authorgravatar for thejoshwolfe@gmail.comJosh Wolfe <thejoshwolfe@gmail.com> 2018-04-29 17:28:11-04:00
committergravatar for thejoshwolfe@gmail.comJosh Wolfe <thejoshwolfe@gmail.com> 2018-04-29 17:28:11-04:00
log2387292f204c59259fc64d7c960d201e808af5a9
tree71a74da3726583f8ca982e3c041124133ff54c97
parent8c567d84f1f14e06286897fe1b2408a1d2fd7d76

move some checks around in utf8Encode logic to be more zig idiomatic

1 files changed, 37 insertions(+), 42 deletions(-)

std/unicode.zig+37-42
......@@ -1,12 +1,11 @@
11const std = @import("./index.zig");
22const debug = std.debug;
33
4// Given a Utf8-Codepoint returns how many (1-4)
5// bytes there are if represented as an array of bytes.
4/// Returns how many bytes the UTF-8 representation would require
5/// for the given codepoint.
66pub fn utf8CodepointSequenceLength(c: u32) !u3 {
77 if (c < 0x80) return u3(1);
88 if (c < 0x800) return u3(2);
9 if (c -% 0xd800 < 0x800) return error.InvalidCodepoint;
109 if (c < 0x10000) return u3(3);
1110 if (c < 0x110000) return u3(4);
1211 return error.CodepointTooLarge;
......@@ -23,45 +22,39 @@ pub fn utf8ByteSequenceLength(first_byte: u8) !u3 {
2322 return error.Utf8InvalidStartByte;
2423}
2524
26/// Encodes a code point back into utf8
27/// c: the code point
28/// out: the out buffer to write to
29/// Notes: out has to have a len big enough for the bytes
30/// however this limit is dependent on the code point
31/// but giving it a minimum of 4 will ensure it will work
32/// for all code points.
33/// Errors: Will return an error if the code point is invalid.
25/// Encodes the given codepoint into a UTF-8 byte sequence.
26/// c: the codepoint.
27/// out: the out buffer to write to. Must have a len >= utf8CodepointSequenceLength(c).
28/// Errors: if c cannot be encoded in UTF-8.
29/// Returns: the number of bytes written to out.
3430pub fn utf8Encode(c: u32, out: []u8) !u3 {
35 if (utf8CodepointSequenceLength(c)) |length| {
36 debug.assert(out.len >= length);
37 switch (length) {
38 // The pattern for each is the same
39 // - Increasing the initial shift by 6 each time
40 // - Each time after the first shorten the shifted
41 // value to a max of 0b111111 (63)
42 1 => out[0] = u8(c), // Can just do 0 + codepoint for initial range
43 2 => {
44 out[0] = u8(0b11000000 | (c >> 6));
45 out[1] = u8(0b10000000 | (c & 0b111111));
46 },
47 3 => {
48 out[0] = u8(0b11100000 | (c >> 12));
49 out[1] = u8(0b10000000 | ((c >> 6) & 0b111111));
50 out[2] = u8(0b10000000 | (c & 0b111111));
51 },
52 4 => {
53 out[0] = u8(0b11110000 | (c >> 18));
54 out[1] = u8(0b10000000 | ((c >> 12) & 0b111111));
55 out[2] = u8(0b10000000 | ((c >> 6) & 0b111111));
56 out[3] = u8(0b10000000 | (c & 0b111111));
57 },
58 else => unreachable,
59 }
60
61 return length;
62 } else |err| {
63 return err;
31 const length = try utf8CodepointSequenceLength(c);
32 debug.assert(out.len >= length);
33 switch (length) {
34 // The pattern for each is the same
35 // - Increasing the initial shift by 6 each time
36 // - Each time after the first shorten the shifted
37 // value to a max of 0b111111 (63)
38 1 => out[0] = u8(c), // Can just do 0 + codepoint for initial range
39 2 => {
40 out[0] = u8(0b11000000 | (c >> 6));
41 out[1] = u8(0b10000000 | (c & 0b111111));
42 },
43 3 => {
44 if (0xd800 <= c and c <= 0xdfff) return error.Utf8CannotEncodeSurrogateHalf;
45 out[0] = u8(0b11100000 | (c >> 12));
46 out[1] = u8(0b10000000 | ((c >> 6) & 0b111111));
47 out[2] = u8(0b10000000 | (c & 0b111111));
48 },
49 4 => {
50 out[0] = u8(0b11110000 | (c >> 18));
51 out[1] = u8(0b10000000 | ((c >> 12) & 0b111111));
52 out[2] = u8(0b10000000 | ((c >> 6) & 0b111111));
53 out[3] = u8(0b10000000 | (c & 0b111111));
54 },
55 else => unreachable,
6456 }
57 return length;
6558}
6659
6760/// Decodes the UTF-8 codepoint encoded in the given slice of bytes.
......@@ -249,8 +242,10 @@ test "utf8 encode" {
249242
250243test "utf8 encode error" {
251244 var array: [4]u8 = undefined;
252 testErrorEncode(0xFFFFFF, array[0..], error.CodepointTooLarge);
253 testErrorEncode(0xd900, array[0..], error.InvalidCodepoint);
245 testErrorEncode(0xd800, array[0..], error.Utf8CannotEncodeSurrogateHalf);
246 testErrorEncode(0xdfff, array[0..], error.Utf8CannotEncodeSurrogateHalf);
247 testErrorEncode(0x110000, array[0..], error.CodepointTooLarge);
248 testErrorEncode(0xffffffff, array[0..], error.CodepointTooLarge);
254249}
255250
256251fn testErrorEncode(codePoint: u32, array: []u8, expectedErr: error) void {