1#!./perl 2# 3# This is a home for regular expression tests that do not fit into 4# the format supported by re/regexp.t. If you want to add a test 5# that does fit that format, add it to re/re_tests, not here. 6 7BEGIN { 8 chdir 't' if -d 't'; 9 require './test.pl'; 10 set_up_inc(qw '../lib .'); 11 require './charset_tools.pl'; 12 skip_all_if_miniperl("miniperl can't load Tie::Hash::NamedCapture, need for %+ and %-"); 13} 14 15use strict; 16use warnings; 17use 5.010; 18our ($REGMARK, $REGERROR); 19 20sub run_tests; 21 22$| = 1; 23 24run_tests() unless caller; 25 26# 27# Tests start here. 28# 29sub run_tests { 30 31 { 32 # Japhy -- added 03/03/2001 33 () = (my $str = "abc") =~ /(...)/; 34 $str = "def"; 35 is($1, "abc", 'Changing subject does not modify $1'); 36 } 37 38 SKIP: 39 { 40 # The trick is that in EBCDIC the explicit numeric range should 41 # match (as also in non-EBCDIC) but the explicit alphabetic range 42 # should not match. 43 like "\x8e", qr/[\x89-\x91]/, '"\x8e" =~ /[\x89-\x91]/'; 44 like "\xce", qr/[\xc9-\xd1]/, '"\xce" =~ /[\xc9-\xd1]/'; 45 like "\xd0", qr/[\xc9-\xd1]/, '"\xd0" =~ /[\xc9-\xd1]/'; 46 47 skip "Not an EBCDIC platform", 2 unless ord ('i') == 0x89 && 48 ord ('J') == 0xd1; 49 50 # In most places these tests would succeed since \x8e does not 51 # in most character sets match 'i' or 'j' nor would \xce match 52 # 'I' or 'J', but strictly speaking these tests are here for 53 # the good of EBCDIC, so let's test these only there. 54 unlike("\x8e", qr/[i-j]/, '"\x8e" !~ /[i-j]/'); 55 unlike("\xce", qr/[I-J]/, '"\xce" !~ /[I-J]/'); 56 unlike("\xd0", qr/[I-J]/, '"\xd0" !~ /[I-J]/'); 57 } 58 59 { 60 like "\x{ab}", qr/\x{ab}/, '"\x{ab}" =~ /\x{ab}/ '; 61 like "\x{abcd}", qr/\x{abcd}/, '"\x{abcd}" =~ /\x{abcd}/'; 62 } 63 64 { 65 my $message = 'bug id 20001008.001 (#4407)'; 66 67 my $strasse = "stra" . uni_to_native("\337") . "e"; 68 my @x = ("$strasse 138", "$strasse 138"); 69 for (@x) { 70 ok(s/(\d+)\s*([\w\-]+)/$1 . uc $2/e, $message); 71 ok(my ($latin) = /^(.+)(?:\s+\d)/, $message); 72 is($latin, $strasse, $message); 73 ok($latin =~ s/$strasse/stra��e/, $message); 74 # 75 # Previous code follows, but outcommented - there were no tests. 76 # 77 # $latin =~ s/stra\337e/stra��e/; # \303\237 after the 2nd a 78 # use utf8; # needed for the raw UTF-8 79 # $latin =~ s!(s)tr(?:a��|s+e)!$1tr.!; # \303\237 after the a 80 } 81 } 82 83 { 84 my $message = 'Test \x escapes'; 85 ok("ba\xd4c" =~ /([a\xd4]+)/ && $1 eq "a\xd4", $message); 86 ok("ba\xd4c" =~ /([a\xd4]+)/ && $1 eq "a\x{d4}", $message); 87 ok("ba\x{d4}c" =~ /([a\xd4]+)/ && $1 eq "a\x{d4}", $message); 88 ok("ba\x{d4}c" =~ /([a\xd4]+)/ && $1 eq "a\xd4", $message); 89 ok("ba\xd4c" =~ /([a\x{d4}]+)/ && $1 eq "a\xd4", $message); 90 ok("ba\xd4c" =~ /([a\x{d4}]+)/ && $1 eq "a\x{d4}", $message); 91 ok("ba\x{d4}c" =~ /([a\x{d4}]+)/ && $1 eq "a\x{d4}", $message); 92 ok("ba\x{d4}c" =~ /([a\x{d4}]+)/ && $1 eq "a\xd4", $message); 93 } 94 95 { 96 my $message = 'Match code points > 255'; 97 $_ = "abc\x{100}\x{200}\x{300}\x{380}\x{400}defg"; 98 ok(/(.\x{300})./, $message); 99 ok($` eq "abc\x{100}" && length ($`) == 4, $message); 100 ok($& eq "\x{200}\x{300}\x{380}" && length ($&) == 3, $message); 101 ok($' eq "\x{400}defg" && length ($') == 5, $message); 102 ok($1 eq "\x{200}\x{300}" && length ($1) == 2, $message); 103 } 104 105 { 106 my $x = "\x{10FFFD}"; 107 $x =~ s/(.)/$1/g; 108 ok ord($x) == 0x10FFFD && length($x) == 1, "From Robin Houston"; 109 } 110 111 { 112 my %d = ( 113 "7f" => [0, 0, 0], 114 "80" => [1, 1, 0], 115 "ff" => [1, 1, 0], 116 "100" => [0, 1, 1], 117 ); 118 119 while (my ($code, $match) = each %d) { 120 my $message = "Properties of \\x$code"; 121 my $char = eval qq ["\\x{$code}"]; 122 123 is(0 + ($char =~ /[\x80-\xff]/), $$match[0], $message); 124 is(0 + ($char =~ /[\x80-\x{100}]/), $$match[1], $message); 125 is(0 + ($char =~ /[\x{100}]/), $$match[2], $message); 126 } 127 } 128 129 { 130 # From Japhy 131 foreach (qw(c g o)) { 132 warning_like(sub {'' =~ "(?$_)"}, qr/^Useless \(\?$_\)/); 133 warning_like(sub {'' =~ "(?-$_)"}, qr/^Useless \(\?-$_\)/); 134 } 135 136 # Now test multi-error regexes 137 foreach (['(?g-o)', qr/^Useless \(\?g\)/, qr/^Useless \(\?-o\)/], 138 ['(?g-c)', qr/^Useless \(\?g\)/, qr/^Useless \(\?-c\)/], 139 # (?c) means (?g) error won't be thrown 140 ['(?o-cg)', qr/^Useless \(\?o\)/, qr/^Useless \(\?-c\)/], 141 ['(?ogc)', qr/^Useless \(\?o\)/, qr/^Useless \(\?g\)/, 142 qr/^Useless \(\?c\)/], 143 ) { 144 my ($re, @warnings) = @$_; 145 warnings_like(sub {eval "qr/$re/"}, \@warnings, "qr/$re/ warns"); 146 } 147 } 148 149 { 150 my $message = "/x tests"; 151 $_ = "foo"; 152 foreach my $pat (<<" --", <<" --") { 153 /f 154 o\r 155 o 156 \$ 157 /x 158 -- 159 /f 160 o 161 o 162 \$\r 163 /x 164 -- 165 is(eval $pat, 1, $message); 166 is($@, '', $message); 167 } 168 } 169 170 { 171 my $message = "/o feature"; 172 sub test_o {$_ [0] =~ /$_[1]/o; return $1} 173 is(test_o ('abc', '(.)..'), 'a', $message); 174 is(test_o ('abc', '..(.)'), 'a', $message); 175 } 176 177 { 178 # Test basic $^N usage outside of a regex 179 my $message = '$^N usage outside of a regex'; 180 my $x = "abcdef"; 181 ok(($x =~ /cde/ and !defined $^N), $message); 182 ok(($x =~ /(cde)/ and $^N eq "cde"), $message); 183 ok(($x =~ /(c)(d)(e)/ and $^N eq "e"), $message); 184 ok(($x =~ /(c(d)e)/ and $^N eq "cde"), $message); 185 ok(($x =~ /(foo)|(c(d)e)/ and $^N eq "cde"), $message); 186 ok(($x =~ /(c(d)e)|(foo)/ and $^N eq "cde"), $message); 187 ok(($x =~ /(c(d)e)|(abc)/ and $^N eq "abc"), $message); 188 ok(($x =~ /(c(d)e)|(abc)x/ and $^N eq "cde"), $message); 189 ok(($x =~ /(c(d)e)(abc)?/ and $^N eq "cde"), $message); 190 ok(($x =~ /(?:c(d)e)/ and $^N eq "d"), $message); 191 ok(($x =~ /(?:c(d)e)(?:f)/ and $^N eq "d"), $message); 192 ok(($x =~ /(?:([abc])|([def]))*/ and $^N eq "f"), $message); 193 ok(($x =~ /(?:([ace])|([bdf]))*/ and $^N eq "f"), $message); 194 ok(($x =~ /(([ace])|([bd]))*/ and $^N eq "e"), $message); 195 {ok(($x =~ /(([ace])|([bdf]))*/ and $^N eq "f"), $message);} 196 ## Test to see if $^N is automatically localized -- it should now 197 ## have the value set in the previous test. 198 is($^N, "e", '$^N is automatically localized'); 199 200 # Now test inside (?{ ... }) 201 $message = '$^N usage inside (?{ ... })'; 202 our ($y, $z); 203 ok(($x =~ /a([abc])(?{$y=$^N})c/ and $y eq "b"), $message); 204 ok(($x =~ /a([abc]+)(?{$y=$^N})d/ and $y eq "bc"), $message); 205 ok(($x =~ /a([abcdefg]+)(?{$y=$^N})d/ and $y eq "bc"), $message); 206 ok(($x =~ /(a([abcdefg]+)(?{$y=$^N})d)(?{$z=$^N})e/ and $y eq "bc" 207 and $z eq "abcd"), $message); 208 ok(($x =~ /(a([abcdefg]+)(?{$y=$^N})de)(?{$z=$^N})/ and $y eq "bc" 209 and $z eq "abcde"), $message); 210 211 } 212 213 SKIP: 214 { 215 ## Should probably put in tests for all the POSIX stuff, 216 ## but not sure how to guarantee a specific locale...... 217 218 my $message = 'Test [[:cntrl:]]'; 219 my $AllBytes = join "" => map {chr} 0 .. 255; 220 (my $x = $AllBytes) =~ s/[[:cntrl:]]//g; 221 $x = join "", sort { $a cmp $b } 222 map { chr utf8::native_to_unicode(ord $_) } split "", $x; 223 is($x, join("", map {chr} 0x20 .. 0x7E, 0x80 .. 0xFF), $message); 224 225 ($x = $AllBytes) =~ s/[^[:cntrl:]]//g; 226 $x = join "", sort { $a cmp $b } 227 map { chr utf8::native_to_unicode(ord $_) } split "", $x; 228 is($x, (join "", map {chr} 0x00 .. 0x1F, 0x7F), $message); 229 } 230 231 { 232 # With /s modifier UTF8 chars were interpreted as bytes 233 my $message = "UTF-8 chars aren't bytes"; 234 my $a = "Hello \x{263A} World"; 235 my @a = ($a =~ /./gs); 236 is($#a, 12, $message); 237 } 238 239 { 240 no warnings 'digit'; 241 # Check that \x## works. 5.6.1 and 5.005_03 fail some of these. 242 my $x; 243 $x = "\x4e" . "E"; 244 like ($x, qr/^\x4EE$/, "Check only 2 bytes of hex are matched."); 245 246 $x = "\x4e" . "i"; 247 like ($x, qr/^\x4Ei$/, "Check that invalid hex digit stops it (2)"); 248 249 $x = "\x4" . "j"; 250 like ($x, qr/^\x4j$/, "Check that invalid hex digit stops it (1)"); 251 252 $x = "\x0" . "k"; 253 like ($x, qr/^\xk$/, "Check that invalid hex digit stops it (0)"); 254 255 $x = "\x0" . "x"; 256 like ($x, qr/^\xx$/, "\\xx isn't to be treated as \\0"); 257 258 $x = "\x0" . "xa"; 259 like ($x, qr/^\xxa$/, "\\xxa isn't to be treated as \\xa"); 260 261 $x = "\x9" . "_b"; 262 like ($x, qr/^\x9_b$/, "\\x9_b isn't to be treated as \\x9b"); 263 264 # and now again in [] ranges 265 266 $x = "\x4e" . "E"; 267 like ($x, qr/^[\x4EE]{2}$/, "Check only 2 bytes of hex are matched."); 268 269 $x = "\x4e" . "i"; 270 like ($x, qr/^[\x4Ei]{2}$/, "Check that invalid hex digit stops it (2)"); 271 272 $x = "\x4" . "j"; 273 like ($x, qr/^[\x4j]{2}$/, "Check that invalid hex digit stops it (1)"); 274 275 $x = "\x0" . "k"; 276 like ($x, qr/^[\xk]{2}$/, "Check that invalid hex digit stops it (0)"); 277 278 $x = "\x0" . "x"; 279 like ($x, qr/^[\xx]{2}$/, "\\xx isn't to be treated as \\0"); 280 281 $x = "\x0" . "xa"; 282 like ($x, qr/^[\xxa]{3}$/, "\\xxa isn't to be treated as \\xa"); 283 284 $x = "\x9" . "_b"; 285 like ($x, qr/^[\x9_b]{3}$/, "\\x9_b isn't to be treated as \\x9b"); 286 287 # Check that \x{##} works. 5.6.1 fails quite a few of these. 288 289 $x = "\x9b"; 290 like ($x, qr/^\x{9_b}$/, "\\x{9_b} is to be treated as \\x9b"); 291 292 $x = "\x9b" . "y"; 293 like ($x, qr/^\x{9_b}y$/, "\\x{9_b} is to be treated as \\x9b (again)"); 294 295 $x = "\x9b" . "y"; 296 like ($x, qr/^\x{9b_}y$/, "\\x{9b_} is to be treated as \\x9b"); 297 298 $x = "\x9b" . "y"; 299 like ($x, qr/^\x{9_bq}y$/, "\\x{9_bc} is to be treated as \\x9b"); 300 301 $x = "\x0" . "y"; 302 like ($x, qr/^\x{x9b}y$/, "\\x{x9b} is to be treated as \\x0"); 303 304 $x = "\x0" . "y"; 305 like ($x, qr/^\x{0x9b}y$/, "\\x{0x9b} is to be treated as \\x0"); 306 307 $x = "\x9b" . "y"; 308 like ($x, qr/^\x{09b}y$/, "\\x{09b} is to be treated as \\x9b"); 309 310 $x = "\x9b"; 311 like ($x, qr/^[\x{9_b}]$/, "\\x{9_b} is to be treated as \\x9b"); 312 313 $x = "\x9b" . "y"; 314 like ($x, qr/^[\x{9_b}y]{2}$/, 315 "\\x{9_b} is to be treated as \\x9b (again)"); 316 317 $x = "\x9b" . "y"; 318 like ($x, qr/^[\x{9b_}y]{2}$/, "\\x{9b_} is to be treated as \\x9b"); 319 320 $x = "\x9b" . "y"; 321 like ($x, qr/^[\x{9_bq}y]{2}$/, "\\x{9_bc} is to be treated as \\x9b"); 322 323 $x = "\x0" . "y"; 324 like ($x, qr/^[\x{x9b}y]{2}$/, "\\x{x9b} is to be treated as \\x0"); 325 326 $x = "\x0" . "y"; 327 like ($x, qr/^[\x{0x9b}y]{2}$/, "\\x{0x9b} is to be treated as \\x0"); 328 329 $x = "\x9b" . "y"; 330 like ($x, qr/^[\x{09b}y]{2}$/, "\\x{09b} is to be treated as \\x9b"); 331 332 } 333 334 { 335 # High bit bug -- japhy 336 my $x = "ab\200d"; 337 like $x, qr/.*?\200/, "High bit fine"; 338 } 339 340 { 341 # The basic character classes and Unicode 342 like "\x{0100}", qr/\w/, 'LATIN CAPITAL LETTER A WITH MACRON in /\w/'; 343 like "\x{0660}", qr/\d/, 'ARABIC-INDIC DIGIT ZERO in /\d/'; 344 like "\x{1680}", qr/\s/, 'OGHAM SPACE MARK in /\s/'; 345 } 346 347 { 348 my $message = "Folding matches and Unicode"; 349 like("a\x{100}", qr/A/i, $message); 350 like("A\x{100}", qr/a/i, $message); 351 like("a\x{100}", qr/a/i, $message); 352 like("A\x{100}", qr/A/i, $message); 353 like("\x{101}a", qr/\x{100}/i, $message); 354 like("\x{100}a", qr/\x{100}/i, $message); 355 like("\x{101}a", qr/\x{101}/i, $message); 356 like("\x{100}a", qr/\x{101}/i, $message); 357 like("a\x{100}", qr/A\x{100}/i, $message); 358 like("A\x{100}", qr/a\x{100}/i, $message); 359 like("a\x{100}", qr/a\x{100}/i, $message); 360 like("A\x{100}", qr/A\x{100}/i, $message); 361 like("a\x{100}", qr/[A]/i, $message); 362 like("A\x{100}", qr/[a]/i, $message); 363 like("a\x{100}", qr/[a]/i, $message); 364 like("A\x{100}", qr/[A]/i, $message); 365 like("\x{101}a", qr/[\x{100}]/i, $message); 366 like("\x{100}a", qr/[\x{100}]/i, $message); 367 like("\x{101}a", qr/[\x{101}]/i, $message); 368 like("\x{100}a", qr/[\x{101}]/i, $message); 369 } 370 371 { 372 use charnames ':full'; 373 my $message = "Folding 'LATIN LETTER A WITH GRAVE'"; 374 375 my $lower = "\N{LATIN SMALL LETTER A WITH GRAVE}"; 376 my $UPPER = "\N{LATIN CAPITAL LETTER A WITH GRAVE}"; 377 378 like($lower, qr/$UPPER/i, $message); 379 like($UPPER, qr/$lower/i, $message); 380 like($lower, qr/[$UPPER]/i, $message); 381 like($UPPER, qr/[$lower]/i, $message); 382 383 $message = "Folding 'GREEK LETTER ALPHA WITH VRACHY'"; 384 385 $lower = "\N{GREEK CAPITAL LETTER ALPHA WITH VRACHY}"; 386 $UPPER = "\N{GREEK SMALL LETTER ALPHA WITH VRACHY}"; 387 388 like($lower, qr/$UPPER/i, $message); 389 like($UPPER, qr/$lower/i, $message); 390 like($lower, qr/[$UPPER]/i, $message); 391 like($UPPER, qr/[$lower]/i, $message); 392 393 $message = "Folding 'LATIN LETTER Y WITH DIAERESIS'"; 394 395 $lower = "\N{LATIN SMALL LETTER Y WITH DIAERESIS}"; 396 $UPPER = "\N{LATIN CAPITAL LETTER Y WITH DIAERESIS}"; 397 398 like($lower, qr/$UPPER/i, $message); 399 like($UPPER, qr/$lower/i, $message); 400 like($lower, qr/[$UPPER]/i, $message); 401 like($UPPER, qr/[$lower]/i, $message); 402 } 403 404 { 405 use charnames ':full'; 406 my $message = "GREEK CAPITAL LETTER SIGMA vs " . 407 "COMBINING GREEK PERISPOMENI"; 408 409 my $SIGMA = "\N{GREEK CAPITAL LETTER SIGMA}"; 410 my $char = "\N{COMBINING GREEK PERISPOMENI}"; 411 412 warning_is(sub {unlike("_:$char:_", qr/_:$SIGMA:_/i, $message)}, undef, 413 'Did not warn [change a5961de5f4215b5c]'); 414 } 415 416 { 417 my $message = '\X'; 418 use charnames ':full'; 419 420 ok("a!" =~ /^(\X)!/ && $1 eq "a", $message); 421 ok("\xDF!" =~ /^(\X)!/ && $1 eq "\xDF", $message); 422 ok("\x{100}!" =~ /^(\X)!/ && $1 eq "\x{100}", $message); 423 ok("\x{100}\x{300}!" =~ /^(\X)!/ && $1 eq "\x{100}\x{300}", $message); 424 ok("\N{LATIN CAPITAL LETTER E}!" =~ /^(\X)!/ && 425 $1 eq "\N{LATIN CAPITAL LETTER E}", $message); 426 ok("\N{LATIN CAPITAL LETTER E}\N{COMBINING GRAVE ACCENT}!" 427 =~ /^(\X)!/ && 428 $1 eq "\N{LATIN CAPITAL LETTER E}\N{COMBINING GRAVE ACCENT}", $message); 429 430 } 431 432 { 433 my $message = "Final Sigma"; 434 435 my $SIGMA = "\x{03A3}"; # CAPITAL 436 my $Sigma = "\x{03C2}"; # SMALL FINAL 437 my $sigma = "\x{03C3}"; # SMALL 438 439 like($SIGMA, qr/$SIGMA/i, $message); 440 like($SIGMA, qr/$Sigma/i, $message); 441 like($SIGMA, qr/$sigma/i, $message); 442 443 like($Sigma, qr/$SIGMA/i, $message); 444 like($Sigma, qr/$Sigma/i, $message); 445 like($Sigma, qr/$sigma/i, $message); 446 447 like($sigma, qr/$SIGMA/i, $message); 448 like($sigma, qr/$Sigma/i, $message); 449 like($sigma, qr/$sigma/i, $message); 450 451 like($SIGMA, qr/[$SIGMA]/i, $message); 452 like($SIGMA, qr/[$Sigma]/i, $message); 453 like($SIGMA, qr/[$sigma]/i, $message); 454 455 like($Sigma, qr/[$SIGMA]/i, $message); 456 like($Sigma, qr/[$Sigma]/i, $message); 457 like($Sigma, qr/[$sigma]/i, $message); 458 459 like($sigma, qr/[$SIGMA]/i, $message); 460 like($sigma, qr/[$Sigma]/i, $message); 461 like($sigma, qr/[$sigma]/i, $message); 462 463 $message = "More final Sigma"; 464 465 my $S3 = "$SIGMA$Sigma$sigma"; 466 467 ok(":$S3:" =~ /:(($SIGMA)+):/i && $1 eq $S3 && $2 eq $sigma, $message); 468 ok(":$S3:" =~ /:(($Sigma)+):/i && $1 eq $S3 && $2 eq $sigma, $message); 469 ok(":$S3:" =~ /:(($sigma)+):/i && $1 eq $S3 && $2 eq $sigma, $message); 470 471 ok(":$S3:" =~ /:(([$SIGMA])+):/i && $1 eq $S3 && $2 eq $sigma, $message); 472 ok(":$S3:" =~ /:(([$Sigma])+):/i && $1 eq $S3 && $2 eq $sigma, $message); 473 ok(":$S3:" =~ /:(([$sigma])+):/i && $1 eq $S3 && $2 eq $sigma, $message); 474 } 475 476 { 477 use charnames ':full'; 478 my $message = "Parlez-Vous " . 479 "Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais?"; 480 481 ok("Fran\N{LATIN SMALL LETTER C}ais" =~ /Fran.ais/ && 482 $& eq "Francais", $message); 483 ok("Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais" =~ /Fran.ais/ && 484 $& eq "Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais", $message); 485 ok("Fran\N{LATIN SMALL LETTER C}ais" =~ /Fran\Xais/ && 486 $& eq "Francais", $message); 487 ok("Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais" =~ /Fran\Xais/ && 488 $& eq "Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais", $message); 489 ok("Franc\N{COMBINING CEDILLA}ais" =~ /Fran\Xais/ && 490 $& eq "Franc\N{COMBINING CEDILLA}ais", $message); 491 ok("Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais" =~ 492 /Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais/ && 493 $& eq "Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais", $message); 494 ok("Franc\N{COMBINING CEDILLA}ais" =~ /Franc\N{COMBINING CEDILLA}ais/ && 495 $& eq "Franc\N{COMBINING CEDILLA}ais", $message); 496 497 my @f = ( 498 ["Fran\N{LATIN SMALL LETTER C}ais", "Francais"], 499 ["Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais", 500 "Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais"], 501 ["Franc\N{COMBINING CEDILLA}ais", "Franc\N{COMBINING CEDILLA}ais"], 502 ); 503 foreach my $entry (@f) { 504 my ($subject, $match) = @$entry; 505 ok($subject =~ /Fran(?:c\N{COMBINING CEDILLA}?| 506 \N{LATIN SMALL LETTER C WITH CEDILLA})ais/x && 507 $& eq $match, $message); 508 } 509 } 510 511 { 512 my $message = "Lingering (and useless) UTF8 flag doesn't mess up /i"; 513 my $pat = "ABcde"; 514 my $str = "abcDE\x{100}"; 515 chop $str; 516 like($str, qr/$pat/i, $message); 517 518 $pat = "ABcde\x{100}"; 519 $str = "abcDE"; 520 chop $pat; 521 like($str, qr/$pat/i, $message); 522 523 $pat = "ABcde\x{100}"; 524 $str = "abcDE\x{100}"; 525 chop $pat; 526 chop $str; 527 like($str, qr/$pat/i, $message); 528 } 529 530 { 531 use charnames ':full'; 532 my $message = "LATIN SMALL LETTER SHARP S " . 533 "(\N{LATIN SMALL LETTER SHARP S})"; 534 535 like("\N{LATIN SMALL LETTER SHARP S}", 536 qr/\N{LATIN SMALL LETTER SHARP S}/, $message); 537 like("\N{LATIN SMALL LETTER SHARP S}", 538 qr'\N{LATIN SMALL LETTER SHARP S}', $message); 539 like("\N{LATIN SMALL LETTER SHARP S}", 540 qr/\N{LATIN SMALL LETTER SHARP S}/i, $message); 541 like("\N{LATIN SMALL LETTER SHARP S}", 542 qr'\N{LATIN SMALL LETTER SHARP S}'i, $message); 543 like("\N{LATIN SMALL LETTER SHARP S}", 544 qr/[\N{LATIN SMALL LETTER SHARP S}]/, $message); 545 like("\N{LATIN SMALL LETTER SHARP S}", 546 qr'[\N{LATIN SMALL LETTER SHARP S}]', $message); 547 like("\N{LATIN SMALL LETTER SHARP S}", 548 qr/[\N{LATIN SMALL LETTER SHARP S}]/i, $message); 549 like("\N{LATIN SMALL LETTER SHARP S}", 550 qr'[\N{LATIN SMALL LETTER SHARP S}]'i, $message); 551 552 like("ss", qr /\N{LATIN SMALL LETTER SHARP S}/i, $message); 553 like("ss", qr '\N{LATIN SMALL LETTER SHARP S}'i, $message); 554 like("SS", qr /\N{LATIN SMALL LETTER SHARP S}/i, $message); 555 like("SS", qr '\N{LATIN SMALL LETTER SHARP S}'i, $message); 556 like("ss", qr/[\N{LATIN SMALL LETTER SHARP S}]/i, $message); 557 like("ss", qr'[\N{LATIN SMALL LETTER SHARP S}]'i, $message); 558 like("SS", qr/[\N{LATIN SMALL LETTER SHARP S}]/i, $message); 559 like("SS", qr'[\N{LATIN SMALL LETTER SHARP S}]'i, $message); 560 561 like("\N{LATIN SMALL LETTER SHARP S}", qr/ss/i, $message); 562 like("\N{LATIN SMALL LETTER SHARP S}", qr/SS/i, $message); 563 564 $message = "Unoptimized named sequence in class"; 565 like("ss", qr/[\N{LATIN SMALL LETTER SHARP S}x]/i, $message); 566 like("ss", qr'[\N{LATIN SMALL LETTER SHARP S}x]'i, $message); 567 like("SS", qr/[\N{LATIN SMALL LETTER SHARP S}x]/i, $message); 568 like("SS", qr'[\N{LATIN SMALL LETTER SHARP S}x]'i, $message); 569 like("\N{LATIN SMALL LETTER SHARP S}", 570 qr/[\N{LATIN SMALL LETTER SHARP S}x]/, $message); 571 like("\N{LATIN SMALL LETTER SHARP S}", 572 qr'[\N{LATIN SMALL LETTER SHARP S}x]', $message); 573 like("\N{LATIN SMALL LETTER SHARP S}", 574 qr/[\N{LATIN SMALL LETTER SHARP S}x]/i, $message); 575 like("\N{LATIN SMALL LETTER SHARP S}", 576 qr'[\N{LATIN SMALL LETTER SHARP S}x]'i, $message); 577 } 578 579 { 580 # More whitespace: U+0085, U+2028, U+2029\n"; 581 582 # U+0085, U+00A0 need to be forced to be Unicode, the \x{100} does that. 583 like "<\x{100}" . uni_to_native("\x{0085}") . ">", qr/<\x{100}\s>/, '\x{0085} in \s'; 584 like "<" . uni_to_native("\x{0085}") . ">", qr/<\v>/, '\x{0085} in \v'; 585 like "<\x{100}" . uni_to_native("\x{00A0}") . ">", qr/<\x{100}\s>/, '\x{00A0} in \s'; 586 like "<" . uni_to_native("\x{00A0}") . ">", qr/<\h>/, '\x{00A0} in \h'; 587 my @h = map {sprintf "%05x" => $_} 0x01680, 0x02000 .. 0x0200A, 588 0x0202F, 0x0205F, 0x03000; 589 my @v = map {sprintf "%05x" => $_} 0x02028, 0x02029; 590 591 my @H = map {sprintf "%05x" => $_} 0x01361, 0x0200B, 0x02408, 0x02420, 592 0x0303F, 0xE0020, 0x180E; 593 my @V = map {sprintf "%05x" => $_} 0x0008A .. 0x0008D, 0x00348, 0x10100, 594 0xE005F, 0xE007C, 0x180E; 595 596 for my $hex (@h) { 597 my $str = eval qq ["<\\x{$hex}>"]; 598 like $str, qr/<\s>/, "\\x{$hex} in \\s"; 599 like $str, qr/<\h>/, "\\x{$hex} in \\h"; 600 unlike $str, qr/<\v>/, "\\x{$hex} not in \\v"; 601 } 602 603 for my $hex (@v) { 604 my $str = eval qq ["<\\x{$hex}>"]; 605 like $str, qr/<\s>/, "\\x{$hex} in \\s"; 606 like $str, qr/<\v>/, "\\x{$hex} in \\v"; 607 unlike $str, qr/<\h>/, "\\x{$hex} not in \\h"; 608 } 609 610 for my $hex (@H) { 611 my $str = eval qq ["<\\x{$hex}>"]; 612 like $str, qr/<\S>/, "\\x{$hex} in \\S"; 613 like $str, qr/<\H>/, "\\x{$hex} in \\H"; 614 } 615 616 for my $hex (@V) { 617 my $str = eval qq ["<\\x{$hex}>"]; 618 like $str, qr/<\S>/, "\\x{$hex} in \\S"; 619 like $str, qr/<\V>/, "\\x{$hex} in \\V"; 620 } 621 } 622 623 { 624 # . with /s should work on characters, as opposed to bytes 625 my $message = ". with /s works on characters, not bytes"; 626 627 my $s = "\x{e4}\x{100}"; 628 # This is not expected to match: the point is that 629 # neither should we get "Malformed UTF-8" warnings. 630 warning_is(sub {$s =~ /\G(.+?)\n/gcs}, undef, 631 "No 'Malformed UTF-8' warning"); 632 633 my @c; 634 push @c => $1 while $s =~ /\G(.)/gs; 635 636 local $" = ""; 637 is("@c", $s, $message); 638 639 # Test only chars < 256 640 my $t1 = "Q003\n\n\x{e4}\x{f6}\n\nQ004\n\n\x{e7}"; 641 my $r1 = ""; 642 while ($t1 =~ / \G ( .+? ) \n\s+ ( .+? ) ( $ | \n\s+ ) /xgcs) { 643 $r1 .= $1 . $2; 644 } 645 646 my $t2 = $t1 . "\x{100}"; # Repeat with a larger char 647 my $r2 = ""; 648 while ($t2 =~ / \G ( .+? ) \n\s+ ( .+? ) ( $ | \n\s+ ) /xgcs) { 649 $r2 .= $1 . $2; 650 } 651 $r2 =~ s/\x{100}//; 652 653 is($r1, $r2, $message); 654 } 655 656 { 657 my $message = "Unicode lookbehind"; 658 like("A\x{100}B", qr/(?<=A.)B/, $message); 659 like("A\x{200}\x{300}B", qr/(?<=A..)B/, $message); 660 like("\x{400}AB", qr/(?<=\x{400}.)B/, $message); 661 like("\x{500}\x{600}B", qr/(?<=\x{500}.)B/, $message); 662 663 # Original code also contained: 664 # ok "\x{500\x{600}}B" =~ /(?<=\x{500}.)B/; 665 # but that looks like a typo. 666 } 667 668 { 669 my $message = 'UTF-8 hash keys and /$/'; 670 # http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters 671 # /2002-01/msg01327.html 672 673 my $u = "a\x{100}"; 674 my $v = substr ($u, 0, 1); 675 my $w = substr ($u, 1, 1); 676 my %u = ($u => $u, $v => $v, $w => $w); 677 for (keys %u) { 678 my $m1 = /^\w*$/ ? 1 : 0; 679 my $m2 = $u {$_} =~ /^\w*$/ ? 1 : 0; 680 is($m1, $m2, $message); 681 } 682 } 683 684 { 685 my $message = "No SEGV in s/// and UTF-8"; 686 my $s = "s#\x{100}" x 4; 687 ok($s =~ s/[^\w]/ /g, $message); 688 if ( 1 or $ENV{PERL_TEST_LEGACY_POSIX_CC} ) { 689 is($s, "s \x{100}" x 4, $message); 690 } 691 else { 692 is($s, "s " x 4, $message); 693 } 694 } 695 696 { 697 my $message = "UTF-8 bug (maybe already known?)"; 698 my $u = "foo"; 699 $u =~ s/./\x{100}/g; 700 is($u, "\x{100}\x{100}\x{100}", $message); 701 702 $u = "foobar"; 703 $u =~ s/[ao]/\x{100}/g; 704 is($u, "f\x{100}\x{100}b\x{100}r", $message); 705 706 $u =~ s/\x{100}/e/g; 707 is($u, "feeber", $message); 708 } 709 710 { 711 my $message = "UTF-8 bug with s///"; 712 # check utf8/non-utf8 mixtures 713 # try to force all float/anchored check combinations 714 715 my $c = "\x{100}"; 716 my $subst; 717 for my $re ("xx.*$c", "x.*$c$c", "$c.*xx", "$c$c.*x", 718 "xx.*(?=$c)", "(?=$c).*xx",) { 719 unlike("xxx", qr/$re/, $message); 720 ok(+($subst = "xxx") !~ s/$re//, $message); 721 } 722 for my $re ("xx.*$c*", "$c*.*xx") { 723 like("xxx", qr/$re/, $message); 724 ok(+($subst = "xxx") =~ s/$re//, $message); 725 is($subst, "", $message); 726 } 727 for my $re ("xxy*", "y*xx") { 728 like("xx$c", qr/$re/, $message); 729 ok(+($subst = "xx$c") =~ s/$re//, $message); 730 is($subst, $c, $message); 731 unlike("xy$c", qr/$re/, $message); 732 ok(+($subst = "xy$c") !~ s/$re//, $message); 733 } 734 for my $re ("xy$c*z", "x$c*yz") { 735 like("xyz", qr/$re/, $message); 736 ok(+($subst = "xyz") =~ s/$re//, $message); 737 is($subst, "", $message); 738 } 739 } 740 741 { 742 # The second half of RT #114808 743 warning_is(sub {'aa' =~ /.+\x{100}/}, undef, 744 'utf8-only floating substr, non-utf8 target, no warning'); 745 } 746 747 { 748 my $message = "qr /.../x"; 749 my $R = qr / A B C # D E/x; 750 ok("ABCDE" =~ $R && $& eq "ABC", $message); 751 ok("ABCDE" =~ /$R/ && $& eq "ABC", $message); 752 ok("ABCDE" =~ m/$R/ && $& eq "ABC", $message); 753 ok("ABCDE" =~ /($R)/ && $1 eq "ABC", $message); 754 ok("ABCDE" =~ m/($R)/ && $1 eq "ABC", $message); 755 } 756 757 { 758 local $\; 759 $_ = 'aaaaaaaaaa'; 760 utf8::upgrade($_); chop $_; $\="\n"; 761 ok /[^\s]+/, 'm/[^\s]/ utf8'; 762 ok /[^\d]+/, 'm/[^\d]/ utf8'; 763 ok +($a = $_, $_ =~ s/[^\s]+/./g), 's/[^\s]/ utf8'; 764 ok +($a = $_, $a =~ s/[^\d]+/./g), 's/[^\s]/ utf8'; 765 } 766 767 { 768 # Subject: Odd regexp behavior 769 # From: Markus Kuhn <Markus.Kuhn@cl.cam.ac.uk> 770 # Date: Wed, 26 Feb 2003 16:53:12 +0000 771 # Message-Id: <E18o4nw-0008Ly-00@wisbech.cl.cam.ac.uk> 772 # To: perl-unicode@perl.org 773 774 my $message = 'Markus Kuhn 2003-02-26'; 775 776 my $x = "\x{2019}\nk"; 777 ok($x =~ s/(\S)\n(\S)/$1 $2/sg, $message); 778 is($x, "\x{2019} k", $message); 779 780 $x = "b\nk"; 781 ok($x =~ s/(\S)\n(\S)/$1 $2/sg, $message); 782 is($x, "b k", $message); 783 784 like("\x{2019}", qr/\S/, $message); 785 } 786 787 { 788 like "\x{100}\n", qr/\x{100}\n$/, "UTF-8 length cache and fbm_compile"; 789 } 790 791 { 792 package Str; 793 use overload q /""/ => sub {${$_ [0]};}; 794 sub new {my ($c, $v) = @_; bless \$v, $c;} 795 796 package main; 797 $_ = Str -> new ("a\x{100}/\x{100}b"); 798 ok join (":", /\b(.)\x{100}/g) eq "a:/", "re_intuit_start and PL_bostr"; 799 } 800 801 { 802 my $re = qq /^([^X]*)X/; 803 utf8::upgrade ($re); 804 like "\x{100}X", qr/$re/, "S_cl_and ANYOF_UNICODE & ANYOF_INVERTED"; 805 my $loc_re = qq /(?l:^([^X]*)X)/; 806 utf8::upgrade ($loc_re); 807 no warnings 'locale'; 808 like "\x{100}X", qr/$loc_re/, "locale, S_cl_and ANYOF_UNICODE & ANYOF_INVERTED"; 809 } 810 811 { 812 like "123\x{100}", qr/^.*1.*23\x{100}$/, 813 'UTF-8 + multiple floating substr'; 814 } 815 816 { 817 my $message = '<20030808193656.5109.1@llama.ni-s.u-net.com>'; 818 819 # LATIN SMALL/CAPITAL LETTER A WITH MACRON 820 like(" \x{101}", qr/\x{100}/i, $message); 821 822 # LATIN SMALL/CAPITAL LETTER A WITH RING BELOW 823 like(" \x{1E01}", qr/\x{1E00}/i, $message); 824 825 # DESERET SMALL/CAPITAL LETTER LONG I 826 like(" \x{10428}", qr/\x{10400}/i, $message); 827 828 # LATIN SMALL/CAPITAL LETTER A WITH RING BELOW + 'X' 829 like(" \x{1E01}x", qr/\x{1E00}X/i, $message); 830 } 831 832 { 833 for (120 .. 130, 240 .. 260) { 834 my $head = 'x' x $_; 835 my $message = q [Don't misparse \x{...} in regexp ] . 836 q [near EXACT char count limit]; 837 for my $tail ('\x{0061}', '\x{1234}', '\x61') { 838 eval qq{like("$head$tail", qr/$head$tail/, \$message)}; 839 is($@, '', $message); 840 } 841 $message = q [Don't misparse \N{...} in regexp ] . 842 q [near EXACT char count limit]; 843 for my $tail ('\N{SNOWFLAKE}') { 844 eval qq {use charnames ':full'; 845 like("$head$tail", qr/$head$tail/, \$message)}; 846 eval qq {use charnames ':full'; 847 like("$head$tail", qr'$head$tail', \$message)}; 848 is($@, '', $message); 849 } 850 } 851 } 852 853 { # TRIE related 854 our @got = (); 855 "words" =~ /(word|word|word)(?{push @got, $1})s$/; 856 is(@got, 1, "TRIE optimisation"); 857 858 @got = (); 859 "words" =~ /(word|word|word)(?{push @got,$1})s$/i; 860 is(@got, 1,"TRIEF optimisation"); 861 862 my @nums = map {int rand 1000} 1 .. 100; 863 my $re = "(" . (join "|", @nums) . ")"; 864 $re = qr/\b$re\b/; 865 866 foreach (@nums) { 867 like $_, qr/$re/, "Trie nums"; 868 } 869 870 $_ = join " ", @nums; 871 @got = (); 872 push @got, $1 while /$re/g; 873 874 my %count; 875 $count {$_} ++ for @got; 876 my $ok = 1; 877 for (@nums) { 878 $ok = 0 if --$count {$_} < 0; 879 } 880 ok $ok, "Trie min count matches"; 881 } 882 883 { 884 # TRIE related 885 # LATIN SMALL/CAPITAL LETTER A WITH MACRON 886 ok "foba \x{101}foo" =~ qr/(foo|\x{100}foo|bar)/i && 887 $1 eq "\x{101}foo", 888 "TRIEF + LATIN SMALL/CAPITAL LETTER A WITH MACRON"; 889 890 # LATIN SMALL/CAPITAL LETTER A WITH RING BELOW 891 ok "foba \x{1E01}foo" =~ qr/(foo|\x{1E00}foo|bar)/i && 892 $1 eq "\x{1E01}foo", 893 "TRIEF + LATIN SMALL/CAPITAL LETTER A WITH RING BELOW"; 894 895 # DESERET SMALL/CAPITAL LETTER LONG I 896 ok "foba \x{10428}foo" =~ qr/(foo|\x{10400}foo|bar)/i && 897 $1 eq "\x{10428}foo", 898 "TRIEF + DESERET SMALL/CAPITAL LETTER LONG I"; 899 900 # LATIN SMALL/CAPITAL LETTER A WITH RING BELOW + 'X' 901 ok "foba \x{1E01}xfoo" =~ qr/(foo|\x{1E00}Xfoo|bar)/i && 902 $1 eq "\x{1E01}xfoo", 903 "TRIEF + LATIN SMALL/CAPITAL LETTER A WITH RING BELOW + 'X'"; 904 905 use charnames ':full'; 906 907 my $s = "\N{LATIN SMALL LETTER SHARP S}"; 908 ok "foba ba$s" =~ qr/(foo|Ba$s|bar)/i && $1 eq "ba$s", 909 "TRIEF + LATIN SMALL LETTER SHARP S =~ ss"; 910 ok "foba ba$s" =~ qr/(Ba$s|foo|bar)/i && $1 eq "ba$s", 911 "TRIEF + LATIN SMALL LETTER SHARP S =~ ss"; 912 ok "foba ba$s" =~ qr/(foo|bar|Ba$s)/i && $1 eq "ba$s", 913 "TRIEF + LATIN SMALL LETTER SHARP S =~ ss"; 914 915 ok "foba ba$s" =~ qr/(foo|Bass|bar)/i && $1 eq "ba$s", 916 "TRIEF + LATIN SMALL LETTER SHARP S =~ ss"; 917 918 ok "foba ba$s" =~ qr/(foo|BaSS|bar)/i && $1 eq "ba$s", 919 "TRIEF + LATIN SMALL LETTER SHARP S =~ SS"; 920 921 ok "foba ba${s}pxySS$s$s" =~ qr/(b(?:a${s}t|a${s}f|a${s}p)[xy]+$s*)/i 922 && $1 eq "ba${s}pxySS$s$s", 923 "COMMON PREFIX TRIEF + LATIN SMALL LETTER SHARP S"; 924 } 925 926 { 927 BEGIN { 928 unshift @INC, 'lib'; 929 } 930 use Cname; # Our custom charname plugin, currently found in 931 # t/lib/Cname.pm 932 933 like 'fooB', qr/\N{foo}[\N{B}\N{b}]/, "Passthrough charname"; 934 my $name = "foo\xDF"; 935 my $result = eval "'A${name}B' =~ /^A\\N{$name}B\$/"; 936 ok !$@ && $result, "Passthrough charname of non-ASCII, Latin1"; 937 eval "qr/\\p{name=foo}/"; 938 like($@, qr/Can't find Unicode property definition "name=foo"/, 939 '\p{name=} doesn\'t see a cumstom charnames translator'); 940 # 941 # Why doesn't must_warn work here? 942 # 943 my $w; 944 local $SIG {__WARN__} = sub {$w .= "@_"}; 945 $result = eval 'q(WARN) =~ /^[\N{WARN}]$/'; 946 ok !$@ && $result && ! $w, '\N{} returning multi-char works'; 947 948 undef $w; 949 eval q [unlike "\0", qr/[\N{EMPTY-STR}XY]/, 950 "Zerolength charname in charclass doesn't match \\\\0"]; 951 ok $w && $w =~ /Ignoring zero length/, 952 'Ignoring zero length \N{} in character class warning'; 953 undef $w; 954 eval q [like 'xy', qr/x[\N{EMPTY-STR} y]/x, 955 'Empty string charname in [] is ignored; finds a following character']; 956 ok $w && $w =~ /Ignoring zero length/, 957 'Ignoring zero length \N{} in character class warning'; 958 undef $w; 959 eval q [like 'x ', qr/x[\N{EMPTY-STR} y]/, 960 'Empty string charname in [] is ignored; finds a following blank under /x']; 961 like $w, qr/Ignoring zero length/, 962 'Ignoring zero length \N{} in character class warning'; 963 964 # EVIL keeps track of its calls, and appends a new character each 965 # time: A AB ABC ABCD ... 966 ok 'AB' =~ /(\N{EVIL})/ && $1 eq 'A', 'Charname caching $1'; 967 like 'ABC', qr/(\N{EVIL})/, 'Charname caching $1'; 968 ok 'ABCD' =~ m'(\N{EVIL})' && $1 eq 'ABC', 'Charname caching $1'; 969 ok 'ABCDE' =~ m'(\N{EVIL})', 'Charname caching $1'; 970 like 'xy', qr/x\N{EMPTY-STR}y/, 971 'Empty string charname produces NOTHING node'; 972 ok 'xy' =~ 'x\N{EMPTY-STR}y', 973 'Empty string charname produces NOTHING node'; 974 like '', qr/\N{EMPTY-STR}/, 975 'Empty string charname produces NOTHING node'; 976 like "\N{LONG-STR}", qr/^\N{LONG-STR}$/, 'Verify that long string works'; 977 like "\N{LONG-STR}", qr/^\N{LONG-STR}$/i, 'Verify under folding that long string works'; 978 979 # perlhacktips points out that these work on both ASCII and EBCDIC 980 like "\xfc", qr/\N{EMPTY-STR}\xdc/i, 'Empty \N{} should change /d to /u'; 981 like "\xfc", qr'\N{EMPTY-STR}\xdc'i, 'Empty \N{} should change /d to /u'; 982 983 eval '/(?[[\N{EMPTY-STR}]])/'; 984 like $@, qr/Zero length \\N\{\}/, 'Verify zero-length return from \N{} correctly fails'; 985 ok "\N{LONG-STR}" =~ /^\N{LONG-STR}$/, 'Verify that long string works'; 986 ok "\N{LONG-STR}" =~ '^\N{LONG-STR}$', 'Verify that long string works'; 987 ok "\N{LONG-STR}" =~ /^\N{LONG-STR}$/i, 'Verify under folding that long string works'; 988 ok "\N{LONG-STR}" =~ m'^\N{LONG-STR}$'i, 'Verify under folding that long string works'; 989 990 undef $w; 991 { 992 () = eval q ["\N{TOO MANY SPACES}"]; 993 like ($@, qr/charnames alias definitions may not contain a sequence of multiple spaces/, "Multiple spaces in a row in a charnames alias is fatal"); 994 eval q [use utf8; () = "\N{TOO MANY SPACES}"]; 995 like ($@, qr/charnames alias definitions may not contain a sequence of multiple spaces/, "... same under utf8"); 996 } 997 998 undef $w; 999 my $Cedilla_Latin1 = "GAR" 1000 . uni_to_native("\xC7") 1001 . "ON"; 1002 my $Cedilla_utf8 = $Cedilla_Latin1; 1003 utf8::upgrade($Cedilla_utf8); 1004 eval qq[is("\\N{$Cedilla_Latin1}", "$Cedilla_Latin1", "A cedilla in character name works")]; 1005 undef $w; 1006 { 1007 use feature 'unicode_eval'; 1008 eval qq[use utf8; is("\\N{$Cedilla_utf8}", "$Cedilla_utf8", "... same under 'use utf8': they work")]; 1009 } 1010 1011 undef $w; 1012 my $NBSP_Latin1 = "NBSP" 1013 . uni_to_native("\xA0") 1014 . "SEPARATED" 1015 . uni_to_native("\xA0") 1016 . "SPACE"; 1017 my $NBSP_utf8 = $NBSP_Latin1; 1018 utf8::upgrade($NBSP_utf8); 1019 () = eval qq[is("\\N{$NBSP_Latin1}", "$NBSP_Latin1"]; 1020 like ($@, qr/Invalid character in \\N\{...}/, "A NO-BREAK SPACE in a charnames alias is fatal"); 1021 undef $w; 1022 { 1023 use feature 'unicode_eval'; 1024 eval qq[use utf8; is("\\N{$NBSP_utf8}"]; 1025 like ($@, qr/Invalid character in \\N\{...}/, "A NO-BREAK SPACE in a charnames alias is fatal"); 1026 } 1027 1028 { 1029 BEGIN { no strict; *CnameTest:: = *{"_charnames\0A::" } } 1030 package CnameTest { sub translator { pop } } 1031 BEGIN { $^H{charnames} = \&CnameTest::translator } 1032 undef $w; 1033 () = eval q ["\N{TOO MANY SPACES}"]; 1034 like ($@, qr/charnames alias definitions may not contain a sequence of multiple spaces/, 1035 'translators in _charnames\0* packages get validated'); 1036 } 1037 1038 # If remove the limitation in regcomp code these should work 1039 # differently 1040 undef $w; 1041 eval q [like "\N{TOO-LONG-STR}" =~ /^\N{TOO-LONG-STR}$/, 'Verify that what once was too long a string works']; 1042 eval 'q() =~ /\N{4F}/'; 1043 ok $@ && $@ =~ /Invalid character/, 'Verify that leading digit in name gives error'; 1044 eval 'q() =~ /\N{COM,MA}/'; 1045 ok $@ && $@ =~ /Invalid character/, 'Verify that comma in name gives error'; 1046 $name = "A" . uni_to_native("\x{D7}") . "O"; 1047 eval "q(W) =~ /\\N{$name}/"; 1048 ok $@ && $@ =~ /Invalid character/, 'Verify that latin1 symbol in name gives error'; 1049 my $utf8_name = "7 CITIES OF GOLD"; 1050 utf8::upgrade($utf8_name); 1051 eval "use utf8; q(W) =~ /\\N{$utf8_name}/"; 1052 ok $@ && $@ =~ /Invalid character/, 'Verify that leading digit in utf8 name gives error'; 1053 $utf8_name = "SHARP #"; 1054 utf8::upgrade($utf8_name); 1055 eval "use utf8; q(W) =~ /\\N{$utf8_name}/"; 1056 ok $@ && $@ =~ /Invalid character/, 'Verify that ASCII symbol in utf8 name gives error'; 1057 $utf8_name = "A HOUSE " . uni_to_native("\xF7") . " AGAINST ITSELF"; 1058 utf8::upgrade($utf8_name); 1059 eval "use utf8; q(W) =~ /\\N{$utf8_name}/"; 1060 ok $@ && $@ =~ /Invalid character/, 'Verify that latin1 symbol in utf8 name gives error'; 1061 $utf8_name = "\x{664} HORSEMEN}"; 1062 eval "use utf8; q(W) =~ /\\N{$utf8_name}/"; 1063 ok $@ && $@ =~ /Invalid character/, 'Verify that leading above Latin1 digit in utf8 name gives error'; 1064 $utf8_name = "A \x{1F4A9} WOULD SMELL AS SWEET}"; 1065 eval "use utf8; q(W) =~ /\\N{$utf8_name}/"; 1066 ok $@ && $@ =~ /Invalid character/, 'Verify that above Latin1 symbol in utf8 name gives error'; 1067 1068 undef $w; 1069 $name = "A" . uni_to_native("\x{D1}") . "O"; 1070 eval "q(W) =~ /\\N{$name}/"; 1071 ok ! $w, 'Verify that latin1 letter in name doesnt give warning'; 1072 1073 # This tests the code path that restarts the parse when the recursive 1074 # call to S_reg() from within S_grok_bslash_N() discovers that the 1075 # pattern needs to be recalculated as UTF-8. use eval to avoid 1076 # needing literal Unicode in this source file: 1077 my $r = eval "qr/\\N{\x{100}\x{100}}/"; 1078 isnt $r, undef, "Generated regex for multi-char UTF-8 charname" 1079 or diag($@); 1080 like "\x{100}\x{100}", $r, "which matches"; 1081 } 1082 1083 { 1084 use charnames ':full'; 1085 1086 unlike 'aabc', qr/a\N{PLUS SIGN}b/, '/a\N{PLUS SIGN}b/ against aabc'; 1087 like 'a+bc', qr/a\N{PLUS SIGN}b/, '/a\N{PLUS SIGN}b/ against a+bc'; 1088 1089 like ' A B', qr/\N{SPACE}\N{U+0041}\N{SPACE}\N{U+0042}/, 1090 'Intermixed named and unicode escapes'; 1091 like "\N{SPACE}\N{U+0041}\N{SPACE}\N{U+0042}", 1092 qr/\N{SPACE}\N{U+0041}\N{SPACE}\N{U+0042}/, 1093 'Intermixed named and unicode escapes'; 1094 like "\N{SPACE}\N{U+0041}\N{SPACE}\N{U+0042}", 1095 qr/[\N{SPACE}\N{U+0041}][\N{SPACE}\N{U+0042}]/, 1096 'Intermixed named and unicode escapes'; 1097 like "\0", qr/^\N{NULL}$/, 'Verify that \N{NULL} works; is not confused with an error'; 1098 } 1099 1100 { 1101 our $brackets; 1102 $brackets = qr{ 1103 { (?> [^{}]+ | (??{ $brackets }) )* } 1104 }x; 1105 1106 unlike "{b{c}d", qr/^((??{ $brackets }))/, "Bracket mismatch"; 1107 1108 SKIP: { 1109 our @stack = (); 1110 my @expect = qw( 1111 stuff1 1112 stuff2 1113 <stuff1>and<stuff2> 1114 right 1115 <right> 1116 <<right>> 1117 <<<right>>> 1118 <<stuff1>and<stuff2>><<<<right>>>> 1119 ); 1120 1121 local $_ = '<<<stuff1>and<stuff2>><<<<right>>>>>'; 1122 ok /^(<((?:(?>[^<>]+)|(?1))*)>(?{push @stack, $2 }))$/, 1123 "Recursion matches"; 1124 is(@stack, @expect, "Right amount of matches") 1125 or skip "Won't test individual results as count isn't equal", 1126 0 + @expect; 1127 my $idx = 0; 1128 foreach my $expect (@expect) { 1129 is($stack [$idx], $expect, 1130 "Expecting '$expect' at stack pos #$idx"); 1131 $idx ++; 1132 } 1133 } 1134 } 1135 1136 { 1137 my $s = '123453456'; 1138 $s =~ s/(?<digits>\d+)\k<digits>/$+{digits}/; 1139 ok $s eq '123456', 'Named capture (angle brackets) s///'; 1140 $s = '123453456'; 1141 $s =~ s/(?'digits'\d+)\k'digits'/$+{digits}/; 1142 ok $s eq '123456', 'Named capture (single quotes) s///'; 1143 } 1144 1145 { 1146 my @ary = ( 1147 pack('U', utf8::unicode_to_native(0x00F1)), # n-tilde 1148 '_'.pack('U', utf8::unicode_to_native(0x00F1)), # _ + n-tilde 1149 'c'.pack('U', 0x0327), # c + cedilla 1150 pack('U*', utf8::unicode_to_native(0x00F1), 0x0327),# n-tilde + cedilla 1151 pack('U', 0x0391), # ALPHA 1152 pack('U', 0x0391).'2', # ALPHA + 2 1153 pack('U', 0x0391).'_', # ALPHA + _ 1154 ); 1155 1156 for my $uni (@ary) { 1157 my ($r1, $c1, $r2, $c2) = eval qq { 1158 use utf8; 1159 scalar ("..foo foo.." =~ /(?'${uni}'foo) \\k'${uni}'/), 1160 \$+{${uni}}, 1161 scalar ("..bar bar.." =~ /(?<${uni}>bar) \\k<${uni}>/), 1162 \$+{${uni}}; 1163 }; 1164 ok $r1, "Named capture UTF (?'')"; 1165 ok defined $c1 && $c1 eq 'foo', "Named capture UTF \%+"; 1166 ok $r2, "Named capture UTF (?<>)"; 1167 ok defined $c2 && $c2 eq 'bar', "Named capture UTF \%+"; 1168 } 1169 } 1170 1171 { 1172 my $s = 'foo bar baz'; 1173 my @res; 1174 if ('1234' =~ /(?<A>1)(?<B>2)(?<A>3)(?<B>4)/) { 1175 foreach my $name (sort keys(%-)) { 1176 my $ary = $- {$name}; 1177 foreach my $idx (0 .. $#$ary) { 1178 push @res, "$name:$idx:$ary->[$idx]"; 1179 } 1180 } 1181 } 1182 my @expect = qw (A:0:1 A:1:3 B:0:2 B:1:4); 1183 is("@res", "@expect", "Check %-"); 1184 eval' 1185 no warnings "uninitialized"; 1186 print for $- {this_key_doesnt_exist}; 1187 '; 1188 ok !$@,'lvalue $- {...} should not throw an exception'; 1189 } 1190 1191 { 1192 # \c\ followed by _ 1193 unlike "x\c_y", qr/x\c\_y/, '\_ in a pattern'; 1194 like "x\c\_y", qr/x\c\_y/, '\_ in a pattern'; 1195 1196 # \c\ followed by other characters 1197 for my $c ("z", "\0", "!", chr(254), chr(256)) { 1198 my $targ = "a" . uni_to_native("\034") . "$c"; 1199 my $reg = "a\\c\\$c"; 1200 ok eval ("qq/$targ/ =~ /$reg/"), "\\c\\ in pattern"; 1201 } 1202 } 1203 1204 { # Test the (*PRUNE) pattern 1205 our $count = 0; 1206 'aaab' =~ /a+b?(?{$count++})(*FAIL)/; 1207 is($count, 9, "Expect 9 for no (*PRUNE)"); 1208 $count = 0; 1209 'aaab' =~ /a+b?(*PRUNE)(?{$count++})(*FAIL)/; 1210 is($count, 3, "Expect 3 with (*PRUNE)"); 1211 local $_ = 'aaab'; 1212 $count = 0; 1213 1 while /.(*PRUNE)(?{$count++})(*FAIL)/g; 1214 is($count, 4, "/.(*PRUNE)/"); 1215 $count = 0; 1216 'aaab' =~ /a+b?(??{'(*PRUNE)'})(?{$count++})(*FAIL)/; 1217 is($count, 3, "Expect 3 with (*PRUNE)"); 1218 local $_ = 'aaab'; 1219 $count = 0; 1220 1 while /.(??{'(*PRUNE)'})(?{$count++})(*FAIL)/g; 1221 is($count, 4, "/.(*PRUNE)/"); 1222 } 1223 1224 { # Test the (*SKIP) pattern 1225 our $count = 0; 1226 'aaab' =~ /a+b?(*SKIP)(?{$count++})(*FAIL)/; 1227 is($count, 1, "Expect 1 with (*SKIP)"); 1228 local $_ = 'aaab'; 1229 $count = 0; 1230 1 while /.(*SKIP)(?{$count++})(*FAIL)/g; 1231 is($count, 4, "/.(*SKIP)/"); 1232 $_ = 'aaabaaab'; 1233 $count = 0; 1234 our @res = (); 1235 1 while /(a+b?)(*SKIP)(?{$count++; push @res,$1})(*FAIL)/g; 1236 is($count, 2, "Expect 2 with (*SKIP)"); 1237 is("@res", "aaab aaab", "Adjacent (*SKIP) works as expected"); 1238 } 1239 1240 { # Test the (*SKIP) pattern 1241 our $count = 0; 1242 'aaab' =~ /a+b?(*MARK:foo)(*SKIP)(?{$count++})(*FAIL)/; 1243 is($count, 1, "Expect 1 with (*SKIP)"); 1244 local $_ = 'aaab'; 1245 $count = 0; 1246 1 while /.(*MARK:foo)(*SKIP)(?{$count++})(*FAIL)/g; 1247 is($count, 4, "/.(*SKIP)/"); 1248 $_ = 'aaabaaab'; 1249 $count = 0; 1250 our @res = (); 1251 1 while /(a+b?)(*MARK:foo)(*SKIP)(?{$count++; push @res,$1})(*FAIL)/g; 1252 is($count, 2, "Expect 2 with (*SKIP)"); 1253 is("@res", "aaab aaab", "Adjacent (*SKIP) works as expected"); 1254 } 1255 1256 { # Test the (*SKIP) pattern 1257 our $count = 0; 1258 'aaab' =~ /a*(*MARK:a)b?(*MARK:b)(*SKIP:a)(?{$count++})(*FAIL)/; 1259 is($count, 3, "Expect 3 with *MARK:a)b?(*MARK:b)(*SKIP:a)"); 1260 local $_ = 'aaabaaab'; 1261 $count = 0; 1262 our @res = (); 1263 1 while 1264 /(a*(*MARK:a)b?)(*MARK:x)(*SKIP:a)(?{$count++; push @res,$1})(*FAIL)/g; 1265 is($count, 5, "Expect 5 with (*MARK:a)b?)(*MARK:x)(*SKIP:a)"); 1266 is("@res", "aaab b aaab b ", 1267 "Adjacent (*MARK:a)b?)(*MARK:x)(*SKIP:a) works as expected"); 1268 } 1269 1270 { # Test the (*COMMIT) pattern 1271 our $count = 0; 1272 'aaabaaab' =~ /a+b?(*COMMIT)(?{$count++})(*FAIL)/; 1273 is($count, 1, "Expect 1 with (*COMMIT)"); 1274 local $_ = 'aaab'; 1275 $count = 0; 1276 1 while /.(*COMMIT)(?{$count++})(*FAIL)/g; 1277 is($count, 1, "/.(*COMMIT)/"); 1278 $_ = 'aaabaaab'; 1279 $count = 0; 1280 our @res = (); 1281 1 while /(a+b?)(*COMMIT)(?{$count++; push @res,$1})(*FAIL)/g; 1282 is($count, 1, "Expect 1 with (*COMMIT)"); 1283 is("@res", "aaab", "Adjacent (*COMMIT) works as expected"); 1284 1285 unlike("1\n2a\n", qr/^\d+(*COMMIT)\w+/m, "COMMIT and anchors"); 1286 } 1287 1288 { 1289 # Test named commits and the $REGERROR var 1290 local $REGERROR; 1291 for my $name ('', ':foo') { 1292 for my $pat ("(*PRUNE$name)", 1293 ($name ? "(*MARK$name)" : "") . "(*SKIP$name)", 1294 "(*COMMIT$name)") { 1295 for my $suffix ('(*FAIL)', '') { 1296 'aaaab' =~ /a+b$pat$suffix/; 1297 is($REGERROR, 1298 ($suffix ? ($name ? 'foo' : "1") : ""), 1299 "Test $pat and \$REGERROR $suffix"); 1300 } 1301 } 1302 } 1303 } 1304 1305 { 1306 # Test named commits and the $REGERROR var 1307 package Fnorble; 1308 our $REGERROR; 1309 local $REGERROR; 1310 for my $name ('', ':foo') { 1311 for my $pat ("(*PRUNE$name)", 1312 ($name ? "(*MARK$name)" : "") . "(*SKIP$name)", 1313 "(*COMMIT$name)") { 1314 for my $suffix ('(*FAIL)','') { 1315 'aaaab' =~ /a+b$pat$suffix/; 1316 ::is($REGERROR, 1317 ($suffix ? ($name ? 'foo' : "1") : ""), 1318 "Test $pat and \$REGERROR $suffix"); 1319 } 1320 } 1321 } 1322 } 1323 1324 { 1325 # Test named commits and the $REGERROR var 1326 my $message = '$REGERROR'; 1327 local $REGERROR; 1328 for my $word (qw (bar baz bop)) { 1329 $REGERROR = ""; 1330 "aaaaa$word" =~ 1331 /a+(?:bar(*COMMIT:bar)|baz(*COMMIT:baz)|bop(*COMMIT:bop))(*FAIL)/; 1332 is($REGERROR, $word, $message); 1333 } 1334 } 1335 1336 { 1337 #Mindnumbingly simple test of (*THEN) 1338 for ("ABC","BAX") { 1339 ok /A (*THEN) X | B (*THEN) C/x, "Simple (*THEN) test"; 1340 } 1341 } 1342 1343 { 1344 my $message = "Relative Recursion"; 1345 my $parens = qr/(\((?:[^()]++|(?-1))*+\))/; 1346 local $_ = 'foo((2*3)+4-3) + bar(2*(3+4)-1*(2-3))'; 1347 my ($all, $one, $two) = ('', '', ''); 1348 ok(m/foo $parens \s* \+ \s* bar $parens/x, $message); 1349 is($1, '((2*3)+4-3)', $message); 1350 is($2, '(2*(3+4)-1*(2-3))', $message); 1351 is($&, 'foo((2*3)+4-3) + bar(2*(3+4)-1*(2-3))', $message); 1352 is($&, $_, $message); 1353 } 1354 1355 { 1356 my $spaces=" "; 1357 local $_ = join 'bar', $spaces, $spaces; 1358 our $count = 0; 1359 s/(?>\s+bar)(?{$count++})//g; 1360 is($_, $spaces, "SUSPEND final string"); 1361 is($count, 1, "Optimiser should have prevented more than one match"); 1362 } 1363 1364 { 1365 # From Message-ID: <877ixs6oa6.fsf@k75.linux.bogus> 1366 my $dow_name = "nada"; 1367 my $parser = "(\$dow_name) = \$time_string =~ /(D\x{e9}\\ " . 1368 "C\x{e9}adaoin|D\x{e9}\\ Sathairn|\\w+|\x{100})/"; 1369 my $time_string = "D\x{e9} C\x{e9}adaoin"; 1370 eval $parser; 1371 ok !$@, "Test Eval worked"; 1372 is($dow_name, $time_string, "UTF-8 trie common prefix extraction"); 1373 } 1374 1375 { 1376 my $v; 1377 ($v = 'bar') =~ /(\w+)/g; 1378 $v = 'foo'; 1379 is("$1", 'bar', 1380 '$1 is safe after /g - may fail due to specialized config in pp_hot.c'); 1381 } 1382 1383 { 1384 my $message = "http://nntp.perl.org/group/perl.perl5.porters/118663"; 1385 my $qr_barR1 = qr/(bar)\g-1/; 1386 like("foobarbarxyz", $qr_barR1, $message); 1387 like("foobarbarxyz", qr/foo${qr_barR1}xyz/, $message); 1388 like("foobarbarxyz", qr/(foo)${qr_barR1}xyz/, $message); 1389 like("foobarbarxyz", qr/(foo)(bar)\g{-1}xyz/, $message); 1390 like("foobarbarxyz", qr/(foo${qr_barR1})xyz/, $message); 1391 like("foobarbarxyz", qr/(foo(bar)\g{-1})xyz/, $message); 1392 } 1393 1394 { 1395 my $message = '$REGMARK'; 1396 our @r = (); 1397 local $REGMARK; 1398 local $REGERROR; 1399 like('foofoo', qr/foo (*MARK:foo) (?{push @r,$REGMARK}) /x, $message); 1400 is("@r","foo", $message); 1401 is($REGMARK, "foo", $message); 1402 unlike('foofoo', qr/foo (*MARK:foo) (*FAIL) /x, $message); 1403 is($REGMARK, '', $message); 1404 is($REGERROR, 'foo', $message); 1405 } 1406 1407 { 1408 my $message = '\K test'; 1409 my $x; 1410 $x = "abc.def.ghi.jkl"; 1411 $x =~ s/.*\K\..*//; 1412 is($x, "abc.def.ghi", $message); 1413 1414 $x = "one two three four"; 1415 $x =~ s/o+ \Kthree//g; 1416 is($x, "one two four", $message); 1417 1418 $x = "abcde"; 1419 $x =~ s/(.)\K/$1/g; 1420 is($x, "aabbccddee", $message); 1421 } 1422 1423 { 1424 sub kt { 1425 return '4' if $_[0] eq '09028623'; 1426 } 1427 # Nested EVAL using PL_curpm (via $1 or friends) 1428 my $re; 1429 our $grabit = qr/ ([0-6][0-9]{7}) (??{ kt $1 }) [890] /x; 1430 $re = qr/^ ( (??{ $grabit }) ) $ /x; 1431 my @res = '0902862349' =~ $re; 1432 is(join ("-", @res), "0902862349", 1433 'PL_curpm is set properly on nested eval'); 1434 1435 our $qr = qr/ (o) (??{ $1 }) /x; 1436 ok 'boob'=~/( b (??{ $qr }) b )/x && 1, "PL_curpm, nested eval"; 1437 } 1438 1439 { 1440 use charnames ":full"; 1441 like "\N{ROMAN NUMERAL ONE}", qr/\p{Alphabetic}/, "I =~ Alphabetic"; 1442 like "\N{ROMAN NUMERAL ONE}", qr/\p{Uppercase}/, "I =~ Uppercase"; 1443 unlike "\N{ROMAN NUMERAL ONE}", qr/\p{Lowercase}/, "I !~ Lowercase"; 1444 like "\N{ROMAN NUMERAL ONE}", qr/\p{IDStart}/, "I =~ ID_Start"; 1445 like "\N{ROMAN NUMERAL ONE}", qr/\p{IDContinue}/, "I =~ ID_Continue"; 1446 like "\N{SMALL ROMAN NUMERAL ONE}", qr/\p{Alphabetic}/, "i =~ Alphabetic"; 1447 unlike "\N{SMALL ROMAN NUMERAL ONE}", qr/\p{Uppercase}/, "i !~ Uppercase"; 1448 like "\N{SMALL ROMAN NUMERAL ONE}", qr/\p{Uppercase}/i, "i =~ Uppercase under /i"; 1449 unlike "\N{SMALL ROMAN NUMERAL ONE}", qr/\p{Titlecase}/, "i !~ Titlecase"; 1450 like "\N{SMALL ROMAN NUMERAL ONE}", qr/\p{Titlecase}/i, "i =~ Titlecase under /i"; 1451 like "\N{ROMAN NUMERAL ONE}", qr/\p{Lowercase}/i, "I =~ Lowercase under /i"; 1452 1453 like "\N{SMALL ROMAN NUMERAL ONE}", qr/\p{Lowercase}/, "i =~ Lowercase"; 1454 like "\N{SMALL ROMAN NUMERAL ONE}", qr/\p{IDStart}/, "i =~ ID_Start"; 1455 like "\N{SMALL ROMAN NUMERAL ONE}", qr/\p{IDContinue}/, "i =~ ID_Continue" 1456 } 1457 1458 { # More checking that /i works on the few properties that it makes a 1459 # difference. Uppercase, Lowercase, and Titlecase were done in the 1460 # block above 1461 like "A", qr/\p{PosixUpper}/, "A =~ PosixUpper"; 1462 like "A", qr/\p{PosixUpper}/i, "A =~ PosixUpper under /i"; 1463 unlike "A", qr/\p{PosixLower}/, "A !~ PosixLower"; 1464 like "A", qr/\p{PosixLower}/i, "A =~ PosixLower under /i"; 1465 unlike "a", qr/\p{PosixUpper}/, "a !~ PosixUpper"; 1466 like "a", qr/\p{PosixUpper}/i, "a =~ PosixUpper under /i"; 1467 like "a", qr/\p{PosixLower}/, "a =~ PosixLower"; 1468 like "a", qr/\p{PosixLower}/i, "a =~ PosixLower under /i"; 1469 1470 like uni_to_native("\xC0"), qr/\p{XPosixUpper}/, "\\xC0 =~ XPosixUpper"; 1471 like uni_to_native("\xC0"), qr/\p{XPosixUpper}/i, "\\xC0 =~ XPosixUpper under /i"; 1472 unlike uni_to_native("\xC0"), qr/\p{XPosixLower}/, "\\xC0 !~ XPosixLower"; 1473 like uni_to_native("\xC0"), qr/\p{XPosixLower}/i, "\\xC0 =~ XPosixLower under /i"; 1474 unlike uni_to_native("\xE0"), qr/\p{XPosixUpper}/, "\\xE0 !~ XPosixUpper"; 1475 like uni_to_native("\xE0"), qr/\p{XPosixUpper}/i, "\\xE0 =~ XPosixUpper under /i"; 1476 like uni_to_native("\xE0"), qr/\p{XPosixLower}/, "\\xE0 =~ XPosixLower"; 1477 like uni_to_native("\xE0"), qr/\p{XPosixLower}/i, "\\xE0 =~ XPosixLower under /i"; 1478 1479 like uni_to_native("\xC0"), qr/\p{UppercaseLetter}/, "\\xC0 =~ UppercaseLetter"; 1480 like uni_to_native("\xC0"), qr/\p{UppercaseLetter}/i, "\\xC0 =~ UppercaseLetter under /i"; 1481 unlike uni_to_native("\xC0"), qr/\p{LowercaseLetter}/, "\\xC0 !~ LowercaseLetter"; 1482 like uni_to_native("\xC0"), qr/\p{LowercaseLetter}/i, "\\xC0 =~ LowercaseLetter under /i"; 1483 unlike uni_to_native("\xC0"), qr/\p{TitlecaseLetter}/, "\\xC0 !~ TitlecaseLetter"; 1484 like uni_to_native("\xC0"), qr/\p{TitlecaseLetter}/i, "\\xC0 =~ TitlecaseLetter under /i"; 1485 unlike uni_to_native("\xE0"), qr/\p{UppercaseLetter}/, "\\xE0 !~ UppercaseLetter"; 1486 like uni_to_native("\xE0"), qr/\p{UppercaseLetter}/i, "\\xE0 =~ UppercaseLetter under /i"; 1487 like uni_to_native("\xE0"), qr/\p{LowercaseLetter}/, "\\xE0 =~ LowercaseLetter"; 1488 like uni_to_native("\xE0"), qr/\p{LowercaseLetter}/i, "\\xE0 =~ LowercaseLetter under /i"; 1489 unlike uni_to_native("\xE0"), qr/\p{TitlecaseLetter}/, "\\xE0 !~ TitlecaseLetter"; 1490 like uni_to_native("\xE0"), qr/\p{TitlecaseLetter}/i, "\\xE0 =~ TitlecaseLetter under /i"; 1491 unlike "\x{1C5}", qr/\p{UppercaseLetter}/, "\\x{1C5} !~ UppercaseLetter"; 1492 like "\x{1C5}", qr/\p{UppercaseLetter}/i, "\\x{1C5} =~ UppercaseLetter under /i"; 1493 unlike "\x{1C5}", qr/\p{LowercaseLetter}/, "\\x{1C5} !~ LowercaseLetter"; 1494 like "\x{1C5}", qr/\p{LowercaseLetter}/i, "\\x{1C5} =~ LowercaseLetter under /i"; 1495 like "\x{1C5}", qr/\p{TitlecaseLetter}/, "\\x{1C5} =~ TitlecaseLetter"; 1496 like "\x{1C5}", qr/\p{TitlecaseLetter}/i, "\\x{1C5} =~ TitlecaseLetter under /i"; 1497 } 1498 1499 { 1500 # requirement of Unicode Technical Standard #18, 1.7 Code Points 1501 # cf. http://www.unicode.org/reports/tr18/#Supplementary_Characters 1502 for my $u (0x7FF, 0x800, 0xFFFF, 0x10000) { 1503 no warnings 'utf8'; # oops 1504 my $c = chr $u; 1505 my $x = sprintf '%04X', $u; 1506 like "A${c}B", qr/A[\0-\x{10000}]B/, "Unicode range - $x"; 1507 } 1508 } 1509 1510 { 1511 no warnings 'uninitialized'; 1512 my $res = ""; 1513 1514 if ('1' =~ /(?|(?<digit>1)|(?<digit>2))/) { 1515 $res = "@{$- {digit}}"; 1516 } 1517 is($res, "1 ", 1518 "Check that repeated named captures in branch reset (?|...) work as expected"); 1519 if ('2' =~ /(?|(?<digit>1)|(?<digit>2))/) { 1520 $res = "@{$- {digit}}"; 1521 } 1522 is($res, " 2", 1523 "Check that repeated named captures in branch reset (?|...) work as expected"); 1524 1525 $res = ""; 1526 if ('11' =~ /(?|(?<digit>1)|(?<digit>2))(?&digit)/) { 1527 $res = "@{$- {digit}}"; 1528 } 1529 is($res, "1 ", 1530 "Check that (?&..) to a buffer inside a (?|...) goes to the leftmost"); 1531 } 1532 1533 { 1534 use warnings; 1535 my $message = "ASCII pattern that really is UTF-8"; 1536 my @w; 1537 local $SIG {__WARN__} = sub {push @w, "@_"}; 1538 my $c = qq (\x{DF}); 1539 like($c, qr/${c}|\x{100}/, $message); 1540 is("@w", '', $message); 1541 } 1542 1543 { 1544 my $message = "Corruption of match results of qr// across scopes"; 1545 my $qr = qr/(fo+)(ba+r)/; 1546 'foobar' =~ /$qr/; 1547 is("$1$2", "foobar", $message); 1548 { 1549 'foooooobaaaaar' =~ /$qr/; 1550 is("$1$2", 'foooooobaaaaar', $message); 1551 } 1552 is("$1$2", "foobar", $message); 1553 } 1554 1555 { 1556 my $message = "HORIZWS"; 1557 local $_ = "\t \r\n \n \t".chr(11)."\n"; 1558 s/\H/H/g; 1559 s/\h/h/g; 1560 is($_, "hhHHhHhhHH", $message); 1561 $_ = "\t \r\n \n \t" . chr (11) . "\n"; 1562 utf8::upgrade ($_); 1563 s/\H/H/g; 1564 s/\h/h/g; 1565 is($_, "hhHHhHhhHH", $message); 1566 } 1567 1568 { 1569 # Various whitespace special patterns 1570 my @h = map {chr utf8::unicode_to_native($_) } 1571 0x09, 0x20, 0xa0, 0x1680, 0x2000, 1572 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, 0x2006, 1573 0x2007, 0x2008, 0x2009, 0x200a, 0x202f, 0x205f, 1574 0x3000; 1575 my @v = map {chr utf8::unicode_to_native($_) } 1576 0x0a, 0x0b, 0x0c, 0x0d, 0x85, 0x2028, 1577 0x2029; 1578 my @lb = (uni_to_native("\x0D\x0A"), 1579 map {chr utf8::unicode_to_native($_) } 1580 0x0A .. 0x0D, 0x85, 0x2028, 0x2029); 1581 foreach my $t ([\@h, qr/\h/, qr/\h+/], 1582 [\@v, qr/\v/, qr/\v+/], 1583 [\@lb, qr/\R/, qr/\R+/],) { 1584 my $ary = shift @$t; 1585 foreach my $pat (@$t) { 1586 foreach my $str (@$ary) { 1587 my $temp_str = $str; 1588 $temp_str = display($temp_str); 1589 ok $str =~ /($pat)/, $temp_str . " =~ /($pat)"; 1590 my $temp_1 = $1; 1591 is($1, $str, "\$1='" . display($temp_1) . "' eq '" . $temp_str . "' after ($pat)"); 1592 utf8::upgrade ($str); 1593 ok $str =~ /($pat)/, "Upgraded " . $temp_str . " =~ /($pat)/"; 1594 is($1, $str, "\$1='" . display($temp_1) . "' eq '" . $temp_str . "'(upgraded) after ($pat)"); 1595 } 1596 } 1597 } 1598 } 1599 1600 { 1601 # Check that \\xDF match properly in its various forms 1602 # Test that \xDF matches properly. this is pretty hacky stuff, 1603 # but its actually needed. The malarky with '-' is to prevent 1604 # compilation caching from playing any role in the test. 1605 my @df = (chr utf8::unicode_to_native(0xDF), '-', chr utf8::unicode_to_native(0xDF)); 1606 utf8::upgrade ($df [2]); 1607 my @strs = ('ss', 'sS', 'Ss', 'SS', chr utf8::unicode_to_native(0xDF)); 1608 my @ss = map {("$_", "$_")} @strs; 1609 utf8::upgrade ($ss [$_ * 2 + 1]) for 0 .. $#strs; 1610 1611 for my $ssi (0 .. $#ss) { 1612 for my $dfi (0 .. $#df) { 1613 my $pat = $df [$dfi]; 1614 my $str = $ss [$ssi]; 1615 my $utf_df = ($dfi > 1) ? 'utf8' : ''; 1616 my $utf_ss = ($ssi % 2) ? 'utf8' : ''; 1617 my $sstr; # We hard-code the ebcdic value below to avoid 1618 # perturbing the test 1619 ($sstr = $str) =~ s/\xDF/\\xDF/ if $::IS_ASCII; 1620 ($sstr = $str) =~ s/\x59/\\x59/ if $::IS_EBCDIC; 1621 1622 if ($utf_df || $utf_ss || length ($ss [$ssi]) == 1) { 1623 my $ret = $str =~ /$pat/i; 1624 next if $pat eq '-'; 1625 if ($::IS_ASCII) { 1626 ok $ret, "\"$sstr\" =~ /\\xDF/i " . 1627 "(str is @{[$utf_ss||'latin']}, pat is " . 1628 "@{[$utf_df||'latin']})"; 1629 } 1630 else { 1631 ok $ret, "\"$sstr\" =~ /\\x59/i " . 1632 "(str is @{[$utf_ss||'latin']}, pat is " . 1633 "@{[$utf_df||'latin']})"; 1634 } 1635 } 1636 else { 1637 my $ret = $str !~ /$pat/i; 1638 next if $pat eq '-'; 1639 if ($::IS_EBCDIC) { 1640 ok $ret, "\"$sstr\" !~ /\\x59/i " . 1641 "(str is @{[$utf_ss||'latin']}, pat is " . 1642 "@{[$utf_df||'latin']})"; 1643 } 1644 else { 1645 ok $ret, "\"$sstr\" !~ /\\xDF/i " . 1646 "(str is @{[$utf_ss||'latin']}, pat is " . 1647 "@{[$utf_df||'latin']})"; 1648 } 1649 } 1650 } 1651 } 1652 } 1653 1654 { 1655 my $message = "BBC(Bleadperl Breaks CPAN) Today: String::Multibyte"; 1656 my $re = qr/(?:[\x00-\xFF]{4})/; 1657 my $hyp = "\0\0\0-"; 1658 my $esc = "\0\0\0\\"; 1659 1660 my $str = "$esc$hyp$hyp$esc$esc"; 1661 my @a = ($str =~ /\G(?:\Q$esc$esc\E|\Q$esc$hyp\E|$re)/g); 1662 1663 is(@a,3, $message); 1664 local $" = "="; 1665 is("@a","$esc$hyp=$hyp=$esc$esc", $message); 1666 } 1667 1668 { 1669 # Test for keys in %+ and %- 1670 my $message = 'Test keys in %+ and %-'; 1671 no warnings 'uninitialized'; 1672 local $_ = "abcdef"; 1673 /(?<foo>a)|(?<foo>b)/; 1674 is((join ",", sort keys %+), "foo", $message); 1675 is((join ",", sort keys %-), "foo", $message); 1676 is((join ",", sort values %+), "a", $message); 1677 is((join ",", sort map "@$_", values %-), "a ", $message); 1678 /(?<bar>a)(?<bar>b)(?<quux>.)/; 1679 is((join ",", sort keys %+), "bar,quux", $message); 1680 is((join ",", sort keys %-), "bar,quux", $message); 1681 is((join ",", sort values %+), "a,c", $message); # leftmost 1682 is((join ",", sort map "@$_", values %-), "a b,c", $message); 1683 /(?<un>a)(?<deux>c)?/; # second buffer won't capture 1684 is((join ",", sort keys %+), "un", $message); 1685 is((join ",", sort keys %-), "deux,un", $message); 1686 is((join ",", sort values %+), "a", $message); 1687 is((join ",", sort map "@$_", values %-), ",a", $message); 1688 } 1689 1690 { 1691 # length() on captures, the numbered ones end up in Perl_magic_len 1692 local $_ = "aoeu " . uni_to_native("\xe6") . "var ook"; 1693 /^ \w+ \s (?<eek>\S+)/x; 1694 1695 is(length $`, 0, q[length $`]); 1696 is(length $', 4, q[length $']); 1697 is(length $&, 9, q[length $&]); 1698 is(length $1, 4, q[length $1]); 1699 is(length $+{eek}, 4, q[length $+{eek} == length $1]); 1700 } 1701 1702 { 1703 my $ok = -1; 1704 1705 $ok = exists ($-{x}) ? 1 : 0 if 'bar' =~ /(?<x>foo)|bar/; 1706 is($ok, 1, '$-{x} exists after "bar"=~/(?<x>foo)|bar/'); 1707 is(scalar (%+), 0, 'scalar %+ == 0 after "bar"=~/(?<x>foo)|bar/'); 1708 is(scalar (%-), 1, 'scalar %- == 1 after "bar"=~/(?<x>foo)|bar/'); 1709 1710 $ok = -1; 1711 $ok = exists ($+{x}) ? 1 : 0 if 'bar' =~ /(?<x>foo)|bar/; 1712 is($ok, 0, '$+{x} not exists after "bar"=~/(?<x>foo)|bar/'); 1713 is(scalar (%+), 0, 'scalar %+ == 0 after "bar"=~/(?<x>foo)|bar/'); 1714 is(scalar (%-), 1, 'scalar %- == 1 after "bar"=~/(?<x>foo)|bar/'); 1715 1716 $ok = -1; 1717 $ok = exists ($-{x}) ? 1 : 0 if 'foo' =~ /(?<x>foo)|bar/; 1718 is($ok, 1, '$-{x} exists after "foo"=~/(?<x>foo)|bar/'); 1719 is(scalar (%+), 1, 'scalar %+ == 1 after "foo"=~/(?<x>foo)|bar/'); 1720 is(scalar (%-), 1, 'scalar %- == 1 after "foo"=~/(?<x>foo)|bar/'); 1721 1722 $ok = -1; 1723 $ok = exists ($+{x}) ? 1 : 0 if 'foo'=~/(?<x>foo)|bar/; 1724 is($ok, 1, '$+{x} exists after "foo"=~/(?<x>foo)|bar/'); 1725 } 1726 1727 { 1728 local $_; 1729 ($_ = 'abc') =~ /(abc)/g; 1730 $_ = '123'; 1731 is("$1", 'abc', "/g leads to unsafe match vars: $1"); 1732 1733 fresh_perl_is(<<'EOP', ">abc<\n", {}, 'mention $&'); 1734$&; 1735my $x; 1736($x='abc')=~/(abc)/g; 1737$x='123'; 1738print ">$1<\n"; 1739EOP 1740 1741 fresh_perl_is(<<'EOP', ">abc<\n", {}, 'no mention of $&'); 1742my $x; 1743($x='abc')=~/(abc)/g; 1744$x='123'; 1745print ">$1<\n"; 1746EOP 1747 } 1748 1749 { 1750 # Message-ID: <20070818091501.7eff4831@r2d2> 1751 my $str = ""; 1752 for (0 .. 5) { 1753 my @x; 1754 $str .= "@x"; # this should ALWAYS be the empty string 1755 'a' =~ /(a|)/; 1756 push @x, 1; 1757 } 1758 is(length $str, 0, "Trie scope error, string should be empty"); 1759 $str = ""; 1760 my @foo = ('a') x 5; 1761 for (@foo) { 1762 my @bar; 1763 $str .= "@bar"; 1764 s/a|/push @bar, 1/e; 1765 } 1766 is(length $str, 0, "Trie scope error, string should be empty"); 1767 } 1768 1769 { 1770# more TRIE/AHOCORASICK problems with mixed utf8 / latin-1 and case folding 1771 for my $ord (160 .. 255) { 1772 my $chr = utf8::unicode_to_native($ord); 1773 my $chr_byte = chr($chr); 1774 my $chr_utf8 = chr($chr); utf8::upgrade($chr_utf8); 1775 my $rx = qr{$chr_byte|X}i; 1776 like($chr_utf8, $rx, "utf8/latin, codepoint $chr"); 1777 } 1778 } 1779 1780 { 1781 our $a = 3; "" =~ /(??{ $a })/; 1782 our $b = $a; 1783 is($b, $a, "Copy of scalar used for postponed subexpression"); 1784 } 1785 1786 { 1787 our @ctl_n = (); 1788 our @plus = (); 1789 our $nested_tags; 1790 $nested_tags = qr{ 1791 < 1792 (\w+) 1793 (?{ 1794 push @ctl_n,$^N; 1795 push @plus,$+; 1796 }) 1797 > 1798 (??{$nested_tags})* 1799 </\s* \w+ \s*> 1800 }x; 1801 1802 my $match = '<bla><blubb></blubb></bla>' =~ m/^$nested_tags$/; 1803 ok $match, 'nested construct matches'; 1804 is("@ctl_n", "bla blubb", '$^N inside of (?{}) works as expected'); 1805 is("@plus", "bla blubb", '$+ inside of (?{}) works as expected'); 1806 } 1807 1808 SKIP: { 1809 # XXX: This set of tests is essentially broken, POSIX character classes 1810 # should not have differing definitions under Unicode. 1811 # There are property names for that. 1812 skip "Tests assume ASCII", 4 unless $::IS_ASCII; 1813 1814 my @notIsPunct = grep {/[[:punct:]]/ and not /\p{IsPunct}/} 1815 map {chr} 0x20 .. 0x7f; 1816 is(join ('', @notIsPunct), '$+<=>^`|~', 1817 '[:punct:] disagrees with IsPunct on Symbols'); 1818 1819 my @isPrint = grep {not /[[:print:]]/ and /\p{IsPrint}/} 1820 map {chr} 0 .. 0x1f, 0x7f .. 0x9f; 1821 is(join ('', @isPrint), "", 1822 'IsPrint agrees with [:print:] on control characters'); 1823 1824 my @isPunct = grep {/[[:punct:]]/ != /\p{IsPunct}/} 1825 map {chr} 0x80 .. 0xff; 1826 is(join ('', @isPunct), "\xa1\xa7\xab\xb6\xb7\xbb\xbf", # �� �� �� �� �� 1827 'IsPunct disagrees with [:punct:] outside ASCII'); 1828 1829 my @isPunctLatin1 = eval q { 1830 grep {/[[:punct:]]/u != /\p{IsPunct}/} map {chr} 0x80 .. 0xff; 1831 }; 1832 skip "Eval failed ($@)", 1 if $@; 1833 skip "PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS set to 0", 1 1834 if !$ENV{PERL_TEST_LEGACY_POSIX_CC}; 1835 is(join ('', @isPunctLatin1), '', 1836 'IsPunct agrees with [:punct:] with explicit Latin1'); 1837 } 1838 1839 { 1840 # Tests for [#perl 71942] 1841 our $count_a; 1842 our $count_b; 1843 1844 my $c = 0; 1845 for my $re ( 1846# [ 1847# should match?, 1848# input string, 1849# re 1, 1850# re 2, 1851# expected values of count_a and count_b, 1852# ] 1853 [ 1854 0, 1855 "xababz", 1856 qr/a+(?{$count_a++})b?(*COMMIT)(*FAIL)/, 1857 qr/a+(?{$count_b++})b?(*COMMIT)z/, 1858 1, 1859 ], 1860 [ 1861 0, 1862 "xababz", 1863 qr/a+(?{$count_a++})b?(*COMMIT)\s*(*FAIL)/, 1864 qr/a+(?{$count_b++})b?(*COMMIT)\s*z/, 1865 1, 1866 ], 1867 [ 1868 0, 1869 "xababz", 1870 qr/a+(?{$count_a++})(?:b|)?(*COMMIT)(*FAIL)/, 1871 qr/a+(?{$count_b++})(?:b|)?(*COMMIT)z/, 1872 1, 1873 ], 1874 [ 1875 0, 1876 "xababz", 1877 qr/a+(?{$count_a++})b{0,6}(*COMMIT)(*FAIL)/, 1878 qr/a+(?{$count_b++})b{0,6}(*COMMIT)z/, 1879 1, 1880 ], 1881 [ 1882 0, 1883 "xabcabcz", 1884 qr/a+(?{$count_a++})(bc){0,6}(*COMMIT)(*FAIL)/, 1885 qr/a+(?{$count_b++})(bc){0,6}(*COMMIT)z/, 1886 1, 1887 ], 1888 [ 1889 0, 1890 "xabcabcz", 1891 qr/a+(?{$count_a++})(bc*){0,6}(*COMMIT)(*FAIL)/, 1892 qr/a+(?{$count_b++})(bc*){0,6}(*COMMIT)z/, 1893 1, 1894 ], 1895 1896 1897 [ 1898 0, 1899 "aaaabtz", 1900 qr/a+(?{$count_a++})b?(*PRUNE)(*FAIL)/, 1901 qr/a+(?{$count_b++})b?(*PRUNE)z/, 1902 4, 1903 ], 1904 [ 1905 0, 1906 "aaaabtz", 1907 qr/a+(?{$count_a++})b?(*PRUNE)\s*(*FAIL)/, 1908 qr/a+(?{$count_b++})b?(*PRUNE)\s*z/, 1909 4, 1910 ], 1911 [ 1912 0, 1913 "aaaabtz", 1914 qr/a+(?{$count_a++})(?:b|)(*PRUNE)(*FAIL)/, 1915 qr/a+(?{$count_b++})(?:b|)(*PRUNE)z/, 1916 4, 1917 ], 1918 [ 1919 0, 1920 "aaaabtz", 1921 qr/a+(?{$count_a++})b{0,6}(*PRUNE)(*FAIL)/, 1922 qr/a+(?{$count_b++})b{0,6}(*PRUNE)z/, 1923 4, 1924 ], 1925 [ 1926 0, 1927 "aaaabctz", 1928 qr/a+(?{$count_a++})(bc){0,6}(*PRUNE)(*FAIL)/, 1929 qr/a+(?{$count_b++})(bc){0,6}(*PRUNE)z/, 1930 4, 1931 ], 1932 [ 1933 0, 1934 "aaaabctz", 1935 qr/a+(?{$count_a++})(bc*){0,6}(*PRUNE)(*FAIL)/, 1936 qr/a+(?{$count_b++})(bc*){0,6}(*PRUNE)z/, 1937 4, 1938 ], 1939 1940 [ 1941 0, 1942 "aaabaaab", 1943 qr/a+(?{$count_a++;})b?(*SKIP)(*FAIL)/, 1944 qr/a+(?{$count_b++;})b?(*SKIP)z/, 1945 2, 1946 ], 1947 [ 1948 0, 1949 "aaabaaab", 1950 qr/a+(?{$count_a++;})b?(*SKIP)\s*(*FAIL)/, 1951 qr/a+(?{$count_b++;})b?(*SKIP)\s*z/, 1952 2, 1953 ], 1954 [ 1955 0, 1956 "aaabaaab", 1957 qr/a+(?{$count_a++;})(?:b|)(*SKIP)(*FAIL)/, 1958 qr/a+(?{$count_b++;})(?:b|)(*SKIP)z/, 1959 2, 1960 ], 1961 [ 1962 0, 1963 "aaabaaab", 1964 qr/a+(?{$count_a++;})b{0,6}(*SKIP)(*FAIL)/, 1965 qr/a+(?{$count_b++;})b{0,6}(*SKIP)z/, 1966 2, 1967 ], 1968 [ 1969 0, 1970 "aaabcaaabc", 1971 qr/a+(?{$count_a++;})(bc){0,6}(*SKIP)(*FAIL)/, 1972 qr/a+(?{$count_b++;})(bc){0,6}(*SKIP)z/, 1973 2, 1974 ], 1975 [ 1976 0, 1977 "aaabcaaabc", 1978 qr/a+(?{$count_a++;})(bc*){0,6}(*SKIP)(*FAIL)/, 1979 qr/a+(?{$count_b++;})(bc*){0,6}(*SKIP)z/, 1980 2, 1981 ], 1982 1983 1984 [ 1985 0, 1986 "aaddbdaabyzc", 1987 qr/a (?{$count_a++;}) (*MARK:T1) (a*) .*? b? (*SKIP:T1) (*FAIL) \s* c \1 /x, 1988 qr/a (?{$count_b++;}) (*MARK:T1) (a*) .*? b? (*SKIP:T1) z \s* c \1 /x, 1989 4, 1990 ], 1991 [ 1992 0, 1993 "aaddbdaabyzc", 1994 qr/a (?{$count_a++;}) (*MARK:T1) (a*) .*? b? (*SKIP:T1) \s* (*FAIL) \s* c \1 /x, 1995 qr/a (?{$count_b++;}) (*MARK:T1) (a*) .*? b? (*SKIP:T1) \s* z \s* c \1 /x, 1996 4, 1997 ], 1998 [ 1999 0, 2000 "aaddbdaabyzc", 2001 qr/a (?{$count_a++;}) (*MARK:T1) (a*) .*? (?:b|) (*SKIP:T1) (*FAIL) \s* c \1 /x, 2002 qr/a (?{$count_b++;}) (*MARK:T1) (a*) .*? (?:b|) (*SKIP:T1) z \s* c \1 /x, 2003 4, 2004 ], 2005 [ 2006 0, 2007 "aaddbdaabyzc", 2008 qr/a (?{$count_a++;}) (*MARK:T1) (a*) .*? b{0,6} (*SKIP:T1) (*FAIL) \s* c \1 /x, 2009 qr/a (?{$count_b++;}) (*MARK:T1) (a*) .*? b{0,6} (*SKIP:T1) z \s* c \1 /x, 2010 4, 2011 ], 2012 [ 2013 0, 2014 "aaddbcdaabcyzc", 2015 qr/a (?{$count_a++;}) (*MARK:T1) (a*) .*? (bc){0,6} (*SKIP:T1) (*FAIL) \s* c \1 /x, 2016 qr/a (?{$count_b++;}) (*MARK:T1) (a*) .*? (bc){0,6} (*SKIP:T1) z \s* c \1 /x, 2017 4, 2018 ], 2019 [ 2020 0, 2021 "aaddbcdaabcyzc", 2022 qr/a (?{$count_a++;}) (*MARK:T1) (a*) .*? (bc*){0,6} (*SKIP:T1) (*FAIL) \s* c \1 /x, 2023 qr/a (?{$count_b++;}) (*MARK:T1) (a*) .*? (bc*){0,6} (*SKIP:T1) z \s* c \1 /x, 2024 4, 2025 ], 2026 2027 2028 [ 2029 0, 2030 "aaaaddbdaabyzc", 2031 qr/a (?{$count_a++;}) (a?) (*MARK:T1) (a*) .*? b? (*MARK:T1) (*SKIP:T1) (*FAIL) \s* c \1 /x, 2032 qr/a (?{$count_b++;}) (a?) (*MARK:T1) (a*) .*? b? (*MARK:T1) (*SKIP:T1) z \s* c \1 /x, 2033 2, 2034 ], 2035 [ 2036 0, 2037 "aaaaddbdaabyzc", 2038 qr/a (?{$count_a++;}) (a?) (*MARK:T1) (a*) .*? b? (*MARK:T1) (*SKIP:T1) \s* (*FAIL) \s* c \1 /x, 2039 qr/a (?{$count_b++;}) (a?) (*MARK:T1) (a*) .*? b? (*MARK:T1) (*SKIP:T1) \s* z \s* c \1 /x, 2040 2, 2041 ], 2042 [ 2043 0, 2044 "aaaaddbdaabyzc", 2045 qr/a (?{$count_a++;}) (a?) (*MARK:T1) (a*) .*? (?:b|) (*MARK:T1) (*SKIP:T1) (*FAIL) \s* c \1 /x, 2046 qr/a (?{$count_b++;}) (a?) (*MARK:T1) (a*) .*? (?:b|) (*MARK:T1) (*SKIP:T1) z \s* c \1 /x, 2047 2, 2048 ], 2049 [ 2050 0, 2051 "aaaaddbdaabyzc", 2052 qr/a (?{$count_a++;}) (a?) (*MARK:T1) (a*) .*? b{0,6} (*MARK:T1) (*SKIP:T1) (*FAIL) \s* c \1 /x, 2053 qr/a (?{$count_b++;}) (a?) (*MARK:T1) (a*) .*? b{0,6} (*MARK:T1) (*SKIP:T1) z \s* c \1 /x, 2054 2, 2055 ], 2056 [ 2057 0, 2058 "aaaaddbcdaabcyzc", 2059 qr/a (?{$count_a++;}) (a?) (*MARK:T1) (a*) .*? (bc){0,6} (*MARK:T1) (*SKIP:T1) (*FAIL) \s* c \1 /x, 2060 qr/a (?{$count_b++;}) (a?) (*MARK:T1) (a*) .*? (bc){0,6} (*MARK:T1) (*SKIP:T1) z \s* c \1 /x, 2061 2, 2062 ], 2063 [ 2064 0, 2065 "aaaaddbcdaabcyzc", 2066 qr/a (?{$count_a++;}) (a?) (*MARK:T1) (a*) .*? (bc*){0,6} (*MARK:T1) (*SKIP:T1) (*FAIL) \s* c \1 /x, 2067 qr/a (?{$count_b++;}) (a?) (*MARK:T1) (a*) .*? (bc*){0,6} (*MARK:T1) (*SKIP:T1) z \s* c \1 /x, 2068 2, 2069 ], 2070 2071 2072 [ 2073 0, 2074 "AbcdCBefgBhiBqz", 2075 qr/(A (.*) (?{ $count_a++ }) C? (*THEN) | A D) (*FAIL)/x, 2076 qr/(A (.*) (?{ $count_b++ }) C? (*THEN) | A D) z/x, 2077 1, 2078 ], 2079 [ 2080 0, 2081 "AbcdCBefgBhiBqz", 2082 qr/(A (.*) (?{ $count_a++ }) C? (*THEN) | A D) \s* (*FAIL)/x, 2083 qr/(A (.*) (?{ $count_b++ }) C? (*THEN) | A D) \s* z/x, 2084 1, 2085 ], 2086 [ 2087 0, 2088 "AbcdCBefgBhiBqz", 2089 qr/(A (.*) (?{ $count_a++ }) (?:C|) (*THEN) | A D) (*FAIL)/x, 2090 qr/(A (.*) (?{ $count_b++ }) (?:C|) (*THEN) | A D) z/x, 2091 1, 2092 ], 2093 [ 2094 0, 2095 "AbcdCBefgBhiBqz", 2096 qr/(A (.*) (?{ $count_a++ }) C{0,6} (*THEN) | A D) (*FAIL)/x, 2097 qr/(A (.*) (?{ $count_b++ }) C{0,6} (*THEN) | A D) z/x, 2098 1, 2099 ], 2100 [ 2101 0, 2102 "AbcdCEBefgBhiBqz", 2103 qr/(A (.*) (?{ $count_a++ }) (CE){0,6} (*THEN) | A D) (*FAIL)/x, 2104 qr/(A (.*) (?{ $count_b++ }) (CE){0,6} (*THEN) | A D) z/x, 2105 1, 2106 ], 2107 [ 2108 0, 2109 "AbcdCBefgBhiBqz", 2110 qr/(A (.*) (?{ $count_a++ }) (CE*){0,6} (*THEN) | A D) (*FAIL)/x, 2111 qr/(A (.*) (?{ $count_b++ }) (CE*){0,6} (*THEN) | A D) z/x, 2112 1, 2113 ], 2114 ) { 2115 $c++; 2116 $count_a = 0; 2117 $count_b = 0; 2118 2119 my $match_a = ($re->[1] =~ $re->[2]) || 0; 2120 my $match_b = ($re->[1] =~ $re->[3]) || 0; 2121 2122 is($match_a, $re->[0], "match a " . ($re->[0] ? "succeeded" : "failed") . " ($c)"); 2123 is($match_b, $re->[0], "match b " . ($re->[0] ? "succeeded" : "failed") . " ($c)"); 2124 is($count_a, $re->[4], "count a ($c)"); 2125 is($count_b, $re->[4], "count b ($c)"); 2126 } 2127 } 2128 2129 { # Bleadperl v5.13.8-292-gf56b639 breaks NEZUMI/Unicode-LineBreak-1.011 2130 # \xdf in lookbehind failed to compile as is multi-char fold 2131 my $message = "Lookbehind with \\xdf matchable compiles"; 2132 my $r = eval 'qr{ 2133 (?u: (?<=^url:) | 2134 (?<=[/]) (?=[^/]) | 2135 (?<=[^-.]) (?=[-~.,_?\#%=&]) | 2136 (?<=[=&]) (?=.) 2137 )}iox'; 2138 is($@, '', $message); 2139 object_ok($r, 'Regexp', $message); 2140 } 2141 2142 # RT #82610 2143 like 'foo/file.fob', qr,^(?=[^\.])[^/]*/(?=[^\.])[^/]*\.fo[^/]$,; 2144 2145 { # This was failing unless an explicit /d was added 2146 my $E0 = uni_to_native("\xE0"); 2147 utf8::upgrade($E0); 2148 my $p = qr/[_$E0]/i; 2149 like(uni_to_native("\xC0"), qr/$p/, "Verify \"\\xC0\" =~ /[\\xE0_]/i; pattern in utf8"); 2150 } 2151 2152 like "x", qr/\A(?>(?:(?:)A|B|C?x))\z/, 2153 "Check TRIE does not overwrite EXACT following NOTHING at start - RT #111842"; 2154 2155 { 2156 my $single = "z"; 2157 my $upper = "\x{390}"; # Fold is 3 chars. 2158 my $multi = CORE::fc($upper); 2159 2160 my $failed = 0; 2161 2162 # Try forcing a node to be split, with a multi-char fold at the 2163 # boundary 2164 for my $repeat (1 .. 300) { 2165 my $string = $single x $repeat; 2166 my $lhs = $string . $upper; 2167 if ($lhs !~ m/$string$multi/i) { 2168 $failed = $repeat; 2169 last; 2170 } 2171 } 2172 ok(! $failed, "Matched multi-char fold across EXACTFish node boundaries; if failed, was at count $failed"); 2173 2174 $failed = 0; 2175 for my $repeat (1 .. 300) { 2176 my $string = $single x $repeat; 2177 my $lhs = $string . "\N{LATIN SMALL LIGATURE FFI}"; 2178 if ($lhs !~ m/${string}ff\N{LATIN SMALL LETTER I}/i) { 2179 $failed = $repeat; 2180 last; 2181 } 2182 } 2183 ok(! $failed, "Matched multi-char fold across EXACTFish node boundaries; if failed, was at count $failed"); 2184 2185 $failed = 0; 2186 for my $repeat (1 .. 300) { 2187 my $string = $single x $repeat; 2188 my $lhs = $string . "\N{LATIN SMALL LIGATURE FFL}"; 2189 if ($lhs !~ m/${string}ff\N{U+6c}/i) { 2190 $failed = $repeat; 2191 last; 2192 } 2193 } 2194 ok(! $failed, "Matched multi-char fold across EXACTFish node boundaries; if failed, was at count $failed"); 2195 2196 # This tests that under /d matching that an 'ss' split across two 2197 # parts of a node doesn't end up turning into something that matches 2198 # \xDF unless it is in utf8. 2199 $failed = 0; 2200 $single = 'a'; # Is non-terminal multi-char fold char 2201 for my $repeat (1 .. 300) { 2202 my $string = $single x $repeat; 2203 my $lhs = "$string\N{LATIN SMALL LETTER SHARP S}"; 2204 utf8::downgrade($lhs); 2205 $string .= "s"; 2206 if ($lhs =~ m/${string}s/di) { 2207 $failed = $repeat; 2208 last; 2209 } 2210 } 2211 ok(! $failed, "Matched multi-char fold 'ss' across EXACTF node boundaries; if failed, was at count $failed"); 2212 2213 for my $non_finals ("t", "ft", "ift", "sift") { 2214 my $base_pat = $non_finals . "enKalt"; # (The tail is taken from 2215 # the trouble ticket, is 2216 # arbitrary) 2217 for my $utf8 ("non-UTF-8", "UTF-8") { 2218 2219 # Try at different lengths to be sure to get a node boundary 2220 for my $repeat (120 .. 270) { # [perl #133756] 2221 my $head = ("b" x $repeat) . "\xDC"; 2222 my $pat = $base_pat; 2223 utf8::upgrade($pat) if $utf8 eq "UTF-8"; 2224 $pat = $head . $pat; 2225 my $text = $head . $base_pat; 2226 2227 if ($text !~ /$pat/i) { 2228 $failed = $repeat; 2229 last; 2230 } 2231 } 2232 2233 ok(! $failed, "A non-final fold character " 2234 . (length($non_finals) - 1) 2235 . " characters from the end of an EXACTFish" 2236 . " $utf8 pattern works; if failed, was at count $failed"); 2237 } 2238 } 2239 } 2240 2241 { 2242 fresh_perl_is('print eval "\"\x{101}\" =~ /[[:lower:]]/", "\n"; print eval "\"\x{100}\" =~ /[[:lower:]]/i", "\n";', 2243 "1\n1", # Both re's should match 2244 {}, 2245 "get [:lower:] swash in first eval; test under /i in second"); 2246 } 2247 2248 { 2249 fresh_perl_is(<<'EOF', 2250 my $s = "\x{41c}"; 2251 $s =~ /(.*)/ or die; 2252 $ls = lc $1; 2253 print $ls eq lc $s ? "good\n" : "bad: [$ls]\n"; 2254EOF 2255 "good\n", 2256 {}, 2257 "swash triggered by lc() doesn't corrupt \$1" 2258 ); 2259 } 2260 2261 { 2262 #' RT #119075 2263 no warnings 'regexp'; # Silence "has useless greediness modifier" 2264 local $@; 2265 eval { /a{0}?/; }; 2266 ok(! $@, 2267 "PCRE regression test: No 'Quantifier follows nothing in regex' warning"); 2268 2269 } 2270 2271 { 2272 unlike("\xB5", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada"); 2273 like("\xB6", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada"); 2274 unlike("\xB7", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada"); 2275 like("\xB5", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada"); 2276 unlike("\xB6", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada"); 2277 like("\xB7", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada"); 2278 2279 unlike("_\xB5", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada"); 2280 like("_\xB6", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada"); 2281 unlike("_\xB7", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada"); 2282 like("_\xB5", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada"); 2283 unlike("_\xB6", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada"); 2284 like("_\xB7", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada"); 2285 } 2286 2287 # These are defined later, so won't be known at regex compile time above 2288 sub IsMyRuntimeProperty { 2289 return "B6\n"; 2290 } 2291 2292 sub IsntMyRuntimeProperty { 2293 return "!B6\n"; 2294 } 2295 2296 { # [perl 121777] 2297 my $regex; 2298 { package Some; 2299 # define a Unicode propertyIs_q 2300 sub Is_q 2301 { 2302 sprintf '%x', ord 'q' 2303 } 2304 $regex = qr/\p{Is_q}/; 2305 2306 # If we uncomment the following line, prior to the patch that 2307 # fixed this, everything would work because we would have expanded 2308 # the property by the time the regex in the 'like' below got 2309 # compiled. 2310 #'q' =~ $regex; 2311 } 2312 2313 like('q', $regex, 'User-defined property matches outside package'); 2314 2315 package Some { 2316 main::like('abcq', qr/abc$regex/, 'Run-time compiled in-package user-defined property matches'); 2317 } 2318 } 2319 2320 { # From Lingua::Stem::UniNE; no ticket filed but related to #121778 2321 use utf8; 2322 my $word = '����������'; 2323 $word =~ s{ (?: 2324 ���� # definite articles for nouns: 2325 | ���� # ��� masculine 2326 | ���� # ��� feminine 2327 | ���� # ��� neutral 2328 | ���� # ��� plural 2329 ) $ }{}x; 2330 is($word, '������', "Handles UTF8 trie correctly"); 2331 } 2332 2333 { # [perl #122460] 2334 my $a = "rdvark"; 2335 $a =~ /(?{})(?=[A-Za-z0-9_])a*?/g; 2336 is (pos $a, 0, "optimizer correctly thinks (?=...) is 0-length"); 2337 } 2338 2339 { # [perl #123417] multi-char \N{...} tripping roundly 2340 use Cname; 2341 my $qr = qr$(\N{foo})$; 2342 "afoot" =~ eval "qr/$qr/"; 2343 is "$1" || $@, "foo", 'multichar \N{...} stringified and retoked'; 2344 } 2345 2346 is (scalar split(/\b{sb}/, "Don't think twice. It's all right."), 2347 2, '\b{wb} splits sentences correctly'); 2348 2349 ok "my/dir/audio_07.mp3" =~ 2350 qr/(.*)\/(.*)\/(.*)\.(?<=(?=(?:\.(?!\d+\b)\w{1,4}$)$)\.)(.*)$()/, 2351 "[perl #133948]"; 2352 2353 2354 # !!! NOTE! Keep the following tests last -- they may crash perl 2355 2356 print "# Tests that follow may crash perl\n"; 2357 { 2358 eval '/\k/'; 2359 like $@, qr/\QSequence \k... not terminated in regex;\E/, 2360 'Lone \k not allowed'; 2361 } 2362 2363 { 2364 my $message = "Substitution with lookahead (possible segv)"; 2365 $_ = "ns1ns1ns1"; 2366 s/ns(?=\d)/ns_/g; 2367 is($_, "ns_1ns_1ns_1", $message); 2368 $_ = "ns1"; 2369 s/ns(?=\d)/ns_/; 2370 is($_, "ns_1", $message); 2371 $_ = "123"; 2372 s/(?=\d+)|(?<=\d)/!Bang!/g; 2373 is($_, "!Bang!1!Bang!2!Bang!3!Bang!", $message); 2374 } 2375 2376 { 2377 # Earlier versions of Perl said this was fatal. 2378 my $message = "U+0FFFF shouldn't crash the regex engine"; 2379 no warnings 'utf8'; 2380 my $a = eval "chr(65535)"; 2381 use warnings; 2382 my $warning_message; 2383 local $SIG{__WARN__} = sub { $warning_message = $_[0] }; 2384 eval $a =~ /[a-z]/; 2385 ok(1, $message); # If it didn't crash, it worked. 2386 } 2387 2388 { # Was looping 2389 watchdog(10); 2390 like("\x{00DF}", qr/[\x{1E9E}_]*/i, "\"\\x{00DF}\" =~ /[\\x{1E9E}_]*/i was looping"); 2391 watchdog(0); 2392 } 2393 2394 { # Bug #90536, caused failed assertion 2395 unlike("s\N{U+DF}", qr/^\x{00DF}/i, "\"s\\N{U+DF}\", qr/^\\x{00DF}/i"); 2396 } 2397 2398 # User-defined Unicode properties to match above-Unicode code points 2399 sub Is_31_Bit_Super { return "110000\t7FFFFFFF\n" } 2400 sub Is_Portable_Super { return '!utf8::Any' } # Matches beyond 32 bits 2401 2402 { # Assertion was failing on on 64-bit platforms; just didn't work on 32. 2403 no warnings qw(non_unicode portable); 2404 use Config; 2405 2406 # We use 'ok' instead of 'like' because the warnings are lexically 2407 # scoped, and want to turn them off, so have to do the match in this 2408 # scope. 2409 if ($Config{uvsize} < 8) { 2410 like(chr(0x7FFF_FFFE), qr/\p{Is_31_Bit_Super}/, 2411 "chr(0x7FFF_FFFE) can match a Unicode property"); 2412 like(chr(0x7FFF_FFFF), qr/\p{Is_31_Bit_Super}/, 2413 "chr(0x7FFF_FFFF) can match a Unicode property"); 2414 my $p = qr/^[\x{7FFF_FFFF}]$/; 2415 like(chr(0x7FFF_FFFF), qr/$p/, 2416 "chr(0x7FFF_FFFF) can match itself in a [class]"); 2417 like(chr(0x7FFF_FFFF), qr/$p/, # Tests any caching 2418 "chr(0x7FFF_FFFF) can match itself in a [class] subsequently"); 2419 } 2420 else { 2421 no warnings 'overflow'; 2422 like(chr(0x7FFF_FFFF_FFFF_FFFE), qr/\p{Is_Portable_Super}/, 2423 "chr(0x7FFF_FFFF_FFFF_FFFE) can match a Unicode property"); 2424 like(chr(0x7FFF_FFFF_FFFF_FFFF), qr/^\p{Is_Portable_Super}$/, 2425 "chr(0x7FFF_FFFF_FFFF_FFFF) can match a Unicode property"); 2426 2427 my $p = eval 'qr/^\x{7FFF_FFFF_FFFF_FFFF}$/'; 2428 like(chr(0x7FFF_FFFF_FFFF_FFFF), qr/$p/, 2429 "chr(0x7FFF_FFFF_FFFF_FFFF) can match itself in a [class]"); 2430 like(chr(0x7FFF_FFFF_FFFF_FFFF), qr/$p/, # Tests any caching 2431 "chr(0x7FFF_FFFF_FFFF_FFFF) can match itself in a [class] subsequently"); 2432 2433 # This test is because something was declared as 32 bits, but 2434 # should have been cast to 64; only a problem where 2435 # sizeof(STRLEN) != sizeof(UV) 2436 unlike(chr(0x7FFF_FFFF_FFFF_FFFE), qr/\p{Is_31_Bit_Super}/, 2437 "chr(0x7FFF_FFFF_FFFF_FFFE) shouldn't match a range ending in 0x7FFF_FFFF"); 2438 } 2439 } 2440 2441 { # [perl #112530], the code below caused a panic 2442 sub InFoo { "a\tb\n9\ta\n" } 2443 like(chr(0xA), qr/\p{InFoo}/, 2444 "Overlapping ranges in user-defined properties"); 2445 } 2446 2447 { # [perl #125990], the final 2 tests below each caused a panic. 2448 # The \0's are not necessary; it could be a printable character 2449 # instead, but were in the ticket, so using them. 2450 my $sharp_s = chr utf8::unicode_to_native(0xdf); 2451 my $string = ("\0" x 8) 2452 . ($sharp_s x 3) 2453 . ("\0" x 42) 2454 . "��"; 2455 my $folded_string = ("\0" x 8) 2456 . ("ss" x 3) 2457 . ("\0" x 42) 2458 . "��"; 2459 utf8::downgrade($string); 2460 utf8::downgrade($folded_string); 2461 2462 use Cname; 2463 like($string, qr/$string/i, "LATIN SMALL SHARP S matches itself under /id"); 2464 unlike($folded_string, qr/$string/i, "LATIN SMALL SHARP S doesn't match 'ss' under /di"); 2465 like($folded_string, qr/\N{EMPTY-STR}$string/i, "\\N{} earlier than LATIN SMALL SHARP S transforms /di into /ui, matches 'ss'"); 2466 like($folded_string, qr/$string\N{EMPTY-STR}/i, "\\N{} after LATIN SMALL SHARP S transforms /di into /ui, matches 'ss'"); 2467 } 2468 2469 { # [perl #126606 crashed the interpreter 2470 use Cname; 2471 like("sS", qr/\N{EMPTY-STR}Ss|/i, '\N{} with empty branch alternation works'); 2472 like("sS", qr'\N{EMPTY-STR}Ss|'i, '\N{} with empty branch alternation works'); 2473 } 2474 2475 { # Regexp:Grammars was broken: 2476 # http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2013-06/msg01290.html 2477 fresh_perl_like('use warnings; "abc" =~ qr{(?&foo){0}abc(?<foo>)}', 2478 qr/Quantifier unexpected on zero-length expression/, 2479 {}, 2480 'No segfault on qr{(?&foo){0}abc(?<foo>)}'); 2481 } 2482 2483 SKIP: 2484 { # [perl #125826] buffer overflow in TRIE_STORE_REVCHAR 2485 # (during compilation, so use a fresh perl) 2486 $Config{uvsize} == 8 2487 or skip("need large code-points for this test", 1); 2488 2489 fresh_perl_is('/\x{E000000000}|/ and print qq(ok\n)', "ok\n", {}, 2490 "buffer overflow in TRIE_STORE_REVCHAR"); 2491 } 2492 2493 { 2494 fresh_perl_like('use warnings; s 0(?(?!00000000000000000000000000��000000)\500000000 0000000000000000000000000000000000000000000000000000��00000000000000000000000000000000 0', 2495 qr/Switch \(\?\(condition\)\.\.\. not terminated/, 2496 {}, 2497 'No segfault [perl #126886]'); 2498 } 2499 2500 { 2501 # [perl 130010] Downstream application texinfo started to report panics 2502 # as of commit a5540cf. 2503 2504 runperl( prog => 'A::xx(); package A; sub InFullwidth{ return qq|\n| } sub xx { split /[^\s\p{InFullwidth}]/, q|x| }' ); 2505 ok(! $?, "User-defined pattern did not cause panic [perl 130010]"); 2506 } 2507 2508 { # [perl #133999] Previously assertion failure 2509 fresh_perl_like('0 =~ /\p{nv:(\B(*COMMIT)C+)}/', 2510 qr/No Unicode property value wildcard matches/, 2511 {}, 2512 "Assertion failure with *COMMIT and wildcard property"); 2513 } 2514 2515 { # [perl #134029] Previously assertion failure 2516 fresh_perl_like('qr/\p{upper:]}|\337(?|ss)|)(?0/', 2517 qr/Unicode property wildcard not terminated/, 2518 {}, 2519 "Assertion failure with single character wildcard"); 2520 } 2521 2522 { # [perl #134034] Previously assertion failure 2523 fresh_perl_is('use utf8; q!��������������������������������������!=~/(?li)\b{wb}\B(*COMMIT)0/;', 2524 "", {}, "*COMMIT caused positioning beyond EOS"); 2525 } 2526 2527 { # [GH #17486] Previously assertion failure 2528 fresh_perl_is('0=~/(?iaa)ss\337(?0)|/', 2529 "", {}, "EXACTFUP node isn't changed into something else"); 2530 } 2531 2532 { # [GH #17593] 2533 fresh_perl_is('qr/((?+2147483647))/', 2534 "Invalid reference to group in regex; marked by <--" 2535 . " HERE in m/((?+2147483647) <-- HERE )/ at - line 1.", 2536 {}, "integer overflow, undefined behavior in ASAN"); 2537 fresh_perl_is('qr/((?-2147483647))/', 2538 "Reference to nonexistent group in regex; marked by <--" 2539 . " HERE in m/((?-2147483647) <-- HERE )/ at - line 1.", 2540 {}, "Large negative relative capture group"); 2541 fresh_perl_is('qr/((?+18446744073709551615))/', 2542 "Invalid reference to group in regex; marked by <--" 2543 . " HERE in m/((?+18446744073709551615 <-- HERE ))/ at -" 2544 . " line 1.", 2545 {}, "Too large relative group number"); 2546 fresh_perl_is('qr/((?-18446744073709551615))/', 2547 "Invalid reference to group in regex; marked by <--" 2548 . " HERE in m/((?-18446744073709551615 <-- HERE ))/ at -" 2549 . " line 1.", 2550 {}, "Too large negative relative group number"); 2551 } 2552 2553 { # GH #17734, ASAN use after free 2554 fresh_perl_like('no warnings "experimental::uniprop_wildcards"; 2555 my $re = q<[[\p{name=/[Y-]+Z/}]]>; 2556 eval { "\N{BYZANTINE MUSICAL SYMBOL PSILI}" 2557 =~ /$re/ }; print $@ if $@; print "Done\n";', 2558 qr/Done/, 2559 {}, "GH #17734"); 2560 } 2561 2562 { # GH $17278 assertion fails 2563 fresh_perl_is('use locale; 2564 my $A_grave = "\N{LATIN CAPITAL LETTER A WITH GRAVE}"; 2565 my $a_grave = "\N{LATIN SMALL LETTER A WITH GRAVE}"; 2566 2567 my $z="q!$a_grave! =~ m!(?^i)[$A_grave]!"; 2568 print eval $z, "\n";', 2569 1, 2570 {}, "GH #17278"); 2571 } 2572 2573 for my $try ( 1 .. 10 ) { 2574 # GH $19350 assertion fails - run 10 times as this bug is a heisenbug 2575 # and does not always fail, but should fail at least once in 10 tries. 2576 fresh_perl_is('use re Debug=>"ALL";qr{(?{a})(?<b>\g{c}})', 2577 <<'EOF_DEBUG_OUT', 2578Assembling pattern from 2 elements 2579Compiling REx "(?{a})(?<b>\g{c}" 2580Starting parse and generation 2581<(?{a})(?<b>>...| 1| reg 2582 | | brnc 2583 | | piec 2584 | | atom 2585<?{a})(?<b>\>...| | reg 2586<(?<b>\g{c}> | 4| piec 2587 | | atom 2588<?<b>\g{c}> | | reg 2589 | | Setting open paren #1 to 4 2590<\g{c}> | 6| brnc 2591 | | piec 2592 | | atom 2593<> | 9| tail~ OPEN1 'b' (4) -> REFN 2594 | | Setting close paren #1 to 9 2595 | 11| lsbr~ tying lastbr REFN <1> (6) to ender CLOSE1 'b' (9) offset 3 2596 | | tail~ REFN <1> (6) -> CLOSE 2597Unmatched ( in regex; marked by <-- HERE in m/(?{a})( <-- HERE ?<b>\g{c}/ at - line 1. 2598Freeing REx: "(?{a})(?<b>\g{c}" 2599EOF_DEBUG_OUT 2600 {rtrim_result=>1}, 2601 "Github Issue #19350, assert fail in " 2602 . "Debug => 'ALL' from malformed qr// (heisenbug try $try)"); 2603 } 2604 { # Related to GH $19350 but segfaults instead of asserts, and does so reliably, not randomly. 2605 # use re Debug => "PARSE" is similar to "ALL", but does not include the optimize info, so we 2606 # do not need to deal with normlazing memory addresses in the output. 2607 fresh_perl_is( 2608 'use re Debug=>"PARSE";qr{(?<b>\g{c})(?<c>x)(?&b)}', 2609 <<'EOF_DEBUG_OUT', 2610Assembling pattern from 1 elements 2611Compiling REx "(?<b>\g{c})(?<c>x)(?&b)" 2612Starting parse and generation 2613<(?<b>\g{c})>...| 1| reg 2614 | | brnc 2615 | | piec 2616 | | atom 2617<?<b>\g{c})(>...| | reg 2618<\g{c})(?<c>>...| 3| brnc 2619 | | piec 2620 | | atom 2621<)(?<c>x)(?&b)> | 6| tail~ OPEN1 'b' (1) -> REFN 2622 | 8| lsbr~ tying lastbr REFN <1> (3) to ender CLOSE1 'b' (6) offset 3 2623 | | tail~ REFN <1> (3) -> CLOSE 2624<(?<c>x)(?&b)> | | piec 2625 | | atom 2626<?<c>x)(?&b)> | | reg 2627<x)(?&b)> | 10| brnc 2628 | | piec 2629 | | atom 2630<)(?&b)> | 12| tail~ OPEN2 'c' (8) -> EXACT 2631 | 14| lsbr~ tying lastbr EXACT <x> (10) to ender CLOSE2 'c' (12) offset 2 2632 | | tail~ EXACT <x> (10) -> CLOSE 2633<(?&b)> | | tail~ OPEN1 'b' (1) 2634 | | ~ REFN <1> (3) 2635 | | ~ CLOSE1 'b' (6) -> OPEN 2636 | | piec 2637 | | atom 2638<?&b)> | | reg 2639<> | 17| tail~ OPEN2 'c' (8) 2640 | | ~ EXACT <x> (10) 2641 | | ~ CLOSE2 'c' (12) -> GOSUB 2642 | 18| lsbr~ tying lastbr OPEN1 'b' (1) to ender END (17) offset 16 2643 | | tail~ OPEN1 'b' (1) 2644 | | ~ REFN <1> (3) 2645 | | ~ CLOSE1 'b' (6) 2646 | | ~ OPEN2 'c' (8) 2647 | | ~ EXACT <x> (10) 2648 | | ~ CLOSE2 'c' (12) 2649 | | ~ GOSUB1[+0:14] 'b' (14) -> END 2650Need to redo parse 2651Freeing REx: "(?<b>\g{c})(?<c>x)(?&b)" 2652Starting parse and generation 2653<(?<b>\g{c})>...| 1| reg 2654 | | brnc 2655 | | piec 2656 | | atom 2657<?<b>\g{c})(>...| | reg 2658<\g{c})(?<c>>...| 3| brnc 2659 | | piec 2660 | | atom 2661<)(?<c>x)(?&b)> | 6| tail~ OPEN1 'b' (1) -> REFN 2662 | 8| lsbr~ tying lastbr REFN2 'c' <1> (3) to ender CLOSE1 'b' (6) offset 3 2663 | | tail~ REFN2 'c' <1> (3) -> CLOSE 2664<(?<c>x)(?&b)> | | piec 2665 | | atom 2666<?<c>x)(?&b)> | | reg 2667<x)(?&b)> | 10| brnc 2668 | | piec 2669 | | atom 2670<)(?&b)> | 12| tail~ OPEN2 'c' (8) -> EXACT 2671 | 14| lsbr~ tying lastbr EXACT <x> (10) to ender CLOSE2 'c' (12) offset 2 2672 | | tail~ EXACT <x> (10) -> CLOSE 2673<(?&b)> | | tail~ OPEN1 'b' (1) 2674 | | ~ REFN2 'c' <1> (3) 2675 | | ~ CLOSE1 'b' (6) -> OPEN 2676 | | piec 2677 | | atom 2678<?&b)> | | reg 2679<> | 17| tail~ OPEN2 'c' (8) 2680 | | ~ EXACT <x> (10) 2681 | | ~ CLOSE2 'c' (12) -> GOSUB 2682 | 18| lsbr~ tying lastbr OPEN1 'b' (1) to ender END (17) offset 16 2683 | | tail~ OPEN1 'b' (1) 2684 | | ~ REFN2 'c' <1> (3) 2685 | | ~ CLOSE1 'b' (6) 2686 | | ~ OPEN2 'c' (8) 2687 | | ~ EXACT <x> (10) 2688 | | ~ CLOSE2 'c' (12) 2689 | | ~ GOSUB1[+0:14] 'b' (14) -> END 2690Required size 17 nodes 2691first at 3 2692Freeing REx: "(?<b>\g{c})(?<c>x)(?&b)" 2693EOF_DEBUG_OUT 2694 {rtrim_result=>1}, 2695 "Related to Github Issue #19350, forward \\g{x} pattern segv under use re Debug => 'PARSE'"); 2696 } 2697 2698 { # perl-security#140, read/write past buffer end 2699 fresh_perl_like('qr/\p{utf8::perl x}/', 2700 qr/Illegal user-defined property name "utf8::perl x" in regex/, 2701 {}, "perl-security#140"); 2702 fresh_perl_is('qr/\p{utf8::_perl_surrogate}/', "", 2703 {}, "perl-security#140"); 2704 } 2705 2706 { # GH 20009 2707 my $x = "awesome quotes"; 2708 utf8::upgrade($x); 2709 $x =~ s/^[\x{0301}\x{030C}]+//; 2710 } 2711 2712 2713 # !!! NOTE that tests that aren't at all likely to crash perl should go 2714 # a ways above, above these last ones. There's a comment there that, like 2715 # this comment, contains the word 'NOTE' 2716 2717 done_testing(); 2718} # End of sub run_tests 2719 27201; 2721