#!./perl
#
-# This is a home for regular expression tests that don't fit into
+# This is a home for regular expression tests that do not fit into
# the format supported by re/regexp.t. If you want to add a test
# that does fit that format, add it to re/re_tests, not here.
BEGIN {
chdir 't' if -d 't';
@INC = ('../lib','.');
- require './test.pl';
+ require './test.pl'; require './charset_tools.pl';
skip_all_if_miniperl("miniperl can't load Tie::Hash::NamedCapture, need for %+ and %-");
}
# should not match.
ok "\x8e" =~ /[\x89-\x91]/, '"\x8e" =~ /[\x89-\x91]/';
ok "\xce" =~ /[\xc9-\xd1]/, '"\xce" =~ /[\xc9-\xd1]/';
+ ok "\xd0" =~ /[\xc9-\xd1]/, '"\xd0" =~ /[\xc9-\xd1]/';
skip "Not an EBCDIC platform", 2 unless ord ('i') == 0x89 &&
ord ('J') == 0xd1;
# the good of EBCDIC, so let's test these only there.
unlike("\x8e", qr/[i-j]/, '"\x8e" !~ /[i-j]/');
unlike("\xce", qr/[I-J]/, '"\xce" !~ /[I-J]/');
+ unlike("\xd0", qr/[I-J]/, '"\xd0" !~ /[I-J]/');
}
{
ok "<\x{100}\x{00A0}>" =~ /<\x{100}\s>/, '\x{00A0} in \s';
ok "<\x{00A0}>" =~ /<\h>/, '\x{00A0} in \h';
}
- my @h = map {sprintf "%05x" => $_} 0x01680, 0x0180E, 0x02000 .. 0x0200A,
+ my @h = map {sprintf "%05x" => $_} 0x01680, 0x02000 .. 0x0200A,
0x0202F, 0x0205F, 0x03000;
my @v = map {sprintf "%05x" => $_} 0x02028, 0x02029;
my @H = map {sprintf "%05x" => $_} 0x01361, 0x0200B, 0x02408, 0x02420,
- 0x0303F, 0xE0020;
+ 0x0303F, 0xE0020, 0x180E;
my @V = map {sprintf "%05x" => $_} 0x0008A .. 0x0008D, 0x00348, 0x10100,
- 0xE005F, 0xE007C;
+ 0xE005F, 0xE007C, 0x180E;
for my $hex (@h) {
my $str = eval qq ["<\\x{$hex}>"];
use Cname;
ok 'fooB' =~ /\N{foo}[\N{B}\N{b}]/, "Passthrough charname";
+ my $name = "foo\xDF";
+ my $result = eval "'A${name}B' =~ /^A\\N{$name}B\$/";
+ ok !$@ && $result, "Passthrough charname of non-ASCII, Latin1";
#
# Why doesn't must_warn work here?
#
ok "\N{LONG-STR}" =~ /^\N{LONG-STR}$/, 'Verify that long string works';
ok "\N{LONG-STR}" =~ /^\N{LONG-STR}$/i, 'Verify under folding that long string works';
+ eval '/(?[[\N{EMPTY-STR}]])/';
+ ok $@ && $@ =~ /Zero length \\N\{}/;
+
+ undef $w;
+ {
+ () = eval q ["\N{TOO MANY SPACES}"];
+ like ($@, qr/charnames alias definitions may not contain a sequence of multiple spaces/, "Multiple spaces in a row in a charnames alias is fatal");
+ eval q [use utf8; () = "\N{TOO MANY SPACES}"];
+ like ($@, qr/charnames alias definitions may not contain a sequence of multiple spaces/, "... same under utf8");
+ }
+
+ undef $w;
+ {
+ () = eval q ["\N{TRAILING SPACE }"];
+ like ($@, qr/charnames alias definitions may not contain trailing white-space/, "Trailing white-space in a charnames alias is fatal");
+ eval q [use utf8; () = "\N{TRAILING SPACE }"];
+ like ($@, qr/charnames alias definitions may not contain trailing white-space/, "... same under utf8");
+ }
+
+ undef $w;
+ my $Cedilla_Latin1 = "GAR"
+ . latin1_to_native("\xC7")
+ . "ON";
+ my $Cedilla_utf8 = $Cedilla_Latin1;
+ utf8::upgrade($Cedilla_utf8);
+ eval qq[is("\\N{$Cedilla_Latin1}", "$Cedilla_Latin1", "A cedilla in character name works")];
+ undef $w;
+ {
+ use feature 'unicode_eval';
+ eval qq[use utf8; is("\\N{$Cedilla_utf8}", "$Cedilla_utf8", "... same under 'use utf8': they work")];
+ }
+
+ undef $w;
+ my $NBSP_Latin1 = "NBSP"
+ . latin1_to_native("\xA0")
+ . "SEPARATED"
+ . latin1_to_native("\xA0")
+ . "SPACE";
+ my $NBSP_utf8 = $NBSP_Latin1;
+ utf8::upgrade($NBSP_utf8);
+ eval qq[is("\\N{$NBSP_Latin1}", "$NBSP_Latin1", "An NBSP in character name works")];
+ like ($w, qr/NO-BREAK SPACE in a charnames alias definition is deprecated/, "... but returns a deprecation warning");
+ undef $w;
+ {
+ use feature 'unicode_eval';
+ eval qq[use utf8; is("\\N{$NBSP_utf8}", "$NBSP_utf8", "Same under 'use utf8': they work")];
+ like ($w, qr/NO-BREAK SPACE in a charnames alias definition is deprecated/, "... but return a deprecation warning");
+ }
+ {
+ # disable lexical warnings
+ BEGIN { ${^WARNING_BITS} = undef; $^W = 0 }
+ undef $w;
+ () = eval qq["\\N{$NBSP_Latin1}"];
+ like ($w, qr/NO-BREAK SPACE in a charnames alias definition is deprecated/, "And returns a deprecation warning outside of lexical warnings");
+ undef $w;
+ use feature 'unicode_eval';
+ eval qq[use utf8; () = "\\N{$NBSP_utf8}"];
+ like ($w, qr/NO-BREAK SPACE in a charnames alias definition is deprecated/, "... same under utf8");
+ }
+ {
+ no warnings 'deprecated';
+ undef $w;
+ eval qq["\\N{$NBSP_Latin1}"];
+ ok (! defined $w, "... and no warning if warnings are off");
+ use feature 'unicode_eval';
+ eval qq[use utf8; "\\N{$NBSP_utf8}"];
+ ok (! defined $w, "... same under 'use utf8'");
+ }
+ {
+ use warnings FATAL=>'deprecated';
+ () = eval qq["\\N{$NBSP_Latin1}"];
+ like ($@, qr/NO-BREAK SPACE in a charnames alias definition is deprecated/, "... the warning can be fatal");
+ use feature 'unicode_eval';
+ eval qq[use utf8; () = "\\N{$NBSP_utf8}"];
+ like ($@, qr/NO-BREAK SPACE in a charnames alias definition is deprecated/, "... same under utf8");
+ }
+
+ {
+ BEGIN { no strict; *CnameTest:: = *{"_charnames\0A::" } }
+ package CnameTest { sub translator { pop } }
+ BEGIN { $^H{charnames} = \&CnameTest::translator }
+ undef $w;
+ () = eval q ["\N{TOO MANY SPACES}"];
+ like ($@, qr/charnames alias definitions may not contain a sequence of multiple spaces/,
+ 'translators in _charnames\0* packages get validated');
+ }
+
# If remove the limitation in regcomp code these should work
# differently
undef $w;
ok $@ && $@ =~ /Invalid character/, 'Verify that leading digit in name gives error';
eval 'q() =~ /\N{COM,MA}/';
ok $@ && $@ =~ /Invalid character/, 'Verify that comma in name gives error';
- my $name = "A\x{D7}O";
+ $name = "A\x{D7}O";
eval "q(W) =~ /\\N{$name}/";
ok $@ && $@ =~ /Invalid character/, 'Verify that latin1 symbol in name gives error';
+ my $utf8_name = "7 CITIES OF GOLD";
+ utf8::upgrade($utf8_name);
+ eval "use utf8; q(W) =~ /\\N{$utf8_name}/";
+ ok $@ && $@ =~ /Invalid character/, 'Verify that leading digit in utf8 name gives error';
+ $utf8_name = "SHARP #";
+ utf8::upgrade($utf8_name);
+ eval "use utf8; q(W) =~ /\\N{$utf8_name}/";
+ ok $@ && $@ =~ /Invalid character/, 'Verify that ASCII symbol in utf8 name gives error';
+ $utf8_name = "A HOUSE \xF7 AGAINST ITSELF";
+ utf8::upgrade($utf8_name);
+ eval "use utf8; q(W) =~ /\\N{$utf8_name}/";
+ ok $@ && $@ =~ /Invalid character/, 'Verify that latin1 symbol in utf8 name gives error';
+ $utf8_name = "\x{664} HORSEMEN}";
+ eval "use utf8; q(W) =~ /\\N{$utf8_name}/";
+ ok $@ && $@ =~ /Invalid character/, 'Verify that leading above Latin1 digit in utf8 name gives error';
+ $utf8_name = "A \x{1F4A9} WOULD SMELL AS SWEET}";
+ eval "use utf8; q(W) =~ /\\N{$utf8_name}/";
+ ok $@ && $@ =~ /Invalid character/, 'Verify that above Latin1 symbol in utf8 name gives error';
+
undef $w;
$name = "A\x{D1}O";
eval "q(W) =~ /\\N{$name}/";
ok ! $w, 'Verify that latin1 letter in name doesnt give warning';
+ # This tests the code path that restarts the parse when the recursive
+ # call to S_reg() from within S_grok_bslash_N() discovers that the
+ # pattern needs to be recalculated as UTF-8. use eval to avoid
+ # needing literal Unicode in this source file:
+ my $r = eval "qr/\\N{\x{100}\x{100}}/";
+ isnt $r, undef, "Generated regex for multi-char UTF-8 charname"
+ or diag($@);
+ ok "\x{100}\x{100}" =~ $r, "which matches";
}
{
1 while /(a+b?)(*COMMIT)(?{$count++; push @res,$1})(*FAIL)/g;
is($count, 1, "Expect 1 with (*COMMIT)");
is("@res", "aaab", "Adjacent (*COMMIT) works as expected");
+
+ ok("1\n2a\n" !~ /^\d+(*COMMIT)\w+/m, "COMMIT and anchors");
}
{
{
# Various whitespace special patterns
- my @h = map {chr $_} 0x09, 0x20, 0xa0, 0x1680, 0x180e, 0x2000,
+ my @h = map {chr $_} 0x09, 0x20, 0xa0, 0x1680, 0x2000,
0x2001, 0x2002, 0x2003, 0x2004, 0x2005, 0x2006,
0x2007, 0x2008, 0x2009, 0x200a, 0x202f, 0x205f,
0x3000;
{
# Test for keys in %+ and %-
my $message = 'Test keys in %+ and %-';
- no warnings 'uninitialized';
+ no warnings 'uninitialized', 'deprecated', 'experimental::lexical_topic';
my $_ = "abcdef";
/(?<foo>a)|(?<foo>b)/;
is((join ",", sort keys %+), "foo", $message);
{
# length() on captures, the numbered ones end up in Perl_magic_len
+ no warnings 'deprecated', 'experimental::lexical_topic';
my $_ = "aoeu \xe6var ook";
/^ \w+ \s (?<eek>\S+)/x;
'IsPunct disagrees with [:punct:] outside ASCII');
my @isPunctLatin1 = eval q {
- use encoding 'latin1';
- grep {/[[:punct:]]/ != /\p{IsPunct}/} map {chr} 0x80 .. 0xff;
+ grep {/[[:punct:]]/u != /\p{IsPunct}/} map {chr} 0x80 .. 0xff;
};
skip "Eval failed ($@)", 1 if $@;
skip "PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS set to 0", 1
{ # This was failing unless an explicit /d was added
my $p = qr/[\xE0_]/i;
utf8::upgrade($p);
- like("\xC0", $p, "Verify \"\\xC0\" =~ /[\\xE0_]/i; pattern in utf8");
+ like("\xC0", qr/$p/, "Verify \"\\xC0\" =~ /[\\xE0_]/i; pattern in utf8");
}
ok "x" =~ /\A(?>(?:(?:)A|B|C?x))\z/,
}
}
ok(! $failed, "Matched multi-char fold across EXACTFish node boundaries; if failed, was at count $failed");
+
+ # This tests that under /d matching that an 'ss' split across two
+ # parts of a node doesn't end up turning into something that matches
+ # \xDF unless it is in utf8.
+ $failed = 0;
+ $single = 'a'; # Is non-terminal multi-char fold char
+ for my $repeat (1 .. 300) {
+ my $string = $single x $repeat;
+ my $lhs = "$string\N{LATIN SMALL LETTER SHARP S}";
+ utf8::downgrade($lhs);
+ $string .= "s";
+ if ($lhs =~ m/${string}s/di) {
+ $failed = $repeat;
+ last;
+ }
+ }
+ ok(! $failed, "Matched multi-char fold 'ss' across EXACTF node boundaries; if failed, was at count $failed");
+ }
+
+ {
+ fresh_perl_is('print eval "\"\x{101}\" =~ /[[:lower:]]/", "\n"; print eval "\"\x{100}\" =~ /[[:lower:]]/i", "\n";',
+ "1\n1", # Both re's should match
+ {},
+ "get [:lower:] swash in first eval; test under /i in second");
+ }
+
+ {
+ #' RT #119075
+ no warnings 'regexp'; # Silence "has useless greediness modifier"
+ local $@;
+ eval { /a{0}?/; };
+ ok(! $@,
+ "PCRE regression test: No 'Quantifier follows nothing in regex' warning");
+
+ }
+
+ {
+ unlike("\xB5", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada");
+ like("\xB6", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada");
+ unlike("\xB7", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada");
+ like("\xB5", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada");
+ unlike("\xB6", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada");
+ like("\xB7", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada");
+
+ unlike("_\xB5", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada");
+ like("_\xB6", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada");
+ unlike("_\xB7", qr/^_?\p{IsMyRuntimeProperty}\z/, "yadayada");
+ like("_\xB5", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada");
+ unlike("_\xB6", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada");
+ like("_\xB7", qr/^_?\P{IsMyRuntimeProperty}\z/, "yadayada");
+ }
+
+ # These are defined later, so won't be known at regex compile time above
+ sub IsMyRuntimeProperty {
+ return "B6\n";
+ }
+
+ sub IsntMyRuntimeProperty {
+ return "!B6\n";
+ }
+
+ { # [perl 121777]
+ my $regex;
+ { package Some;
+ # define a Unicode propertyIs_q
+ sub Is_q
+ {
+ sprintf '%x', ord 'q'
+ }
+ $regex = qr/\p{Is_q}/;
+
+ # If we uncomment the following line, prior to the patch that
+ # fixed this, everything would work because we would have expanded
+ # the property by the time the regex in the 'like' below got
+ # compiled.
+ #'q' =~ $regex;
+ }
+
+ like('q', $regex, 'User-defined property matches outside package');
+
+ package Some {
+ main::like('abcq', qr/abc$regex/, 'Run-time compiled in-package user-defined property matches');
+ }
+ }
+
+ { # From Lingua::Stem::UniNE; no ticket filed but related to #121778
+ use utf8;
+ my $word = 'рабта';
+ $word =~ s{ (?:
+ ия # definite articles for nouns:
+ | ът # ∙ masculine
+ | та # ∙ feminine
+ | то # ∙ neutral
+ | те # ∙ plural
+ ) $ }{}x;
+ is($word, 'раб', "Handles UTF8 trie correctly");
}
#
my $p = qr/^[\x{FFFF_FFFF}]$/;
ok(chr(0xFFFF_FFFF) =~ $p,
"chr(0xFFFF_FFFF) can match itself in a [class]");
+ ok(chr(0xFFFF_FFFF) =~ $p, # Tests any caching
+ "chr(0xFFFF_FFFF) can match itself in a [class] subsequently");
}
else {
no warnings 'overflow';
my $p = qr/^[\x{FFFF_FFFF_FFFF_FFFF}]$/;
ok(chr(0xFFFF_FFFF_FFFF_FFFF) =~ $p,
"chr(0xFFFF_FFFF_FFFF_FFFF) can match itself in a [class]");
+ ok(chr(0xFFFF_FFFF_FFFF_FFFF) =~ $p, # Tests any caching
+ "chr(0xFFFF_FFFF_FFFF_FFFF) can match itself in a [class] subsequently");
# This test is because something was declared as 32 bits, but
# should have been cast to 64; only a problem where
"Overlapping ranges in user-defined properties");
}
+ { # Regexp:Grammars was broken:
+ # http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2013-06/msg01290.html
+ fresh_perl_like('use warnings; "abc" =~ qr{(?&foo){0}abc(?<foo>)}',
+ qr/Quantifier unexpected on zero-length expression/,
+ {},
+ 'No segfault on qr{(?&foo){0}abc(?<foo>)}');
+ }
+
# !!! NOTE that tests that aren't at all likely to crash perl should go
# a ways above, above these last ones.
https://perl5.git.perl.org / perl5.git / blobdiff