1# $NetBSD: varmod-match.mk,v 1.12 2022/08/24 21:03:57 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# As of 2022-06-11, this expression calls Str_Match 5,242,223 times. 37# Adding another '*?' to the pattern calls Str_Match 41,261,143 times. 38.if ${:U..................................................b:M*?*?*?*?*?a} 39.endif 40 41# To match a dollar sign in a word, double it. 42# 43# This is different from the :S and :C variable modifiers, where a '$' 44# has to be escaped as '\$'. 45.if ${:Ua \$ sign:M*$$*} != "\$" 46. error 47.endif 48 49# In the :M modifier, '\$' does not escape a dollar. Instead it is 50# interpreted as a backslash followed by whatever expression the 51# '$' starts. 52# 53# This differs from the :S, :C and several other variable modifiers. 54${:U*}= asterisk 55.if ${:Ua \$ sign any-asterisk:M*\$*} != "any-asterisk" 56. error 57.endif 58 59# TODO: ${VAR:M(((}}}} 60# TODO: ${VAR:M{{{)))} 61# TODO: ${VAR:M${UNBALANCED}} 62# TODO: ${VAR:M${:U(((\}\}\}}} 63 64.MAKEFLAGS: -d0 65 66# Special characters: 67# * matches 0 or more arbitrary characters 68# ? matches a single arbitrary character 69# \ starts an escape sequence, only outside ranges 70# [ starts a set for matching a single character 71# ] ends a set for matching a single character 72# - in a set, forms a range of characters 73# ^ as the first character in a set, negates the set 74# ( during parsing of the pattern, starts a nesting level 75# ) during parsing of the pattern, ends a nesting level 76# { during parsing of the pattern, starts a nesting level 77# } during parsing of the pattern, ends a nesting level 78# : during parsing of the pattern, finishes the pattern 79# $ during parsing of the pattern, starts a nested expression 80# # in a line except a shell command, starts a comment 81# 82# Pattern parts: 83# * matches 0 or more arbitrary characters 84# ? matches exactly 1 arbitrary character 85# \x matches exactly the character 'x' 86# [...] matches exactly 1 character from the set 87# [^...] matches exactly 1 character outside the set 88# [a-z] matches exactly 1 character from the range 'a' to 'z' 89# 90 91# [] matches never 92.if ${ ab a[]b a[b a b :L:M[]} != "" 93. error 94.endif 95 96# a[]b matches never 97.if ${ ab a[]b a[b a b [ ] :L:Ma[]b} != "" 98. error 99.endif 100 101# [^] matches exactly 1 arbitrary character 102.if ${ ab a[]b a[b a b [ ] :L:M[^]} != "a b [ ]" 103. error 104.endif 105 106# a[^]b matches 'a', then exactly 1 arbitrary character, then 'b' 107.if ${ ab a[]b a[b a b :L:Ma[^]b} != "a[b" 108. error 109.endif 110 111# [Nn0] matches exactly 1 character from the set 'N', 'n', '0' 112.if ${ a b N n 0 Nn0 [ ] :L:M[Nn0]} != "N n 0" 113. error 114.endif 115 116# [a-c] matches exactly 1 character from the range 'a' to 'c' 117.if ${ A B C a b c d [a-c] [a] :L:M[a-c]} != "a b c" 118. error 119.endif 120 121# [c-a] matches the same as [a-c] 122.if ${ A B C a b c d [a-c] [a] :L:M[c-a]} != "a b c" 123. error 124.endif 125 126# [^a-c67] 127# matches a single character, except for 'a', 'b', 'c', '6' or 128# '7' 129.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" 130. error 131.endif 132 133# [\] matches a single backslash 134WORDS= a\b a[\]b ab 135.if ${WORDS:Ma[\]b} != "a\\b" 136. error 137.endif 138 139# : terminates the pattern 140.if ${ A * :L:M:} != "" 141. error 142.endif 143 144# \: matches a colon 145.if ${ ${:U\: \:\:} :L:M\:} != ":" 146. error 147.endif 148 149# ${:U\:} matches a colon 150.if ${ ${:U\:} ${:U\:\:} :L:M${:U\:}} != ":" 151. error 152.endif 153 154# [:] matches never since the ':' starts the next modifier 155# expect+2: Unknown modifier "]" 156# expect+1: Malformed conditional (${ ${:U\:} ${:U\:\:} :L:M[:]} != ":") 157.if ${ ${:U\:} ${:U\:\:} :L:M[:]} != ":" 158. error 159.else 160. error 161.endif 162 163# [\] matches exactly a backslash; no escaping takes place in 164# character ranges 165# Without the 'a' in the below words, the backslash would end a word and thus 166# influence how the string is split into words. 167WORDS= 1\a 2\\a 168.if ${WORDS:M?[\]a} != "1\\a" 169. error 170.endif 171 172# [[-]] May look like it would match a single '[', '\' or ']', but 173# the inner ']' has two roles: it is the upper bound of the 174# character range as well as the closing character of the 175# character list. The outer ']' is just a regular character. 176WORDS= [ ] [] \] ]] 177.if ${WORDS:M[[-]]} != "[] \\] ]]" 178. error 179.endif 180 181# [b[-]a] 182# Same as for '[[-]]': the character list stops at the first 183# ']', and the 'a]' is treated as a literal string. 184WORDS= [a \a ]a []a \]a ]]a [a] \a] ]a] ba] 185.if ${WORDS:M[b[-]a]} != "[a] \\a] ]a] ba]" 186. error 187.endif 188 189# [-] Matches a single '-' since the '-' only becomes part of a 190# character range if it is preceded and followed by another 191# character. 192WORDS= - -] 193.if ${WORDS:M[-]} != "-" 194. error 195.endif 196 197# [ Incomplete empty character list, never matches. 198WORDS= a a[ 199.if ${WORDS:Ma[} != "" 200. error 201.endif 202 203# [^ Incomplete negated empty character list, matches any single 204# character. 205WORDS= a a[ aX 206.if ${WORDS:Ma[^} != "a[ aX" 207. error 208.endif 209 210# [-x1-3 Incomplete character list, matches those elements that can be 211# parsed without lookahead. 212WORDS= - + x xx 0 1 2 3 4 [x1-3 213.if ${WORDS:M[-x1-3} != "- x 1 2 3" 214. error 215.endif 216 217# [^-x1-3 218# Incomplete negated character list, matches any character 219# except those elements that can be parsed without lookahead. 220WORDS= - + x xx 0 1 2 3 4 [x1-3 221.if ${WORDS:M[^-x1-3} != "+ 0 4" 222. error 223.endif 224 225# [\ Incomplete character list containing a single '\'. 226# 227# A word can only end with a backslash if the preceding 228# character is a backslash as well; in all other cases the final 229# backslash would escape the following space, making the space 230# part of the word. Only the very last word of a string can be 231# '\', as there is no following space that could be escaped. 232WORDS= \\ \a ${:Ux\\} 233.if ${WORDS:M?[\]} != "\\\\ x\\" 234. error 235.endif 236 237# [x- Incomplete character list containing an incomplete character 238# range, matches only the 'x'. 239WORDS= [x- x x- y 240.if ${WORDS:M[x-} != "x" 241. error 242.endif 243 244# [^x- Incomplete negated character list containing an incomplete 245# character range; matches each word that does not have an 'x' 246# at the position of the character list. 247# 248# XXX: Even matches strings that are longer than a single 249# character. 250WORDS= [x- x x- y yyyyy 251.if ${WORDS:M[^x-} != "[x- y yyyyy" 252. error 253.endif 254 255 256# The modifier ':tW' prevents splitting at whitespace. Even leading and 257# trailing whitespace is preserved. 258.if ${ plain string :L:tW:M*} != " plain string " 259. error 260.endif 261 262# Without the modifier ':tW', the string is split into words. All whitespace 263# around and between the words is normalized to a single space. 264.if ${ plain string :L:M*} != "plain string" 265. error 266.endif 267 268 269# The pattern can come from a variable expression. For single-letter 270# variables, either the short form or the long form can be used, just as 271# everywhere else. 272PRIMES= 2 3 5 7 11 273n= 2 274.if ${PRIMES:M$n} != "2" 275. error 276.endif 277.if ${PRIMES:M${n}} != "2" 278. error 279.endif 280.if ${PRIMES:M${:U2}} != "2" 281. error 282.endif 283 284 285# Before var.c 1.1031 from 2022-08-24, the following expressions caused an 286# out-of-bounds read beyond the indirect ':M' modifiers. 287.if ${:U:${:UM\\}} # The ':M' pattern need not be unescaped, the 288. error # resulting pattern is '\', it never matches 289.endif # anything. 290.if ${:U:${:UM\\\:\\}} # The ':M' pattern must be unescaped, the 291. error # resulting pattern is ':\', it never matches 292.endif # anything. 293