <- Char.chr(c lsr 8);
+ s.[k+1] <- Char.chr(c land 0xff);
+ done;
+ s
+;;
+
+let make_ucs4 start stop =
+ let s = String.create ((stop - start) * 4) in
+ for i = 0 to stop-start-1 do
+ let k = 4 * i in
+ let c = i + start in
+ s.[k] <- Char.chr(c lsr 24);
+ s.[k+1] <- Char.chr((c lsr 16) land 0xff);
+ s.[k+2] <- Char.chr((c lsr 8) land 0xff);
+ s.[k+3] <- Char.chr(c land 0xff);
+ done;
+ s
+;;
+
+let name_of_encoding enc =
+ match enc with
+ `Enc_iso88591 -> "ISO_8859-1"
+ | `Enc_iso88592 -> "ISO_8859-2"
+ | `Enc_iso88593 -> "ISO_8859-3"
+ | `Enc_iso88594 -> "ISO_8859-4"
+ | `Enc_iso88595 -> "ISO_8859-5"
+ | `Enc_iso88596 -> "ISO_8859-6"
+ | `Enc_iso88597 -> "ISO_8859-7"
+ | `Enc_iso88598 -> "ISO_8859-8"
+ | `Enc_iso88599 -> "ISO_8859-9"
+ | `Enc_iso885910 -> "ISO_8859-10"
+ | `Enc_iso885913 -> "ISO_8859-13"
+ | `Enc_iso885914 -> "ISO_8859-14"
+ | `Enc_iso885915 -> "ISO_8859-15"
+ | `Enc_utf8 -> "UTF-8"
+ | `Enc_ucs4 -> "UCS-4"
+ | `Enc_ucs2 -> "UCS-2"
+ | `Enc_utf16 -> "UTF-16"
+
+ (* Note: GNU-iconv assumes big endian byte order *)
+;;
+
+let iconv_recode_string in_enc out_enc in_s =
+ let in_enc_name = name_of_encoding in_enc in
+ let out_enc_name = name_of_encoding out_enc in
+ let out_s = ref "" in
+
+ let out_ch,in_ch = Unix.open_process ("iconv -f " ^ in_enc_name ^ " -t " ^
+ out_enc_name) in
+ (* Write in_s to in_ch in a new thread: *)
+ ignore
+ (Thread.create
+ (fun () ->
+ output_string in_ch in_s;
+ close_out in_ch;
+ )
+ ()
+ );
+ (* Read the result in the current thread: *)
+ let buf = String.create 1024 in
+ let n = ref 1 in
+ while !n <> 0 do
+ let n' = input out_ch buf 0 1024 in
+ out_s := !out_s ^ String.sub buf 0 n';
+ n := n'
+ done;
+ ignore(Unix.close_process (out_ch,in_ch));
+ !out_s
+;;
+
+let test_iso_and_utf8 enc =
+ let name = name_of_encoding enc in
+ print_string ("Recode: " ^ name ^ " and UTF-8... "); flush stdout;
+ let s = make_iso enc in
+ let s1' = Netconversion.recode_string (enc :> Netconversion.encoding)
+ `Enc_utf8 s in
+ let s2' = iconv_recode_string enc `Enc_utf8 s in
+ assert(s1' = s2');
+ let s1 = Netconversion.recode_string `Enc_utf8
+ (enc :> Netconversion.encoding) s1' in
+ let s2 = iconv_recode_string `Enc_utf8 enc s1' in
+ assert(s1 = s2 && s1 = s);
+ print_endline "OK"; flush stdout
+;;
+
+let test_utf16_and_utf8_0000_d7ff () =
+ print_string "Recode: UTF-16-BE and UTF-8, #0000-#D7FF... ";
+ flush stdout;
+ let s = make_ucs2 0 0xd800 in
+ let s1' = Netconversion.recode_string `Enc_utf16_be `Enc_utf8 s in
+ let s2' = iconv_recode_string `Enc_utf16 `Enc_utf8 s in
+ assert(s1' = s2');
+ let s1 = Netconversion.recode_string `Enc_utf8 `Enc_utf16_be s1' in
+ let s2 = iconv_recode_string `Enc_utf8 `Enc_utf16 s1' in
+ assert(s1 = s2 && s1 = s);
+ print_endline "OK"; flush stdout
+;;
+
+let test_utf16_and_utf8_e000_fffd () =
+ print_string "Recode: UTF-16-BE and UTF-8, #E000-#FFFD... ";
+ flush stdout;
+ let s = make_ucs2 0xe000 0xfffe in
+ let s1' = Netconversion.recode_string `Enc_utf16_be `Enc_utf8 s in
+ let s2' = iconv_recode_string `Enc_utf16 `Enc_utf8 s in
+ assert(s1' = s2');
+ let s1 = Netconversion.recode_string `Enc_utf8 `Enc_utf16_be s1' in
+ let s2 = iconv_recode_string `Enc_utf8 `Enc_utf16 s1' in
+ assert(s1 = s2 && s1 = s);
+ print_endline "OK"; flush stdout
+;;
+
+let test_utf16_and_utf8_10000_10FFFF () =
+ print_string "Recode: UTF-16-BE and UTF-8, #10000-#10FFFF... ";
+ flush stdout;
+ for i = 1 to 16 do
+ let s0 = make_ucs4 (i * 0x10000) (i * 0x10000 + 0x10000) in
+ let s = iconv_recode_string `Enc_ucs4 `Enc_utf16 s0 in
+ let s1' = Netconversion.recode_string `Enc_utf16_be `Enc_utf8 s in
+ let s2' = iconv_recode_string `Enc_utf16 `Enc_utf8 s in
+ assert(s1' = s2');
+ let s1 = Netconversion.recode_string `Enc_utf8 `Enc_utf16_be s1' in
+ let s2 = iconv_recode_string `Enc_utf8 `Enc_utf16 s1' in
+ assert(s1 = s2 && s1 = s);
+ print_string "+"; flush stdout;
+ done;
+ print_endline "OK"; flush stdout
+;;
+
+
+print_endline "Warning: You need the command 'iconv' to run this test!";
+flush stdout;
+test_iso_and_utf8 `Enc_iso88591;
+test_iso_and_utf8 `Enc_iso88592;
+test_iso_and_utf8 `Enc_iso88593;
+test_iso_and_utf8 `Enc_iso88594;
+test_iso_and_utf8 `Enc_iso88595;
+test_iso_and_utf8 `Enc_iso88596;
+test_iso_and_utf8 `Enc_iso88597;
+(* test_iso_and_utf8 `Enc_iso88598; *)
+test_iso_and_utf8 `Enc_iso88599;
+test_iso_and_utf8 `Enc_iso885910;
+(* test_iso_and_utf8 `Enc_iso885913; *)
+(* test_iso_and_utf8 `Enc_iso885914; *)
+(* test_iso_and_utf8 `Enc_iso885915; *)
+test_utf16_and_utf8_0000_d7ff();
+test_utf16_and_utf8_e000_fffd();
+(* This test does not work because iconv does not support the surrogate
+ * representation of UTF-16:
+ * test_utf16_and_utf8_10000_10FFFF();
+ *)
+()
+;;
projects / helm.git / blobdiff