1# $NetBSD: varmod-match.mk,v 1.15 2023/06/23 04:56:54 rillig Exp $ 2# 3# Tests for the :M variable modifier, which filters words that match the 4# given pattern. 5# 6# See ApplyModifier_Match and ModifyWord_Match for the implementation. 7 8.MAKEFLAGS: -dc 9 10NUMBERS= One Two Three Four five six seven 11 12# Only keep words that start with an uppercase letter. 13.if ${NUMBERS:M[A-Z]*} != "One Two Three Four" 14. error 15.endif 16 17# Only keep words that start with a character other than an uppercase letter. 18.if ${NUMBERS:M[^A-Z]*} != "five six seven" 19. error 20.endif 21 22# Only keep words that don't start with s and at the same time end with 23# either of [ex]. 24# 25# This test case ensures that the negation from the first character class 26# does not propagate to the second character class. 27.if ${NUMBERS:M[^s]*[ex]} != "One Three five" 28. error 29.endif 30 31# Before 2020-06-13, this expression called Str_Match 601,080,390 times. 32# Since 2020-06-13, this expression calls Str_Match 1 time. 33.if ${:U****************:M****************b} 34.endif 35 36# Before 2023-06-22, this expression called Str_Match 2,621,112 times. 37# Adding another '*?' to the pattern called Str_Match 20,630,572 times. 38# Adding another '*?' to the pattern called Str_Match 136,405,672 times. 39# Adding another '*?' to the pattern called Str_Match 773,168,722 times. 40# Adding another '*?' to the pattern called Str_Match 3,815,481,072 times. 41# Since 2023-06-22, Str_Match no longer backtracks. 42.if ${:U..................................................b:M*?*?*?*?*?a} 43.endif 44 45# To match a dollar sign in a word, double it. 46# 47# This is different from the :S and :C variable modifiers, where a '$' 48# has to be escaped as '\$'. 49.if ${:Ua \$ sign:M*$$*} != "\$" 50. error 51.endif 52 53# In the :M modifier, '\$' does not escape a dollar. Instead it is 54# interpreted as a backslash followed by whatever expression the 55# '$' starts. 56# 57# This differs from the :S, :C and several other variable modifiers. 58${:U*}= asterisk 59.if ${:Ua \$ sign any-asterisk:M*\$*} != "any-asterisk" 60. error 61.endif 62 63# TODO: ${VAR:M(((}}}} 64# TODO: ${VAR:M{{{)))} 65# TODO: ${VAR:M${UNBALANCED}} 66# TODO: ${VAR:M${:U(((\}\}\}}} 67 68.MAKEFLAGS: -d0 69 70# Special characters: 71# * matches 0 or more arbitrary characters 72# ? matches a single arbitrary character 73# \ starts an escape sequence, only outside ranges 74# [ starts a set for matching a single character 75# ] ends a set for matching a single character 76# - in a set, forms a range of characters 77# ^ as the first character in a set, negates the set 78# ( during parsing of the pattern, starts a nesting level 79# ) during parsing of the pattern, ends a nesting level 80# { during parsing of the pattern, starts a nesting level 81# } during parsing of the pattern, ends a nesting level 82# : during parsing of the pattern, finishes the pattern 83# $ during parsing of the pattern, starts a nested expression 84# # in a line except a shell command, starts a comment 85# 86# Pattern parts: 87# * matches 0 or more arbitrary characters 88# ? matches exactly 1 arbitrary character 89# \x matches exactly the character 'x' 90# [...] matches exactly 1 character from the set 91# [^...] matches exactly 1 character outside the set 92# [a-z] matches exactly 1 character from the range 'a' to 'z' 93# 94 95# [] matches never 96.if ${ ab a[]b a[b a b :L:M[]} != "" 97. error 98.endif 99 100# a[]b matches never 101.if ${ ab a[]b a[b a b [ ] :L:Ma[]b} != "" 102. error 103.endif 104 105# [^] matches exactly 1 arbitrary character 106.if ${ ab a[]b a[b a b [ ] :L:M[^]} != "a b [ ]" 107. error 108.endif 109 110# a[^]b matches 'a', then exactly 1 arbitrary character, then 'b' 111.if ${ ab a[]b a[b a b :L:Ma[^]b} != "a[b" 112. error 113.endif 114 115# [Nn0] matches exactly 1 character from the set 'N', 'n', '0' 116.if ${ a b N n 0 Nn0 [ ] :L:M[Nn0]} != "N n 0" 117. error 118.endif 119 120# [a-c] matches exactly 1 character from the range 'a' to 'c' 121.if ${ A B C a b c d [a-c] [a] :L:M[a-c]} != "a b c" 122. error 123.endif 124 125# [c-a] matches the same as [a-c] 126.if ${ A B C a b c d [a-c] [a] :L:M[c-a]} != "a b c" 127. error 128.endif 129 130# [^a-c67] 131# matches a single character, except for 'a', 'b', 'c', '6' or 132# '7' 133.if ${ A B C a b c d 5 6 7 8 [a-c] [a] :L:M[^a-c67]} != "A B C d 5 8" 134. error 135.endif 136 137# [\] matches a single backslash 138WORDS= a\b a[\]b ab 139.if ${WORDS:Ma[\]b} != "a\\b" 140. error 141.endif 142 143# : terminates the pattern 144.if ${ A * :L:M:} != "" 145. error 146.endif 147 148# \: matches a colon 149.if ${ ${:U\: \:\:} :L:M\:} != ":" 150. error 151.endif 152 153# ${:U\:} matches a colon 154.if ${ ${:U\:} ${:U\:\:} :L:M${:U\:}} != ":" 155. error 156.endif 157 158# [:] matches never since the ':' starts the next modifier 159# expect+3: warning: Unfinished character list in pattern '[' of modifier ':M' 160# expect+2: Unknown modifier "]" 161# expect+1: Malformed conditional (${ ${:U\:} ${:U\:\:} :L:M[:]} != ":") 162.if ${ ${:U\:} ${:U\:\:} :L:M[:]} != ":" 163. error 164.else 165. error 166.endif 167 168# [\] matches exactly a backslash; no escaping takes place in 169# character ranges 170# Without the 'a' in the below words, the backslash would end a word and thus 171# influence how the string is split into words. 172WORDS= 1\a 2\\a 173.if ${WORDS:M?[\]a} != "1\\a" 174. error 175.endif 176 177# [[-]] May look like it would match a single '[', '\' or ']', but 178# the inner ']' has two roles: it is the upper bound of the 179# character range as well as the closing character of the 180# character list. The outer ']' is just a regular character. 181WORDS= [ ] [] \] ]] 182.if ${WORDS:M[[-]]} != "[] \\] ]]" 183. error 184.endif 185 186# [b[-]a] 187# Same as for '[[-]]': the character list stops at the first 188# ']', and the 'a]' is treated as a literal string. 189WORDS= [a \a ]a []a \]a ]]a [a] \a] ]a] ba] 190.if ${WORDS:M[b[-]a]} != "[a] \\a] ]a] ba]" 191. error 192.endif 193 194# [-] Matches a single '-' since the '-' only becomes part of a 195# character range if it is preceded and followed by another 196# character. 197WORDS= - -] 198.if ${WORDS:M[-]} != "-" 199. error 200.endif 201 202# [ Incomplete empty character list, never matches. 203WORDS= a a[ 204# expect+1: warning: Unfinished character list in pattern 'a[' of modifier ':M' 205.if ${WORDS:Ma[} != "" 206. error 207.endif 208 209# [^ Incomplete negated empty character list, matches any single 210# character. 211WORDS= a a[ aX 212# expect+1: warning: Unfinished character list in pattern 'a[^' of modifier ':M' 213.if ${WORDS:Ma[^} != "a[ aX" 214. error 215.endif 216 217# [-x1-3 Incomplete character list, matches those elements that can be 218# parsed without lookahead. 219WORDS= - + x xx 0 1 2 3 4 [x1-3 220# expect+1: warning: Unfinished character list in pattern '[-x1-3' of modifier ':M' 221.if ${WORDS:M[-x1-3} != "- x 1 2 3" 222. error 223.endif 224 225# *[-x1-3 Incomplete character list after a wildcard, matches those 226# words that end with one of the characters from the list. 227WORDS= - + x xx 0 1 2 3 4 00 01 10 11 000 001 010 011 100 101 110 111 [x1-3 228# expect+1: warning: Unfinished character list in pattern '*[-x1-3' of modifier ':M' 229.if ${WORDS:M*[-x1-3} != "- x xx 1 2 3 01 11 001 011 101 111 [x1-3" 230. warning ${WORDS:M*[-x1-3} 231.endif 232 233# [^-x1-3 234# Incomplete negated character list, matches any character 235# except those elements that can be parsed without lookahead. 236WORDS= - + x xx 0 1 2 3 4 [x1-3 237# expect+1: warning: Unfinished character list in pattern '[^-x1-3' of modifier ':M' 238.if ${WORDS:M[^-x1-3} != "+ 0 4" 239. error 240.endif 241 242# [\ Incomplete character list containing a single '\'. 243# 244# A word can only end with a backslash if the preceding 245# character is a backslash as well; in all other cases the final 246# backslash would escape the following space, making the space 247# part of the word. Only the very last word of a string can be 248# '\', as there is no following space that could be escaped. 249WORDS= \\ \a ${:Ux\\} 250.if ${WORDS:M?[\]} != "\\\\ x\\" 251. error 252.endif 253 254# [x- Incomplete character list containing an incomplete character 255# range, matches only the 'x'. 256WORDS= [x- x x- y 257# expect+1: warning: Unfinished character range in pattern '[x-' of modifier ':M' 258.if ${WORDS:M[x-} != "x" 259. error 260.endif 261 262# [^x- Incomplete negated character list containing an incomplete 263# character range; matches each word that does not have an 'x' 264# at the position of the character list. 265# 266# XXX: Even matches strings that are longer than a single 267# character. 268WORDS= [x- x x- y yyyyy 269# expect+1: warning: Unfinished character range in pattern '[^x-' of modifier ':M' 270.if ${WORDS:M[^x-} != "[x- y yyyyy" 271. error 272.endif 273 274 275# The modifier ':tW' prevents splitting at whitespace. Even leading and 276# trailing whitespace is preserved. 277.if ${ plain string :L:tW:M*} != " plain string " 278. error 279.endif 280 281# Without the modifier ':tW', the string is split into words. All whitespace 282# around and between the words is normalized to a single space. 283.if ${ plain string :L:M*} != "plain string" 284. error 285.endif 286 287 288# The pattern can come from a variable expression. For single-letter 289# variables, either the short form or the long form can be used, just as 290# everywhere else. 291PRIMES= 2 3 5 7 11 292n= 2 293.if ${PRIMES:M$n} != "2" 294. error 295.endif 296.if ${PRIMES:M${n}} != "2" 297. error 298.endif 299.if ${PRIMES:M${:U2}} != "2" 300. error 301.endif 302 303 304# Before var.c 1.1031 from 2022-08-24, the following expressions caused an 305# out-of-bounds read beyond the indirect ':M' modifiers. 306.if ${:U:${:UM\\}} # The ':M' pattern need not be unescaped, the 307. error # resulting pattern is '\', it never matches 308.endif # anything. 309.if ${:U:${:UM\\\:\\}} # The ':M' pattern must be unescaped, the 310. error # resulting pattern is ':\', it never matches 311.endif # anything. 312