icuSources/test/testdata/test3.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) 2001-2013, International Business Machines
   5 # * Corporation and others.  All Rights Reserved.
   6 # *******************************************************************************
   7 #
   8 # test3.ucm
   9 #
  10 # Test file for MBCS conversion with three-byte codepage data.
  11 # Also contains extension mappings (m:n).
  12
  13 <code_set_name>     "test3"
  14 <mb_cur_max>        3
  15 <mb_cur_min>        1
  16 <uconv_class>       "MBCS"
  17 <subchar>           \xff
  18 <icu:state>         0, 1:1, 5-9, ff
  19 <icu:state>         2:2
  20 <icu:state>         4, a-f.p
  21
  22 CHARMAP
  23
  24 # fromUnicode result is zero byte from other than U+0000
  25 <U20ac>     \x00 |0
  26
  27 # nothing special
  28 <U0005>     \x05 |0
  29
  30 # extensions
  31 <U00c0>     \x05+\x01\x02\x0d |0
  32 <U00c0>     \x05+\x01\x02\x0e |3
  33 <U00c0>     \x05+\xff |3
  34
  35 # toUnicode result is fallback direct
  36 <U0006>     \x06 |3
  37
  38 # toUnicode result is direct non-BMP code point
  39 <U101234>   \x07 |0
  40 <Ufebcd>    \x08 |3
  41
  42 # extensions
  43 <U101234>+<U50005>+<U60006> \x07+\x00+\x01\x02\x0f+\x09 |0
  44 <U101234>+<U50005>          \x07+\x00+\x01\x02\x0e+\x05 |0
  45 <U101234>+<U60006>          \x07+\x00+\x01\x02\x0f+\x06 |0
  46 <U101234>+<U70007>          \x07+\x00+\x01\x02\x0f |1
  47
  48 #unassigned \x09
  49
  50 # extensions where the first code point is unassigned, for replay testing
  51 #<U00c4><U0300> \x09+\x09 |0
  52 <U00c4><U00c4><U101234><U0005> \x05+\x01\x02\x0c |0
  53
  54 # toUnicode result is surrogate pair: test real pair, single unit, unassigned
  55 <U23456>    \x01\x02\x0a |0
  56 <U000b>     \x01\x02\x0b |0
  57 #unassigned \x01\x02\x0c
  58 <U34567>    \x01\x02\x0d |3
  59 <U000e>     \x01\x02\x0e |3
  60 #unassigned \x01\x02\x0f
  61
  62 # "good one-way" mappings
  63 <U0023>         \x01\x02\x04 |4
  64 <U0023>+<UFE0E> \x01\x02\x04 |4
  65 <U0023>+<UFE0F> \x01\x02\x04 |0
  66
  67 END CHARMAP