tests/yaml/multipass-vs-match.yaml

#!tools/lou_checkyaml
#
# Comparison of match and multi-pass opcodes
# ::::::::::::::::::::::::::::::::::::::::::
#
# Copyright © 2019 by Bert Frees
#
# Copying and distribution of this file, with or without modification,
# are permitted in any medium without royalty provided the copyright
# notice and this notice are preserved. This file is offered as-is,
# without any warranty.
# ..............................................................................................

display: tables/unicode-without-blank.dis

# multipass-vs-match_common.uti:
#   include tables/latinLetterDef8Dots.uti
#   include tables/loweredDigits6Dots.uti
#   space \s 0
#   letter ç 4-14
#   punctuation ! 2346
#   punctuation . 46
#   punctuation - 36
#   sign % 146
#   sign _ 145

# ..............................................................................................

# simple match replace
table: |
  include multipass-vs-match_common.uti
  noback context "abc" @123
table : |
  include multipass-vs-match_common.uti
  noback context ["abc"] @123
table: |
  include multipass-vs-match_common.uti
  match - abc - 123
tests:
  - - "abc"
    - "⠇"

# lookbehind
table: |
  include multipass-vs-match_common.uti
  always x 456
  noback context _"x"["abc"] @123
table: |
  include multipass-vs-match_common.uti
  always x 456
  match x abc - 123
tests:
  - - "xabc"
    - "⠸⠇"

# without the _ it is not a true lookbehind
table: |
  include multipass-vs-match_common.uti
  always x 456
  noback context "x"["abc"] @123
tests:
  - - "xabc"
    - "⠸⠁⠃⠉"

# lookahead
table: |
  include multipass-vs-match_common.uti
  always x 456
  noback context ["abc"]"x" @123
table: |
  include multipass-vs-match_common.uti
  always x 456
  match - abc x 123
tests:
  - - "abcx"
    - "⠇⠸"

# zero-length match (insert)
# cursor advances at least one so that a zero-length match does not result in an endless loop

# with lookahead
table: |
  include multipass-vs-match_common.uti
  always x 456
  noback context []"xx" @123
tests:
  - - "xxx"
    - "⠇⠸⠇⠸⠸"

# with lookbehind
table: |
  include multipass-vs-match_common.uti
  always x 456
  noback context "x"[] @123
tests:
  - - "xxx"
    - "⠭⠇⠭⠇⠭⠇"
    - xfail: got ⠇⠸⠇
table: |
  include multipass-vs-match_common.uti
  always x 456
  noback context "xx"[] @123
tests:
  - - "xxx"
    - "⠭⠭⠇⠭⠇"
    - xfail: got ⠭⠭⠇⠸

# error: characters, dots, attributes, or class swap not found in test part
# table: |
#  include multipass-vs-match_common.uti
#   noback context _"x"[] @123
# tests:
#   - - "xxx"
#     - "⠇⠭⠇⠭⠇⠭"

# beginning of string
table: |
  include multipass-vs-match_common.uti
  noback context `["abc"] @123
table: |
  include multipass-vs-match_common.uti
  noback context `"abc" @123
table: |
  include multipass-vs-match_common.uti
  match ^ abc - 123
table: |
  include multipass-vs-match_common.uti
  match $ abc - 123
tests:
  - - "xabc"
    - "⠭⠁⠃⠉"
  - - "abc"
    - "⠇"

# end of string
table: |
  include multipass-vs-match_common.uti
  noback context ["abc"]~ @123
table: |
  include multipass-vs-match_common.uti
  noback context "abc"~ @123
table: |
  include multipass-vs-match_common.uti
  match - abc $ 123
table: |
  include multipass-vs-match_common.uti
  match - abc ^ 123
tests:
  - - "abcx"
    - "⠁⠃⠉⠭"
  - - "abc"
    - "⠇"

# wildcard
table: |
  include multipass-vs-match_common.uti
  noback context ["abc"]$a"x" @123
table: |
  include multipass-vs-match_common.uti
  match - abc .x 123
tests:
  - - "abcdx"
    - "⠇⠙⠭"
  - - "abc!x"
    - "⠇⠮⠭"

# character classes
table: |
  include multipass-vs-match_common.uti
  noback context ["abc"]$U @123@1 # uppercase letter
  noback context ["abc"]$u @123@2 # lowercase letter
  noback context ["abc"]$l @123@3 # letter
  noback context ["abc"]$p @123@4 # punctuation
  noback context ["abc"]$d @123@5 # digit
  noback context ["abc"]$s @123@6 # space
  noback context ["abc"]$S @123@7 # sign
  noback context ["def"]$sp @456  # space or punctuation
table: |
  include multipass-vs-match_common.uti
  match - abc %u 123-1  # uppercase letter
  match - abc %l 123-2  # lowercase letter
  match - abc %a 123-3  # letter
  match - abc %. 123-4  # punctuation
  match - abc %# 123-5  # digit
  match - abc %_ 123-6  # space
  match - abc %$ 123-7  # sign
  match - def %[_.] 456 # space or punctuation
tests:
  - - "abcX"
    - "⠇⠁⡭"
  - - "abcx"
    - "⠇⠂⠭"
  - - "abcç"
    - "⠇⠄⠈⠉"
  - - "abc!"
    - "⠇⠈⠮"
  - - "abc1"
    - "⠇⠐⠂"
  - - "abc "
    - "⠇⠠ "
  - - "abc%"
    - "⠇⡀⠩"
  - - "def!"
    - "⠸⠮"
  - - "def "
    - "⠸ "
  - - "defx"
    - "⠙⠑⠋⠭"

# character sets / custom character classes
table: |
  include multipass-vs-match_common.uti
  class xy xy
  noback context ["abc"]%xy @123
table: |
  include multipass-vs-match_common.uti
  class xy xy
  noback context ["abc"]$w @123
table: |
  include multipass-vs-match_common.uti
  match - abc [xy] 123
table: |
  include multipass-vs-match_common.uti
  match - abc x|y 123
table: |
  include multipass-vs-match_common.uti
  match - abc (x|y) 123
table: |
  include multipass-vs-match_common.uti
  attribute 0 xy
  match - abc %0 123
table: |
  include multipass-vs-match_common.uti
  class fo _
  class foo _
  class fooo _
  class foooo _
  class fooooo xy # fifth class
  match - abc %0 123
tests:
  - - "abcx"
    - "⠇⠭"
  - - "abcy"
    - "⠇⠽"
  - - "abcz"
    - "⠁⠃⠉⠵"

# alternation
table: |
  include multipass-vs-match_common.uti
  match - abc xy|z 123
table: |
  include multipass-vs-match_common.uti
  noback context ["abc"]"xy" @123
  noback context ["abc"]"z" @123
tests:
  - - "abcx"
    - "⠁⠃⠉⠭"
  - - "abcxy"
    - "⠇⠭⠽"
  - - "abcz"
    - "⠇⠵"

# repitition

# single character
table: |
  include multipass-vs-match_common.uti
  match - abc x+y 123
table: |
  include multipass-vs-match_common.uti
  class x x
  noback context ["abc"]%x."y" @123
tests:
  - - "abcy"
    - "⠁⠃⠉⠽"
  - - "abcxy"
    - "⠇⠭⠽"
  - - "abcxxy"
    - "⠇⠭⠭⠽"

# character set/class
table: |
  include multipass-vs-match_common.uti
  match - abc [xy]+z 123
table: |
  include multipass-vs-match_common.uti
  match - abc (x|y)+z 123
table: |
  include multipass-vs-match_common.uti
  class xy xy
  noback context ["abc"]%xy."z" @123
tests:
  - - "abcxz"
    - "⠇⠭⠵"
  - - "abcxyz"
    - "⠇⠭⠽⠵"
  - - "abcyyxz"
    - "⠇⠽⠽⠭⠵"

# group of characters
table: |
  include multipass-vs-match_common.uti
  match - abc (xy)+z 123
tests:
  - - "abcxyz"
    - "⠇⠭⠽⠵"
  - - "abcxyxz"
    - "⠁⠃⠉⠭⠽⠭⠵"
  - - "abcxyxyz"
    - "⠇⠭⠽⠭⠽⠵"

# zero or more
table: |
  include multipass-vs-match_common.uti
  match - abc x*y 123
table: |
  include multipass-vs-match_common.uti
  class x x
  noback context ["abc"]"y" @123
  noback context ["abc"]%x."y" @123
tests:
  - - "abcy"
    - "⠇⠽"
  - - "abcxy"
    - "⠇⠭⠽"
  - - "abcxxy"
    - "⠇⠭⠭⠽"

# optional
table: |
  include multipass-vs-match_common.uti
  match - abc x?y 123
table: |
  include multipass-vs-match_common.uti
  noback context ["abc"]"y" @123
  noback context ["abc"]"xy" @123
tests:
  - - "abcy"
    - "⠇⠽"
  - - "abcxy"
    - "⠇⠭⠽"
  - - "abcxxy"
    - "⠁⠃⠉⠭⠭⠽"

# limiting repitition (not more than two)
table: |
  include multipass-vs-match_common.uti
  class x x
  noback context ["abc"]%x1-2"y" @123
tests:
  - - "abcy"
    - "⠁⠃⠉⠽"
  - - "abcxy"
    - "⠇⠭⠽"
  - - "abcxxy"
    - "⠇⠭⠭⠽"
  - - "abcxxxy"
    - "⠁⠃⠉⠭⠭⠭⠽"

# exact repitition
table: |
  include multipass-vs-match_common.uti
  class x x
  noback context ["abc"]%x2"y" @123
tests:
  - - "abcxy"
    - "⠁⠃⠉⠭⠽"
  - - "abcxxy"
    - "⠇⠭⠭⠽"
  - - "abcxxxy"
    - "⠁⠃⠉⠭⠭⠭⠽"

# three or more
table: |
  include multipass-vs-match_common.uti
  class x x
  noback context ["abc"]%x2%x."y" @123
tests:
  - - "abcxxy"
    - "⠁⠃⠉⠭⠭⠽"
  - - "abcxxxy"
    - "⠇⠭⠭⠭⠽"
  - - "abcxxxxy"
    - "⠇⠭⠭⠭⠭⠽"

# negation
table: |
  include multipass-vs-match_common.uti
  match - abc !x 123
table: |
  include multipass-vs-match_common.uti
  noback context ["abc"]!"x" @123
table: |
  include multipass-vs-match_common.uti
  class x x
  noback context ["abc"]!%x @123
tests:
  - - "abcx"
    - "⠁⠃⠉⠭"
  - - "abcy"
    - "⠇⠽"
  - - "abc"
    - "⠁⠃⠉"

# negated group
table: |
  include multipass-vs-match_common.uti
  match - abc !(xy) 123
tests:
  - - "abcx"
    - "⠁⠃⠉⠭"
  - - "abcxy"
    - "⠁⠃⠉⠭⠽"

# replace with nothing
table: |
  include multipass-vs-match_common.uti
  noback context "abc" ?
tests:
  - - "xabcx"
    - "⠭⠭"

# character classes in match
table: |
  include multipass-vs-match_common.uti
  class xy xy
  noback context $s  @1
  noback context $d. @3
  noback context %xy @2
tests:
  - - " "
    - "⠁"
  - - "y"
    - "⠂"
  - - "12345"
    - "⠄"