NEO VM does not distinguish between empty and nil slices. Supporting
this is not easy and requires changing lots of other opcodes.
Pointers are not supported anyway and slices can be checked for
emptiness by inspecting their `len`.
Set default value also for complex (struct, map) types.
Note: because VM does not distinguish empty array and nil array,
comparison 'a == nil' can't be handled properly without substantial
effort. This will be fixed in NEO3.
Previously we could generate dynamic appcall with a kludge of
AppCall([]byte{/* 20 zeroes */, realScriptHash, args...)
Now there is a separate function for this.
Sequence points is a way to map a specific instruction offset
from a compiled contract to a text span in a source file.
This commit implements mapping only for `return` statements.
Further improvements are straight-forward.
Previously, struct variables were initialize with VM's nil value
which is of primitive type. Thus SETITEM used for struct's field
updating wasn't working.
Previously this declarations were ignored which resulted
in runtime errors, because VM's nil is an element of primitive type
and can't be converted to an array.
When `return` or `break` statement is encountered inside
a for/range/switch statement, top stack items can be auxilliary.
They need to be cleaned up before returning from the function.
Current implementation of short-circuting is just plain wrong
as it uses `last` or `before-last` labels which meaning depend
on context. It doesn't even handle simple assignements like
`a := x == 1 && y == 2`.
This commit makes all jumps in such conditions local
and adds tests.
Closes#699, #700.
While initializing a struct, it is a top item on ALTSTACK.
This means that if we need to load a local variable,
DUPFROMALTSTACK won't longer push an array of locals on stack
but rather a currently initializing struct.
Closes#656.
Old implementation could view 0x62 byte in
a script as a JMP instruction irregardless of whether it is
a real opcode or a part of a parameter of another instruction.
In this commit instructions are decoded together with parameters
during jump label rewriting.
Because the constants are loaded directly via `emitLoadConst`, there is no need to store
them in an array of locals. It can have a big overhead, because it
is done at the beginning of every function.
Append should leave it's result on top of the stack.
Thus we need to transform top of the stack:
(top) a . b --> (top) a . b . b
It can be done with just OVER + SWAP.
