extra/pcre/pcre-tests.factor

   1 USING: accessors arrays assocs continuations http.client kernel
   2 literals math math.parser ranges pcre pcre.ffi pcre.private
   3 random sequences system tools.test ;
   4 QUALIFIED: regexp
   5 QUALIFIED: splitting
   6 IN: pcre.tests
   7
   8 { { "Bords" "words" "word" } } [
   9     "Bords, words, word." { ", " ", " "." } split-subseqs
  10 ] unit-test
  11
  12 { { { 3 "day" } { 2 "month" } { 1 "year" } } } [
  13     "(?P<year>\\d{4})-(?P<month>\\d{2})-(?P<day>\\d{2})"
  14     <compiled-pcre> nametable>>
  15 ] unit-test
  16
  17 CONSTANT: iso-date "(?P<year>\\d{4})-(?P<month>\\d{2})-(?P<day>\\d{2})"
  18
  19 ! On windows the erroffset appears to be set to 0 despite there being
  20 ! nothing wrong with the regexp.
  21 { t } [
  22     "foo" (pcre) 3array rest { { f -1 } { f 0 } } member?
  23 ] unit-test
  24
  25 { { 1 2 3 } } [
  26     iso-date <pcre>
  27     { "year" "month" "day" } [ pcre_get_stringnumber ] with map
  28 ] unit-test
  29
  30 { t } [
  31     "foo" <compiled-pcre> PCRE_UTF8 has-option?
  32 ] unit-test
  33
  34 ! This option is not present on old PCRE versions.
  35 { t } [
  36     "foo" <compiled-pcre> version 8.10 >
  37     [ PCRE_UCP has-option? ] [ PCRE_UCP has-option? not ] if
  38 ] unit-test
  39
  40 os unix? [ [ 10 ] [ PCRE_CONFIG_NEWLINE pcre-config ] unit-test ] when
  41
  42 ! In this day and age, not supporting utf-8 is broken.
  43 { 1 } [ PCRE_CONFIG_UTF8 pcre-config ] unit-test
  44
  45 { 1 } [ PCRE_CONFIG_UNICODE_PROPERTIES pcre-config ] unit-test
  46
  47 ! Ok if these options throw if the pcre library is to old to support
  48 ! these configuration parameters.
  49 { t } [
  50     [ PCRE_CONFIG_UTF16 pcre-config ] [ what>> ] recover
  51     { 0 $ PCRE_CONFIG_UTF16 } member?
  52 ] unit-test
  53 { t } [
  54     [ PCRE_CONFIG_UTF32 pcre-config ] [ what>> ] recover
  55     { 0 $ PCRE_CONFIG_UTF32 } member?
  56 ] unit-test
  57
  58 { 33 }
  59 [
  60     [ "foo" <pcre> f 33 pcre-fullinfo ] [ what>> ] recover
  61 ] unit-test
  62
  63 ! Tests for findall
  64 {
  65     { { f "1999-01-12" } { "year" "1999" } { "month" "01" } { "day" "12" } }
  66 } [
  67     "1999-01-12" iso-date <compiled-pcre> findall first
  68 ] unit-test
  69
  70 { 3 } [
  71     "2003-10-09 1999-09-01 1514-10-20" iso-date <compiled-pcre> findall length
  72 ] unit-test
  73
  74 { 5 } [ "abcdef" "[a-e]" findall length ] unit-test
  75
  76 { 3 } [ "foo bar baz" "foo|bar|baz" findall length ] unit-test
  77
  78 { 3 } [ "örjan är åtta" "[åäö]" findall length ] unit-test
  79
  80 { 3 } [ "ÅÄÖ" "\\p{Lu}" findall length ] unit-test
  81
  82 { 3 } [ "foobar" "foo(?=bar)" findall first first second length ] unit-test
  83
  84 { { { { f ", " } } { { f ", " } } { { f "." } } } } [
  85     "Words, words, word." "\\W+" findall
  86 ] unit-test
  87
  88 { { ", " ", " "." } } [
  89     "Words, words, word." "\\W+" findall [ first second ] map
  90 ] unit-test
  91
  92 : long-string ( -- x )
  93     10000 [ CHAR: a CHAR: z [a..b] random ] "" replicate-as ;
  94
  95 ! Performance
  96 { 0 } [ long-string ".{0,15}foobar.{0,10}" findall length ] unit-test
  97
  98 ! Empty matches, corner case behavior is copied from pcredemo.c
  99 { { { { f "foo" } } { { f "" } } } }
 100 [ "foo" ".*" findall ] unit-test
 101
 102 { { { { f "" } } { { f "" } } { { f "" } } } }
 103 [ "foo" "B*" findall ] unit-test
 104
 105 ! Empty matches in strings with multi-byte characters are tricky.
 106 { { { { f "" } } { { f "" } } { { f "" } } { { f "" } } } }
 107 [ "öööö" "x*" findall ] unit-test
 108
 109 ! Tests for matches?
 110 { t } [ "örjan" "örjan" matches? ] unit-test
 111
 112 { t } [ "abcö" "\\p{Ll}{4}" matches? ] unit-test
 113
 114 ! This used to work in 8.36, but might have changed in later versions.
 115 ! See: https://bugs.exim.org/show_bug.cgi?id=1875
 116 version 8.36 <= [
 117     { t t } [
 118         "(?s)." <compiled-pcre> PCRE_DOTALL has-option?
 119         "(?i)x" <compiled-pcre> PCRE_CASELESS has-option?
 120     ] unit-test
 121 ] when
 122
 123 { f } [ "\n" "." matches? ] unit-test
 124 { t } [ "\n" "(?s)." matches? ] unit-test
 125
 126 { f t } [
 127     "hello\nthere" "^.*$" <compiled-pcre> matches?
 128     "hello\nthere" "(?s)^.*$" <compiled-pcre> matches?
 129 ] unit-test
 130
 131 ! Modes off by default
 132 { f f } [
 133     ! Caseless mode
 134     "x" <compiled-pcre> PCRE_CASELESS has-option?
 135     ! Dotall mode
 136     "." <compiled-pcre> PCRE_DOTALL has-option?
 137 ] unit-test
 138
 139 ! Backreferences
 140 { { t f } } [
 141     { "response and responsibility" "sense and responsibility" }
 142     [ "(sens|respons)e and \\1ibility" matches? ] map
 143 ] unit-test
 144
 145 { { t t f } } [
 146     { "rah rah" "RAH RAH" "RAH rah" } [ "((?i)rah)\\s+\\1" matches? ] map
 147 ] unit-test
 148
 149 ! Splitting
 150 { { { "Words" "words" "word" } { "Words" "words" "word" } } } [
 151     "Words, words, word." { "\\W+" "[,. ]" } [ split ] with map
 152 ] unit-test
 153
 154 ! Bigger tests
 155 { t } [
 156     "https://factorcode.org/" http-get nip
 157     "href=\"(?P<link>[^\"]+)\"" findall [ "link" of ] map sequence?
 158 ] unit-test
 159
 160 ! Test that the regexp syntax works.
 161 { t } [ "1234abcd" regexp:R/ ^\d+\w+$/ matches? ] unit-test