Finding Minimal Pass Sequences for Attribute Grammars

Alblas, Henk

doi:10.1137/0214063

Cited by 3 publications

(4 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proof consists of reducing the shortest common superseqluence problem to this problem. We note that in [3] a polynomial algorithm is given for computing minimal pass sequences.…”

Section: Time Complexitymentioning

confidence: 99%

Passes, sweeps, and visits in attribute grammars

Engelfriet

Filé

1989

J. ACM

View full text Add to dashboard Cite

Theoretical results are presented on multi-pass (both left-to-right and alternating), multisweep, and multi-visit attribute grammars. For each of these, a pure type and a simple type are distinguished: The pure attribute grammars are defined by nondeterministic attribute evaluators, and the simple ones by the corresponding (usual) deterministic evaluators. The time complexity of deciding membership in these classes of attribute grammars is studied. In general, this is harder for the pure classes than for the simple ones, for which it is either polynomial or NP-complete. The expressive power of the eight classes is compared by studying the translations they can compute. It is shown that sweeps are more powerful than passes, and visits are more powerful than sweeps.

show abstract

“…The proof consists of reducing the shortest common superseqluence problem to this problem. We note that in [3] a polynomial algorithm is given for computing minimal pass sequences.…”

Section: Time Complexitymentioning

confidence: 99%

Passes, sweeps, and visits in attribute grammars

Engelfriet

Filé

1989

J. ACM

View full text Add to dashboard Cite

show abstract

“…16, where terminal symbols are deleted. The replacement of the conditional statement by its then part or its else part may cause instances of attribute mad to get different values and this, in its turn, may require the recomputation of instances of /-pool and s-pool, needed for additional transformations within stats [2]. The situation pictured in Fig.…”

Section: Transformation Rulesmentioning

confidence: 99%

“…Observe that no re-evaluation rules have been included for i of compound [-1] and i of compound [2]. Apparently their values can be copied from the corresponding attribute instances in itt.…”

Section: Definition 34mentioning

confidence: 99%

“…Now, we add another tree transformation rule which specifies the replacement of a conditional statement, in case of a compile-time evaluable condition, by its then part or its else part. [2] into (stat, (compound, begin, stats [2], end)) eval mad of compound :=mad of stats [2]; mad of stat.-=mod of stats [2] ; /-pool of compound :=/-pool of stat; s-pool of compound ==s-pool of stats [2] ; s-pool of stat ==s-pool of stats [2] end.…”

Section: Transformation Rulesmentioning

confidence: 99%

See 1 more Smart Citation