1 #!/opt/lua-5.0.2/bin/lua
2 -- there are four string-like ctype closures:
3 -- unicode.ascii, latin1, utf8 and grapheme
5 -- ascii and latin1 are single-byte like string,
6 -- but use the unicode table for upper/lower and character classes
7 -- ascii does not touch bytes > 127 on upper/lower
9 -- ascii or latin1 can be used as locale-independent string replacement.
10 -- (There is a compile switch to do this automatically for ascii).
12 -- UTF-8 operates on UTF-8 sequences as of RFC 3629:
13 -- 1 byte 0-7F, 2 byte 80-7FF, 3 byte 800-FFFF, 4 byte 1000-10FFFF
14 -- (not exclusing UTF-16 surrogate characters)
15 -- Any byte not part of such a sequence is treated as it's (Latin-1) value.
17 -- Grapheme takes care of grapheme clusters, which are characters followed by
18 -- "grapheme extension" characters (Mn+Me) like combining diacritical marks.
22 -- sub(str, start [,end=-1])
23 -- byte(str, start [,end=-1])
29 -- same as in string: rep, format, dump
30 -- TODO: use char count with %s in format? (sub does the job)
31 -- TODO: grapheme.byte: only first code of any cluster?
33 -- find, gfind, gsub: done, but need thorough testing ...:
34 -- ascii does not match them on any %class (but on ., literals and ranges)
35 -- behaviour of %class with class not ASCII is undefined
36 -- frontier %f currently disabled -- should we?
38 -- character classes are:
39 -- %a L* (Lu+Ll+Lt+Lm+Lo)
43 -- %n N* (Nd+Nl+No, new)
44 -- %p P* (Pc+Pd+Ps+Pe+Pi+Pf+Po)
45 -- %s Z* (Zs+Zl+Zp) plus the controls 9-13 (HT,LF,VT,FF,CR)
47 -- %w %a+%n+Pc (e.g. '_')
50 -- c.f. http://unicode.org/Public/UNIDATA/UCD.html#General_Category_Values
51 -- http://unicode.org/Public/UNIDATA/UnicodeData.txt
53 -- NOTE: find positions are in bytes for all ctypes!
54 -- use ascii.sub to cut found ranges!
55 -- this is a) faster b) more reliable
57 -- UTF-8 behaviour: match is by codes, code ranges are supported
59 -- grapheme behaviour: any %class, '.' and range match includes
60 -- any following grapheme extensions.
61 -- Ranges apply to single code points only.
62 -- If a [] enumeration contains a grapheme cluster,
63 -- this matches only the exact same cluster.
64 -- However, a literal single 'o' standalone or in an [] enumeration
65 -- will match just that 'o', even if it has a extension in the string.
66 -- Consequently, grapheme match positions are not always cluster positions.
69 local unicode = require("unicode")
70 local utf8 = unicode.utf8
71 unicode.string = string -- for tests unicode[ctype]
72 local sprintf = string.format
73 local function printf (fmt, ...) return print(sprintf(fmt, ...)) end
75 local function check (test, ok, got)
76 if ok == got then return printf("ok %s = %s",test,ok) end
77 return printf("NOK %s = %s GOT '%s'",test, ok, got or "<nil>")
79 local function checka (test, ok, ...)
82 return check(test, ok, table.concat(arg, ","))
86 local function testlen (str,bytes,codes,chars)
87 codes = codes or bytes
88 chars = chars or codes
89 return check(sprintf("len '%s'", str),
90 sprintf("%d/%d/%d", bytes, codes, chars),
91 sprintf("%d/%d/%d", string.len(str), utf8.len(str), unicode.grapheme.len(str)))
94 -- 176 = 00B0;DEGREE SIGN -- UTF-8: C2,B0 = \194\176
95 -- 196 = 00C4;LATIN CAPITAL LETTER A WITH DIAERESIS
96 -- 214 = 00D6;LATIN CAPITAL LETTER O WITH DIAERESIS
97 -- 776 = 0308;COMBINING DIAERESIS -- UTF-8: CC,88 = \204\136
98 testlen("A\tB",3) -- plain Latin-1
99 testlen("\176\196\214",3) -- plain Latin-1
100 testlen("\196\176\214",3,2) -- C4,B0 is valid seq 0130 I WITH DOT ABOVE
101 testlen("\192\178",2) -- C0,B2 is bad seq for 2
102 testlen("°ÄÖ",6,3) -- simple Latin-1 chars in UTF-8
103 testlen("\204\136A\204\136O\204\136",8,5,3) -- decomposed (with broken lead)
106 local function testsub (ctype,ok,str,start,e)
107 return check(sprintf("%s.sub('%s',%d,%d)", ctype, str, start, e), ok,
108 unicode[ctype].sub(str,start,e))
110 testsub("ascii","BCD","ABCDE",2,4)
111 testsub("utf8","BCD","ABCDE",2,4)
112 testsub("latin1","Ä","°ÄÖ",3,4)
113 testsub("utf8","Ä","°ÄÖ",2,2)
114 testsub("utf8","ÄÖ","°ÄÖ",2,-1)
115 testsub("utf8","\204\136","A\204\136O\204\136",2,2) -- decomposed
116 testsub("grapheme","O\204\136","A\204\136O\204\136",2,2) -- decomposed
119 local function testbyte (ctype, ok, str, ...)
120 return checka(sprintf("%s.byte('%s',%s)",ctype,str,table.concat({...}, ",")),
121 ok, unicode[ctype].byte(str, ...))
123 testbyte("string","194,176","Ä°Ö",3,4) -- the UTF-8 seq for °
124 testbyte("ascii","194,176","Ä°Ö",3,4)
125 testbyte("utf8","176,214","Ä°Ö",2,3) -- code points for °,Ö
126 testbyte("utf8","65,776","\204\136A\204\136O\204\136",2,3) -- decomposed
127 testbyte("grapheme","65,776","\204\136A\204\136O\204\136",2) -- decomposed
130 local function testchar (ctype, ok, ...)
131 return check(sprintf("%s.char(%s)",ctype,table.concat({...}, ",")),
132 ok, unicode[ctype].char(...))
134 testchar("ascii", "AB", 65,66)
135 testchar("ascii", "\176", 176)
136 testchar("utf8", "\194\176", 176)
139 local function testcase (ctype,str,up,lo)
140 check(sprintf("%s.lower('%s')", ctype, str), lo, unicode[ctype].lower(str))
141 check(sprintf("%s.upper('%s')", ctype, str), up, unicode[ctype].upper(str))
143 -- upper/lower also fixes plain Latin
144 testcase("utf8","Ab\196üo\204\136","ABÄÜO\204\136","abäüo\204\136")
145 testcase("ascii","Ab\196üo\204\136","AB\196üO\204\136","ab\196üo\204\136")
146 testcase("latin1","Ab\196","AB\196","ab\228")
149 local function testrev (ctype,ok,str)
150 return check(sprintf("%s.reverse('%s')",ctype,str),
151 ok, unicode[ctype].reverse(str))
153 testrev("ascii","b\136\204oa\176\194ba","ab°ao\204\136b");
154 testrev("utf8","b\204\136oa°ba","ab°ao\204\136b");
155 testrev("grapheme","bo\204\136a°ba","ab°ao\204\136b");
159 local function testfind (ctype,ok,str,pat)
160 return checka(sprintf("%s.find('%s','%s')",ctype,str,pat),
161 ok, unicode[ctype].find(str, pat))
163 testfind("ascii","1,1","e=mc2","%a")
164 testfind("ascii","3,4","e=mc2","%a%a")
165 testfind("ascii","5,5","e=mc2","%d")
166 testfind("ascii","","Ä","%a")
167 testfind("ascii","1,2","Ä","%A*")
168 testfind("latin1","1,1","Ä","%a")
169 testfind("utf8","1,2","Ä","%a")
170 testfind("utf8","1,1","o\204\136","%a*")
171 testfind("utf8","2,3","o\204\136","%A")
172 testfind("utf8","1,1","o\204\136",".")
173 testfind("grapheme","1,3","o\204\136","%a*")
174 testfind("grapheme","2,3","o\204\136","%A") -- didn't expect this?
175 testfind("grapheme","1,3","o\204\136",".")
176 testfind("utf8","4,5","ÜHÄPPY","[À-Ö]")
177 testfind("utf8","4,5","ÜHÄPPY","[Ä-]")
178 testfind("utf8","7,7","ÜHÄP-PY","[ä-]")
179 testfind("ascii","1,4","abcdef","%a*d")
180 testfind("utf8","1,10","äöüßü","%a*ü")
181 testfind("utf8","1,6","äöüß","%a*ü")
182 testfind("utf8","4,5,Ä","ÜHÄPPY","([À-Ö])")
183 testfind("utf8","1,5,ÜHÄ","ÜHÄ_PPY","([%w]+)")
184 testfind("utf8","1,9,ÜHÄ_PPY","ÜHÄ_PPY","([%w_]+)")
187 local function testgsub (ctype,ok,str,pat,repl)
188 return check(sprintf("%s.gsub('%s','%s','%s')",ctype,str,pat,repl),
189 ok, unicode[ctype].gsub(str,pat,repl))
191 testgsub("ascii","hello hello world world","hello world", "(%w+)", "%1 %1")
192 testgsub("ascii","world hello Lua from",
193 "hello world from Lua", "(%w+)%s*(%w+)", "%2 %1")
194 testgsub("ascii","l helö wöfr rldöL müä",
195 "hellö wörld fröm Lüä", "(%w+)%s*(%w+)", "%2 %1")
196 testgsub("utf8","wörld hellö Lüä fröm",
197 "hellö wörld fröm Lüä", "(%w+)%s*(%w+)", "%2 %1")
198 testgsub("utf8","HÜppÄ","HÄppÜ","([À-Ö])(%l*)(%u)","%3%2%1")
202 for i=0,65535 do if i ~= utf8.byte(utf8.char(i)) then fail=fail+1 end end
203 check("code-decode failures", 0, fail)
206 for i=192,65535,64 do
208 io.write(sprintf("%04x\\%3d\\%3d ",i, 224+k/64, 128+math.mod(k,64)))
210 io.write(utf8.char(j))