Simplify minterm generation in Regex NonBacktracking engine#61232
Simplify minterm generation in Regex NonBacktracking engine#61232olsaarik merged 2 commits intodotnet:mainfrom
Conversation
PartitionTree had a holdover from when it was tracking the predicates that made up each minterm, which caused it to use order n too much memory for some patterns. Nodes with the same predicate as the parent are no longer created. This also significantly simplified the algorithm.
|
Tagging subscribers to this area: @eerhardt, @dotnet/area-system-text-regularexpressions Issue DetailsPartitionTree had a holdover from when it was tracking the predicates that made up each minterm, which caused it to use order n too much memory for some patterns: nodes with the same predicate as the parent are no longer created. This also significantly simplified the algorithm. The effect is particularly evident in patterns with lots of non-overlapping predicates. For the pattern "qwertyuuiopasdfghjklzxcvbnm" this dropped the number of PartitionTree nodes created from 378 to 53.
|
| @@ -41,25 +40,17 @@ public MintermGenerator(IBooleanAlgebra<TPredicate> algebra) | |||
| /// <returns>all minterms of the given predicate sequence</returns> | |||
| public List<TPredicate> GenerateMinterms(params TPredicate[] preds) | |||
There was a problem hiding this comment.
As long as you're improving things here, does this need to be params or could that be removed?
There was a problem hiding this comment.
Good question, I don't think it has to be. I'll change it to match what this is called with.
stephentoub
left a comment
stephentoub
left a comment
There was a problem hiding this comment.
Less code and cheaper. Yay :)
Also remove the deduplication in there, since its now optional and the only call site already passes in a unique set of predicates.
ghost
locked as resolved and limited conversation to collaborators
2 participants
PartitionTree had a holdover from when it was tracking the predicates that made up each minterm, which caused it to use order n too much memory for some patterns: nodes with the same predicate as the parent are no longer created. This also significantly simplified the algorithm.
The effect is particularly evident in patterns with lots of non-overlapping predicates. For the pattern "qwertyuuiopasdfghjklzxcvbnm" this dropped the number of PartitionTree nodes created from 378 to 53.