icuSources/test/testdata/test4.ucm

   1 # Copyright (C) 2016 and later: Unicode, Inc. and others.
   2 # License & terms of use: http://www.unicode.org/copyright.html
   3 # *******************************************************************************
   4 # * Copyright (C) 2001-2003, International Business Machines
   5 # * Corporation and others.  All Rights Reserved.
   6 # *******************************************************************************
   7 #
   8 # test4.ucm
   9 #
  10 # Test file for MBCS conversion with four-byte codepage data.
  11
  12 <code_set_name>     "test4"
  13 <mb_cur_max>        4
  14 <mb_cur_min>        1
  15 <uconv_class>       "MBCS"
  16
  17 # both subchars are single-byters, which does not make sense
  18 # but works - adding subchar1 for tests but don't want to
  19 # change old tests for a new subchar -- markus 20031028
  20 <subchar>           \xff
  21 <subchar1>          \xe1
  22 <icu:state>         0, 1:1, 5-9, e1, ff
  23 <icu:state>         2:2
  24 <icu:state>         3:3
  25 <icu:state>         a-f.p, ff
  26
  27 CHARMAP
  28
  29 # fromUnicode result is zero byte from other than U+0000
  30 <U20ac>     \x00 |0
  31
  32 # fallback from non-zero to zero possible with extension table
  33 <U20ad>     \x00 |1
  34
  35 # nothing special
  36 <U0005>     \x05 |0
  37
  38 # toUnicode result is fallback direct
  39 <U0006>     \x06 |3
  40
  41 # toUnicode result is direct non-BMP code point
  42 <U101234>   \x07 |0
  43 <Ufebcd>    \x08 |3
  44
  45 #unassigned \x09
  46
  47 # toUnicode result is surrogate pair: test real pair, single unit, unassigned
  48 <U23456>    \x01\x02\x03\x0a |0
  49 <U000b>     \x01\x02\x03\x0b |0
  50 #unassigned \x01\x02\x03\x0c
  51 <U34567>    \x01\x02\x03\x0d |3
  52 <U000e>     \x01\x02\x03\x0e |3
  53 #unassigned \x01\x02\x03\x0f
  54
  55 # <subchar1> non-mapping
  56 <U50005>    \xe1 |2
  57 # add a mapping that turns the above's Unicode side into a prefix
  58 <U50005><U60006> \x06 |1
  59
  60 # many bytes, and bytes per UChar
  61 <U30ab><U309a> \x01\x02\x03\x0a\x01\x02\x03\x0b\x01\x02\x03\x0c\x01\x02\x03\x0d\x01\x02\x03\x0e\x01\x02\x03\x0f\x01\x02\x03\x0a\x05\x06\x07 |0
  62
  63 # many UChars, and UChars per byte
  64 <U304b><U309a><U304d><U309a><U304f><U309a><U3051><U309a><U3053><U309a><U30ab><U309a><U30ad><U309a><U30af><U309a><U30b1><U309a><U0300> \x08\x09 |0
  65
  66 END CHARMAP