Circular attribute grammars with remote attribute references and their evaluators

Abstract

Attribute grammars (AGs) are a suitable formalism for the development of language processing systems. However, for languages including unrestricted labeled jumps, such as "goto' in C, the optimizers in compilers are difficult to write in AGs. This is due to two problems that few previous researchers could deal with simultaneously, i.e., references of attribute values on distant nodes and circularity in attribute dependency. This paper proposes circular remote attribute grammars (CRAGs), an extension of AGs that allows (1) direct relations between two distant attribute instances through pointers referring to other nodes in the derivation tree, and (2) circular dependencies, under certain conditions including those that arise from remote references. This extension gives AG programmers a natural means of describing language processors and programming environments for languages that include any type of jump structure. We also show a method of constructing an efficient evaluator for CRAGs called a mostly static evaluator. The performance of the proposed evaluator has been measured and compared with dynamic and static evaluators.

Circular reference attributed grammars — their evaluation and applications

“…Sasaki and Sassa [26] also have a refined mostly static version of their originally completely static technique. The basic idea is here to have several versions of attribute evaluation sequences, one for each possible pattern of remote dependency edges.…”

“…His scheme is also limited to traditional AGs without remote references since it depends on deriving the attribute dependencies statically from the productions of the grammar. Sasaki and Sassa [26] have elaborated on the technique of Farrow in the presence of remote references. However, these references are not considered to be a part of the attribute grammar and must be evaluated separately in an initial phase.…”

“…They also constructed an exact algorithm with exponential time complexity and showed that finding the circular attributes is a harder problem than testing for circularity. The problem of testing reference attributed grammars for circularity can be addressed as in [26] by introducing all possible remote edges with lots of potential cycles in the dependency graph as a result. Most abstract syntax trees will, however, not contain any cycles and the iterations performed by the generated evaluator are unnecessary.…”

“…Boyland has implemented a similar combination in his APS system [2]. Sasaki and Sassa present Circular Remote Attribute Grammars (also abbreviated CRAGs), which on the surface is similar to our CRAGs [26]. However, Sasaki and Sassa assume that the remote links are computed separately outside the attribute grammar.…”

“…This is guaranteed if the values for each attribute on a cycle can be organized in a lattice of finite height and if all the semantic functions involved in computing these attributes are monotonic on the respective lattices. Several authors (e.g., [9,15,26]) have shown how the possibility of circular definitions of attributes allows simple AG specifications for some well-known problems from different areas. Examples include data-flow analysis, code optimizations, and properties of circuits in a hierarchical VLSI design system.…”