X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=t%2Fio%2Futf8.t;h=22f392ac788be0d223af6e2c308fe6ec1ca14eb4;hb=ced497e2fca8b0ac1628855f422776e9bf331e65;hp=38d7007c784e2184d177541c4cf223e1b8c0ba89;hpb=e0e62c2a4ed4c005b9afcfe9fe3e1838fcea785d;p=p5sagit%2Fp5-mst-13.2.git diff --git a/t/io/utf8.t b/t/io/utf8.t index 38d7007..22f392a 100755 --- a/t/io/utf8.t +++ b/t/io/utf8.t @@ -3,127 +3,132 @@ BEGIN { chdir 't' if -d 't'; @INC = '../lib'; - unless (exists $open::layers{'perlio'}) { + unless (find PerlIO::Layer 'perlio') { print "1..0 # Skip: not perlio\n"; exit 0; } } +no utf8; # needed for use utf8 not griping about the raw octets + +require "./test.pl"; + +plan(tests => 55); + $| = 1; -print "1..25\n"; open(F,"+>:utf8",'a'); print F chr(0x100).'£'; -print '#'.tell(F)."\n"; -print "not " unless tell(F) == 4; -print "ok 1\n"; +cmp_ok( tell(F), '==', 4, tell(F) ); print F "\n"; -print '#'.tell(F)."\n"; -print "not " unless tell(F) >= 5; -print "ok 2\n"; +cmp_ok( tell(F), '>=', 5, tell(F) ); seek(F,0,0); -print "not " unless getc(F) eq chr(0x100); -print "ok 3\n"; -print "not " unless getc(F) eq "£"; -print "ok 4\n"; -print "not " unless getc(F) eq "\n"; -print "ok 5\n"; +is( getc(F), chr(0x100) ); +is( getc(F), "£" ); +is( getc(F), "\n" ); seek(F,0,0); binmode(F,":bytes"); -print "not " unless getc(F) eq chr(0xc4); -print "ok 6\n"; -print "not " unless getc(F) eq chr(0x80); -print "ok 7\n"; -print "not " unless getc(F) eq chr(0xc2); -print "ok 8\n"; -print "not " unless getc(F) eq chr(0xa3); -print "ok 9\n"; -print "not " unless getc(F) eq "\n"; -print "ok 10\n"; +my $chr = chr(0xc4); +if (ord('A') == 193) { $chr = chr(0x8c); } # EBCDIC +is( getc(F), $chr ); +$chr = chr(0x80); +if (ord('A') == 193) { $chr = chr(0x41); } # EBCDIC +is( getc(F), $chr ); +$chr = chr(0xc2); +if (ord('A') == 193) { $chr = chr(0x80); } # EBCDIC +is( getc(F), $chr ); +$chr = chr(0xa3); +if (ord('A') == 193) { $chr = chr(0x44); } # EBCDIC +is( getc(F), $chr ); +is( getc(F), "\n" ); seek(F,0,0); binmode(F,":utf8"); -print "not " unless scalar() eq "\x{100}£\n"; -print "ok 11\n"; +is( scalar(), "\x{100}£\n" ); seek(F,0,0); $buf = chr(0x200); $count = read(F,$buf,2,1); -print "not " unless $count == 2; -print "ok 12\n"; -print "not " unless $buf eq "\x{200}\x{100}£"; -print "ok 13\n"; +cmp_ok( $count, '==', 2 ); +is( $buf, "\x{200}\x{100}£" ); close(F); { -$a = chr(300); # This *is* UTF-encoded -$b = chr(130); # This is not. + $a = chr(300); # This *is* UTF-encoded + $b = chr(130); # This is not. -open F, ">:utf8", 'a' or die $!; -print F $a,"\n"; -close F; + open F, ">:utf8", 'a' or die $!; + print F $a,"\n"; + close F; -open F, "<:utf8", 'a' or die $!; -$x = ; -chomp($x); -print "not " unless $x eq chr(300); -print "ok 14\n"; - -open F, "a" or die $!; # Not UTF -$x = ; -chomp($x); -print "not " unless $x eq chr(196).chr(172); -print "ok 15\n"; -close F; + open F, "<:utf8", 'a' or die $!; + $x = ; + chomp($x); + is( $x, chr(300) ); -open F, ">:utf8", 'a' or die $!; -binmode(F); # we write a "\n" and then tell() - avoid CRLF issues. -print F $a; -my $y; -{ my $x = tell(F); - { use bytes; $y = length($a);} - print "not " unless $x == $y; - print "ok 16\n"; -} + open F, "a" or die $!; # Not UTF + binmode(F, ":bytes"); + $x = ; + chomp($x); + $chr = chr(196).chr(172); + if (ord('A') == 193) { $chr = chr(141).chr(83); } # EBCDIC + is( $x, $chr ); + close F; -{ # Check byte length of $b -use bytes; my $y = length($b); -print "not " unless $y == 1; -print "ok 17\n"; -} + open F, ">:utf8", 'a' or die $!; + binmode(F); # we write a "\n" and then tell() - avoid CRLF issues. + binmode(F,":utf8"); # turn UTF-8-ness back on + print F $a; + my $y; + { my $x = tell(F); + { use bytes; $y = length($a);} + cmp_ok( $x, '==', $y ); + } -print F $b,"\n"; # Don't upgrades $b + { # Check byte length of $b + use bytes; my $y = length($b); + cmp_ok( $y, '==', 1 ); + } -{ # Check byte length of $b -use bytes; my $y = length($b); -print "not ($y) " unless $y == 1; -print "ok 18\n"; -} + print F $b,"\n"; # Don't upgrades $b -{ my $x = tell(F); - { use bytes; $y += 3;} - print "not ($x,$y) " unless $x == $y; - print "ok 19\n"; -} + { # Check byte length of $b + use bytes; my $y = length($b); + cmp_ok( $y, '==', 1 ); + } -close F; + { + my $x = tell(F); + { use bytes; if (ord('A')==193){$y += 2;}else{$y += 3;}} # EBCDIC ASCII + cmp_ok( $x, '==', $y ); + } -open F, "a" or die $!; # Not UTF -$x = ; -chomp($x); -printf "not (%vd) ", $x unless $x eq v196.172.194.130; -print "ok 20\n"; + close F; -open F, "<:utf8", "a" or die $!; -$x = ; -chomp($x); -close F; -printf "not (%vd) ", $x unless $x eq chr(300).chr(130); -print "ok 21\n"; - -# Now let's make it suffer. -open F, ">", "a" or die $!; -eval { print F $a; }; -print "not " unless $@ and $@ =~ /Wide character in print/i; -print "ok 22\n"; + open F, "a" or die $!; # Not UTF + binmode(F, ":bytes"); + $x = ; + chomp($x); + $chr = v196.172.194.130; + if (ord('A') == 193) { $chr = v141.83.130; } # EBCDIC + is( $x, $chr, sprintf('(%vd)', $x) ); + + open F, "<:utf8", "a" or die $!; + $x = ; + chomp($x); + close F; + is( $x, chr(300).chr(130), sprintf('(%vd)', $x) ); + + open F, ">", "a" or die $!; + binmode(F, ":bytes:"); + + # Now let's make it suffer. + my $w; + { + use warnings 'utf8'; + local $SIG{__WARN__} = sub { $w = $_[0] }; + print F $a; + ok( (!$@)); + like($w, qr/Wide character in print/i ); + } } # Hm. Time to get more evil. @@ -134,29 +139,218 @@ print F chr(130)."\n"; close F; open F, "<", "a" or die $!; +binmode(F, ":bytes"); $x = ; chomp $x; -print "not " unless $x eq v196.172.130; -print "ok 23\n"; +$chr = v196.172.130; +if (ord('A') == 193) { $chr = v141.83.130; } # EBCDIC +is( $x, $chr ); # Right. open F, ">:utf8", "a" or die $!; print F $a; close F; open F, ">>", "a" or die $!; +binmode(F, ":bytes"); print F chr(130)."\n"; close F; open F, "<", "a" or die $!; +binmode(F, ":bytes"); $x = ; chomp $x; -print "not " unless $x eq v196.172.130; -print "ok 24\n"; +SKIP: { + skip("Defaulting to UTF-8 output means that we can't generate a mangled file") + if $UTF8_OUTPUT; + is( $x, $chr ); +} # Now we have a deformed file. -open F, "<:utf8", "a" or die $!; -$x = ; chomp $x; -{ local $SIG{__WARN__} = sub { print "ok 25\n"; }; -eval { sprintf "%vd\n", $x; } + +SKIP: { + if (ord('A') == 193) { + skip("EBCDIC doesn't complain", 2); + } else { + my @warnings; + open F, "<:utf8", "a" or die $!; + $x = ; chomp $x; + local $SIG{__WARN__} = sub { push @warnings, $_[0]; }; + eval { sprintf "%vd\n", $x }; + is (scalar @warnings, 1); + like ($warnings[0], qr/Malformed UTF-8 character \(unexpected continuation byte 0x82, with no preceding start byte/); + } } +close F; unlink('a'); +open F, ">:utf8", "a"; +@a = map { chr(1 << ($_ << 2)) } 0..5; # 0x1, 0x10, .., 0x100000 +unshift @a, chr(0); # ... and a null byte in front just for fun +print F @a; +close F; + +my $c; + +# read() should work on characters, not bytes +open F, "<:utf8", "a"; +$a = 0; +my $failed; +for (@a) { + unless (($c = read(F, $b, 1) == 1) && + length($b) == 1 && + ord($b) == ord($_) && + tell(F) == ($a += bytes::length($b))) { + print '# ord($_) == ', ord($_), "\n"; + print '# ord($b) == ', ord($b), "\n"; + print '# length($b) == ', length($b), "\n"; + print '# bytes::length($b) == ', bytes::length($b), "\n"; + print '# tell(F) == ', tell(F), "\n"; + print '# $a == ', $a, "\n"; + print '# $c == ', $c, "\n"; + $failed++; + last; + } +} +close F; +is($failed, undef); + +{ + # Check that warnings are on on I/O, and that they can be muffled. + + local $SIG{__WARN__} = sub { $@ = shift }; + + undef $@; + open F, ">a"; + binmode(F, ":bytes"); + print F chr(0x100); + close(F); + + like( $@, 'Wide character in print' ); + + undef $@; + open F, ">:utf8", "a"; + print F chr(0x100); + close(F); + + isnt( defined $@ ); + + undef $@; + open F, ">a"; + binmode(F, ":utf8"); + print F chr(0x100); + close(F); + + isnt( defined $@ ); + + no warnings 'utf8'; + + undef $@; + open F, ">a"; + print F chr(0x100); + close(F); + + isnt( defined $@ ); + + use warnings 'utf8'; + + undef $@; + open F, ">a"; + binmode(F, ":bytes"); + print F chr(0x100); + close(F); + + like( $@, 'Wide character in print' ); +} + +{ + open F, ">:bytes","a"; print F "\xde"; close F; + + open F, "<:bytes", "a"; + my $b = chr 0x100; + $b .= ; + is( $b, chr(0x100).chr(0xde), "21395 '.= <>' utf8 vs. bytes" ); + close F; +} + +{ + open F, ">:utf8","a"; print F chr 0x100; close F; + + open F, "<:utf8", "a"; + my $b = "\xde"; + $b .= ; + is( $b, chr(0xde).chr(0x100), "21395 '.= <>' bytes vs. utf8" ); + close F; +} + +{ + my @a = ( [ 0x007F, "bytes" ], + [ 0x0080, "bytes" ], + [ 0x0080, "utf8" ], + [ 0x0100, "utf8" ] ); + my $t = 34; + for my $u (@a) { + for my $v (@a) { + # print "# @$u - @$v\n"; + open F, ">a"; + binmode(F, ":" . $u->[1]); + print F chr($u->[0]); + close F; + + open F, "[1]); + + my $s = chr($v->[0]); + utf8::upgrade($s) if $v->[1] eq "utf8"; + + $s .= ; + is( $s, chr($v->[0]) . chr($u->[0]), 'rcatline utf8' ); + close F; + $t++; + } + } + # last test here 49 +} + +{ + # [perl #23428] Somethings rotten in unicode semantics + open F, ">a"; + binmode F, ":utf8"; + syswrite(F, $a = chr(0x100)); + close F; + is( ord($a), 0x100, '23428 syswrite should not downgrade scalar' ); + like( $a, qr/^\w+/, '23428 syswrite should not downgrade scalar' ); +} + +# sysread() and syswrite() tested in lib/open.t since Fcntl is used + +{ + # on a :utf8 stream should complain immediately with -w + # if it finds bad UTF-8 (:encoding(utf8) works this way) + use warnings 'utf8'; + undef $@; + local $SIG{__WARN__} = sub { $@ = shift }; + open F, ">a"; + binmode F; + my ($chrE4, $chrF6) = (chr(0xE4), chr(0xF6)); + if (ord('A') == 193) # EBCDIC + { ($chrE4, $chrF6) = (chr(0x43), chr(0xEC)); } + print F "foo", $chrE4, "\n"; + print F "foo", $chrF6, "\n"; + close F; + open F, "<:utf8", "a"; + undef $@; + my $line = ; + my ($chrE4, $chrF6) = ("E4", "F6"); + if (ord('A') == 193) { ($chrE4, $chrF6) = ("43", "EC"); } # EBCDIC + like( $@, qr/utf8 "\\x$chrE4" does not map to Unicode .+ line 1/, + "<:utf8 readline must warn about bad utf8"); + undef $@; + $line .= ; + like( $@, qr/utf8 "\\x$chrF6" does not map to Unicode .+ line 2/, + "<:utf8 rcatline must warn about bad utf8"); + close F; +} + +END { + 1 while unlink "a"; + 1 while unlink "b"; +}