We describe a new compiler for Standard ML called TIL, that is based on four technologies: intensional polymorphism, tag-free garbage collection, conventional functional language optimization, and loop optimization. We use intensional polymorphism and tag-free garbage collection to provide specialized representations, even though SML is a polymorphic language. We use conventional functional language optimization to reduce the cost of intensional polymorphism, and loop optimization to generate good code for recursive functions. We present an example of TIL compiling an SML function to machine code, and compare the performance of TIL code against that of a widely used compiler, Standard ML of New Jersey.

Cayenne is a Haskell-like language. The main difference between Haskell and Cayenne is that Cayenne has dependent types, i.e., the result type of a function may depend on the argument value, and types of record components (which can be types or values) may depend on other components. Cayenne also combines the syntactic categories for value expressions and type expressions; thus reducing the number of language concepts. Having dependent types and combined type and value expressions makes the language very powerful. It is powerful enough that a special module concept is unnecessary; ordinary records suffice. It is also powerful enough to encode predicate logic at the type level, allowing types to be used as specifications of programs. However, this power comes at a cost: type checking of Cayenne is undecidable. While this may appear to be a steep price to pay, it seems to work well in practice.

Java is typically compiled into an intermediate language, JVML, that is interpreted by the Java Virtual Machine. Because mobile JVML code is not always trusted, a bytecode verifier enforces static constraints that prevent various dynamic errors. Given the importance of the bytecode verifier for security, its current descriptions are inadequate. This paper proposes using typing rules to describe the bytecode verifier because they are more precise than prose, clearer than code, and easier to reason about than either. JVML has a subroutine construct used for the compilation of Java’s try-finally statement. Subroutines are a major source of complexity for the bytecode verifier because they are not obviously last-in/first-out and because they require a kind of polymorphism. Focusing on subroutines, we isolate an interesting, small subset of JVML. We give typing rules for this subset and prove their correctness. Our type system constitutes a sound basis for bytecode verification and a rational reconstruction of a delicate part of Sun’s bytecode verifier. 1 Bytecode verification and typing rules The Java language is typically compiled into an intermediate language that is interpreted by the Java Virtual Machine (VM) [LY96]. This intermediate language, which we call JVML, is an object-oriented language similar to Java. Its features include packages, classes with single inheritance, and interfaces with multiple inheritance. However, unlike method bodies in Java, method bodies in JVML are sequences of bytecode instructions. These instructions are fairly high-level but, compared to the structured statements used in Java, they are more compact and easier to interpret. JVML code is often shipped across networks to Java VMs embedded in web browsers and other applications. Mobile JVML code is not always trusted by the VM that receives it. Therefore, a bytecode verifier enforces static constraints on mobile JVML code. These constraints rule out type errors (such as dereferencing an integer), access control violations (such as accessing a private method from outside its class),

Today's type-safe low-level languages rely on garbage collection to recycle heap-allocated objects safely. We present LTAL, a safe, low-level, yet simple language that "stands on its own": it guarantees safe execution within a fixed memory space, without relying on external run-time support. We demonstrate the expressiveness of LTAL by giving a type-preserving compiler for the functional core of ML. But this independence comes at a steep price: LTAL's type system imposes a draconian discipline of linearity that ensures that memory can be reused safely, but prohibits any useful kind of sharing. We present the results of experiments with a prototype LTAL system that show just how high the price of linearity can be.

Intensional polymorphism, the ability to dispatch to di#erent routines based on types at run time, enables a variety of advanced implementation techniques for polymorphic languages, including tag-free garbage collection, unboxed function arguments, polymorphic marshalling, and flattened data structures. To date, languages that support intensional polymorphism have required a type-passing (as opposed to type-erasure) interpretation where types are constructed and passed to polymorphic functions at run time. Unfortunately, type-passing su#ers from a number of drawbacks: it requires duplication of run-time constructs at the term and type levels, it prevents abstraction, and it severely complicates polymorphic closure conversion.

Phantom types are data types with type constraints associated with dierent cases. Examples of phantom types include typed type representations and typed higher-order abstract syntax trees. These types can be used to support typed generic functions, dynamic typing, and staged compilation in higher-order, statically typed languages such as Haskell or Standard ML. In our system, type constraints can be equations between type constructors as well as type functions of higher-order kinds. We prove type soundness and decidability for a Haskell-like language extended by phantom types.

The FLINT project at Yale aims to build a state-of-the-art systems environment for modern typesafe languages. One important component of the FLINT system is a high-performance type-directed compiler for SML'97 (extended with higher-order modules). The FLINT/ML compiler provides several new capabilities that are not available in other type-based compilers: ffl type-directed compilation is carried over across the higher-order module boundaries; ffl recursive and mutable data objects can use unboxed representations without incurring expensive runtime cost on heavily polymorphic code; ffl parameterized modules (functors) can be selectively specialized, just as normal polymorphic functions; ffl new type representations are used to reduce the cost of type manipulation thus the compilation time. This paper gives an overview of these novel aspects, and a preliminary report on the current status of the implementation. 1 Introduction The FLINT project at Yale aims to build a state-of-the-ar...

Note: This document accompanies the paper “Practical type inference for arbitrary-rank types ” [6]. Prior reading of the main paper is required. 1 Contents

Run-time type dispatch enables a variety of advanced optimization techniques for polymorphic languages, including tag-free garbage collection, unboxed function arguments, and flattened data structures. However, modern type-preserving compilers transform types between stages of compilation, making type dispatch prohibitively complex at low levels of typed compilation. It is crucial therefore for type analysis at these low levels to refer to the types of previous stages. Unfortunately, no current intermediate language supports this facility. To fill this gap, we present the language LX, which provides a rich language of type constructors supporting type analysis (possibly of previous-stage types) as a programming idiom. This language is quite flexible, supporting a variety of other applications such as analysis of quantified types, analysis with incomplete type information, and type classes. We also show that LX is compatible with a type-erasure semantics. 1 Introduction Type-directed co...

Statically typed languages with Hindley-Milner polymorphism have long been compiled using inefficient and fully boxed data representations. Recently, several new compilation methods have been proposed to support more efficient and unboxed multi-word representations. Unfortunately, none of these techniques is fully satisfactory. For example, Leroy's coercion-based approach does not handle recursive data types and mutable types well. The type-passing approach (proposed by Harper and Morrisett) handles all data objects, but it involves extensive runtime type analysis and code manipulations. This paper presents a new flexible representation analysis technique that combines the best of both approaches. Our new scheme supports unboxed representations for recursive and mutable types, yet it only requires little runtime type analysis. In fact, we show that there is a continuum of possibilities between the coercion-based approach and the type-passing approach. By varying the amount of boxing an...