1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
|
BEGIN {
chdir 't' if -d 't';
@INC = '../lib';
require './test.pl';
}
plan tests => 24;
my $a = chr(0x100);
is(ord($a), 0x100, "ord sanity check");
is(length($a), 1, "length sanity check");
is(substr($a, 0, 1), "\x{100}", "substr sanity check");
is(index($a, "\x{100}"), 0, "index sanity check");
is(rindex($a, "\x{100}"), 0, "rindex sanity check");
is(bytes::length($a), 2, "bytes::length sanity check");
is(bytes::chr(0x100), chr(0), "bytes::chr sanity check");
{
use bytes;
my $b = chr(0x100); # affected by 'use bytes'
is(ord($b), 0, "chr truncates under use bytes");
is(length($b), 1, "length truncated under use bytes");
is(bytes::ord($b), 0, "bytes::ord truncated under use bytes");
is(bytes::length($b), 1, "bytes::length truncated under use bytes");
is(bytes::substr($b, 0, 1), "\0", "bytes::substr truncated under use bytes");
}
my $c = chr(0x100);
my $c2 = chr(0x2c7); # a unicode character that doesn't fold
utf8::encode(my $c2_utf8 = $c2);
{
use bytes;
if (ord('A') == 193) { # EBCDIC?
is(ord($c), 0x8c, "ord under use bytes looks at the 1st byte");
} else {
is(ord($c), 0xc4, "ord under use bytes looks at the 1st byte");
}
is(length($c), 2, "length under use bytes looks at bytes");
is(bytes::length($c), 2, "bytes::length under use bytes looks at bytes");
if (ord('A') == 193) { # EBCDIC?
is(bytes::ord($c), 0x8c, "bytes::ord under use bytes looks at the 1st byte");
} else {
is(bytes::ord($c), 0xc4, "bytes::ord under use bytes looks at the 1st byte");
}
# In z/OS \x41,\x8c are the codepoints corresponding to \x80,\xc4 respectively under ASCII platform
if (ord('A') == 193) { # EBCDIC?
is(bytes::substr($c, 0, 1), "\x8c", "bytes::substr under use bytes looks at bytes");
is(bytes::index($c, "\x41"), 1, "bytes::index under use bytes looks at bytes");
is(bytes::rindex($c, "\x8c"), 0, "bytes::rindex under use bytes looks at bytes");
}
else{
is(bytes::substr($c, 0, 1), "\xc4", "bytes::substr under use bytes looks at bytes");
is(bytes::index($c, "\x80"), 1, "bytes::index under use bytes looks at bytes");
is(bytes::rindex($c, "\xc4"), 0, "bytes::rindex under use bytes looks at bytes");
}
# [perl #117355] [lu]cfirst don't respect 'use bytes'
# and if there's other tests for lc/uc under bytes I didn't find them
is(lc($c2), $c2_utf8, "lc under use bytes returns bytes");
is(uc($c2), $c2_utf8, "uc under use bytes returns bytes");
is(lcfirst($c2), $c2_utf8, "lcfirst under use bytes returns bytes");
is(ucfirst($c2), $c2_utf8, "unfirst under use bytes returns bytes");
}
{
fresh_perl_like ('use bytes; bytes::moo()',
qr/Undefined subroutine bytes::moo/, {stderr=>1},
"Check Carp is loaded for AUTOLOADing errors")
}
|