mirror/zig
1
0
Fork 0
mirror of https://github.com/ziglang/zig.git synced 2025-12-06 06:13:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

fixes and improvements for parseCharLiteral

Browse Source
This commit is contained in:
dbandstra 2020-08-19 13:10:05 -07:00 committed by Andrew Kelley
parent 083c0f1ceb
commit a237283d90
1 changed files with 47 additions and 39 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

86
lib/std/zig.zig
View File

@ -84,7 +84,7 @@ pub fn binNameAlloc(
/// Slice must be valid utf8 starting and ending with "'" and exactly one codepoint in between.
pub fn parseCharLiteral(
slice: []const u8,
bad_index: *usize, // populated if error.InvalidCharacter is returned)
bad_index: *usize, // populated if error.InvalidCharacter is returned
) error{InvalidCharacter}!u32 {
std.debug.assert(slice.len >= 3 and slice[0] == '\'' and slice[slice.len - 1] == '\'');
@ -101,41 +101,8 @@ pub fn parseCharLiteral(
bad_index.* = slice.len - 2;
return error.InvalidCharacter;
}
var value: u32 = 0;
for (slice[3..5]) |c, i| {
switch (slice[3]) {
'0'...'9' => {
value *= 16;
value += c - '0';
},
'a'...'f' => {
value *= 16;
value += c - 'a';
},
'A'...'F' => {
value *= 16;
value += c - 'a';
},
else => {
bad_index.* = i;
return error.InvalidCharacter;
},
}
}
return value;
},
'u' => {
if (slice.len < 6 or slice[3] != '{') {
bad_index.* = 2;
return error.InvalidCharacter;
}
var value: u32 = 0;
for (slice[4..]) |c, i| {
if (value > 0x10ffff) {
bad_index.* = i;
return error.InvalidCharacter;
}
switch (c) {
'0'...'9' => {
value *= 16;
@ -143,25 +110,56 @@ pub fn parseCharLiteral(
},
'a'...'f' => {
value *= 16;
value += c - 'a';
value += c - 'a' + 10;
},
'A'...'F' => {
value *= 16;
value += c - 'A';
value += c - 'A' + 10;
},
'}' => break,
else => {
bad_index.* = i;
bad_index.* = 3 + i;
return error.InvalidCharacter;
},
}
}
return value;
},
'u' => {
if (slice.len < "'\\u{0}'".len or slice[3] != '{' or slice[slice.len - 2] != '}') {
bad_index.* = 2;
return error.InvalidCharacter;
}
var value: u32 = 0;
for (slice[4 .. slice.len - 2]) |c, i| {
switch (c) {
'0'...'9' => {
value *= 16;
value += c - '0';
},
'a'...'f' => {
value *= 16;
value += c - 'a' + 10;
},
'A'...'F' => {
value *= 16;
value += c - 'A' + 10;
},
else => {
bad_index.* = 4 + i;
return error.InvalidCharacter;
},
}
if (value > 0x10ffff) {
bad_index.* = 4 + i;
return error.InvalidCharacter;
}
}
return value;
},
else => {
bad_index.* = 2;
return error.InvalidCharacter;
}
},
}
}
return std.unicode.utf8Decode(slice[1 .. slice.len - 1]) catch unreachable;
@ -172,13 +170,23 @@ test "parseCharLiteral" {
std.testing.expectEqual(try parseCharLiteral("'a'", &bad_index), 'a');
std.testing.expectEqual(try parseCharLiteral("'ä'", &bad_index), 'ä');
std.testing.expectEqual(try parseCharLiteral("'\\x00'", &bad_index), 0);
std.testing.expectEqual(try parseCharLiteral("'\\x4f'", &bad_index), 0x4f);
std.testing.expectEqual(try parseCharLiteral("'\\x4F'", &bad_index), 0x4f);
std.testing.expectEqual(try parseCharLiteral("'ぁ'", &bad_index), 0x3041);
std.testing.expectEqual(try parseCharLiteral("'\\u{0}'", &bad_index), 0);
std.testing.expectEqual(try parseCharLiteral("'\\u{3041}'", &bad_index), 0x3041);
std.testing.expectEqual(try parseCharLiteral("'\\u{7f}'", &bad_index), 0x7f);
std.testing.expectEqual(try parseCharLiteral("'\\u{7FFF}'", &bad_index), 0x7FFF);
std.testing.expectError(error.InvalidCharacter, parseCharLiteral("'\\x0'", &bad_index));
std.testing.expectError(error.InvalidCharacter, parseCharLiteral("'\\x000'", &bad_index));
std.testing.expectError(error.InvalidCharacter, parseCharLiteral("'\\y'", &bad_index));
std.testing.expectError(error.InvalidCharacter, parseCharLiteral("'\\u'", &bad_index));
std.testing.expectError(error.InvalidCharacter, parseCharLiteral("'\\uFFFF'", &bad_index));
std.testing.expectError(error.InvalidCharacter, parseCharLiteral("'\\u{}'", &bad_index));
std.testing.expectError(error.InvalidCharacter, parseCharLiteral("'\\u{FFFFFF}'", &bad_index));
std.testing.expectError(error.InvalidCharacter, parseCharLiteral("'\\u{FFFF'", &bad_index));
std.testing.expectError(error.InvalidCharacter, parseCharLiteral("'\\u{FFFF}x'", &bad_index));
}
test "" {