t/t7816-grep-binary-pattern.sh

   1 #!/bin/sh
   2
   3 test_description='git grep with a binary pattern files'
   4
   5 . ./lib-gettext.sh
   6
   7 nul_match_internal () {
   8         matches=$1
   9         prereqs=$2
  10         lc_all=$3
  11         extra_flags=$4
  12         flags=$5
  13         pattern=$6
  14         pattern_human=$(echo "$pattern" | sed 's/Q/<NUL>/g')
  15
  16         if test "$matches" = 1
  17         then
  18                 test_expect_success $prereqs "LC_ALL='$lc_all' git grep $extra_flags -f f $flags '$pattern_human' a" "
  19                         printf '$pattern' | q_to_nul >f &&
  20                         LC_ALL='$lc_all' git grep $extra_flags -f f $flags a
  21                 "
  22         elif test "$matches" = 0
  23         then
  24                 test_expect_success $prereqs "LC_ALL='$lc_all' git grep $extra_flags -f f $flags '$pattern_human' a" "
  25                         >stderr &&
  26                         printf '$pattern' | q_to_nul >f &&
  27                         test_must_fail env LC_ALL=\"$lc_all\" git grep $extra_flags -f f $flags a 2>stderr &&
  28                         test_i18ngrep ! 'This is only supported with -P under PCRE v2' stderr
  29                 "
  30         elif test "$matches" = P
  31         then
  32                 test_expect_success $prereqs "error, PCRE v2 only: LC_ALL='$lc_all' git grep -f f $flags '$pattern_human' a" "
  33                         >stderr &&
  34                         printf '$pattern' | q_to_nul >f &&
  35                         test_must_fail env LC_ALL=\"$lc_all\" git grep -f f $flags a 2>stderr &&
  36                         test_i18ngrep 'This is only supported with -P under PCRE v2' stderr
  37                 "
  38         else
  39                 test_expect_success "PANIC: Test framework error. Unknown matches value $matches" 'false'
  40         fi
  41 }
  42
  43 nul_match () {
  44         matches=$1
  45         matches_pcre2=$2
  46         matches_pcre2_locale=$3
  47         flags=$4
  48         pattern=$5
  49         pattern_human=$(echo "$pattern" | sed 's/Q/<NUL>/g')
  50
  51         nul_match_internal "$matches" "" "C" "" "$flags" "$pattern"
  52         nul_match_internal "$matches_pcre2" "LIBPCRE2" "C" "-P" "$flags" "$pattern"
  53         nul_match_internal "$matches_pcre2_locale" "LIBPCRE2,GETTEXT_LOCALE" "$is_IS_locale" "-P" "$flags" "$pattern"
  54 }
  55
  56 test_expect_success 'setup' "
  57         echo 'binaryQfileQm[*]cQ*æQð' | q_to_nul >a &&
  58         git add a &&
  59         git commit -m.
  60 "
  61
  62 # Simple fixed-string matching that can use kwset (no -i && non-ASCII)
  63 nul_match P P P '-F' 'yQf'
  64 nul_match P P P '-F' 'yQx'
  65 nul_match P P P '-Fi' 'YQf'
  66 nul_match P P P '-Fi' 'YQx'
  67 nul_match P P 1 '' 'yQf'
  68 nul_match P P 0 '' 'yQx'
  69 nul_match P P 1 '' 'æQð'
  70 nul_match P P P '-F' 'eQm[*]c'
  71 nul_match P P P '-Fi' 'EQM[*]C'
  72
  73 # Regex patterns that would match but shouldn't with -F
  74 nul_match P P P '-F' 'yQ[f]'
  75 nul_match P P P '-F' '[y]Qf'
  76 nul_match P P P '-Fi' 'YQ[F]'
  77 nul_match P P P '-Fi' '[Y]QF'
  78 nul_match P P P '-F' 'æQ[ð]'
  79 nul_match P P P '-F' '[æ]Qð'
  80
  81 # The -F kwset codepath can't handle -i && non-ASCII...
  82 nul_match P 1 1 '-i' '[æ]Qð'
  83
  84 # ...PCRE v2 only matches non-ASCII with -i casefolding under UTF-8
  85 # semantics
  86 nul_match P P P '-Fi' 'ÆQ[Ð]'
  87 nul_match P 0 1 '-i'  'ÆQ[Ð]'
  88 nul_match P 0 1 '-i'  '[Æ]QÐ'
  89 nul_match P 0 1 '-i' '[Æ]Qð'
  90 nul_match P 0 1 '-i' 'ÆQÐ'
  91
  92 # \0 in regexes can only work with -P & PCRE v2
  93 nul_match P P 1 '' 'yQ[f]'
  94 nul_match P P 1 '' '[y]Qf'
  95 nul_match P P 1 '-i' 'YQ[F]'
  96 nul_match P P 1 '-i' '[Y]Qf'
  97 nul_match P P 1 '' 'æQ[ð]'
  98 nul_match P P 1 '' '[æ]Qð'
  99 nul_match P P 1 '-i' 'ÆQ[Ð]'
 100 nul_match P P 1 '' 'eQm.*cQ'
 101 nul_match P P 1 '-i' 'EQM.*cQ'
 102 nul_match P P 0 '' 'eQm[*]c'
 103 nul_match P P 0 '-i' 'EQM[*]C'
 104
 105 # Assert that we're using REG_STARTEND and the pattern doesn't match
 106 # just because it's cut off at the first \0.
 107 nul_match P P 0 '-i' 'NOMATCHQð'
 108 nul_match P P 0 '-i' '[Æ]QNOMATCH'
 109 nul_match P P 0 '-i' '[æ]QNOMATCH'
 110
 111 # Ensure that the matcher doesn't regress to something that stops at
 112 # \0
 113 nul_match P P P '-F' 'yQ[f]'
 114 nul_match P P P '-Fi' 'YQ[F]'
 115 nul_match P P 0 '' 'yQNOMATCH'
 116 nul_match P P 0 '' 'QNOMATCH'
 117 nul_match P P 0 '-i' 'YQNOMATCH'
 118 nul_match P P 0 '-i' 'QNOMATCH'
 119 nul_match P P P '-F' 'æQ[ð]'
 120 nul_match P P P '-Fi' 'ÆQ[Ð]'
 121 nul_match P P 1 '-i' 'ÆQ[Ð]'
 122 nul_match P P 0 '' 'yQNÓMATCH'
 123 nul_match P P 0 '' 'QNÓMATCH'
 124 nul_match P P 0 '-i' 'YQNÓMATCH'
 125 nul_match P P 0 '-i' 'QNÓMATCH'
 126
 127 test_done