1
0
mirror of https://github.com/KarolS/millfork.git synced 2024-07-06 00:28:55 +00:00
millfork/docs/lang/types.md
2019-07-15 13:52:05 +02:00

174 lines
5.7 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

[< back to index](../README.md)
# Types
Millfork puts extra limitations on which types can be used in which contexts.
## Numeric types
* `byte` 1-byte value of undefined signedness, defaulting to unsigned
* `word` 2-byte value of undefined signedness, defaulting to unsigned
(alias: `int16`)
* `int24` 3-byte value of undefined signedness, defaulting to unsigned
(alias: `farword`; this alias is deprecated and will be removed in the future)
* `long` 4-byte value of undefined signedness, defaulting to unsigned
(alias: `int32`)
* `int40`, `int48`,... `int128` even larger types
* `sbyte` signed 1-byte value
* `ubyte` unsigned 1-byte value
* `pointer` raw pointers; the same as `word`, but variables of this type default to be zero-page-allocated
and you can index `pointer`-typed expressions.
You can create pointer values by suffixing `.addr` to the name of a variable, function or array.
You can access single bytes of variables by using the following notations:
* for 2-byte-sized variables: `.lo` for the least significant byte and `.hi` for the most significant byte
* for larger variables: `.b0` for the least significant byte and then `.b1`, `.b2` and so on
You can also access words that are parts of variables:
* for 3-byte-sized variables: `.loword` is the word formed from `.b1` and `.b0` and `.hiword` is the word formed from `.b2` and `.b1`
* for 4-byte-sized variables: `.loword` is the word formed from `.b1` and `.b0` and `.hiword` is the word formed from `.b3` and `.b2`
Numeric types can be converted automatically:
* from a smaller type to a bigger type (`byte`→`word`)
* from a type of undefined signedness to a type of defined signedness (`byte`→`sbyte`)
* from a type of defined signedness to a type of undefined signedness (`sbyte`→`byte`)
## Typed pointers
For every type `T`, there is a pointer type defined called `pointer.T`.
Unlike raw pointers, they are not subject to arithmetic.
If the type `T` is of size 1, you can index the pointer like a raw pointer.
If the type `T` is of size 2, you can index the pointer only with the constant 0.
Examples:
pointer.t p
p.raw // expression of type pointer, pointing to the same location in memory as 'p'
p.lo // equivalent to 'p.raw.lo'
p.hi // equivalent to 'p.raw.lo'
p[0] // valid only if the type 't' is of size 1 or 2, accesses the pointed element
p[i] // valid only if the type 't' is of size 1, equivalent to 't(p.raw[i])'
p->x // valid only if the type 't' has a field called 'x', accesses the field 'x' of the pointed element
p->x.y[0]->z[0][6] // you can stack it
## `nullptr`
There is a 2-byte constant `nullptr` that can be assigned to any 2-byte pointer type.
Its actual value is defined using the feature `NULLPTR`, by default it's 0.
`nullptr` isn't directly assignable to non-pointer types.
## Boolean types
TODO
## Special types
* `void` a unit type containing no information, can be only used as a return type for a function.
## Enumerations
Enumeration is a 1-byte type that represents a set of values:
enum <name> { <variants, separated by commas or newlines> }
The first variant has value 0. Every next variant has a value increased by 1 compared to a previous one.
Alternatively, a variant can be given a custom constant value, which will change the sequence.
If there is at least one variant and no variant is given a custom constant value,
then the enumeration is considered _plain_. Plain enumeration types can be used as array keys.
For plain enumerations, a constant `<name>.count` is defined,
equal to the number of variants in the enumeration.
Assigment between numeric types and enumerations is not possible without an explicit type cast:
enum E { EA, EB }
byte b
E e
e = EA // ok
e = b // won't compile
b = e // won't compile
b = byte(e) // ok
e = E(b) // ok
array a[E] // E is plain, array has size 2
a[0] // won't compile
a[EB] // ok
Plain enumerations have their variants equal to `byte(0)` to `byte(<name>.count - 1)`.
Tip: You can use an enumeration with no variants as a strongly checked alternative byte type,
as there are no checks on values when converting bytes to enumeration values and vice versa.
## Structs
Struct is a compound type containing multiple fields of various types:
struct <name> { <field definitions (type and name), separated by commas or newlines>}
A struct is represented in memory as a contiguous area of variables laid out one after another.
Struct can have a maximum size of 255 bytes. Larger structs are not supported.
You can access a field of a struct with the dot:
struct point { word x, word y }
point p
p.x = 3
p.y.lo = 4
Offsets are available as `structname.fieldname.offset`:
pointer ptr
ptr = p.addr
ptr += point.y.offset
// ptr points now at p.y
// alternatively:
ptr = p.y.addr
You can create constant expressions of struct types using so-called struct constructors, e.g.:
point(5,6)
All arguments to the constructor must be constant.
## Unions
union <name> { <field definitions (type and name), separated by commas or newlines>}
Unions are pretty similar to structs, with the difference that all fields of the union
start at the same point in memory and therefore overlap each other.
struct point { byte x, byte y }
union point_or_word { point p, word w }
point_or_word u
u.p.x = 0
u.p.y = 0
if u.w == 0 { ok() }
Offset constants are also available, but they're obviously all zero.
Unions currently do not have an equivalent of struct constructors. This may be improved on in the future.