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
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
|
#
#
# Nimrod's Runtime Library
# (c) Copyright 2012 Andreas Rumpf
#
# See the file "copying.txt", included in this
# distribution, for details about the copyright.
#
## Nimrod support for C/C++'s `wide strings`:idx:. This is part of the system
## module! Do not import it directly!
type
TUtf16Char* = distinct int16
WideCString* = ptr array[0.. 1_000_000, TUtf16Char]
proc len*(w: WideCString): int =
## returns the length of a widestring. This traverses the whole string to
## find the binary zero end marker!
while int16(w[result]) != 0'i16: inc result
when true:
const
UNI_REPLACEMENT_CHAR = TUtf16Char(0xFFFD'i16)
UNI_MAX_BMP = 0x0000FFFF
UNI_MAX_UTF16 = 0x0010FFFF
UNI_MAX_UTF32 = 0x7FFFFFFF
UNI_MAX_LEGAL_UTF32 = 0x0010FFFF
halfShift = 10
halfBase = 0x0010000
halfMask = 0x3FF
UNI_SUR_HIGH_START = 0xD800
UNI_SUR_HIGH_END = 0xDBFF
UNI_SUR_LOW_START = 0xDC00
UNI_SUR_LOW_END = 0xDFFF
template ones(n: expr): expr = ((1 shl n)-1)
template fastRuneAt(s: cstring, i: int, result: expr, doInc = true) =
## Returns the unicode character ``s[i]`` in `result`. If ``doInc == true``
## `i` is incremented by the number of bytes that have been processed.
bind ones
if ord(s[i]) <=% 127:
result = ord(s[i])
when doInc: inc(i)
elif ord(s[i]) shr 5 == 0b110:
#assert(ord(s[i+1]) shr 6 == 0b10)
result = (ord(s[i]) and (ones(5))) shl 6 or (ord(s[i+1]) and ones(6))
when doInc: inc(i, 2)
elif ord(s[i]) shr 4 == 0b1110:
#assert(ord(s[i+1]) shr 6 == 0b10)
#assert(ord(s[i+2]) shr 6 == 0b10)
result = (ord(s[i]) and ones(4)) shl 12 or
(ord(s[i+1]) and ones(6)) shl 6 or
(ord(s[i+2]) and ones(6))
when doInc: inc(i, 3)
elif ord(s[i]) shr 3 == 0b11110:
#assert(ord(s[i+1]) shr 6 == 0b10)
#assert(ord(s[i+2]) shr 6 == 0b10)
#assert(ord(s[i+3]) shr 6 == 0b10)
result = (ord(s[i]) and ones(3)) shl 18 or
(ord(s[i+1]) and ones(6)) shl 12 or
(ord(s[i+2]) and ones(6)) shl 6 or
(ord(s[i+3]) and ones(6))
when doInc: inc(i, 4)
else:
result = 0xFFFD
when doInc: inc(i)
iterator runes(s: cstring): int =
var
i = 0
result: int
while s[i] != '\0':
fastRuneAt(s, i, result, true)
yield result
proc allocWideCString*(source: cstring, L: int): WideCString =
## free after usage with `dealloc`.
result = cast[wideCString](alloc(L * 4 + 2))
var d = 0
for ch in runes(source):
if ch <=% UNI_MAX_BMP:
if ch >=% UNI_SUR_HIGH_START and ch <=% UNI_SUR_LOW_END:
result[d] = UNI_REPLACEMENT_CHAR
else:
result[d] = TUtf16Char(toU16(ch))
elif ch >% UNI_MAX_UTF16:
result[d] = UNI_REPLACEMENT_CHAR
else:
let ch = ch -% halfBase
result[d] = TUtf16Char(toU16((ch shr halfShift) +% UNI_SUR_HIGH_START))
inc d
result[d] = TUtf16Char(toU16((ch and halfMask) +% UNI_SUR_LOW_START))
inc d
result[d] = TUtf16Char(0'i16)
proc allocWideCString*(s: cstring): WideCString =
## free after usage with `dealloc`.
if s.isNil: return nil
when not defined(c_strlen):
proc c_strlen(a: CString): int {.nodecl, noSideEffect, importc: "strlen".}
let L = cstrlen(s)
result = allocWideCString(s, L)
proc allocWideCString*(s: string): WideCString =
## free after usage with `dealloc`.
result = allocWideCString(s, s.len)
proc `$`*(w: wideCString, estimate: int): string =
result = newStringOfCap(estimate + estimate shr 2)
var i = 0
while w[i].int16 != 0'i16:
var ch = w[i].int
inc i
if ch >=% UNI_SUR_HIGH_START and ch <=% UNI_SUR_HIGH_END:
# If the 16 bits following the high surrogate are in the source buffer...
let ch2 = w[i].int
# If it's a low surrogate, convert to UTF32:
if ch2 >=% UNI_SUR_LOW_START and ch2 <=% UNI_SUR_LOW_END:
ch = ((ch -% UNI_SUR_HIGH_START) shr halfShift) +%
(ch2 -% UNI_SUR_LOW_START) +% halfBase
inc i
if ch <=% 127:
result.add chr(ch)
elif ch <=% 0x07FF:
result.add chr((ch shr 6) or 0b110_00000)
result.add chr((ch and ones(6)) or 0b10_000000)
elif ch <=% 0xFFFF:
result.add chr(ch shr 12 or 0b1110_0000)
result.add chr(ch shr 6 and ones(6) or 0b10_0000_00)
result.add chr(ch and ones(6) or 0b10_0000_00)
elif ch <=% 0x0010FFFF:
result.add chr(ch shr 18 or 0b1111_0000)
result.add chr(ch shr 12 and ones(6) or 0b10_0000_00)
result.add chr(ch shr 6 and ones(6) or 0b10_0000_00)
result.add chr(ch and ones(6) or 0b10_0000_00)
else:
# replacement char:
result.add chr(0xFFFD shr 12 or 0b1110_0000)
result.add chr(0xFFFD shr 6 and ones(6) or 0b10_0000_00)
result.add chr(0xFFFD and ones(6) or 0b10_0000_00)
proc `$`*(s: WideCString): string =
result = s $ 80
else:
const
utf8Encoding = 65001
proc MultiByteToWideChar*(
CodePage: int32,
dwFlags: int32,
lpMultiByteStr: cstring,
cbMultiByte: cint,
lpWideCharStr: WideCString,
cchWideChar: cint): cint {.
stdcall, importc: "MultiByteToWideChar", dynlib: "kernel32".}
proc WideCharToMultiByte*(
CodePage: int32,
dwFlags: int32,
lpWideCharStr: WideCString,
cchWideChar: cint,
lpMultiByteStr: cstring,
cbMultiByte: cint,
lpDefaultChar: cstring=nil,
lpUsedDefaultChar: pointer=nil): cint {.
stdcall, importc: "WideCharToMultiByte", dynlib: "kernel32".}
proc raiseEncodingError() {.noinline, noreturn.} =
raise newException(EOS, "error in unicode conversion")
proc `$`*(s: WideCString, len: int): string =
# special case: empty string: needed because MultiByteToWideChar
# returns 0 in case of error:
if len == 0: return ""
# educated guess of capacity:
var cap = len + len shr 2
result = newStringOfCap(cap)
let m = WideCharToMultiByte(
CodePage = utf8Encoding,
dwFlags = 0'i32,
lpWideCharStr = s,
cchWideChar = cint(len),
lpMultiByteStr = cstring(result),
cbMultiByte = cap)
if m == 0:
# try again; ask for capacity:
cap = WideCharToMultiByte(
CodePage = utf8Encoding,
dwFlags = 0'i32,
lpWideCharStr = s,
cchWideChar = cint(len),
lpMultiByteStr = nil,
cbMultiByte = cint(0))
# and do the conversion properly:
result = newStringOfCap(cap)
let m = WideCharToMultiByte(
CodePage = utf8Encoding,
dwFlags = 0'i32,
lpWideCharStr = s,
cchWideChar = cint(len),
lpMultiByteStr = cstring(result),
cbMultiByte = cap)
if m == 0: raiseEncodingError()
setLen(result, m)
elif m <= cap:
setLen(result, m)
else:
sysAssert(false, "") # cannot happen
proc `$`*(s: WideCString): string =
result = s $ s.len
proc allocWideCString*(s: string): WideCString =
## free after usage with `dealloc`.
let cap = s.len+1
result = cast[wideCString](alloc0(cap * 2))
# special case: empty string: needed because MultiByteToWideChar
# return 0 in case of error:
if s.len == 0: return
# convert to utf-16 LE
let m = MultiByteToWideChar(CodePage = utf8Encoding, dwFlags = 0'i32,
lpMultiByteStr = cstring(s),
cbMultiByte = cint(s.len),
lpWideCharStr = result,
cchWideChar = cint(cap))
if m == 0: raiseEncodingError()
proc allocWideCString*(s: cstring): WideCString =
## free after usage with `dealloc`.
if s.isNil: return nil
when not defined(c_strlen):
proc c_strlen(a: CString): int {.nodecl, noSideEffect, importc: "strlen".}
let len = cstrlen(s)
let cap = len+1
result = cast[wideCString](alloc0(cap * 2))
# special case: empty string: needed because MultiByteToWideChar
# return 0 in case of error:
if s.len == 0: return
# convert to utf-16 LE
let m = MultiByteToWideChar(CodePage = utf8Encoding, dwFlags = 0'i32,
lpMultiByteStr = s,
cbMultiByte = cint(len),
lpWideCharStr = result,
cchWideChar = cint(cap))
if m == 0: raiseEncodingError()
|