t/t7812-grep-icase-non-ascii.sh

   1 #!/bin/sh
   2
   3 test_description='grep icase on non-English locales'
   4
   5 . ./lib-gettext.sh
   6
   7 test_expect_success GETTEXT_LOCALE 'setup' '
   8         test_write_lines "TILRAUN: Halló Heimur!" >file &&
   9         git add file &&
  10         LC_ALL="$is_IS_locale" &&
  11         export LC_ALL
  12 '
  13
  14 test_have_prereq GETTEXT_LOCALE &&
  15 test-tool regex "HALLÓ" "Halló" ICASE &&
  16 test_set_prereq REGEX_LOCALE
  17
  18 test_expect_success REGEX_LOCALE 'grep literal string, no -F' '
  19         git grep -i "TILRAUN: Halló Heimur!" &&
  20         git grep -i "TILRAUN: HALLÓ HEIMUR!"
  21 '
  22
  23 test_expect_success GETTEXT_LOCALE,PCRE 'grep pcre utf-8 icase' '
  24         git grep --perl-regexp    "TILRAUN: H.lló Heimur!" &&
  25         git grep --perl-regexp -i "TILRAUN: H.lló Heimur!" &&
  26         git grep --perl-regexp -i "TILRAUN: H.LLÓ HEIMUR!"
  27 '
  28
  29 test_expect_success GETTEXT_LOCALE,PCRE 'grep pcre utf-8 string with "+"' '
  30         test_write_lines "TILRAUN: Hallóó Heimur!" >file2 &&
  31         git add file2 &&
  32         git grep -l --perl-regexp "TILRAUN: H.lló+ Heimur!" >actual &&
  33         echo file >expected &&
  34         echo file2 >>expected &&
  35         test_cmp expected actual
  36 '
  37
  38 test_expect_success REGEX_LOCALE 'grep literal string, with -F' '
  39         git grep -i -F "TILRAUN: Halló Heimur!" &&
  40         git grep -i -F "TILRAUN: HALLÓ HEIMUR!"
  41 '
  42
  43 test_expect_success REGEX_LOCALE 'grep string with regex, with -F' '
  44         test_write_lines "TILRAUN: Halló Heimur [abc]!" >file3 &&
  45         git add file3 &&
  46         git grep -i -F "TILRAUN: Halló Heimur [abc]!" file3
  47 '
  48
  49 test_expect_success REGEX_LOCALE 'pickaxe -i on non-ascii' '
  50         git commit -m first &&
  51         git log --format=%f -i -S"TILRAUN: HALLÓ HEIMUR!" >actual &&
  52         echo first >expected &&
  53         test_cmp expected actual
  54 '
  55
  56 test_expect_success GETTEXT_LOCALE,LIBPCRE2 'PCRE v2: setup invalid UTF-8 data' '
  57         printf "\\200\\n" >invalid-0x80 &&
  58         echo "ævar" >expected &&
  59         cat expected >>invalid-0x80 &&
  60         git add invalid-0x80 &&
  61
  62         # Test for PCRE2_MATCH_INVALID_UTF bug
  63         # https://bugs.exim.org/show_bug.cgi?id=2642
  64         printf "\\345Aæ\\n" >invalid-0xe5 &&
  65         git add invalid-0xe5
  66 '
  67
  68 test_expect_success GETTEXT_LOCALE,LIBPCRE2 'PCRE v2: grep ASCII from invalid UTF-8 data' '
  69         git grep -h "var" invalid-0x80 >actual &&
  70         test_cmp expected actual &&
  71         git grep -h "(*NO_JIT)var" invalid-0x80 >actual &&
  72         test_cmp expected actual
  73 '
  74
  75 test_expect_success GETTEXT_LOCALE,LIBPCRE2 'PCRE v2: grep ASCII from invalid UTF-8 data (PCRE2 bug #2642)' '
  76         git grep -h "Aæ" invalid-0xe5 >actual &&
  77         test_cmp invalid-0xe5 actual &&
  78         git grep -h "(*NO_JIT)Aæ" invalid-0xe5 >actual &&
  79         test_cmp invalid-0xe5 actual
  80 '
  81
  82 test_expect_success GETTEXT_LOCALE,LIBPCRE2 'PCRE v2: grep non-ASCII from invalid UTF-8 data' '
  83         git grep -h "æ" invalid-0x80 >actual &&
  84         test_cmp expected actual &&
  85         git grep -h "(*NO_JIT)æ" invalid-0x80 >actual &&
  86         test_cmp expected actual
  87 '
  88
  89 test_expect_success GETTEXT_LOCALE,LIBPCRE2 'PCRE v2: grep non-ASCII from invalid UTF-8 data (PCRE2 bug #2642)' '
  90         git grep -h "Aæ" invalid-0xe5 >actual &&
  91         test_cmp invalid-0xe5 actual &&
  92         git grep -h "(*NO_JIT)Aæ" invalid-0xe5 >actual &&
  93         test_cmp invalid-0xe5 actual
  94 '
  95
  96 test_lazy_prereq PCRE2_MATCH_INVALID_UTF '
  97         test-tool pcre2-config has-PCRE2_MATCH_INVALID_UTF
  98 '
  99
 100 test_expect_success GETTEXT_LOCALE,LIBPCRE2 'PCRE v2: grep non-ASCII from invalid UTF-8 data with -i' '
 101         test_might_fail git grep -hi "Æ" invalid-0x80 >actual &&
 102         test_might_fail git grep -hi "(*NO_JIT)Æ" invalid-0x80 >actual
 103 '
 104
 105 test_expect_success GETTEXT_LOCALE,LIBPCRE2,PCRE2_MATCH_INVALID_UTF 'PCRE v2: grep non-ASCII from invalid UTF-8 data with -i' '
 106         git grep -hi "Æ" invalid-0x80 >actual &&
 107         test_cmp expected actual &&
 108         git grep -hi "(*NO_JIT)Æ" invalid-0x80 >actual &&
 109         test_cmp expected actual
 110 '
 111
 112 test_expect_success GETTEXT_LOCALE,LIBPCRE2,PCRE2_MATCH_INVALID_UTF 'PCRE v2: grep non-ASCII from invalid UTF-8 data with -i (PCRE2 bug #2642)' '
 113         git grep -hi "Æ" invalid-0xe5 >actual &&
 114         test_cmp invalid-0xe5 actual &&
 115         git grep -hi "(*NO_JIT)Æ" invalid-0xe5 >actual &&
 116         test_cmp invalid-0xe5 actual &&
 117
 118         # Only the case of grepping the ASCII part in a way that
 119         # relies on -i fails
 120         git grep -hi "aÆ" invalid-0xe5 >actual &&
 121         test_cmp invalid-0xe5 actual &&
 122         git grep -hi "(*NO_JIT)aÆ" invalid-0xe5 >actual &&
 123         test_cmp invalid-0xe5 actual
 124 '
 125
 126 test_expect_success GETTEXT_LOCALE,LIBPCRE2 'PCRE v2: grep non-literal ASCII from UTF-8' '
 127         git grep --perl-regexp -h -o -e ll. file >actual &&
 128         echo "lló" >expected &&
 129         test_cmp expected actual
 130 '
 131
 132 test_done