t/t7816-grep-binary-pattern.sh

   1 #!/bin/sh
   2
   3 test_description='git grep with a binary pattern files'
   4
   5 TEST_PASSES_SANITIZE_LEAK=true
   6 . ./lib-gettext.sh
   7
   8 nul_match_internal () {
   9         matches=$1
  10         prereqs=$2
  11         lc_all=$3
  12         extra_flags=$4
  13         flags=$5
  14         pattern=$6
  15         pattern_human=$(echo "$pattern" | sed 's/Q/<NUL>/g')
  16
  17         if test "$matches" = 1
  18         then
  19                 test_expect_success $prereqs "LC_ALL='$lc_all' git grep $extra_flags -f f $flags '$pattern_human' a" "
  20                         printf '$pattern' | q_to_nul >f &&
  21                         LC_ALL='$lc_all' git grep $extra_flags -f f $flags a
  22                 "
  23         elif test "$matches" = 0
  24         then
  25                 test_expect_success $prereqs "LC_ALL='$lc_all' git grep $extra_flags -f f $flags '$pattern_human' a" "
  26                         >stderr &&
  27                         printf '$pattern' | q_to_nul >f &&
  28                         test_must_fail env LC_ALL=\"$lc_all\" git grep $extra_flags -f f $flags a 2>stderr &&
  29                         test_i18ngrep ! 'This is only supported with -P under PCRE v2' stderr
  30                 "
  31         elif test "$matches" = P
  32         then
  33                 test_expect_success $prereqs "error, PCRE v2 only: LC_ALL='$lc_all' git grep -f f $flags '$pattern_human' a" "
  34                         >stderr &&
  35                         printf '$pattern' | q_to_nul >f &&
  36                         test_must_fail env LC_ALL=\"$lc_all\" git grep -f f $flags a 2>stderr &&
  37                         test_i18ngrep 'This is only supported with -P under PCRE v2' stderr
  38                 "
  39         else
  40                 test_expect_success "PANIC: Test framework error. Unknown matches value $matches" 'false'
  41         fi
  42 }
  43
  44 nul_match () {
  45         matches=$1
  46         matches_pcre2=$2
  47         matches_pcre2_locale=$3
  48         flags=$4
  49         pattern=$5
  50         pattern_human=$(echo "$pattern" | sed 's/Q/<NUL>/g')
  51
  52         nul_match_internal "$matches" "" "C" "" "$flags" "$pattern"
  53         nul_match_internal "$matches_pcre2" "LIBPCRE2" "C" "-P" "$flags" "$pattern"
  54         nul_match_internal "$matches_pcre2_locale" "LIBPCRE2,GETTEXT_LOCALE" "$is_IS_locale" "-P" "$flags" "$pattern"
  55 }
  56
  57 test_expect_success 'setup' "
  58         echo 'binaryQfileQm[*]cQ*æQð' | q_to_nul >a &&
  59         git add a &&
  60         git commit -m.
  61 "
  62
  63 # Simple fixed-string matching
  64 nul_match P P P '-F' 'yQf'
  65 nul_match P P P '-F' 'yQx'
  66 nul_match P P P '-Fi' 'YQf'
  67 nul_match P P P '-Fi' 'YQx'
  68 nul_match P P 1 '' 'yQf'
  69 nul_match P P 0 '' 'yQx'
  70 nul_match P P 1 '' 'æQð'
  71 nul_match P P P '-F' 'eQm[*]c'
  72 nul_match P P P '-Fi' 'EQM[*]C'
  73
  74 # Regex patterns that would match but shouldn't with -F
  75 nul_match P P P '-F' 'yQ[f]'
  76 nul_match P P P '-F' '[y]Qf'
  77 nul_match P P P '-Fi' 'YQ[F]'
  78 nul_match P P P '-Fi' '[Y]QF'
  79 nul_match P P P '-F' 'æQ[ð]'
  80 nul_match P P P '-F' '[æ]Qð'
  81
  82 # Matching pattern and subject case with -i
  83 nul_match P 1 1 '-i' '[æ]Qð'
  84
  85 # ...PCRE v2 only matches non-ASCII with -i casefolding under UTF-8
  86 # semantics
  87 nul_match P P P '-Fi' 'ÆQ[Ð]'
  88 nul_match P 0 1 '-i'  'ÆQ[Ð]'
  89 nul_match P 0 1 '-i'  '[Æ]QÐ'
  90 nul_match P 0 1 '-i' '[Æ]Qð'
  91 nul_match P 0 1 '-i' 'ÆQÐ'
  92
  93 # \0 in regexes can only work with -P & PCRE v2
  94 nul_match P P 1 '' 'yQ[f]'
  95 nul_match P P 1 '' '[y]Qf'
  96 nul_match P P 1 '-i' 'YQ[F]'
  97 nul_match P P 1 '-i' '[Y]Qf'
  98 nul_match P P 1 '' 'æQ[ð]'
  99 nul_match P P 1 '' '[æ]Qð'
 100 nul_match P P 1 '-i' 'ÆQ[Ð]'
 101 nul_match P P 1 '' 'eQm.*cQ'
 102 nul_match P P 1 '-i' 'EQM.*cQ'
 103 nul_match P P 0 '' 'eQm[*]c'
 104 nul_match P P 0 '-i' 'EQM[*]C'
 105
 106 # Assert that we're using REG_STARTEND and the pattern doesn't match
 107 # just because it's cut off at the first \0.
 108 nul_match P P 0 '-i' 'NOMATCHQð'
 109 nul_match P P 0 '-i' '[Æ]QNOMATCH'
 110 nul_match P P 0 '-i' '[æ]QNOMATCH'
 111
 112 # Ensure that the matcher doesn't regress to something that stops at
 113 # \0
 114 nul_match P P P '-F' 'yQ[f]'
 115 nul_match P P P '-Fi' 'YQ[F]'
 116 nul_match P P 0 '' 'yQNOMATCH'
 117 nul_match P P 0 '' 'QNOMATCH'
 118 nul_match P P 0 '-i' 'YQNOMATCH'
 119 nul_match P P 0 '-i' 'QNOMATCH'
 120 nul_match P P P '-F' 'æQ[ð]'
 121 nul_match P P P '-Fi' 'ÆQ[Ð]'
 122 nul_match P P 1 '-i' 'ÆQ[Ð]'
 123 nul_match P P 0 '' 'yQNÓMATCH'
 124 nul_match P P 0 '' 'QNÓMATCH'
 125 nul_match P P 0 '-i' 'YQNÓMATCH'
 126 nul_match P P 0 '-i' 'QNÓMATCH'
 127
 128 test_done