This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
In utf8decode.t, test that we get the expected Unicode character(s)
[perl5.git] / t / op / utf8decode.t
CommitLineData
a9917092
JH
1#!./perl
2
3BEGIN {
4 chdir 't' if -d 't';
5 @INC = '../lib';
680218c4 6 require './test.pl';
daf0f78e
NIS
7}
8
9{
10 my $wide = v256;
11 use bytes;
ffbc6a93
JH
12 my $ordwide = ord($wide);
13 printf "# under use bytes ord(v256) = 0x%02x\n", $ordwide;
680218c4
NC
14 skip_all('UTF-EBCDIC (not UTF-8) used here') if $ordwide == 140;
15
16 if ($ordwide != 196) {
ffbc6a93
JH
17 printf "# v256 starts with 0x%02x\n", $ordwide;
18 }
a9917092
JH
19}
20
3b0e0cb6 21no utf8;
ffc61ed2 22
e1a67460
NC
23foreach (<DATA>) {
24 if (/^(?:\d+(?:\.\d+)?)\s/ || /^#/) {
25 # print "# $_\n";
04f3c608 26 } elsif (/^(\d+\.\d+\.\d+[bu]?)\s+(y|n|N-?\d+)\s+([0-9a-f]{1,8}(?:,[0-9a-f]{1,8})*|-)\s+(\d+)\s+([0-9a-f]{2}(?::[0-9a-f]{2})*)\s+(\d+|-)(?:\s+(.+))?$/) {
c03ceae6
NC
27 my ($id, $okay, $Unicode, $byteslen, $hex, $charslen, $experr) =
28 ($1, $2, $3, $4, $5, $6, $7);
e1a67460
NC
29 my @hex = split(/:/, $hex);
30 is(scalar @hex, $byteslen, 'Amount of hex tallies with byteslen');
31 my $octets = join '', map {chr hex $_} @hex;
32 is(length $octets, $byteslen, 'Number of octets tallies with byteslen');
e1a67460 33 if ($okay eq 'y') {
04f3c608
NC
34 my @chars = map {hex $_} split ',', $Unicode;
35 is(scalar @chars, $charslen, 'Amount of hex tallies with charslen');
36 my @got;
37 warning_is(sub {@got = unpack 'C0U*', $octets}, undef,
c03ceae6 38 "No warnings expected for $id");
04f3c608 39 is("@got", "@chars", 'Got expected Unicode characters');
f96a66c2
NC
40 } elsif ($okay eq 'n') {
41 isnt($experr, '', "Expected warning for $id provided");
42 warnings_like(sub {unpack 'C0U*', $octets}, [qr/$experr/],
43 "Only expected warning for $id");
44 } elsif ($okay !~ /^N(-?\d+)/) {
e1a67460 45 is($okay, 'n', "Confused test description for $id");
e1a67460 46 } else {
f96a66c2
NC
47 my $expect = $1;
48 my @warnings;
c03ceae6
NC
49
50 {
51 local $SIG{__WARN__} = sub {
52 print "# $id: @_";
f96a66c2 53 push @warnings, "@_";
c03ceae6
NC
54 };
55 unpack 'C0U*', $octets;
56 }
57
f96a66c2
NC
58 isnt($experr, '', "Expected first warning for $id provided");
59 like($warnings[0], qr/$experr/, "Expected first warning for $id seen");
60 local $::TODO;
61 if ($expect < 0) {
62 $expect = -$expect;
63 $::TODO = "Markus Kuhn states that $expect invalid sequences should be signalled";
64 }
65 is(scalar @warnings, $expect, "Expected number of warnings for $id seen");
e1a67460
NC
66 }
67 } else {
68 fail("unknown format '$_'");
69 }
70}
71
72done_testing();
73
a9917092
JH
74# This table is based on Markus Kuhn's UTF-8 Decode Stress Tester,
75# http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt,
daf0f78e 76# version dated 2000-09-02.
a9917092 77
e1a67460 78__DATA__
a9917092 791 Correct UTF-8
04f3c608 801.1.1 y 3ba,1f79,3c3,3bc,3b5 11 ce:ba:e1:bd:b9:cf:83:ce:bc:ce:b5 5
daf0f78e 812 Boundary conditions
a9917092 822.1 First possible sequence of certain length
bb6a3342
NC
832.1.1 y 0 1 00 1
842.1.2 y 80 2 c2:80 1
852.1.3 y 800 3 e0:a0:80 1
862.1.4 y 10000 4 f0:90:80:80 1
872.1.5 y 200000 5 f8:88:80:80:80 1
882.1.6 y 4000000 6 fc:84:80:80:80:80 1
a9917092 892.2 Last possible sequence of certain length
bb6a3342
NC
902.2.1 y 7f 1 7f 1
912.2.2 y 7ff 2 df:bf 1
8567041c 92# The ffff is legal by default since 872c91ae155f6880
bb6a3342
NC
932.2.3 y ffff 3 ef:bf:bf 1 character 0xffff
942.2.4 y 1fffff 4 f7:bf:bf:bf 1
952.2.5 y 3ffffff 5 fb:bf:bf:bf:bf 1
962.2.6 y 7fffffff 6 fd:bf:bf:bf:bf:bf 1
a9917092 972.3 Other boundary conditions
bb6a3342
NC
982.3.1 y d7ff 3 ed:9f:bf 1
992.3.2 y e000 3 ee:80:80 1
1002.3.3 y fffd 3 ef:bf:bd 1
1012.3.4 y 10ffff 4 f4:8f:bf:bf 1
1022.3.5 y 110000 4 f4:90:80:80 1
a9917092
JH
1033 Malformed sequences
1043.1 Unexpected continuation bytes
bb6a3342
NC
1053.1.1 n - 1 80 - unexpected continuation byte 0x80
1063.1.2 n - 1 bf - unexpected continuation byte 0xbf
f96a66c2
NC
1073.1.3 N2 - 2 80:bf - unexpected continuation byte 0x80
1083.1.4 N3 - 3 80:bf:80 - unexpected continuation byte 0x80
1093.1.5 N4 - 4 80:bf:80:bf - unexpected continuation byte 0x80
1103.1.6 N5 - 5 80:bf:80:bf:80 - unexpected continuation byte 0x80
1113.1.7 N6 - 6 80:bf:80:bf:80:bf - unexpected continuation byte 0x80
1123.1.8 N7 - 7 80:bf:80:bf:80:bf:80 - unexpected continuation byte 0x80
1133.1.9 N64 - 64 80:81:82:83:84:85:86:87:88:89:8a:8b:8c:8d:8e:8f:90:91:92:93:94:95:96:97:98:99:9a:9b:9c:9d:9e:9f:a0:a1:a2:a3:a4:a5:a6:a7:a8:a9:aa:ab:ac:ad:ae:af:b0:b1:b2:b3:b4:b5:b6:b7:b8:b9:ba:bb:bc:bd:be:bf - unexpected continuation byte 0x80
a9917092 1143.2 Lonely start characters
f96a66c2
NC
1153.2.1 N32 - 64 c0:20:c1:20:c2:20:c3:20:c4:20:c5:20:c6:20:c7:20:c8:20:c9:20:ca:20:cb:20:cc:20:cd:20:ce:20:cf:20:d0:20:d1:20:d2:20:d3:20:d4:20:d5:20:d6:20:d7:20:d8:20:d9:20:da:20:db:20:dc:20:dd:20:de:20:df:20 - unexpected non-continuation byte 0x20, immediately after start byte 0xc0
1163.2.2 N16 - 32 e0:20:e1:20:e2:20:e3:20:e4:20:e5:20:e6:20:e7:20:e8:20:e9:20:ea:20:eb:20:ec:20:ed:20:ee:20:ef:20 - unexpected non-continuation byte 0x20, immediately after start byte 0xe0
1173.2.3 N8 - 16 f0:20:f1:20:f2:20:f3:20:f4:20:f5:20:f6:20:f7:20 - unexpected non-continuation byte 0x20, immediately after start byte 0xf0
1183.2.4 N4 - 8 f8:20:f9:20:fa:20:fb:20 - unexpected non-continuation byte 0x20, immediately after start byte 0xf8
1193.2.5 N2 - 4 fc:20:fd:20 - unexpected non-continuation byte 0x20, immediately after start byte 0xfc
a9917092 1203.3 Sequences with last continuation byte missing
bb6a3342
NC
1213.3.1 n - 1 c0 - 1 byte, need 2
1223.3.2 n - 2 e0:80 - 2 bytes, need 3
1233.3.3 n - 3 f0:80:80 - 3 bytes, need 4
1243.3.4 n - 4 f8:80:80:80 - 4 bytes, need 5
1253.3.5 n - 5 fc:80:80:80:80 - 5 bytes, need 6
1263.3.6 n - 1 df - 1 byte, need 2
1273.3.7 n - 2 ef:bf - 2 bytes, need 3
1283.3.8 n - 3 f7:bf:bf - 3 bytes, need 4
1293.3.9 n - 4 fb:bf:bf:bf - 4 bytes, need 5
1303.3.10 n - 5 fd:bf:bf:bf:bf - 5 bytes, need 6
a9917092 1313.4 Concatenation of incomplete sequences
f96a66c2 1323.4.1 N-10 - 30 c0:e0:80:f0:80:80:f8:80:80:80:fc:80:80:80:80:df:ef:bf:f7:bf:bf:fb:bf:bf:bf:fd:bf:bf:bf:bf - unexpected non-continuation byte 0xe0, immediately after start byte 0xc0
a9917092 1333.5 Impossible bytes
bb6a3342
NC
1343.5.1 n - 1 fe - byte 0xfe
1353.5.2 n - 1 ff - byte 0xff
f96a66c2 1363.5.3 N4 - 4 fe:fe:ff:ff - byte 0xfe
a9917092
JH
1374 Overlong sequences
1384.1 Examples of an overlong ASCII character
bb6a3342
NC
1394.1.1 n - 2 c0:af - 2 bytes, need 1
1404.1.2 n - 3 e0:80:af - 3 bytes, need 1
1414.1.3 n - 4 f0:80:80:af - 4 bytes, need 1
1424.1.4 n - 5 f8:80:80:80:af - 5 bytes, need 1
1434.1.5 n - 6 fc:80:80:80:80:af - 6 bytes, need 1
a9917092 1444.2 Maximum overlong sequences
bb6a3342
NC
1454.2.1 n - 2 c1:bf - 2 bytes, need 1
1464.2.2 n - 3 e0:9f:bf - 3 bytes, need 2
1474.2.3 n - 4 f0:8f:bf:bf - 4 bytes, need 3
1484.2.4 n - 5 f8:87:bf:bf:bf - 5 bytes, need 4
1494.2.5 n - 6 fc:83:bf:bf:bf:bf - 6 bytes, need 5
a9917092 1504.3 Overlong representation of the NUL character
bb6a3342
NC
1514.3.1 n - 2 c0:80 - 2 bytes, need 1
1524.3.2 n - 3 e0:80:80 - 3 bytes, need 1
1534.3.3 n - 4 f0:80:80:80 - 4 bytes, need 1
1544.3.4 n - 5 f8:80:80:80:80 - 5 bytes, need 1
1554.3.5 n - 6 fc:80:80:80:80:80 - 6 bytes, need 1
a9917092
JH
1565 Illegal code positions
1575.1 Single UTF-16 surrogates
04f3c608
NC
1585.1.1 y d800 3 ed:a0:80 1 UTF-16 surrogate 0xd800
1595.1.2 y db7f 3 ed:ad:bf 1 UTF-16 surrogate 0xdb7f
1605.1.3 y db80 3 ed:ae:80 1 UTF-16 surrogate 0xdb80
1615.1.4 y dbff 3 ed:af:bf 1 UTF-16 surrogate 0xdbff
1625.1.5 y dc00 3 ed:b0:80 1 UTF-16 surrogate 0xdc00
1635.1.6 y df80 3 ed:be:80 1 UTF-16 surrogate 0xdf80
1645.1.7 y dfff 3 ed:bf:bf 1 UTF-16 surrogate 0xdfff
a9917092 1655.2 Paired UTF-16 surrogates
04f3c608
NC
1665.2.1 y d800,dc00 6 ed:a0:80:ed:b0:80 2 UTF-16 surrogates 0xd800, dc00
1675.2.2 y d800,dfff 6 ed:a0:80:ed:bf:bf 2 UTF-16 surrogates 0xd800, dfff
1685.2.3 y db7f,dc00 6 ed:ad:bf:ed:b0:80 2 UTF-16 surrogates 0xdb7f, dc00
1695.2.4 y db7f,dfff 6 ed:ad:bf:ed:bf:bf 2 UTF-16 surrogates 0xdb7f, dfff
1705.2.5 y db80,dc00 6 ed:ae:80:ed:b0:80 2 UTF-16 surrogates 0xdb80, dc00
1715.2.6 y db80,dfff 6 ed:ae:80:ed:bf:bf 2 UTF-16 surrogates 0xdb80, dfff
1725.2.7 y dbff,dc00 6 ed:af:bf:ed:b0:80 2 UTF-16 surrogates 0xdbff, dc00
1735.2.8 y dbff,dfff 6 ed:af:bf:ed:bf:bf 2 UTF-16 surrogates 0xdbff, dfff
a9917092 1745.3 Other illegal code positions
04f3c608 1755.3.1 y fffe 3 ef:bf:be 1 byte order mark 0xfffe
8567041c 176# The ffff is legal by default since 872c91ae155f6880
04f3c608 1775.3.2 y ffff 3 ef:bf:bf 1 character 0xffff