unicode 0.3.1.1-x86-mingw32 → 0.4.0.1-x86-mingw32
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/README +7 -6
- data/Rakefile +8 -1
- data/ext/unicode/unicode.c +57 -2
- data/ext/unicode/unidata.map +1870 -2
- data/lib/unicode/1.8/unicode_native.so +0 -0
- data/lib/unicode/1.9/unicode_native.so +0 -0
- data/tools/normtest.rb +111 -0
- data/unicode.gemspec +2 -2
- metadata +6 -5
|
Binary file
|
|
Binary file
|
data/tools/normtest.rb
ADDED
|
@@ -0,0 +1,111 @@
|
|
|
1
|
+
#! /usr/local/bin/ruby -KU
|
|
2
|
+
|
|
3
|
+
## Conformance test with NormaliztionTest.txt
|
|
4
|
+
## Copyrigth 2010 yoshidam
|
|
5
|
+
|
|
6
|
+
require 'unicode'
|
|
7
|
+
|
|
8
|
+
TESTFILE = "NormalizationTest.txt"
|
|
9
|
+
|
|
10
|
+
def from_hex(str)
|
|
11
|
+
ret = ""
|
|
12
|
+
chars = str.split(" ")
|
|
13
|
+
chars.each do |c|
|
|
14
|
+
ret << [c.hex].pack("U")
|
|
15
|
+
end
|
|
16
|
+
return ret
|
|
17
|
+
end
|
|
18
|
+
|
|
19
|
+
def to_hex(str)
|
|
20
|
+
ret = ""
|
|
21
|
+
str = str.unpack('U*')
|
|
22
|
+
str.each do |c|
|
|
23
|
+
ret += sprintf("%04X ", c)
|
|
24
|
+
end
|
|
25
|
+
ret
|
|
26
|
+
end
|
|
27
|
+
|
|
28
|
+
open(TESTFILE) do |f|
|
|
29
|
+
while l = f.gets
|
|
30
|
+
next if l =~ /^#/
|
|
31
|
+
l.chomp
|
|
32
|
+
if l =~ /^@/
|
|
33
|
+
puts l
|
|
34
|
+
next
|
|
35
|
+
end
|
|
36
|
+
c1, c2, c3, c4, c5 = l.split(';')
|
|
37
|
+
code = c1
|
|
38
|
+
c1 = from_hex(c1)
|
|
39
|
+
c2 = from_hex(c2)
|
|
40
|
+
c3 = from_hex(c3)
|
|
41
|
+
c4 = from_hex(c4)
|
|
42
|
+
c5 = from_hex(c5)
|
|
43
|
+
## NFC TEST
|
|
44
|
+
if c2 == Unicode.nfc(c1) && c2 == Unicode.nfc(c2) &&
|
|
45
|
+
c2 == Unicode.nfc(c3) &&
|
|
46
|
+
c4 == Unicode.nfc(c4) && c4 == Unicode.nfc(c4)
|
|
47
|
+
##puts "NFC OK: " + code
|
|
48
|
+
else
|
|
49
|
+
puts "NFC NG: " + to_hex(c1)
|
|
50
|
+
printf(" c2=%s NFC(c1)=%s NFC(c2)=%s NFC(c3)=%s\n",
|
|
51
|
+
to_hex(c2),
|
|
52
|
+
to_hex(Unicode.nfc(c1)),
|
|
53
|
+
to_hex(Unicode.nfc(c2)),
|
|
54
|
+
to_hex(Unicode.nfc(c3)))
|
|
55
|
+
printf(" c4=%s NFC(c4)=%s NFC(c5)=%s\n",
|
|
56
|
+
to_hex(c4),
|
|
57
|
+
to_hex(Unicode.nfc(c4)),
|
|
58
|
+
to_hex(Unicode.nfc(c5)))
|
|
59
|
+
end
|
|
60
|
+
|
|
61
|
+
## NFD TEST
|
|
62
|
+
if c3 == Unicode.nfd(c1) && c3 == Unicode.nfd(c2) &&
|
|
63
|
+
c3 == Unicode.nfd(c3) &&
|
|
64
|
+
c5 == Unicode.nfd(c4) && c5 == Unicode.nfd(c5)
|
|
65
|
+
##puts "NFD OK: " + code
|
|
66
|
+
else
|
|
67
|
+
puts "NFD NG: " + to_hex(c1)
|
|
68
|
+
printf(" c3=%s NFD(c1)=%s NFD(c2)=%s NFD(c3)=%s\n",
|
|
69
|
+
to_hex(c3),
|
|
70
|
+
to_hex(Unicode.nfd(c1)),
|
|
71
|
+
to_hex(Unicode.nfd(c2)),
|
|
72
|
+
to_hex(Unicode.nfd(c3)))
|
|
73
|
+
printf(" c5=%s NFD(c4)=%s NFD(c5)=%s\n",
|
|
74
|
+
to_hex(c5),
|
|
75
|
+
to_hex(Unicode.nfd(c4)),
|
|
76
|
+
to_hex(Unicode.nfd(c5)))
|
|
77
|
+
end
|
|
78
|
+
|
|
79
|
+
## NFKC TEST
|
|
80
|
+
if c4 == Unicode.nfkc(c1) && c4 == Unicode.nfkc(c2) &&
|
|
81
|
+
c4 == Unicode.nfkc(c3) &&
|
|
82
|
+
c4 == Unicode.nfkc(c4) && c4 == Unicode.nfkc(c5)
|
|
83
|
+
##puts "NFKC OK: " + code
|
|
84
|
+
else
|
|
85
|
+
puts "NFKC NG: " + to_hex(c1)
|
|
86
|
+
printf(" c4=%s NFKC(c1)=%s NFKC(c2)=%s NFKC(c3)=%s NFKC(c4)=%s NFKC(c5)=%s\n",
|
|
87
|
+
to_hex(c4),
|
|
88
|
+
to_hex(Unicode.nfkc(c1)),
|
|
89
|
+
to_hex(Unicode.nfkc(c2)),
|
|
90
|
+
to_hex(Unicode.nfkc(c3)),
|
|
91
|
+
to_hex(Unicode.nfkc(c4)),
|
|
92
|
+
to_hex(Unicode.nfkc(c5)))
|
|
93
|
+
end
|
|
94
|
+
|
|
95
|
+
## NFKD TEST
|
|
96
|
+
if c5 == Unicode.nfkd(c1) && c5 == Unicode.nfkd(c2) &&
|
|
97
|
+
c5 == Unicode.nfkd(c3) &&
|
|
98
|
+
c5 == Unicode.nfkd(c4) && c5 == Unicode.nfkd(c5)
|
|
99
|
+
##puts "NFKD OK: " + code
|
|
100
|
+
else
|
|
101
|
+
puts "NFKD NG: " + to_hex(c1)
|
|
102
|
+
printf(" c5=%s NFKD(c1)=%s NFKD(c2)=%s NFKD(c3)=%s NFKD(c4)=%s NFKD(c5)=%s\n",
|
|
103
|
+
to_hex(c5),
|
|
104
|
+
to_hex(Unicode.nfkd(c1)),
|
|
105
|
+
to_hex(Unicode.nfkd(c2)),
|
|
106
|
+
to_hex(Unicode.nfkd(c3)),
|
|
107
|
+
to_hex(Unicode.nfkd(c4)),
|
|
108
|
+
to_hex(Unicode.nfkd(c5)))
|
|
109
|
+
end
|
|
110
|
+
end
|
|
111
|
+
end
|
data/unicode.gemspec
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
Gem::Specification.new { |s|
|
|
2
2
|
s.name = %q{unicode}
|
|
3
|
-
s.version = %q{0.
|
|
4
|
-
s.date = %q{2010-
|
|
3
|
+
s.version = %q{0.4.0.1}
|
|
4
|
+
s.date = %q{2010-10-14}
|
|
5
5
|
s.summary = %q{Unicode normalization library.}
|
|
6
6
|
s.require_paths = %w[lib]
|
|
7
7
|
s.author = %q{Yoshida Masato}
|
metadata
CHANGED
|
@@ -1,14 +1,14 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: unicode
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
hash:
|
|
4
|
+
hash: 109
|
|
5
5
|
prerelease: false
|
|
6
6
|
segments:
|
|
7
7
|
- 0
|
|
8
|
-
-
|
|
9
|
-
-
|
|
8
|
+
- 4
|
|
9
|
+
- 0
|
|
10
10
|
- 1
|
|
11
|
-
version: 0.
|
|
11
|
+
version: 0.4.0.1
|
|
12
12
|
platform: x86-mingw32
|
|
13
13
|
authors:
|
|
14
14
|
- Yoshida Masato
|
|
@@ -16,7 +16,7 @@ autorequire:
|
|
|
16
16
|
bindir: bin
|
|
17
17
|
cert_chain: []
|
|
18
18
|
|
|
19
|
-
date: 2010-
|
|
19
|
+
date: 2010-10-14 00:00:00 -03:00
|
|
20
20
|
default_executable:
|
|
21
21
|
dependencies: []
|
|
22
22
|
|
|
@@ -42,6 +42,7 @@ files:
|
|
|
42
42
|
- test/test.rb
|
|
43
43
|
- tools/README
|
|
44
44
|
- tools/mkunidata.rb
|
|
45
|
+
- tools/normtest.rb
|
|
45
46
|
- unicode.gemspec
|
|
46
47
|
- lib/unicode/1.8/unicode_native.so
|
|
47
48
|
- lib/unicode/1.9/unicode_native.so
|