M2000 Interpreter

See Introduction to M2000 for installation and information about M2000

M2000 start as an experimental interpreted language, using a Module in Module idea (like a Procedure in Procedure) where each inner Module is closed for modification, but open for expansion, by replacing code at run time. Code executed in one pass. There is a pre-pass to determine the major type of an expression, a number or a string. Look this paradigm: We call inner Beta in two stages. Ins second stage we change inner Theta with Theta2. This is the decoration of Beta with Theta as Theta2. This is a temporary decoration because Beta after execution erase any new identifier including Theta. So each time we call Beta, statement Module Theta make this module unless a decoration stop it.

Module Beta {       Module Theta (x){             Print "This is Theta, we get x=";x       }       For i=1 to 3 : Theta i : Next i } Beta Module Theta2 (x) {             Print "This is Theta2, we get x=";x } Beta ; Theta as Theta2

As we can see from code, some statements are like BASIC, except the use of curly brackets {}. Check the code below. We have a Module (as a Procedure), where we define some functions and some variables, as entities. These entities defined in every call to CheckIt, and erased when execution return from CheckIt. By default every parameter in M2000 pass by value. We define two functions, Alfa() where we set type for parameter x and ExpType$() where we not set type for x, and we wish to return the name of type when we make a call. Using DEF we can define variables and functions (in one line).

Module CheckIt {       \\ We can't call something out of this module       Function Alfa(X as double) {             =X**2       }       Print Alfa(5) ' we pass a double type number       Print Alfa(12121221121.1212@) ' we pass a Decimal type       \\ M get type from first assingment       M=Alfa(3)       Print Type$(M) ' Double       \\ Def can create once Z (second time raise error)       Def Z as Double       \\ If a variable get a type then any number convert to that type before assign to it       Z=121212112.11212@       Print Type$(Z)       Def Currency Z1=212.12, Z2=323233.12, Z3=223212323323.12       Print Z1, Z2, Z3       Def ExpType$(x)=Type$(x)       Print ExpType$(Z1+Z2) ' Currency       Print ExpType$(Z1*Z2) ' Currency       Print ExpType$(Z2*Z3) ' Double (Currency can't hold this number)       \\ Integer, Long, Single, Currency, Decimal       Print ExpType$(10%), ExpType$(10&), ExpType$(10~), ExpType$(10#), ExpType$(10@)       \\ Without symbol is a Double       Print ExpType$(10) } \\ We call it CheckIt \\ Now Funtion Alfa and all variables erased. \\ We can call it again CheckIt

For small programs, for learning algorithms there is no need to use types, except the major types, Numeric and String, plus one more: Integer. Integers can be number types like Long and Integer types, or can be any numeric type using without decimal part. So Integer variables are variables with % at the end of their name. String Variable need a $ at the end of name.

Module CheckInt {       A%=1212212.12@       Print A% ' 1212212       A%=1212212.52@       Print A% ' 1212213       Print Type$(A%) ' Decimal       A%++       Print A% ' 1212214       Print Type$(A%) ' Decimal       B%=122121212.12 ' make Integer from Double       Print B% ' 122121212       Print Type$(B%) ' Double       B%++       Print Type$(B%) ' Double       Print B% ' 122121213 } CheckInt

So we say about Integer Variables, and no Integer Numeric Type. Like in Basic, M2000 is not case sensitive (except for labels), so A% and a% is the same. We may have A, A$ and A% as three variables, or A(), A%(), A$() as arrays or and functions. We can use name(* ) to call function and not array if we have each with same name.

Dim A(10)=1 Def A(x)=x**2 Print A(3), A(* 4)

By default all variables are local. M2000 uses heap to store local variables, not stack. We can use global variables, but a local definition hide a global one. We can make global variables for temporary use (at return from module which we define a global, this global erased). A global definition in an already defined global variable with same name, hide for temporary use the old global variable.

It is easy to make modules, with modules, functions, variables, arrays. We can use subroutines when we want code in a module or a function to used more than once. We can use local variables inside subroutines. In subroutines we can use module's variables, functions, subroutines. Until here we see something like BASIC, except that variables are local to modules and functions, but can be used by subroutines. Also we have see that we can replace a module in a module when we call it. With this replacement we can use a predefined logic with some "terminals", modules that get parameters and do final things. We can make global modules and functions in local modules and functions, but these exist until creator module or function exit.

M2000 can be used for more advanced programming. Modules and functions can be members of Groups. Groups are values. So a function can return a Group. Modules and functions get parameters in a stack. Modules get the parent stack with parameters, and functions get a new stack with parameters. Modules return from call and are responsible to clear as needed the stack. Functions drop the stack after the call from an expression. A module can return entities in stack, if we want to do that.

Groups are objects for M2000. We can use them as values, and if we want we can use pointers for groups (but we can make a lot of programs without using pointers to groups). Without using pointer, a Group may be local with a name in a Module or can be in a container like array, inventory (a type of map) and stack objects, without name. Groups may have private and public members, may have properties as read only, may get value and return value, and may have operators. Members of Groups can be anything, including other groups. Groups may have events too. We can make groups by using CLASS statement or by using own function which return a Group or a pointer to Group.

A special object is the lambda object, which may have a number of closures, and have two faces, a variable and a function:

Module  Delta {       \\ Make a group like a lambda       \\ A Class make a Global Function       \\ (in a group a Class definition make a group member, not a global function)       Class Alfa {       Private:             Z=0       Public:             Value (x) {             \\ see dot before Z             \\ this is the same as This.Z                   =X**2+.Z             }       Class:             \\ After Class: definitions are lost when we copy group             \\ this is the constructor             \\ here only a Read .Z used             Module Alfa(.Z) {}       }       \\ M, N, K and P get type by first value on assignment       M=Alfa(5)       \\ M has no Alfa module, because a copy perform in Alfa()       Print M(3)=14       \\ Make a lambda       N=Lambda Z=5 (x)->X**2+Z       Print N(3)=14       K=N ' copy lambda N to K       Print N(3)=K(3)       Print Type$(N)="lambda"       \\ we want a copy of M not value of M       P=Group(M )       Print M(3)=P(3)       Print Type$(P) ="Group" } Delta

We can place a lambda function as closure in a lambda function, and we can build a program as one "top" function, and many lambda as closures, which may have other lambda functions. Closures can be change values from lambda but not from outside unless they are pointers (so we may have same pointers outside of lambda, or in another lambda. A lambda variable can change function and closures in a new assignment. We can store lambda state (if we don't use pointers in closures) in a new variable which take as first value lambda (and then can take only lambda functions)

Until now we see modules/functions/subs for procedural programming, Groups for OOP, Groups as lambda functions and lambda functions for functional programming. We can use events for groups and for COM objects, including GUI objects.

Module Beta {       Group WithEvents Alpha {             Event "One"             Module DoIt (x) {                   \\ if no service function founded, no error happen                   Call Event "One", x             }       }       Alpha.Doit 50       \\  nothing because we didn't define yet the service function       M=10       X=5       Function Alpha_One (New X){             \\ Event functions called "local", with scope on module             Print M+X             M++       }       \\ Now function Alpha_One() exist       Alpha.Doit 20 ' 30       Print X=5 ' delete New, in event service function to get 20       Print M=11       \\In M2000 we can change definitions       Function Alpha_One (New X){             \\ Event functions called "local", with scope on module             Print M+X             M+=100       }       Alpha.Doit 20 ' 31       Print M=111 } Beta

Until now we see that programming with events can be used in M2000. There is an Event object, which we see later, among the light events in Groups using WithEvents, and we can use COM Events to handle objects like MS Word.

Modules may have Threads, part of modules that can be executed in intervals, can be halted, or can be released, and can be erased. Each thread has own stack and may have own static variables (Modules and Functions also may have static variables), but can use modules variables and functions/modules/subs. Threads can run concurrent (thread return to task manager after execution of a statement or a block of statements) or sequential (a thread has to exit from interval to start other thread)

Module Zeta {       k=10       Thread {            k++       } As M interval 50       \\ This is the main task (a thread)       Main.Task 100 {             Print k             If keypress(32) then Exit       } } Zeta

M2000 Interpreter can work with Structures and Buffers. Buffer is a memory block and Structure can define types of memory, and a buffer can hold arrays of as structure. Each structure can be made from other structures or arrays of structures, and may have unions. We can use buffers for code execution (writing machine code), using standard buffers (not executable), for data holder. Also we can use Buffers calling external dll and passing by address.

Module Kappa {       Function Theta(x) {              Structure Points_single {                   x as single                   y as single             }             Buffer Clear Points as Points_single*x             =Points       }       P=Theta(100)       Print Len(P)=800 ' 2x4x100 bytes       \\ return to buffer multiple values using index!offset       Return P, 0!x:=1212.12, 0!y:=21.1212       \\       Print Eval(P, 0!x)=1212.12~, Eval(P, 0!y)=21.1212~ } Kappa