In imperative synchronous languages, a statement is called schizophrenic if it is executed more than once in a single clock. When a schizophrenic statement is translated into a circuit, the circuit can behave abnormally because of the multiple executions. To solve the problems caused by schizophrenic statements, compilers duplicate the statements to avoid multiple executions. Esterel is an imperative synchronous language. Schizophrenic statements in Esterel are considered to occur due to the instantaneous reentrance of local signal declarations or parallel statements. However, if the corresponding circuit of a schizophrenic statement behaves normally, it is harmless and thus curing is not necessary. In this paper, we identify the conditions under which a schizophrenic statement of the Esterel program must be cured during circuit translation. We also propose an algorithm to detect schizophrenic statements that have to be cured on the control flow graphs (CFGs) of source codes. Our algorithm detects all schizophrenic statements that have to be cured and results in fewer false alarms on the benchmark programs used in the previous work. It is simple and based on the CFG of a program, implying that it can be merged into existing compilers easily.

Esterel is an imperative synchronous language for control-dominant reactive systems. The combination of imperative structures and the perfect synchrony hypothesis often result in schizophrenic statements. Previous studies explain the characteristics of schizophrenia as the instantaneous reentrance to block statements: local signal declarations and parallel statements. In practice, however, most instantly-reentered block statements do not cause any problems in Esterel compilation. In this paper, we refine schizophrenic problems in terms of signal emissions, and suggest an algorithm to detect harmful schizophrenia using reachability on control flow graphs (CFGs) in Esterel. Our algorithm performs well in analyzing practical programs. Moreover, it can be easily applied to existing compilers.

k-transformable grammars have been conjectured to be the uppermost class of LL(k) covering transformable grammars. PLR(k) grammars have been known as a well characterized subclass of k-transformable grammars. Being contrary to those claims, this paper shows that some PLR(k) grammars are not k-transformable, and so k-transformable grammars are not the true uppermost. A powerful LL(k) covering transformation is suggested in this paper. It is a generalization of the transformations of k-transformable grammars and PLR(k) grammars. A remarkable aspect of the new transforming process is the deterministic property, where "deterministic" means that the transformation is obtained in a single process without requiring any heuristic, unlike k-transformable grammars' transformation for which a heuristic is required. The transformable grammar class is shown to be larger than k-transformable grammars and PLR(k) grammars.

When the compiler encounters an error symbol in an erroneous input, the local error-repair method repairs the input by either inserting a repair string before the error symbol or deleting the error symbol. Although the extended FMQ of Fischer et al. and the method of McKenzie et al. report the improved quality of diagnostic messages, they suffer from redundant parse stack configurations.This article proposes an efficient LR error-recovery method, with validation-removing repairs that give the same validation result as a previously considered, lower-cost repair. Moreover, its execution speed is proportional to the length of the stack configuration. The algorithm is implemented on a Bison, GNU LALR(1), parser generating system. Experimental results are presented.

The traditional LALR analysis is reexamined using a new operator and an associated graph. An improved method that allows factoring out a crucial part of the computation for defining states of LR(0) canonical collection and for computing LALR(1) lookahead sets is presented. This factorization leads to significantly improved algorithms with respect to execution time as well as storage requirements. Experimental results including comparison with other known methods are presented.