xref: /openbsd/gnu/usr.bin/perl/t/io/utf8.t (revision 3bef86f7)
1#!./perl
2
3BEGIN {
4    chdir 't' if -d 't';
5    require './test.pl';
6    set_up_inc('../lib');
7	require './charset_tools.pl';
8}
9skip_all_without_perlio();
10
11no utf8; # needed for use utf8 not griping about the raw octets
12
13
14plan(tests => 62);
15
16$| = 1;
17
18my $a_file = tempfile();
19
20open(F,"+>:utf8",$a_file);
21print F chr(0x100).'�';
22cmp_ok( tell(F), '==', 4, tell(F) );
23print F "\n";
24cmp_ok( tell(F), '>=', 5, tell(F) );
25seek(F,0,0);
26is( getc(F), chr(0x100) );
27is( getc(F), "�" );
28is( getc(F), "\n" );
29seek(F,0,0);
30binmode(F,":bytes");
31
32# Byte representation of these characters
33my $U_100 = byte_utf8a_to_utf8n("\xc4\x80");
34my $POUND_SIGN = byte_utf8a_to_utf8n("\xc2\xa3");
35
36my $chr = substr($U_100, 0, 1);
37is( getc(F), $chr );
38$chr = substr($U_100, 1, 1);
39is( getc(F), $chr );
40$chr = substr($POUND_SIGN, 0, 1);
41is( getc(F), $chr );
42$chr = substr($POUND_SIGN, 1, 1);
43is( getc(F), $chr );
44is( getc(F), "\n" );
45seek(F,0,0);
46binmode(F,":utf8");
47is( scalar(<F>), "\x{100}�\n" );
48seek(F,0,0);
49$buf = chr(0x200);
50$count = read(F,$buf,2,1);
51cmp_ok( $count, '==', 2 );
52is( $buf, "\x{200}\x{100}�" );
53close(F);
54
55{
56    $a = chr(300); # This *is* UTF-encoded
57    $b = chr(130); # This is not.
58
59    open F, ">:utf8", $a_file or die $!;
60    print F $a,"\n";
61    close F;
62
63    open F, "<:utf8", $a_file or die $!;
64    $x = <F>;
65    chomp($x);
66    is( $x, chr(300) );
67
68    open F, $a_file or die $!; # Not UTF
69    binmode(F, ":bytes");
70    $x = <F>;
71    chomp($x);
72    $chr = byte_utf8a_to_utf8n(chr(196).chr(172));
73    is( $x, $chr );
74    close F;
75
76    open F, ">:utf8", $a_file or die $!;
77    binmode(F);  # we write a "\n" and then tell() - avoid CRLF issues.
78    binmode(F,":utf8"); # turn UTF-8-ness back on
79    print F $a;
80    my $y;
81    { my $x = tell(F);
82      { use bytes; $y = length($a);}
83      cmp_ok( $x, '==', $y );
84  }
85
86    { # Check byte length of $b
87	use bytes; my $y = length($b);
88	cmp_ok( $y, '==', 1 );
89    }
90
91    print F $b,"\n"; # Don't upgrade $b
92
93    { # Check byte length of $b
94	use bytes; my $y = length($b);
95	cmp_ok( $y, '==', 1 );
96    }
97
98    {
99	my $x = tell(F);
100	{ use bytes; if ($::IS_EBCDIC){$y += 2;}else{$y += 3;}} # EBCDIC ASCII
101	cmp_ok( $x, '==', $y );
102    }
103
104    close F;
105
106    open F, $a_file or die $!; # Not UTF
107    binmode(F, ":bytes");
108    $x = <F>;
109    chomp($x);
110    $chr = v196.172.194.130;
111    if ($::IS_EBCDIC) { $chr = v141.83.130; } # EBCDIC
112    is( $x, $chr, sprintf('(%vd)', $x) );
113
114    open F, "<:utf8", $a_file or die $!;
115    $x = <F>;
116    chomp($x);
117    close F;
118    is( $x, chr(300).chr(130), sprintf('(%vd)', $x) );
119
120    open F, ">", $a_file or die $!;
121    binmode(F, ":bytes:");
122
123    # Now let's make it suffer.
124    my $w;
125    {
126	use warnings 'utf8';
127	local $SIG{__WARN__} = sub { $w = $_[0] };
128	print F $a;
129        ok( (!$@));
130	like($w, qr/Wide character in print/i );
131    }
132}
133
134# Hm. Time to get more evil.
135open F, ">:utf8", $a_file or die $!;
136print F $a;
137binmode(F, ":bytes");
138print F chr(130)."\n";
139close F;
140
141open F, "<", $a_file or die $!;
142binmode(F, ":bytes");
143$x = <F>; chomp $x;
144$chr = v196.172.130;
145if ($::IS_EBCDIC) { $chr = v141.83.130; } # EBCDIC
146is( $x, $chr );
147
148# Right.
149open F, ">:utf8", $a_file or die $!;
150print F $a;
151close F;
152open F, ">>", $a_file or die $!;
153binmode(F, ":bytes");
154print F chr(130)."\n";
155close F;
156
157open F, "<", $a_file or die $!;
158binmode(F, ":bytes");
159$x = <F>; chomp $x;
160SKIP: {
161    skip("Defaulting to UTF-8 output means that we can't generate a mangled file")
162	if $UTF8_OUTPUT;
163    is( $x, $chr );
164}
165
166# Now we have a deformed file.
167
168SKIP: {
169    if ($::IS_EBCDIC) {
170	skip("EBCDIC The file isn't deformed in UTF-EBCDIC", 2);
171    } else {
172	my @warnings;
173	open F, "<:utf8", $a_file or die $!;
174	$x = <F>; chomp $x;
175	local $SIG{__WARN__} = sub { push @warnings, $_[0]; };
176	eval { sprintf "%vd\n", $x };
177	is (scalar @warnings, 1);
178	like ($warnings[0], qr/Malformed UTF-8 character: \\x82 \(unexpected continuation byte 0x82, with no preceding start byte/);
179    }
180}
181
182close F;
183unlink($a_file);
184
185open F, ">:utf8", $a_file;
186@a = map { chr(1 << ($_ << 2)) } 0..5; # 0x1, 0x10, .., 0x100000
187unshift @a, chr(0); # ... and a null byte in front just for fun
188print F @a;
189close F;
190
191my $c;
192
193# read() should work on characters, not bytes
194open F, "<:utf8", $a_file;
195$a = 0;
196my $failed;
197for (@a) {
198    unless (($c = read(F, $b, 1) == 1)  &&
199            length($b)           == 1  &&
200            ord($b)              == ord($_) &&
201            tell(F)              == ($a += bytes::length($b))) {
202        print '# ord($_)           == ', ord($_), "\n";
203        print '# ord($b)           == ', ord($b), "\n";
204        print '# length($b)        == ', length($b), "\n";
205        print '# bytes::length($b) == ', bytes::length($b), "\n";
206        print '# tell(F)           == ', tell(F), "\n";
207        print '# $a                == ', $a, "\n";
208        print '# $c                == ', $c, "\n";
209	$failed++;
210        last;
211    }
212}
213close F;
214is($failed, undef);
215
216{
217    # Check that warnings are on on I/O, and that they can be muffled.
218
219    local $SIG{__WARN__} = sub { $@ = shift };
220
221    undef $@;
222    open F, ">$a_file";
223    binmode(F, ":bytes");
224    print F chr(0x100);
225    close(F);
226
227    like( $@, qr/Wide character in print/ );
228
229    undef $@;
230    open F, ">:utf8", $a_file;
231    print F chr(0x100);
232    close(F);
233
234    isnt( defined $@, !0 );
235
236    undef $@;
237    open F, ">$a_file";
238    binmode(F, ":utf8");
239    print F chr(0x100);
240    close(F);
241
242    isnt( defined $@, !0 );
243
244    no warnings 'utf8';
245
246    undef $@;
247    open F, ">$a_file";
248    print F chr(0x100);
249    close(F);
250
251    isnt( defined $@, !0 );
252
253    use warnings 'utf8';
254
255    undef $@;
256    open F, ">$a_file";
257    binmode(F, ":bytes");
258    print F chr(0x100);
259    close(F);
260
261    like( $@, qr/Wide character in print/ );
262}
263
264{
265    open F, ">:bytes",$a_file; print F "\xde"; close F;
266
267    open F, "<:bytes", $a_file;
268    my $b = chr 0x100;
269    $b .= <F>;
270    is( $b, chr(0x100).chr(0xde), "21395 '.= <>' utf8 vs. bytes" );
271    close F;
272}
273
274{
275    open F, ">:utf8",$a_file; print F chr 0x100; close F;
276
277    open F, "<:utf8", $a_file;
278    my $b = "\xde";
279    $b .= <F>;
280    is( $b, chr(0xde).chr(0x100), "21395 '.= <>' bytes vs. utf8" );
281    close F;
282}
283
284{
285    my @a = ( [ 0x007F, "bytes" ],
286	      [ 0x0080, "bytes" ],
287	      [ 0x0080, "utf8"  ],
288	      [ 0x0100, "utf8"  ] );
289    my $t = 34;
290    for my $u (@a) {
291	for my $v (@a) {
292	    # print "# @$u - @$v\n";
293	    open F, ">$a_file";
294	    binmode(F, ":" . $u->[1]);
295	    print F chr($u->[0]);
296	    close F;
297
298	    open F, "<$a_file";
299	    binmode(F, ":" . $u->[1]);
300
301	    my $s = chr($v->[0]);
302	    utf8::upgrade($s) if $v->[1] eq "utf8";
303
304	    $s .= <F>;
305	    is( $s, chr($v->[0]) . chr($u->[0]), 'rcatline utf8' );
306	    close F;
307	    $t++;
308	}
309    }
310    # last test here 49
311}
312
313{
314    # [perl #23428] Somethings rotten in unicode semantics
315    open F, ">$a_file";
316    binmode F;
317    $a = "A";
318    utf8::upgrade($a);
319    syswrite(F, $a);
320    close F;
321    ok(utf8::is_utf8($a), '23428 syswrite should not downgrade scalar' );
322}
323
324{
325    # <FH> on a :utf8 stream should complain immediately with -w
326    # if it finds bad UTF-8 (:encoding(utf8) works this way)
327    use warnings 'utf8';
328    undef $@;
329    local $SIG{__WARN__} = sub { $@ = shift };
330    open F, ">$a_file";
331    binmode F;
332    my ($chrE4, $chrF6) = (chr(0xE4), chr(0xF6));
333    if ($::IS_EBCDIC)	# EBCDIC
334    { ($chrE4, $chrF6) = (chr(0x43), chr(0xEC)); }
335    print F "foo", $chrE4, "\n";
336    print F "foo", $chrF6, "\n";
337    close F;
338    open F, "<:utf8", $a_file;
339    undef $@;
340    my $line = <F>;
341    my ($chrE4, $chrF6) = ("E4", "F6");
342    if ($::IS_EBCDIC) { ($chrE4, $chrF6) = ("43", "EC"); } # EBCDIC
343    like( $@, qr/utf8 "\\x$chrE4" does not map to Unicode .+ <F> line 1/,
344	  "<:utf8 readline must warn about bad utf8");
345    undef $@;
346    $line .= <F>;
347    like( $@, qr/utf8 "\\x$chrF6" does not map to Unicode .+ <F> line 2/,
348	  "<:utf8 rcatline must warn about bad utf8");
349    close F;
350}
351
352{
353    # fixed record reads
354    open F, ">:utf8", $a_file;
355    print F "foo\xE4";
356    print F "bar\xFE";
357    print F "\xC0\xC8\xCC\xD2";
358    print F "a\xE4ab";
359    print F "a\xE4a";
360    close F;
361    open F, "<:utf8", $a_file;
362    local $/ = \4;
363    my $line = <F>;
364    is($line, "foo\xE4", "readline with \$/ = \\4");
365    $line .= <F>;
366    is($line, "foo\xE4bar\xFE", "rcatline with \$/ = \\4");
367    $line = <F>;
368    is($line, "\xC0\xC8\xCC\xD2", "readline with several encoded characters");
369    $line = <F>;
370    is($line, "a\xE4ab", "readline with another boundary condition");
371    $line = <F>;
372    is($line, "a\xE4a", "readline with boundary condition");
373    close F;
374
375    # badly encoded at EOF
376    open F, ">:raw", $a_file;
377    print F "foo\xEF\xAC"; # truncated \x{FB04} small ligature ffl
378    close F;
379
380    use warnings 'utf8';
381    open F, "<:utf8", $a_file;
382    undef $@;
383    local $SIG{__WARN__} = sub { $@ = shift };
384    $line = <F>;
385
386    like( $@, qr/utf8 "\\xEF" does not map to Unicode .+ <F> chunk 1/,
387	  "<:utf8 readline (fixed) must warn about bad utf8");
388    close F;
389}
390
391# getc should reset the utf8 flag and not be affected by previous
392# return values
393SKIP: {
394    skip "no PerlIO::scalar on miniperl", 2, if is_miniperl();
395    open my $fh, "<:raw",  \($buf = chr 255);
396    open my $uh, "<:utf8", \($uuf = $U_100);
397    for([$uh,chr 256], [$fh,chr 255]) {
398	is getc $$_[0], $$_[1],
399	  'getc returning non-utf8 after utf8';
400    }
401}
402