Data Types

Basic Types

All data types are stored in memory with little-endian byte order.

s8

A signed 8-bit integer.

u8

An unsigned 8-bit integer.

s16

A signed 16-bit integer.

u16

An unsigned 16-bit integer.

s32

A signed 32-bit integer.

u32

An unsigned 32-bit integer.

f32

A 32-bit IEEE 754 floating point value.

Alias Types

bool

An alias of u8. A boolean representing true or false. Valid values are 0 and 1.

Structs and Arrays

Arrays and Structs are used to group data together and is common for data points, headers and plenty of other types. Arrays and Structs are byte aligned and tightly packed. This means there are no padding bytes between members.

A Struct lets us make composite types and group data together with names. This makes it easier for us to reason about data.

Unlike in C or C++ we can use variables in our array sizes. This help us define structures cleanly, however when it comes to implementing these structures, you will need to use containers of some kind.

In constract to the tiddlyhost wiki, our types have sizes implicitly known. This means instead of every Struct telling us the individual byte sizes and types, we'll just tell you the type name, and the size can be inferred based on the type. This makes arrays much more intuitive too.

Given this structure, lets assume a, b, and c are related. We can simply store them together in an array.

Placeholder Types

Placeholder types denote data of unknown size. Ideally we should never need these, however we don't fully understand the Thumper structures so this is needed to make notes to ourselves in a semantic way.

Unbounded

The unbound placeholder ... may have any value from 0 to +Inf. This is the most extreme placeholder type may select any amount of data.

Bounded

The bounded placeholder a..b may select any value in the range in the closed range of a to b inclusive.

For example u8[4..6] may select 4, 5, or 6 bytes.

Composite Types and Structures

Its common to see arrays with a dynamic size. This will be called a vector. Pull from C++ terms. Shortened to vec in structs.

vec<T>

Composite aliases

str

A string in Thumper is simply a size that refers to the following byte count, then the bytes themselves. Strings are an alias of vec<T> where T = u8. When expanded follows this structure:

f32vec2

An array of 2 f32. Typically used to represent texture coordinates in meshes.

• The first offset is commonly aliased as x, r or s.
• The second offset is commonly aliased as y, g or t.

f32vec3

An array of 3 f32. Typically used to represent positions and normal vectors.

• The first offset is commonly aliased as x, or r.
• The second offset is commonly aliased as y, g.
• The third offset is commonly aliased as z, b.

u8vec4

An array of 4 u8. Typically used to represent colors.

• The first offset is commonly aliased as x, or r.
• The second offset is commonly aliased as y, g.
• The third offset is commonly aliased as z, b.
• The fourth offset is commonly aliased as w, a.

Data Point Types

boolDataPoint

f32DataPoint