diff options
author | Norbert Preining <preining@logic.at> | 2010-03-01 01:54:19 +0000 |
---|---|---|
committer | Norbert Preining <preining@logic.at> | 2010-03-01 01:54:19 +0000 |
commit | 904fd0757fe037dbfbf156b31f72e5ff5c7cd796 (patch) | |
tree | 36f000ab754854574aad17c01d9cd9ac739f1053 /Master/tlpkg/tlperl.straw/lib/XML/SAX/PurePerl/EncodingDetect.pm | |
parent | 402bd194f686177d2dfca24f7c4766434c514141 (diff) |
commit more files of the tlperl.straw dir, still not complete
git-svn-id: svn://tug.org/texlive/trunk@17244 c570f23f-e606-0410-a88d-b1316a301751
Diffstat (limited to 'Master/tlpkg/tlperl.straw/lib/XML/SAX/PurePerl/EncodingDetect.pm')
-rwxr-xr-x | Master/tlpkg/tlperl.straw/lib/XML/SAX/PurePerl/EncodingDetect.pm | 105 |
1 files changed, 105 insertions, 0 deletions
diff --git a/Master/tlpkg/tlperl.straw/lib/XML/SAX/PurePerl/EncodingDetect.pm b/Master/tlpkg/tlperl.straw/lib/XML/SAX/PurePerl/EncodingDetect.pm new file mode 100755 index 00000000000..6f1c8eadcbd --- /dev/null +++ b/Master/tlpkg/tlperl.straw/lib/XML/SAX/PurePerl/EncodingDetect.pm @@ -0,0 +1,105 @@ +# $Id: EncodingDetect.pm,v 1.6 2007-02-07 09:33:50 grant Exp $ + +package XML::SAX::PurePerl; # NB, not ::EncodingDetect! + +use strict; + +sub encoding_detect { + my ($parser, $reader) = @_; + + my $error = "Invalid byte sequence at start of file"; + + my $data = $reader->data; + if ($data =~ /^\x00\x00\xFE\xFF/) { + # BO-UCS4-be + $reader->move_along(4); + $reader->set_encoding('UCS-4BE'); + return; + } + elsif ($data =~ /^\x00\x00\xFF\xFE/) { + # BO-UCS-4-2143 + $reader->move_along(4); + $reader->set_encoding('UCS-4-2143'); + return; + } + elsif ($data =~ /^\x00\x00\x00\x3C/) { + $reader->set_encoding('UCS-4BE'); + return; + } + elsif ($data =~ /^\x00\x00\x3C\x00/) { + $reader->set_encoding('UCS-4-2143'); + return; + } + elsif ($data =~ /^\x00\x3C\x00\x00/) { + $reader->set_encoding('UCS-4-3412'); + return; + } + elsif ($data =~ /^\x00\x3C\x00\x3F/) { + $reader->set_encoding('UTF-16BE'); + return; + } + elsif ($data =~ /^\xFF\xFE\x00\x00/) { + # BO-UCS-4LE + $reader->move_along(4); + $reader->set_encoding('UCS-4LE'); + return; + } + elsif ($data =~ /^\xFF\xFE/) { + $reader->move_along(2); + $reader->set_encoding('UTF-16LE'); + return; + } + elsif ($data =~ /^\xFE\xFF\x00\x00/) { + $reader->move_along(4); + $reader->set_encoding('UCS-4-3412'); + return; + } + elsif ($data =~ /^\xFE\xFF/) { + $reader->move_along(2); + $reader->set_encoding('UTF-16BE'); + return; + } + elsif ($data =~ /^\xEF\xBB\xBF/) { # UTF-8 BOM + $reader->move_along(3); + $reader->set_encoding('UTF-8'); + return; + } + elsif ($data =~ /^\x3C\x00\x00\x00/) { + $reader->set_encoding('UCS-4LE'); + return; + } + elsif ($data =~ /^\x3C\x00\x3F\x00/) { + $reader->set_encoding('UTF-16LE'); + return; + } + elsif ($data =~ /^\x3C\x3F\x78\x6D/) { + # $reader->set_encoding('UTF-8'); + return; + } + elsif ($data =~ /^\x3C\x3F\x78/) { + # $reader->set_encoding('UTF-8'); + return; + } + elsif ($data =~ /^\x3C\x3F/) { + # $reader->set_encoding('UTF-8'); + return; + } + elsif ($data =~ /^\x3C/) { + # $reader->set_encoding('UTF-8'); + return; + } + elsif ($data =~ /^[\x20\x09\x0A\x0D]+\x3C[^\x3F]/) { + # $reader->set_encoding('UTF-8'); + return; + } + elsif ($data =~ /^\x4C\x6F\xA7\x94/) { + $reader->set_encoding('EBCDIC'); + return; + } + + warn("Unable to recognise encoding of this document"); + return; +} + +1; + |