mirror of
https://gitlab.com/freepascal.org/fpc/source.git
synced 2025-10-31 08:31:49 +01:00
aee5380ae0
o support for the new codepage-aware ansistrings in the jvm branch
o empty ansistrings are now always represented by a nil pointer rather than
by an empty string, because an empty string also has a code page which
can confuse code (although this will make ansistrings harder to use
in Java code)
o more string helpers code shared between the general and jvm rtl
o support for indexbyte/word in the jvm rtl (warning: first parameter
is an open array rather than an untyped parameter there, so
indexchar(pcharvar^,10,0) will be equivalent to
indexchar[pcharvar^],10,0) there, which is different from what is
intended; changing it to an untyped parameter wouldn't help though)
o default() support is not yet complete
o calling fpcres is currently broken due to limitations in
sysutils.executeprocess() regarding handling unix quoting and
the compiler using the same command lines for scripts and directly
calling external programs
o compiling the Java compiler currently requires adding ALLOW_WARNINGS=1
to the make command line
git-svn-id: branches/jvmbackend@20887 -
223 lines
7.5 KiB
PHP
223 lines
7.5 KiB
PHP
{ source: http://download.java.net/jdk8/docs/technotes/guides/intl/encoding.doc.html
|
|
}
|
|
|
|
type
|
|
twin2javacs = record
|
|
cp: word;
|
|
name: unicodestring;
|
|
end;
|
|
(*
|
|
* Code Page Identifiers
|
|
* http://msdn.microsoft.com/en-us/library/dd317756.aspx
|
|
*)
|
|
const
|
|
win2javacs_arr: array[0..154] of twin2javacs =
|
|
(* (cp:; name:'x-MacDingbat'), not supported ? *)
|
|
(* (cp:; name:'x-MacSymbol'), not supported ? *)
|
|
((cp:037; name:'IBM037'),
|
|
(cp:437; name:'IBM437'),
|
|
(cp:500; name:'IBM500'),
|
|
(cp:737; name:'x-IBM737'),
|
|
(cp:775; name:'IBM775'),
|
|
(cp:850; name:'IBM850'),
|
|
(cp:852; name:'IBM852'),
|
|
(cp:855; name:'IBM855'),
|
|
(cp:857; name:'IBM857'),
|
|
(cp:858; name:'IBM00858'),
|
|
(cp:860; name:'IBM860'),
|
|
(cp:861; name:'IBM861'),
|
|
(cp:862; name:'IBM862'),
|
|
(cp:863; name:'IBM863'),
|
|
(cp:864; name:'IBM864'),
|
|
(cp:865; name:'IBM865'),
|
|
(cp:866; name:'IBM866'),
|
|
(cp:868; name:'IBM868'),
|
|
(cp:869; name:'IBM869'),
|
|
(cp:870; name:'IBM870'),
|
|
(cp:874; name:'x-windows-874'),
|
|
(cp:918; name:'IBM918'),
|
|
(cp:932; name:'Shift_JIS'),
|
|
(cp:932; name:'x-PCK'),
|
|
(cp:936; name:'GB2312'),
|
|
(cp:936; name:'x-mswin-936'),
|
|
(cp:942; name:'x-IBM942'),
|
|
(cp:943; name:'x-IBM943'),
|
|
(cp:948; name:'x-IBM948'),
|
|
(cp:949; name:'x-IBM949'),
|
|
(cp:949; name:'x-windows-949'),
|
|
(cp:950; name:'Big5'),
|
|
(cp:950; name:'x-IBM950'),
|
|
(cp:950; name:'x-windows-950'),
|
|
(cp:951; name:'x-Big5-HKSCS-2001'),
|
|
(cp:951; name:'Big5-HKSCS'),
|
|
(cp:951; name:'x-MS950-HKSCS'),
|
|
(cp:951; name:'x-MS950-HKSCS-XP'),
|
|
(cp:964; name:'x-IBM964'),
|
|
(cp:970; name:'x-IBM970'),
|
|
(cp:1006; name:'x-IBM1006'),
|
|
(cp:1025; name:'x-IBM1025'),
|
|
(cp:1026; name:'IBM1026'),
|
|
(cp:1046; name:'x-IBM1046'),
|
|
(cp:1047; name:'IBM1047'),
|
|
(cp:1097; name:'x-IBM1097'),
|
|
(cp:1098; name:'x-IBM1098'),
|
|
(cp:1112; name:'x-IBM1112'),
|
|
(cp:1122; name:'x-IBM1122'),
|
|
(cp:1123; name:'x-IBM1123'),
|
|
(cp:1124; name:'x-IBM1124'),
|
|
(cp:1140; name:'IBM01140'),
|
|
(cp:1141; name:'IBM01141'),
|
|
(cp:1142; name:'IBM01142'),
|
|
(cp:1143; name:'IBM01143'),
|
|
(cp:1144; name:'IBM01144'),
|
|
(cp:1145; name:'IBM01145'),
|
|
(cp:1146; name:'IBM01146'),
|
|
(cp:1147; name:'IBM01147'),
|
|
(cp:1148; name:'IBM01148'),
|
|
(cp:1149; name:'IBM01149'),
|
|
(cp:1200; name:'UTF-16LE'),
|
|
(cp:1200; name:'x-UTF-16LE-BOM'),
|
|
(cp:1201; name:'UTF-16'),
|
|
(cp:1201; name:'UTF-16BE'),
|
|
(cp:1250; name:'windows-1250'),
|
|
(cp:1251; name:'windows-1251'),
|
|
(cp:1252; name:'windows-1252'),
|
|
(cp:1253; name:'windows-1253'),
|
|
(cp:1254; name:'windows-1254'),
|
|
(cp:1255; name:'windows-1255'),
|
|
(cp:1256; name:'windows-1256'),
|
|
(cp:1257; name:'windows-1257'),
|
|
(cp:1258; name:'windows-1258'),
|
|
(cp:1259; name:'windows-31j'),
|
|
(cp:1361; name:'x-Johab'),
|
|
(cp:1381; name:'x-IBM1381'),
|
|
(cp:1383; name:'x-IBM1383'),
|
|
(cp:10000; name:'MacRoman'),
|
|
(cp:10004; name:'x-MacArabic'),
|
|
(cp:10005; name:'x-MacHebrew'),
|
|
(cp:10006; name:'x-MacGreek'),
|
|
(cp:10007; name:'x-MacCyrillic'),
|
|
(cp:10010; name:'x-MacRomania'),
|
|
(cp:10017; name:'x-MacUkraine'),
|
|
(cp:10021; name:'x-MacThai'),
|
|
(cp:10029; name:'x-MacCentralEurope'),
|
|
(cp:10079; name:'x-MacIceland'),
|
|
(cp:10081; name:'x-MacTurkish'),
|
|
(cp:10082; name:'x-MacCroatian'),
|
|
(cp:12000; name:'UTF-32LE'),
|
|
(cp:12000; name:'X-UTF-32LE-BOM'),
|
|
(cp:12001; name:'UTF-32'),
|
|
(cp:12001; name:'UTF-32BE'),
|
|
(cp:12001; name:'X-UTF-32BE-BOM'),
|
|
(cp:20127; name:'US-ASCII'),
|
|
(cp:20273; name:'IBM273'),
|
|
(cp:20277; name:'IBM277'),
|
|
(cp:20278; name:'IBM278'),
|
|
(cp:20280; name:'IBM280'),
|
|
(cp:20284; name:'IBM284'),
|
|
(cp:20285; name:'IBM285'),
|
|
(cp:20297; name:'IBM297'),
|
|
(cp:20420; name:'IBM420'),
|
|
(cp:20424; name:'IBM424'),
|
|
(cp:20833; name:'x-IBM833'),
|
|
(cp:20834; name:'x-IBM834'),
|
|
(cp:20838; name:'IBM-Thai'),
|
|
(cp:20856; name:'x-IBM856'),
|
|
(cp:20866; name:'KOI8-R'),
|
|
(cp:20871; name:'IBM871'),
|
|
(cp:20874; name:'x-IBM874'),
|
|
(cp:20875; name:'x-IBM875'),
|
|
(cp:20921; name:'x-IBM921'),
|
|
(cp:20922; name:'x-IBM922'),
|
|
(cp:20932; name:'EUC-JP'),
|
|
(cp:20932; name:'x-JIS0208'),
|
|
(cp:20932; name:'x-JISAutoDetect'),
|
|
(cp:21866; name:'KOI8-U'),
|
|
(cp:28591; name:'ISO-8859-1'),
|
|
(cp:28592; name:'ISO-8859-2'),
|
|
(cp:28593; name:'ISO-8859-3'),
|
|
(cp:28594; name:'ISO-8859-4'),
|
|
(cp:28595; name:'ISO-8859-5'),
|
|
(cp:28596; name:'ISO-8859-6'),
|
|
(cp:28597; name:'ISO-8859-7'),
|
|
(cp:28598; name:'ISO-8859-8'),
|
|
(cp:28599; name:'ISO-8859-9'),
|
|
(cp:28601; name:'TIS-620'),
|
|
(cp:28601; name:'x-iso-8859-11'),
|
|
(cp:28603; name:'ISO-8859-13'),
|
|
(cp:28605; name:'ISO-8859-15'),
|
|
(cp:33722; name:'x-IBM33722'),
|
|
(cp:50220; name:'x-windows-50220'),
|
|
(cp:50220; name:'x-windows-iso2022jp'),
|
|
(cp:50221; name:'JIS_X0201'),
|
|
(cp:50221; name:'x-windows-50221'),
|
|
(cp:50222; name:'ISO-2022-JP'),
|
|
(cp:50222; name:'ISO-2022-JP-2'), (* not exact, Windows does not support ISO-2022-JP-2 *)
|
|
(cp:50222; name:'JIS_X0212-1990'),
|
|
(cp:50225; name:'ISO-2022-KR'),
|
|
(cp:50227; name:'ISO-2022-CN'),
|
|
(cp:50227; name:'x-ISO-2022-CN-CNS'),
|
|
(cp:50229; name:'x-ISO-2022-CN-GB'),
|
|
(cp:50930; name:'x-IBM930'),
|
|
(cp:50933; name:'x-IBM933'),
|
|
(cp:50935; name:'x-IBM935'),
|
|
(cp:50937; name:'x-IBM937'),
|
|
(cp:50939; name:'x-IBM939'),
|
|
(cp:51932; name:'x-MS932_0213'),
|
|
(cp:51932; name:'x-SJIS_0213'),
|
|
(cp:51949; name:'EUC-KR'),
|
|
(cp:54936; name:'GB18030'),
|
|
(cp:57002; name:'x-ISCII91'),
|
|
(cp:65001; name:'UTF-8'));
|
|
|
|
|
|
function win2javacs(cp: word): unicodestring;
|
|
var
|
|
l, h, i, ccp: longint;
|
|
begin
|
|
l:=low(win2javacs_arr);
|
|
h:=high(win2javacs_arr);
|
|
repeat
|
|
i:=(l+h+1) shr 1;
|
|
ccp:=win2javacs_arr[i].cp;
|
|
if cp=ccp then
|
|
break;
|
|
if cp>=ccp then
|
|
l:=i
|
|
else
|
|
h:=i-1;
|
|
until l>=h;
|
|
if cp=win2javacs_arr[i].cp then
|
|
begin
|
|
{ the array has been ordered so that in case multiple alias names
|
|
exist, the first entry for the cp is the most commonly supported
|
|
one
|
|
}
|
|
while (i>low(win2javacs_arr)) and
|
|
(win2javacs_arr[i-1].cp=cp) do
|
|
dec(i);
|
|
result:=win2javacs_arr[i].name;
|
|
end
|
|
else
|
|
{ or better raise an error? }
|
|
result:='<unsupported>';
|
|
end;
|
|
|
|
|
|
function javacs2win(cpname: unicodestring): word;
|
|
var
|
|
i: longint;
|
|
begin
|
|
{ simple linear scan, not a common operation and hence not worth
|
|
building a separate array for }
|
|
for i:=low(win2javacs_arr) to high(win2javacs_arr) do
|
|
if win2javacs_arr[i].name=cpname then
|
|
begin
|
|
result:=win2javacs_arr[i].cp;
|
|
exit;
|
|
end;
|
|
{ rawbytestring (or better raise an error?) }
|
|
result:=65535;
|
|
end;
|
|
|