Concepts of Programming Language Design — Assignment An abstract machine for MinHs Version 1.0 Due date: Friday, 19/01/2024 Initial spec released In this assignment you will implement an evaluator for MinHS programs in form of an abstract machine, based on the E-machine. This abstract machine is part of an evaluator of MinHs, a small functional language similar to ML and Haskell. MinHs is fully typed, with types specified by the programmer. The assignment consists of a base component, worth 60% of the overall marks for the assignment, and five additional components, each worth another 8%. • Task 1 (60%) Implement an interpreter for the MinHs language presented in the lectures, using an envi- ronment semantics, including support for recursion and closures. • Task 2 (8%) Extend the abstract machine to support partial application of prim-ops and the list con- structor Cons. • Task 3 (8%) Extend the abstract machine to support n-ary functions • Task 4 (8%) Extend the abstract machine to support multiple bindings in the one let form. • Task 5 (8%) Extend the abstract machine to support let bindings that take parameters, definining non- recursive functions. • Task 6 (8%) Extend the abstract machine to support infinite lists. Each of these parts is explained in detail below. The front end of the interpreter (lexer, parser, type checker) is provided for you, along with the implementation of the evaluate and evalE functions (found in the file Evaluator.hs). The function evaluate requires a program as argument, and returns an object of type Value, and evalE takes an expression as argument. You will need to extend the set of constructors for Value, but not the implementation for evaluate and evalE. The return value of evaluate is used to check the correctness of your assignment. History 12/12/2023 Overview 1 The types of the functions msInFinalState, msGetValue and msStep are provided and may not be changed. The definition of the type MachineState is currently empty, and you have to find a suitable definition modelling the different states of the machine. The control stack and the environment are part of the machine state. For the environment, you can use the predefined type VEnv, but you have to define your own data type to model the control stack. Apart from Evaluator.hs, no other files can be modified! The function msStep performs one single evaluation step in the E-machine, with an explicit control stack and environment, and may not be recursive. The function evalE acts as driver and calls msStep on the machine state until it ends up in a final state. Given a closed expression e, then passing the expression to evalE should result in the value as specified by the big step semantics given in Section 2. You can assume the typechecker has done its job and will only give you correct programs to evaluate. The type checker will, in general, rule out invalid programs, so the abstract machine does not have to consider them. The dynamic (big step) semantics of MinHs is given in Section 2, and you should use this as a specification for the behaviour of your abstract machine for the core tasks. That is, iff an expression e under the environment Γ evaluates to a value v according to the big step semantics, then it should evaluate, under the same environment and an intially empty control stack, to the same value v and an empty control stack in a finite number of steps in your E-machine implementation. For the subset of the language we discussed in the lecture, you can directly use the E-machine rules. For the remaining language constructs, you still need to find the corresponding rules. For the advanced tasks, you have to come up with the formal sematics from the informal description yourself. Ask in the MS Teams assignment chat if you have any questions about the assignment. 1 Task1 This is the core part of the assignment. You are to implement an abstract machine for MinHs. The following expressions must be handled: • variables. v,u • integer constants. 1,2,.. • boolean constants. True,False • some primitive arithmetic and boolean operations. +,*,<,<=,.. • constructors for lists. Nil, Cons • destructors for lists. head, tail • inspectors for lists. null • function application. f x • ifethene1 elsee2 • let x :: τx = e1; y :: τy = e2; ... in e2 • recfun f :: (τ1 -> τ2) x = e expressions These cases are explained in detail below. The abstract syntax defining these syntactic entities is in Syntax.hs. You should understand the data type Exp and Bind well. The big step semantics of a higher-order abstract syntax representation of MinHs is given is this document. Your task is to design a corresponding E-machine semantics (that is, with explicit control stack and environment, not substitution!) and implement it in Haskell. Page 2 If a runtime error occurs, which is possible, abort the execution and use Haskell’s error :: String -> a function to emit a suitable error message (the error code returned by error is non- zero, which is what will be checked for – the actual error message is not important). You do not need to model an error state in the machine. 1.1 Program structure A program in MinHs may evaluate to either an integer, a list of integers, or a boolean, depending on the type assigned to the main function. The main function must always be defined; furthermore, for this task, main is always the only top-level binding. For example: main :: Int = 1 + 2; or main :: Bool = let x :: Int = 1; in if x + ((recfun f :: (Int -> Int) y = y * y) 2) == 0 then True else False; 1.2 Variables, Literals and Constants MinHs is a spartan language. We only have to consider 4 types: Int Bool t1 -> t2 [Int] The only literals you will encounter are integers. The only non-literal constructors are True and False for the Bool type, and Nil and Cons for the [Int] type. 1.3 Function application MinHs is, by virtue of its Haskell implementation, a non-strict language. An argument to a function is only evaluated when needed — when the function tries to inspect its value. This does not add a great deal of complexity to your implementation — it will occur naturally as you will be writing the abstract machine in Haskell, which is also a non-strict language. The result of a function application may in turn be a function. 1.4 Primitive operations You need to implement the following primitive operations: + :: Int -> Int -> Int - :: Int -> Int -> Int * :: Int -> Int -> Int / :: Int -> Int -> Int negate :: Int -> Int > :: Int -> Int -> Bool >= :: Int -> Int -> Bool < :: Int -> Int -> Bool <= :: Int -> Int -> Bool Page 3 == :: Int -> Int -> Bool /= :: Int -> Int -> Bool head :: [Int] -> Int tail :: [Int] -> [Int] null :: [Int] -> Bool These operations are defined over Ints, [Int]s, and Bools, as usual. negate is the primop representation of the unary negation function, i.e. -1. The abstract syntax for primops is defined in Syntax.hs. 1.5 if - then - else MinHs has an if e then e1 else e2 construct. The types of e1 and e2 are the same. The type of e is Bool. 1.6 let For the first task you only need to handle simple let expressions of the kind we have discussed in the lectures. Like these: main :: Int = let x :: Int = 1 + 2; in x; or main :: Int = let f :: (Int -> Int) = recfun f :: (Int -> Int) x = x + x; in f 3; For the base component of the assignment, you do not need to handle let bindings of more than one variable at a time (as is possible in Haskell). Remember, a let may bind a recursive function defined with recfun . 1.7 recfun The recfun expression introduces a new, named function value. It has the form: (recfun f :: (Int -> Bool) x = x + x) A recfun value is a first-class value, and may be bound to a variable with let. The value ‘f’ is bound in the body of the function, so it is possible to write recursive functions: recfun f :: (Int -> Int) x = if x < 10 then f (x+1) else x Be very careful when implementing this construct, as there can be problems when using envi- ronments in a language allowing functions to be returned by functions. Page 4 1.8 Evaluation strategy We have seen in the lecture how it is possible to evaluate expressions via substitution. This is an extremely inefficient way to run a program. In this assignment you are to use an environment instead. You will be penalised for an abstract machine that operates via substitution. The module Env.hs provides a data type suitable for most uses. Consult the lecture notes on how environments are to be used in the E-machine. The strategy is to bind variables to values in the environment, and look them up when requried. In general, you will need to use: empty, lookup and addAll to begin with an empty environ- ment, lookup the environment, or to add a binding to the environment, respectively. As these functions clash with functions in the Prelude, a good idea is to import the module Env qualified: import qualified MinHS.Env as E This makes the functions accessible as E.empty and E.lookup, to disambiguate from the Pre- lude versions. 2 Dynamic Semantics of MinHs Big-step semantics We define a relation ⇓ which relates an environment Γ, which maps variables to values1 and an expression in higher-order abstract syntax e ∈ E to the resultant value v of that expression V . Note that the Haskell representation you will be working on for the assignment is a first-order representation. Our value set for V , initially, consists of: • Machine integers • Boolean values • Lists of integers We will also need to add closures, or function values to our value set to deal with the recfun construct in a sound way. See the section on Function Values for details. Environment The environment Γ maps variables to values, and is used in place of substitution. It is specified as follows: Γ::= ·|Γ,x=v Values bound in the environment are closed – they contain no free variables. This requirement creates a problem with function values created with recfun whose bodies contain variables bound in an outer scope. We must bundle them with their associated environment as a closure. Constants and Boolean Constructors Γ⊢(Numn)⇓n Γ⊢(ConTrue)⇓True Γ⊢(ConFalse)⇓False 1Or, possibly, to an unevaluated computation of the value, but in Haskell these two things are indistinguishable. Page 5 Primitive operations Evaluation of if -expression Variables List constructors and primops Γ⊢e1 ⇓True Γ⊢e2 ⇓v Γ⊢(Ife1 e2 e3)⇓v Γ⊢e1 ⇓False Γ⊢e3 ⇓v Γ⊢(Ife1 e2 e3)⇓v Γ(x) = v Γ⊢x⇓v Γ⊢e1 ⇓v1 Γ⊢e2 ⇓v2 Γ⊢(Pluse1 e2)⇓v1 +v2 Similarly for the other arithmetic and comparison operations (as for the language of arithmetic expressions, and in chapters 7,8 of Harper [2].). Note that division by zero should cause your abstract machine to throw an error using Haskell’s error function. The abstract syntax of the abstract machine re-uses function application to represent applica- tion of primitive operations, so (Plus e1 e2) is actually represented as: (App ((App ((Prim Plus) e1)) e2)) For this first part of the assignment, you may assume that prim-ops are never partially applied — that is, they are fully saturated with arguments, so the term (App ((Prim Plus) e1)) will never occur in isolation. Γ ⊢ e ⇓ ve Γ ⊢ es ⇓ ves Γ⊢Nil⇓[] Γ⊢(Consees)⇓(ve:ves) Γ ⊢ e ⇓ (v:vs) Γ ⊢ e ⇓ (v:vs) Γ ⊢ e ⇓ (v:vs) Γ⊢(Heade)⇓v Γ⊢(Taile)⇓vs Γ⊢(Nulle)⇓False For the first part of the assignment, you may assume that Cons is also never partially applied, as with prim-ops. Variable Bindings with Let Γ ⊢ e1 ⇓ v1 Γ,x=v1⊢e2⇓v2 Γ ⊢ (Let e1 (x.e2)) ⇓ v2 Page 6 Function values To maintain soundness with function values, we need to pair a function with its environment, forming a closure. We introduce the following syntax for function values: ⟨⟨Γ, f.x.e⟩⟩ You will need to decide on a suitable representation of closures as a Haskell data type. Now we can specify how to introduce closed function values: Γ ⊢ (Recfun τ1 τ2 (f.x.e1 )) ⇓ ⟨⟨Γ, f.x.e⟩⟩ Function Application Γ⊢e1 ⇓⟨⟨Γ′,f.x.ef⟩⟩ Γ ⊢ e2 ⇓ v2 Γ′,f=v1,x=v2⊢ef⇓r Γ⊢(Appe1 e2)⇓r 3 Advanced Tasks 3.1 Task 2: Partial Primops In the base part of the assignment, you are allowed to assume that all primitive operations (and the constructor Cons) are fully saturated with arguments. In this task you are to implement partial application of primitive operations (and Cons), which removes this assumption. For example: main :: Int = let inc :: (Int -> Int) = (+) 1; in inc 2; -- returns 3 Note that the expression (+) 1 partially applies the primop Plus to 1, returning a function from Int to Int. You will need to develop a suitable dynamic semantics for such expressions and implement it in your evaluator. The parser and type checker are already capable of dealing with expressions of this form. 3.2 Task 3: n-ary functions In this task you are to implement n-ary functions. In other words, you should modify the abstract machine to handle bindings of functions of more than 1 argument. main :: Bool = let eq :: (Int -> Int -> Bool) = recfun eq :: (Int -> Int -> Bool) x y = x == y; in eq 3 4; We haven’t discussed the semantics of such functions in the lectures so you will need to work out a reasonable dynamic semantics for n-ary functions on your own, based on the semantics for unary functions. The parser and type checker are once again already capable of handling expressions of this form, so the only extension necessary is in the evaluator component. Hint: Is the following example semantically different to the previous? Page 7 main :: Bool = let eq :: (Int -> Int -> Bool) = recfun eq :: (Int -> Int -> Bool) x = recfun eq2 :: (Int -> Bool) y = x == y; in eq 3 4; 3.3 Task 4: Multiple bindings in let In the base part of the assignment, we specify that let expressions contain only one binding. In this task, you are to extend the abstract machine so that let expressions with multiple bindings, like: main :: Int = let a :: Int = 3; b :: Int = 2; in a + b; are evaluated the same way as multiple nested let expressions: main :: Int = let a :: Int = 3; in let b :: Int = 2; in a + b; Once again the only place where extensions need to be made are in the evaluator, as the type checker and parser are already capable of handling multiple let bindings. 3.4 Task 5: let bindings declare functions In this task you are to extend the let construct further so that let bindings can take parameters — defining non-recursive functions. This makes programming in MinHs much less verbose, as recfun is only necessary for defining recursive functions. For example: main :: Int = let y :: Int = 3; in let f :: (Int -> Int) x = x + 1; in f y; -- returns 4 Note that these bindings are non-recursive, so let x = x is a scope error, as the x in the binding is not in scope. 3.5 Task 6: Lazy Lists We re-use the recfun syntax, only without arguments, to construct infinite lists. The recursive reference is provided but no function argument is necessary. When the infitine liste one is passed as argument to a list operation, it should be evaluated to so-called weak head normal form, that is, just far enough to expose its topmost constructor (Cons or Nil). Therefore, the following program should terminate main :: [Int] = let ones :: [Int] = recfun ones :: [Int] = Cons 1 ones; in Cons (head ones) (Cons (head (tail ones)) Nil); and evaluate to (Cons 1 (Cons 1 Nil)) Page 8 4 Testing Your assignments will be tested very rigorously. You are encouraged to test yourself. minhs comes with a regress tester script, and you should add your own tests to this. Simply add test programs names prgname.mhs and a corresponding file with the expected result prgname.out anywhere in the tests subdirectory, and the testscript will find it and run it. The tests that come with this assignment tarball cover the base part (the first 60%) of the assignment only. You are responsible for testing the other parts adequately. 5 Building minhs minhs (the compiler/abstract machine) is written in Haskell, and can be used with either the cabal build tool or the stack build tool. You can use whichever works best on your machine. 5.1 Building with cabal You should be able to build the project by simply invoking: $ cabal build To see the debugging options, run (after building): $ cabal exec minhs-1 To get, for example, the parser output for a file.mhs, type $ cabal exec minhs-1 -- --dump parser-raw file.mhs To run the compiler with a particular file, run: $ cabal exec minhs-1 -- foo.mhs And to run all of our tests, if you are on Mac or Linux, type: (remember to run cabal build first!) $ ./run_tests_cabal.sh If you are on Windows, type: $ .\run_tests_cabal.cmd 5.2 Building with stack Incidentally, the same as in Section 5.1, except replace all occurrences of “cabal” with “stack” (and in particular, use the test script ./run_tests_stack.sh, or the Windows variant, which can be run using .\run_tests_stack.cmd). 6 Submitting Submission will be via Blackboard. Remember that you must submit only the file Evaluator.hs. This means that you must not modify any of the other files. Page 9 7 Late Penalty Unless otherwise stated if you wish to submit an assignment late, you may do so, but a late penalty reducing the maximum available mark applies to every late assignment. The maximum available mark is reduced by 10% if the assignment is one day late, by 25% if it is 2 days late and by 50% if it is 3 days late. Assignments that are late 4 days or more will be awarded zero marks. So if your assignment is worth 88% and you submit it one day late you still get 88%, but if you submit it two days late you get 75%, three days late 50%, and four days late zero. Assignment extensions are only awarded for serious and unforeseeable events. Therefore aim to complete your assignments well before the due date in case of last minute illness, and make regular backups of your work. 8 Plagiarism This assignment is an individual assignment. All work submitted for assessment must be entirely your own work. We regard unacknowledged copying of material, in whole or part, as an extremely serious offence. If you haven’t done so yet, please read the plagiarism policy of UU: https://students.uu.nl/en/practical-information/policies-and-procedures/fraud-and-plagiarism In this course submission plagiarism includes any work derived from another person, or solely or jointly written by and or with someone else, without clear and explicit acknowledgement. Note this includes including unreferenced work from books, the internet, etc. Do not provide or show your assessable work to any other person. If you knowingly provide or show your assessment work to another person for any reason, and work derived from it is subse- quently submitted you will be penalized, even if the work was submitted without your knowledge or consent. This will apply even if your work is submitted by a third party unknown to you. You should keep your work private until submissions have closed. If you are unsure about whether certain activities would constitute plagiarism ask us before engaging in them! References [1] Report on the Programming Language Haskell 98, eds. Simon Peyton Jones, John Hughes, (1999) http://www.haskell.org/onlinereport/ [2] Robert Harper, Programming Languages: Theory and Practice, (Draft of Jan 2003), http: //www-2.cs.cmu.edu/~rwh/plbook/. [3] The Implementation of Functional Programming Languages, Simon Peyton Jones, published by Prentice Hall, 1987. Full text online (as jpg page images). [4] Simon Peyton-Jones, Implementing Functional Languages : a tutorial, 2000. Page 10 A Lexical Structure The lexical structure of MinHS is an small subset of Haskell98. See section 2.2 of the Haskell98 report[1]. The lexical conventions are implemented by the Parsec parser library, which we use for our Parser implementation. B Concrete syntax The concrete syntax is based firstly on Haskell. It provides the usual arithmetic and boolean primitive operations (most of the Int-type primitive operations of GHC). It has conventional let bindings. At the outermost scope, the let is optional. As a result, multiple outer-level bindings are treated as nested let bindings down the page. It is required that a distinguished main function, of atomic type, exist. There is an if-then-else conditional expression. The primitive types of MinHS are Int,Bool and [Int]. MinHS also implements, at least partially, a number of extensions to MinML: inline comments, n-ary functions, infix notation, more primitive numerical operations and a non-mutually recursive, simultaneous let declaration (treated as a nested-let). Function values may be specified with recfun . The concrete syntax is described and implemented in the Parser.hs module, a grammar specified using the Parser combinator library Parsec. Features of Haskell we do not provide: • No nested comments • No layout rule. Thus, semi-colons are required to terminate certain expressions. Consult the grammar. C Abstract syntax The (first-order) abstract syntax is based closely on the MinHs syntax introduced in the lectures. It is implemented in the file Syntax.hs. Extensions to the MinHs abstract syntax take their cue from the Haskell kernel language. Presented below is the abstract syntax, with smatterings of concrete syntax for clarity. D Static semantics The static semantics are based on those of the lecture, and of and MinML, from Bob Harper’s book. They are implemented by the module TypeChecker.hs. D.1 n-ary functions Functions may be declared to take more than 1 argument at a time. E Environments Environments are required by typechecker and possibly by the abstract machine. The typechecker needs to map variables to types, and the abstract machine might need to map variables to functions or values (like a heap). This latter structure is used to provide a fast alternative to substitution. We provide a general environment module, keyed by identifiers, in Env.hs. Environments are generally simpler in MinHs than in real Haskell. We still need to bind variables to partially evaluated functions, however. Page 11 Types τ Literals n ::= Int|Bool|τ ->τ ::= ... |0|1|2|... b ::= True | False Primops o ::= +|-|*|/| Expressions exp Bind bind | >|>=|==|/=|<|<= → (Var x) | (Lit n) | (Con b) | (Apply exp exp) | (Let [bind] exp) | (Letrec [bind] exp) | (Recfun bind) | (If exp exp exp) → (Bind f τ [arg] e) Figure 1: The expression abstract syntax of MinHS F Dynamic semantics The dynamic semantics are described in this document, the lectures, and resemble that of Harper [2]. Implemented in the module Evaluator.hs. F.1 Abstract Machine The abstract machine is the backend that runs by default. It should implement the dynamic semantics of MinHs. G Interfaces The basic types are found in Syntax.hs, which contains definitions for the structure of terms, types, primOps, and others. Printing Most structures in MinHS need to be printed at some point. We have included the prettyprinter library for you to use; follow the example of the printValue function in Evaluate.hs. A number of MinHS types already have a printer function defined in the MinHS.Pretty module. Note that you will not be able to define instances of the Pretty class of the prettyprinter library if you want to use any of the functions defined in MinHS.Pretty, because we are using ANSI terminal colours which require use of the ANSI terminal backend for the library. This forces the ann type parameter of Doc to be AnsiStyle, but the Pretty class requires its methods to be polymorphic in the annotation type. Defining manual top-level functions does work; we have provided prettyList’ as a replacement for prettyList from the Pretty class if you need it. Page 12 Testing ./run_tests_cabal.sh (or if you are using stack: ./run_tests_stack.sh). Remember to cabal build (or stack build) first. Test directories may have an optional ‘Flag’ file, that contains flags you wish to pass to minhs in that directory, or the magic flag, ‘expect-fail’, which inverts the sense in which success is defined by the driver. Page 13