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
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
|
{
Some handly UTF8 function copied from the Lazarus LCL. Surely we can move
this into FPC?
}
unit gfx_UTF8utils;
{$mode objfpc}{$H+}
interface
uses
Classes, SysUtils;
function UTF8Pos(const SearchForText, SearchInText: string): integer;
function UTF8Copy(const s: string; StartCharIndex, CharCount: integer): string;
function UTF8Length(const s: string): integer;
function UTF8Length(p: PChar; ByteCount: integer): integer;
function UTF8CharStart(UTF8Str: PChar; Len, Index: integer): PChar;
function UTF8CharacterLength(p: PChar): integer;
{$Note I believe we need a UTF8Delete and UTF8Insert as well. }
implementation
// returns the character index, where the SearchForText starts in SearchInText
function UTF8Pos(const SearchForText, SearchInText: string): integer;
var
p: LongInt;
begin
p:=System.Pos(SearchForText,SearchInText);
if p>0 then
Result:=UTF8Length(PChar(SearchInText),p-1)+1
else
Result:=0;
end;
// returns substring
function UTF8Copy(const s: string; StartCharIndex, CharCount: integer): string;
var
StartBytePos: PChar;
EndBytePos: PChar;
MaxBytes: PtrInt;
begin
StartBytePos:=UTF8CharStart(PChar(s),length(s),StartCharIndex-1);
if StartBytePos=nil then
Result:=''
else
begin
MaxBytes:=PtrInt(PChar(s)+length(s)-StartBytePos);
EndBytePos:=UTF8CharStart(StartBytePos,MaxBytes,CharCount);
if EndBytePos=nil then
Result:=copy(s,StartBytePos-PChar(s)+1,MaxBytes)
else
Result:=copy(s,StartBytePos-PChar(s)+1,EndBytePos-StartBytePos);
end;
end;
function UTF8Length(const s: string): integer;
begin
Result := UTF8Length(PChar(s),length(s));
end;
function UTF8Length(p: PChar; ByteCount: integer): integer;
var
CharLen: LongInt;
begin
Result:=0;
while (ByteCount>0) do
begin
inc(Result);
CharLen:=UTF8CharacterLength(p);
inc(p,CharLen);
dec(ByteCount,CharLen);
end;
end;
function UTF8CharStart(UTF8Str: PChar; Len, Index: integer): PChar;
var
CharLen: LongInt;
begin
Result:=UTF8Str;
if Result<>nil then
begin
while (Index>0) and (Len>0) do
begin
CharLen:=UTF8CharacterLength(Result);
dec(Len,CharLen);
dec(Index);
inc(Result,CharLen);
end;
if (Index>0) or (Len<0) then
Result:=nil;
end;
end;
function UTF8CharacterLength(p: PChar): integer;
begin
if p<>nil then
begin
if ord(p^)< %11000000 then
begin
// regular single byte character (#0 is a character, this is pascal ;)
Result:=1;
end
else if ((ord(p^) and %11100000) = %11000000) then
begin
// could be 2 byte character
if (ord(p[1]) and %11000000) = %10000000 then
Result:=2
else
Result:=1;
end
else if ((ord(p^) and %11110000) = %11100000) then
begin
// could be 3 byte character
if ((ord(p[1]) and %11000000) = %10000000)
and ((ord(p[2]) and %11000000) = %10000000) then
Result:=3
else
Result:=1;
end
else if ((ord(p^) and %11111000) = %11110000) then
begin
// could be 4 byte character
if ((ord(p[1]) and %11000000) = %10000000)
and ((ord(p[2]) and %11000000) = %10000000)
and ((ord(p[3]) and %11000000) = %10000000) then
Result:=4
else
Result:=1;
end
else
Result:=1;
end
else
Result:=0;
end;
end.
|