1 # This file is part of NIT ( http://www.nitlanguage.org ).
3 # This file is free software, which comes along with NIT. This software is
4 # distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
5 # without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
6 # PARTICULAR PURPOSE. You can modify it is you want, provided this header
7 # is kept unaltered, and a notification of the changes is added.
8 # You are allowed to redistribute it and sell it, alone or is a part of
11 # Native structures for text and bytes
22 #include <libkern/OSByteOrder.h>
23 #define be32toh(x) OSSwapBigToHostInt32(x)
27 #define be16toh(val) (((val) >> 8) | ((val) << 8))
28 #define be32toh(val) ((be16toh((val) << 16) | (be16toh((val) >> 16))))
31 #define be32toh(val) betoh32(val)
36 # Gives the length of the UTF-8 char starting with `self`
38 if self & 0b1000_0000u
8 == 0u8
then
40 else if self & 0b1110_0000u
8 == 0b1100_0000u
8 then
42 else if self & 0b1111_0000u
8 == 0b1110_0000u
8 then
44 else if self & 0b1111_1000u
8 == 0b1111_0000u
8 then
51 # Is `self` a valid UTF-8 sequence start ?
54 # assert 0u8.is_valid_utf8_start
55 # assert 0xC0u8.is_valid_utf8_start
56 # assert 0xE0u8.is_valid_utf8_start
57 # assert 0xF0u8.is_valid_utf8_start
59 fun is_valid_utf8_start
: Bool do
60 if self & 0x80u
8 == 0u8
then return true
61 if self & 0b1110_0000u
8 == 0b1100_0000u
8 then return true
62 if self & 0b1111_0000u
8 == 0b1110_0000u
8 then return true
63 if self & 0b1111_1000u
8 == 0b1111_0000u
8 then return true
69 # Returns the code_point from a utf16 surrogate pair
71 # assert 0xD83DDE02.from_utf16_surr == 0x1F602
72 fun from_utf16_surr
: Int do
73 var hi
= (self & 0xFFFF0000) >> 16
74 var lo
= self & 0xFFFF
76 cp
+= (hi
- 0xD800) << 10
83 # Native strings are simple C char *
84 extern class NativeString `{ char* `}
85 # Creates a new NativeString with a capacity of `length
`
86 new(length: Int) is intern
88 # Returns a char* starting at `index
`.
90 # WARNING: Unsafe for extern code, use only for temporary
91 # pointer manipulation purposes (e.g. write to file or such)
92 fun fast_cstring(index: Int): NativeString is intern
94 # Get char at `index
`.
95 fun [](index: Int): Byte is intern
97 # Set char `item
` at index.
98 fun []=(index: Int, item: Byte) is intern
100 # Copy `self` to `dest
`.
101 fun copy_to(dest: NativeString, length: Int, from: Int, to: Int) is intern
103 redef fun ==(o) is intern do return is_same_instance(o)
105 redef fun !=(o) is intern do return not is_same_instance(o)
107 # Position of the first nul character.
108 fun cstring_length: Int
111 while self[l] != 0u8 do l += 1
115 # Parse `self` as an Int.
116 fun atoi: Int is intern
118 # Parse `self` as a Float.
119 fun atof: Float `{ return atof(self); `}
121 # Gets the UTF-8 char at index `pos`
123 # Index is expressed in Unicode chars
126 # assert "かきく".as(FlatString).items.char_at(0) == 'か'
129 # If the char at position pos is an invalid Unicode char,
130 # the Unicode replacement character � (0xFFFD) will be used.
133 # assert "かきく".as(FlatString).items.char_at(1) == '�'
135 fun char_at
(pos
: Int): Char do
137 if c
& 0x80u
8 == 0u8
then return c
.ascii
138 var b
= fetch_4_hchars
(pos
)
140 if b
& 0xC00000 != 0x800000 then return 0xFFFD.code_point
141 if b
& 0xE0000000 == 0xC0000000 then
142 ret
|= (b
& 0x1F000000) >> 18
143 ret
|= (b
& 0x3F0000) >> 16
144 return ret
.code_point
146 if not b
& 0xC000 == 0x8000 then return 0xFFFD.code_point
147 if b
& 0xF0000000 == 0xE0000000 then
148 ret
|= (b
& 0xF000000) >> 12
149 ret
|= (b
& 0x3F0000) >> 10
150 ret
|= (b
& 0x3F00) >> 8
151 return ret
.code_point
153 if not b
& 0xC0 == 0x80 then return 0xFFFD.code_point
154 if b
& 0xF8000000 == 0xF0000000 then
155 ret
|= (b
.to_i
& 0x7000000) >> 6
156 ret
|= (b
.to_i
& 0x3F0000) >> 4
157 ret
|= (b
.to_i
& 0x3F00) >> 2
159 return ret
.code_point
161 return 0xFFFD.code_point
164 # Gets the byte index of char at position `n` in UTF-8 String
165 fun char_to_byte_index
(n
: Int): Int do return char_to_byte_index_cached
(n
, 0, 0)
167 # Gets the length of the character at position `pos` (1 if invalid sequence)
168 fun length_of_char_at
(pos
: Int): Int do
170 if c
& 0x80u
8 == 0x00u
8 then
172 else if c
& 0xE0u
8 == 0xC0u
8 and self[pos
+ 1] & 0xC0u
8 == 0x80u
8 then
174 else if c
& 0xF0u
8 == 0xE0u
8 and self[pos
+ 1] & 0xC0u
8 == 0x80u
8 and self[pos
+ 2] & 0xC0u
8 == 0x80u
8 then
176 else if c
& 0xF8u
8 == 0xF0u
8 and self[pos
+ 1] & 0xC0u
8 == 0x80u
8 and self[pos
+ 2] & 0xC0u
8 == 0x80u
8 and self[pos
+ 3] & 0xC0u
8 == 0x80u
8 then
183 # Gets the byte index of char at position `n` in UTF-8 String
185 # `char_from` and `byte_from` are cached values to seek from.
187 # NOTE: char_from and byte_from are not guaranteed to be valid cache values
188 # It it up to the client to ensure the validity of the information
189 fun char_to_byte_index_cached
(n
, char_from
, byte_from
: Int): Int do
197 var i
= fetch_4_chars
(ns_i
)
198 if i
& 0x80808080 != 0 then break
203 if dist
== 0 then break
204 ns_i
+= length_of_char_at
(ns_i
)
211 var i
= fetch_4_chars
(ns_i
- 4)
212 if i
& 0x80808080 != 0 then break
217 if dist
== 0 then break
218 ns_i
= find_beginning_of_char_at
(ns_i
- 1)
226 # Gets the char index of byte at position `n` in a UTF-8 String
228 # `char_from` and `byte_from` are cached values to seek from.
230 # NOTE: char_from and byte_from are not guaranteed to be valid cache values
231 # It it up to the client to ensure the validity of the information
232 fun byte_to_char_index_cached
(n
, char_from
, byte_from
: Int): Int do
237 ns_i
+= length_of_char_at
(ns_i
)
242 ns_i
= find_beginning_of_char_at
(ns_i
- 1)
249 # Returns the beginning position of the char at position `pos`
251 # If the char is invalid UTF-8, `pos` is returned as-is
254 # assert "abc".items.find_beginning_of_char_at(2) == 2
255 # assert "か".items.find_beginning_of_char_at(1) == 0
256 # assert [0x41u8, 233u8].to_s.items.find_beginning_of_char_at(1) == 1
258 fun find_beginning_of_char_at
(pos
: Int): Int do
261 if c
& 0x80u
8 == 0x00u
8 then return pos
262 while c
& 0xC0u
8 == 0x80u
8 do
267 if length_of_char_at
(stpos
) >= (endpos
- stpos
+ 1) then return pos
271 # Number of UTF-8 characters in `self` starting at `from`, for a length of `bytelen`
272 fun utf8_length
(from
, bytelen
: Int): Int is intern do
276 while bytelen
>= 4 do
277 var i
= fetch_4_chars
(st
)
278 if i
& 0x80808080 != 0 then break
283 if bytelen
== 0 then break
284 var cln
= length_of_char_at
(st
)
292 # Fetch 4 chars in `self` at `pos`
293 fun fetch_4_chars
(pos
: Int): Int is intern `{ return (long)*((uint32_t*)(self+pos)); `}
295 # Fetch 4 chars in `self` at `pos
`
296 fun fetch_4_hchars(pos: Int): Int is intern `{ return (long)be32toh(*((uint32_t*)(self+pos))); `}
299 # Right shifts `len` bytes of `self` from `sh` bytes starting at position `pos`
300 fun rshift
(sh
, len
, pos
: Int) do
301 copy_to
(self, len
, pos
, pos
+ sh
)
304 # Left shifts `len` bytes of `self` from `sh` bytes starting at position `pos`
305 fun lshift
(sh
, len
, pos
: Int) do
306 copy_to
(self, len
, pos
, pos
- sh
)