9 unless ($has_perlio = find PerlIO::Layer 'perlio') {
11 # Since you don't have perlio you might get failures with UTF-8 locales.
18 no utf8; # Ironic, no?
22 # Think carefully before adding tests here. In general this should be
23 # used only for about three categories of tests:
25 # (1) tests that absolutely require 'use utf8', and since that in general
26 # shouldn't be needed as the utf8 is being obsoleted, this should
27 # have rather few tests. If you want to test Unicode and regexes,
28 # you probably want to go to op/regexp or op/pat; if you want to test
29 # split, go to op/split; pack, op/pack; appending or joining,
30 # op/append or op/join, and so forth
32 # (2) tests that have to do with Unicode tokenizing (though it's likely
33 # that all the other Unicode tests sprinkled around the t/**/*.t are
34 # going to catch that)
36 # (3) complicated tests that simultaneously stress so many Unicode features
37 # that deciding into which other test script the tests should go to
38 # is hard -- maybe consider breaking up the complicated test
47 { use bytes; $a = "\xc3\xa4" }
48 { use utf8; $b = "\xe4" }
54 { use utf8; ok($a ne $b) }
61 my $smiley = "\x{263a}";
63 for my $s ("\x{263a}",
72 my $length_chars = length($s);
74 { use bytes; $length_bytes = length($s) }
75 my @regex_chars = $s =~ m/(.)/g;
76 my $regex_chars = @regex_chars;
77 my @split_chars = split //, $s;
78 my $split_chars = @split_chars;
79 ok("$length_chars/$regex_chars/$split_chars/$length_bytes" eq
83 for my $s ("\x{263a}" . "\x{263a}",
92 my $length_chars = length($s);
94 { use bytes; $length_bytes = length($s) }
95 my @regex_chars = $s =~ m/(.)/g;
96 my $regex_chars = @regex_chars;
97 my @split_chars = split //, $s;
98 my $split_chars = @split_chars;
99 ok("$length_chars/$regex_chars/$split_chars/$length_bytes" eq
107 local $SIG{__WARN__} = sub { print "#($_[0])\n"; $w++ };
108 my $x = eval q/"\\/ . "\x{100}" . q/"/;;
110 ok($w == 0 && $x eq "\x{100}");
117 $result .= '>' . join (',', map {ord} split //, $_) . '<'
123 eval $show or die $@; # We don't expect this sub definition to fail.
124 my $progfile = 'utf' . $$;
125 END {unlink_all $progfile}
127 # If I'm right 60 is '>' in ASCII, ' ' in EBCDIC
128 # 173 is not punctuation in either ASCII or EBCDIC
130 foreach (60, 173, 257, 65532) {
133 # I don't want to use map {ord} and I've no need to hardcode the UTF
135 my $charsubst = $char;
136 $charsubst =~ s/(.)/ord ($1) . ','/ge;
138 # Not testing this one against map {ord}
140 = join " . ", map {sprintf 'chr (%d)', ord $_} split //, $char;
141 push @char, [$_, $char, $charsubst, $char_as_ord];
143 # Now we've done all the UTF8 munching hopefully we're safe
145 ['check our detection program works',
146 'my @a = ("'.chr(60).'\x2A", ""); $b = show @a', qr/^>60,42<><$/],
147 ['check literal 8 bit input',
148 '$a = "' . chr (173) . '"; $b = show $a', qr/^>173<$/],
149 ['check no utf8; makes no change',
150 'no utf8; $a = "' . chr (173) . '"; $b = show $a', qr/^>173<$/],
151 # Now we do the real byte sequences that are valid UTF8
153 ["the utf8 sequence for chr $_->[0]",
154 qq{\$a = "$_->[1]"; \$b = show \$a}, qr/^>$_->[2]<$/],
155 ["no utf8; for the utf8 sequence for chr $_->[0]",
156 qq(no utf8; \$a = "$_->[1]"; \$b = show \$a), qr/^>$_->[2]<$/],
157 ["use utf8; for the utf8 sequence for chr $_->[0]",
158 qq(use utf8; \$a = "$_->[1]"; \$b = show \$a), qr/^>$_->[0]<$/],
160 # Interpolation of hex characters needs to take place now, as we're
161 # testing feeding malformed utf8 into perl. Bug now fixed was an
162 # "out of memory" error. We really need the "" [rather than qq()
163 # or q()] to get the best explosion.
164 ["!Feed malformed utf8 into perl.", <<"BANG",
165 use utf8; %a = ("\xE1\xA0"=>"sterling");
166 print 'start'; printf '%x,', ord \$_ foreach keys %a; print "end\n";
168 qr/^Malformed UTF-8 character \(\d bytes?, need \d, .+\).*start\d+,end$/sm
172 my ($why, $prog, $expect) = @$_;
173 open P, ">$progfile" or die "Can't open '$progfile': $!";
174 binmode(P, ":bytes") if $has_perlio;
175 print P $show, $prog, '; print $b'
176 or die "Print to 'progfile' failed: $!";
177 close P or die "Can't close '$progfile': $!";
178 if ($why =~ s/^!//) {
179 print "# Possible delay...\n";
183 my $result = runperl ( stderr => 1, progfile => $progfile );
184 like ($result, $expect, $why);
187 "# Again! Again! [but this time as eval, and not the explosive one]\n";
188 # and now we've safely done them all as separate files, check that the
189 # evals do the same thing. Hopefully doing it later successfully decouples
190 # the previous tests from anything messy that may go wrong with the evals.
192 my ($why, $prog, $expect) = @$_;
193 next if $why =~ m/^!/; # Goes bang.
194 my $result = eval $prog;
196 print "# prog is $prog\n";
197 print "# \$\@=", _qq($@), "\n";
199 like ($result, $expect, $why);
202 # See what the tokeniser does with hash keys.
203 print "# What does the tokeniser do with utf8 hash keys?\n";
205 # This is the control - I don't expect it to fail
206 ["assign utf8 for chr $_->[0] to a hash",
207 qq(my \$a = "$_->[1]"; my %h; \$h{\$a} = 1;
208 my \$b = show keys %h; \$b .= 'F' unless \$h{$_->[3]}; \$b),
210 ["no utf8; assign utf8 for chr $_->[0] to a hash",
211 qq(no utf8; my \$a = "$_->[1]"; my %h; \$h{\$a} = 1;
212 my \$b = show keys %h; \$b .= 'F' unless \$h{$_->[3]}; \$b),
214 ["use utf8; assign utf8 for chr $_->[0] to a hash",
215 qq(use utf8; my \$a = "$_->[1]"; my %h; \$h{\$a} = 1;
216 my \$b = show keys %h; \$b .= 'F' unless \$h{chr $_->[0]}; \$b),
218 # Now check literal $h{"x"} constructions.
219 ["\$h{\"x\"} construction, where x is utf8 for chr $_->[0]",
220 qq(my \$a = "$_->[1]"; my %h; \$h{"$_->[1]"} = 1;
221 my \$b = show keys %h; \$b .= 'F' unless \$h{$_->[3]}; \$b),
223 ["no utf8; \$h{\"x\"} construction, where x is utf8 for chr $_->[0]",
224 qq(no utf8; my \$a = "$_->[1]"; my %h; \$h{"$_->[1]"} = 1;
225 my \$b = show keys %h; \$b .= 'F' unless \$h{$_->[3]}; \$b),
227 ["use utf8; \$h{\"x\"} construction, where x is utf8 for chr $_->[0]",
228 qq(use utf8; my \$a = "$_->[1]"; my %h; \$h{"$_->[1]"} = 1;
229 my \$b = show keys %h; \$b .= 'F' unless \$h{chr $_->[0]}; \$b),
231 # Now check "x" => constructions.
232 ["assign \"x\"=>1 to a hash, where x is utf8 for chr $_->[0]",
233 qq(my \$a = "$_->[1]"; my %h; %h = ("$_->[1]" => 1);
234 my \$b = show keys %h; \$b .= 'F' unless \$h{$_->[3]}; \$b),
236 ["no utf8; assign \"x\"=>1 to a hash, where x is utf8 for chr $_->[0]",
237 qq(no utf8; my \$a = "$_->[1]"; my %h; %h = ("$_->[1]" => 1);
238 my \$b = show keys %h; \$b .= 'F' unless \$h{$_->[3]}; \$b),
240 ["use utf8; assign \"x\"=>1 to a hash, where x is utf8 for chr $_->[0]",
241 qq(use utf8; my \$a = "$_->[1]"; my %h; %h = ("$_->[1]" => 1);
242 my \$b = show keys %h; \$b .= 'F' unless \$h{chr $_->[0]}; \$b),
244 # Check copies of hashes made from literal utf8 keys
245 ["assign utf8 for chr $_->[0] to a hash, then copy it",
246 qq(my \$a = "$_->[1]"; my %i; \$i{\$a} = 1; my %h = %i;
247 my \$b = show keys %h; \$b .= 'F' unless \$h{$_->[3]}; \$b),
249 ["no utf8; assign utf8 for chr $_->[0] to a hash, then copy it",
250 qq(no utf8; my \$a = "$_->[1]"; my %i; \$i{\$a} = 1;; my %h = %i;
251 my \$b = show keys %h; \$b .= 'F' unless \$h{$_->[3]}; \$b),
253 ["use utf8; assign utf8 for chr $_->[0] to a hash, then copy it",
254 qq(use utf8; my \$a = "$_->[1]"; my %i; \$i{\$a} = 1; my %h = %i;
255 my \$b = show keys %h; \$b .= 'F' unless \$h{chr $_->[0]}; \$b),
259 my ($why, $prog, $expect) = @$_;
261 my $result = eval $prog;
262 like ($result, $expect, $why);
267 # bug fixed by change #17928
268 # separate perl used because we rely on 'strict' not yet loaded;
269 # before the patch, the eval died with an error like:
270 # "my" variable $strict::VERSION can't be in a package
273 skip("Embedded UTF-8 does not work in EBCDIC", 1) if ord("A") == 193;
274 ok('' eq runperl(prog => <<'CODE'), "change #17928");
275 my $code = qq{ my \$\xe3\x83\x95\xe3\x83\xbc = 5; };
287 0 ....... 1 ....... 2 ....... 3 ....... 4 ....... 5 ....... 6 ....... 7 .......
292 push @i, $s = index($a, '6'); # 60
293 push @i, $s = index($a, '.', $s); # next . after 60 is 62
294 push @i, $s = index($a, '5'); # 50
295 push @i, $s = index($a, '.', $s); # next . after 52 is 52
296 push @i, $s = index($a, '7'); # 70
297 push @i, $s = index($a, '.', $s); # next . after 70 is 72
298 push @i, $s = index($a, '4'); # 40
299 push @i, $s = index($a, '.', $s); # next . after 40 is 42
300 is("@i", "60 62 50 52 70 72 40 42", "utf8 heredoc index");
303 push @i, $s = rindex($a, '6'); # 60
304 push @i, $s = rindex($a, '.', $s); # previous . before 60 is 58
305 push @i, $s = rindex($a, '5'); # 50
306 push @i, $s = rindex($a, '.', $s); # previous . before 52 is 48
307 push @i, $s = rindex($a, '7'); # 70
308 push @i, $s = rindex($a, '.', $s); # previous . before 70 is 68
309 push @i, $s = rindex($a, '4'); # 40
310 push @i, $s = rindex($a, '.', $s); # previous . before 40 is 38
311 is("@i", "60 58 50 48 70 68 40 38", "utf8 heredoc rindex");
314 push @i, $s = index($a, '6'); # 60
315 push @i, index($a, '.', $s); # next . after 60 is 62
316 push @i, rindex($a, '.', $s); # previous . before 60 is 58
317 push @i, $s = rindex($a, '5'); # 60
318 push @i, index($a, '.', $s); # next . after 50 is 52
319 push @i, rindex($a, '.', $s); # previous . before 50 is 48
320 push @i, $s = index($a, '7', $s); # 70
321 push @i, index($a, '.', $s); # next . after 70 is 72
322 push @i, rindex($a, '.', $s); # previous . before 70 is 68
323 is("@i", "60 62 58 50 52 48 70 72 68", "utf8 heredoc index and rindex");
327 skip("Embedded UTF-8 does not work in EBCDIC", 1) if ord("A") == 193;
329 is eval qq{q \xc3\xbc test \xc3\xbc . qq\xc2\xb7 test \xc2\xb7},
331 "utf8 quote delimiters [perl #16823]";
334 # Test the "internals".
341 ok( utf8::valid($a), "utf8::valid basic");
342 ok( utf8::valid($b), "utf8::valid beyond");
343 ok( utf8::valid($c), "utf8::valid unicode");
345 ok(!utf8::is_utf8($a), "!utf8::is_utf8 basic");
346 ok(!utf8::is_utf8($b), "!utf8::is_utf8 beyond");
347 ok( utf8::is_utf8($c), "utf8::is_utf8 unicode");
349 is(utf8::upgrade($a), 1, "utf8::upgrade basic");
350 if (ord('A') == 193) { # EBCDIC.
351 is(utf8::upgrade($b), 1, "utf8::upgrade beyond");
353 is(utf8::upgrade($b), 2, "utf8::upgrade beyond");
355 is(utf8::upgrade($c), 2, "utf8::upgrade unicode");
357 is($a, "A", "basic");
358 is($b, "\xFF", "beyond");
359 is($c, "\x{100}", "unicode");
361 ok( utf8::valid($a), "utf8::valid basic");
362 ok( utf8::valid($b), "utf8::valid beyond");
363 ok( utf8::valid($c), "utf8::valid unicode");
365 ok( utf8::is_utf8($a), "utf8::is_utf8 basic");
366 ok( utf8::is_utf8($b), "utf8::is_utf8 beyond");
367 ok( utf8::is_utf8($c), "utf8::is_utf8 unicode");
369 is(utf8::downgrade($a), 1, "utf8::downgrade basic");
370 is(utf8::downgrade($b), 1, "utf8::downgrade beyond");
372 is($a, "A", "basic");
373 is($b, "\xFF", "beyond");
375 ok( utf8::valid($a), "utf8::valid basic");
376 ok( utf8::valid($b), "utf8::valid beyond");
378 ok(!utf8::is_utf8($a), "!utf8::is_utf8 basic");
379 ok(!utf8::is_utf8($b), "!utf8::is_utf8 beyond");
385 is($a, "A", "basic");
386 if (ord('A') == 193) { # EBCDIC.
387 is(length($b), 1, "beyond length");
389 is(length($b), 2, "beyond length");
391 is(length($c), 2, "unicode length");
393 ok(utf8::valid($a), "utf8::valid basic");
394 ok(utf8::valid($b), "utf8::valid beyond");
395 ok(utf8::valid($c), "utf8::valid unicode");
397 # encode() clears the UTF-8 flag (unlike upgrade()).
398 ok(!utf8::is_utf8($a), "!utf8::is_utf8 basic");
399 ok(!utf8::is_utf8($b), "!utf8::is_utf8 beyond");
400 ok(!utf8::is_utf8($c), "!utf8::is_utf8 unicode");
406 is($a, "A", "basic");
407 is($b, "\xFF", "beyond");
408 is($c, "\x{100}", "unicode");
410 ok(utf8::valid($a), "!utf8::valid basic");
411 ok(utf8::valid($b), "!utf8::valid beyond");
412 ok(utf8::valid($c), " utf8::valid unicode");
414 ok(!utf8::is_utf8($a), "!utf8::is_utf8 basic");
415 if (ord('A') == 193) { # EBCDIC.
416 ok( utf8::is_utf8(pack('U',0x0ff)), " utf8::is_utf8 beyond");
418 ok( utf8::is_utf8($b), " utf8::is_utf8 beyond"); # $b stays in UTF-8.
420 ok( utf8::is_utf8($c), " utf8::is_utf8 unicode");
424 eval {utf8::encode("£")};
425 like($@, qr/^Modification of a read-only value attempted/,
426 "utf8::encode should refuse to touch read-only values");
430 # Make sure utf8::decode respects copy-on-write [perl #91834].
431 # Hash keys are the easiest way to test this.
432 my $name = "\x{c3}\x{b3}";
433 my ($k1) = keys %{ { $name=>undef } };
437 my $h = { $k1 => 1, $k2 => 2 };
438 is join('', keys %$h), $k2, 'utf8::decode respects copy-on-write';
442 # Make sure utf8::decode does not modify read-only scalars
445 my $name = "\x{c3}\x{b3}";
446 Internals::SvREADONLY($name, 1);
447 eval { utf8::decode($name) };
448 like $@, qr/^Modification of a read-only/,
449 'utf8::decode respects readonliness';
453 # utf8::decode should stringify refs [perl #91852].
455 package eieifg { use overload '""' => sub { "\x{c3}\x{b3}" },
458 my $name = bless[], eieifg::;
460 is $name, "\xf3", 'utf8::decode flattens references';
464 # What do the utf8::* functions do when given a reference? A test
465 # for a behavior change that made this start dying as of
466 # v5.15.6-407-gc710240 due to a fix for [perl #91852]:
468 # ./miniperl -Ilib -wle 'use strict; print $]; my $s = shift; my $s_ref = \$s; utf8::decode($s_ref); print $$s_ref' hlagh
470 'utf8::is_utf8' => { returns => "hlagh" },
471 'utf8::valid' => { returns => "hlagh" },
472 'utf8::encode' => { error => qr/Can't use string .*? as a SCALAR ref/},
473 'utf8::decode' => { error => qr/Can't use string .*? as a SCALAR ref/},
474 'utf8::upgrade' => { error => qr/Can't use string .*? as a SCALAR ref/ },
475 'utf8::downgrade' => { returns => "hlagh" },
476 'utf8::native_to_unicode' => { returns => "hlagh" },
477 'utf8::unicode_to_native' => { returns => "hlagh" },
479 for my $func (sort keys %expected) { # sort just so it's deterministic wrt diffing *.t output
480 my $code = sprintf q[
487 my $ret = eval $code or my $error = $@;
488 if (my $error_rx = $expected{$func}->{error}) {
489 if (defined $error) {
490 like $error, $error_rx, "The $func function should die with an error matching $error_rx";
492 fail("We were expecting an error when calling the $func function but got a value of '$ret' instead");
494 } elsif (my $returns = $expected{$func}->{returns}) {
495 is($ret, $returns, "The $func function lives and returns '$returns' as expected");
497 die "PANIC: Internal Error"
506 my $b = "123456\xb6";
509 is($b, $a, "utf8::upgrade OffsetOK");
513 fresh_perl_like ('use utf8; utf8::moo()',
514 qr/Undefined subroutine utf8::moo/, {stderr=>1},
515 "Check Carp is loaded for AUTOLOADing errors")
519 # failure of is_utf8_char() without NATIVE_TO_UTF on EBCDIC (0260..027F)
520 ok(utf8::valid(chr(0x250)), "0x250");
521 ok(utf8::valid(chr(0x260)), "0x260");
522 ok(utf8::valid(chr(0x270)), "0x270");
523 ok(utf8::valid(chr(0x280)), "0x280");
528 ok( !utf8::is_utf8( "asd" ), "Wasteful format - qq{}" );
529 ok( !utf8::is_utf8( 'asd' ), "Wasteful format - q{}" );
530 ok( !utf8::is_utf8( qw(asd) ), "Wasteful format - qw{}" );
531 ok( !utf8::is_utf8( (asd => 1)[0] ), "Wasteful format - =>" );
532 ok( !utf8::is_utf8( -asd ), "Wasteful format - -word" );
533 no warnings 'bareword';
534 ok( !utf8::is_utf8( asd:: ), "Wasteful format - word::" );
535 no warnings 'reserved';
537 ok( !utf8::is_utf8( asd ), "Wasteful format - bareword" );
542 (undef, 0x7F, 0x7FF, 0xFFFF, 0x1FFFFF, 0x3FFFFFF, 0x7FFFFFFF);
544 (undef, undef, 0x40, 0x1000, 0x40000, 0x1000000, 0x40000000);
546 foreach my $length (6, 5, 4, 3, 2) {
547 my $high = $highest[$length];
548 while ($high > $highest[$length - 1]) {
549 my $low = $high - $step[$length] + 1;
550 $low = $highest[$length - 1] + 1 if $low <= $highest[$length - 1];
551 ok(utf8::valid(do {no warnings 'utf8'; chr $low}),
552 sprintf "chr %x, length $length is valid", $low);
553 ok(utf8::valid(do {no warnings 'utf8'; chr $high}),
554 sprintf "chr %x, length $length is valid", $high);
555 $high -= $step[$length];
560 # #80190 update pos, and cached length/position-mapping after
561 # utf8 upgrade/downgrade, encode/decode
566 my $s = "A\xc8\x81\xe8\xab\x86\x{100}";
571 is(length($s), 6, "(pos $pos) len before utf8::downgrade");
572 is(pos($s), $pos, "(pos $pos) pos before utf8::downgrade");
574 is(length($s), 6, "(pos $pos) len after utf8::downgrade");
575 is(pos($s), $pos, "(pos $pos) pos after utf8::downgrade");
576 is($s, "A\xc8\x81\xe8\xab\x86","(pos $pos) str after utf8::downgrade");
578 is(length($s), 3, "(pos $pos) len after D; utf8::decode");
579 is(pos($s), undef, "(pos $pos) pos after D; utf8::decode");
580 is($s, "A\x{201}\x{8ac6}", "(pos $pos) str after D; utf8::decode");
582 is(length($s), 6, "(pos $pos) len after D; utf8::encode");
583 is(pos($s), undef, "(pos $pos) pos after D; utf8::encode");
584 is($s, "A\xc8\x81\xe8\xab\x86","(pos $pos) str after D; utf8::encode");
586 $s = "A\xc8\x81\xe8\xab\x86";
589 is(length($s), 6, "(pos $pos) len before utf8::upgrade");
590 is(pos($s), $pos, "(pos $pos) pos before utf8::upgrade");
592 is(length($s), 6, "(pos $pos) len after utf8::upgrade");
593 is(pos($s), $pos, "(pos $pos) pos after utf8::upgrade");
594 is($s, "A\xc8\x81\xe8\xab\x86","(pos $pos) str after utf8::upgrade");
596 is(length($s), 3, "(pos $pos) len after U; utf8::decode");
597 is(pos($s), undef, "(pos $pos) pos after U; utf8::decode");
598 is($s, "A\x{201}\x{8ac6}", "(pos $pos) str after U; utf8::decode");
600 is(length($s), 6, "(pos $pos) len after U; utf8::encode");
601 is(pos($s), undef, "(pos $pos) pos after U; utf8::encode");
602 is($s, "A\xc8\x81\xe8\xab\x86","(pos $pos) str after U; utf8::encode");
605 # [perl #119043] utf8::upgrade should not croak on read-only COWs
607 eval { utf8::upgrade($_) };
608 is $@, "", 'no error with utf8::upgrade on read-only COW';
610 # This one croaks, but not because the scalar is read-only
611 eval "package \x{100};\n" . <<'END'
613 eval { utf8::downgrade($_) };
614 ::like $@, qr/^Wide character/,
615 'right error with utf8::downgrade on read-only COW';