DoxigAlpha

writeSignedFixed

This is an "advanced" function. It allows one to use a fixed amount of memory to store an ILEB128. This defeats the entire purpose of using this data encoding; it will no longer use fewer bytes to store smaller numbers. The advantage of using a fixed width is that it makes fields have a predictable size and so depending on the use case this tradeoff can be worthwhile. An example use case of this is in emitting DWARF info where one wants to make a ILEB128 field "relocatable", meaning that it becomes possible to later go back and patch the number to be a different value without shifting all the following code.

Function parameters

Parameters

#
l:usize
ptr:*[l]u8
int:std.meta.Int(.signed, l * 7)

Read a single unsigned LEB128 value from the given reader as type T,

Functions

#
readUleb128
Read a single unsigned LEB128 value from the given reader as type T,
writeUleb128
Write a single unsigned integer as unsigned LEB128 to the given writer.
readIleb128
Read a single signed LEB128 value from the given reader as type T,
writeIleb128
Write a single signed integer as signed LEB128 to the given writer.
writeUnsignedFixed
This is an "advanced" function.
writeUnsignedExtended
Same as `writeUnsignedFixed` but with a runtime-known length.
writeSignedFixed
This is an "advanced" function.

Source

Implementation

#
pub fn writeSignedFixed(comptime l: usize, ptr: *[l]u8, int: std.meta.Int(.signed, l * 7)) void {
    const T = @TypeOf(int);
    const U = if (@typeInfo(T).int.bits < 8) u8 else T;
    var value: U = @intCast(int);

    comptime var i = 0;
    inline while (i < (l - 1)) : (i += 1) {
        const byte: u8 = @bitCast(@as(i8, @truncate(value)) | -0b1000_0000);
        value >>= 7;
        ptr[i] = byte;
    }
    ptr[i] = @as(u7, @bitCast(@as(i7, @truncate(value))));
}