diplodocus.org Git - nmh/blob - test/scan/test-scan-multibyte

   1 #!/bin/sh
   2 ############################################################
   3 #
   4 # Test scan to see if multibyte support (UTF-8 locale) works
   5 #
   6 # Other tests will get the normal ASCII case, so all we care
   7 # about here is UTF-8 encoded headers (RFC 2047).
   8 #
   9 # Note that this file should be edited via a UTF-8 aware
  10 # editor, since UTF-8 characters are in it.
  11 #
  12 ############################################################
  13
  14 set -e
  15
  16 if test -z "${MH_OBJ_DIR}"; then
  17     srcdir=`dirname "$0"`/../..
  18     MH_OBJ_DIR=`cd "$srcdir" && pwd`; export MH_OBJ_DIR
  19 fi
  20
  21 . "$MH_OBJ_DIR/test/common.sh"
  22
  23 setup_test
  24
  25 if test "${MULTIBYTE_ENABLED}" -ne 1; then
  26     test_skip "configure did not detect multibyte support"
  27 fi
  28
  29 LC_ALL=en_US.UTF-8; export LC_ALL
  30
  31 #
  32 # Create a test message with RFC 2047 headers we can scan
  33 #
  34 # In this Subject header in this message is a "n" with a Combining Diaeresis
  35 # (U+0308).  There is different interpretation of this character with respect
  36 # to wcwidth() (which is supposed to return the column width of a character).
  37 # We use a test program to determine what the output width of U+0308 is
  38 # and adjust our test output appropriately.
  39 #
  40
  41 cat > "${MH_TEST_DIR}/Mail/inbox/11" <<EOF
  42 From: David =?utf-8?q?=EF=AC=86?= Hubbins <hubbins@example.com>
  43 To: Sir Denis =?utf-8?q?Eton=E2=80=93Hogg? <sirdenis@example.com>
  44 Date: Friday, 2 Mar 1984 00:00:00
  45 Subject: =?utf-8?q?Sp=C4=B1n=CC=88al_Tap_=E2=86=92_Tap_into_America!?=
  46
  47 Things are looking great!
  48 EOF
  49
  50 width=`${MH_OBJ_DIR}/test/getcwidth`
  51 if test $? -ne 0; then
  52     echo "getcwidth failed to run"
  53     exit 1
  54 fi
  55
  56 expected="$MH_TEST_DIR/$$.expected"
  57 actual="$MH_TEST_DIR/$$.actual"
  58
  59 if test "$width" -eq 1; then
  60 cat > "$expected" <<EOF
  61   11  03/02 David ﬆ Hubbins    Spın̈al Tap → Tap into America!<<Things are look
  62 EOF
  63 elif test "$width" -eq 0; then
  64 cat > "$expected" <<EOF
  65   11  03/02 David ﬆ Hubbins    Spın̈al Tap → Tap into America!<<Things are looki
  66 EOF
  67 else
  68     echo "Unsupported width for combining diaeresis: $width"
  69     exit 1
  70 fi
  71
  72 scan -width 80 +inbox 11 > $actual || exit 1
  73 set -x
  74 oldfailed="${failed:-0}"
  75 check "$expected" "$actual"
  76 if test "$oldfailed" -ne "${failed:-0}"; then
  77         echo "Complete UTF-8 width table for BMP"
  78         ${MH_OBJ_DIR}/test/getcwidth --dump
  79 fi
  80 set +x
  81
  82 #
  83 # Check decoding with an invalid multibyte sequence.  We skip this test
  84 # if we don't have iconv support, since it requires converting from one
  85 # character set to another.  Be sure we created the test file, though, because
  86 # it's required for the test right after it.
  87 #
  88
  89 cat >`mhpath new` <<EOF
  90 From: Test12 <test12@example.com>
  91 To: Some User <user@example.com>
  92 Date: Mon, 31 Dec 2012 00:00:00
  93 Message-Id: 12@test.nmh
  94 Subject: =?UTF-8?B?MjAxMyBOZXcgWWVhcuKAmXMgRGVhbHMhIFN0YXJ0IHRoZSB5ZWFy?=
  95         =?UTF-8?B?IHJpZ2h0IHdpdGggYmlnIHNhdmluZ3M=?=
  96
  97 This message has an encoded Subject with an invalid character for the
  98 ISO-8859-1 character set, but it (U+2019) is valid UTF-8.
  99 EOF
 100
 101 cat >"$expected" <<EOF
 102   12  12/31 Test12             2013 New Year?s Deals! Start the year right
 103 EOF
 104
 105 if test "$ICONV_ENABLED" -eq 1; then
 106     LC_CTYPE=ISO-8859-1 MM_CHARSET=ISO-8859-1 scan -width 75 last >"$actual"
 107     check "$expected" "$actual"
 108 fi
 109
 110 # check scan width with a valid multibyte sequence
 111 cat >"$expected" <<EOF
 112   12  12/31 Test12             2013 New Year’s Deals! Start the year right
 113 EOF
 114
 115 LC_CTYPE=en_US.UTF-8 MM_CHARSET=UTF-8 scan -width 75 last >"$actual"
 116 check "$expected" "$actual"
 117
 118
 119 exit $failed