Expand RegexNode atomicity test to loops at the end of alternation branches #1769
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…anches Given an expression like `(abcd*|efgh*)ijk`, we will now see that if the `i` after the alternation doesn't match and we backtrack into the alternation, we won't need to backtrack into the `d*`, as it can't give back anything that would allow the `i` to match.
stephentoub
added
area-System.Text.RegularExpressions
tenet-performance
danmoseley
reviewed
Jan 17, 2020
| } | ||
|
|
||
| return false; | ||
| } | ||
| } | ||
| } | ||
| } |
There was a problem hiding this comment.
Aside, I noticed on line 832, this
// If this node is a one/notone/setloop, see if it overlaps with its successor in the concatenation.
// If it doesn't, then we can upgrade it to being a one/notone/setloopatomic.
maybe should be clarified to say
// If this node is a oneloop/notoneloop/setloop, see if it overlaps with its successor in the concatenation.
// If it doesn't, then we can upgrade it to being a oneloopatomic/notoneloopatomic/setloopatomic.
since one and notone are node types that are distinct to oneloop/notoneloop.
There was a problem hiding this comment.
Eh, it's everywhere eg
/// Simple optimization. If an atomic subexpression contains only a one/notone/set loop,
/// change it to be an atomic one/notone/set loop and remove the atomic node.
There was a problem hiding this comment.
Ok. I can see how it might not be clear; in my head I see the grouping, e.g. as if it were {one/notone/set}loop, but I can clean it up. I'll do so in my next PR.
danmoseley
reviewed
Jan 17, 2020
| // skip Groups, as they should have already been reduced away. | ||
| // If there's a concatenation, we can jump to the last element of it. | ||
| while (node.Type == Capture || node.Type == Concatenate) | ||
| static void ProcessNode(RegexNode node, RegexNode subsequent, int maxDepth = 20) |
There was a problem hiding this comment.
Nit, might be better to pass in 20 above rather than default it here, so any future calls don't forget to decide what to pass in.
There was a problem hiding this comment.
Ok. I'll address that subsequently.
danmoseley
approved these changes
Jan 17, 2020
ghost
locked as resolved and limited conversation to collaborators
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Given an expression like
(abcd*|efgh*)ijk, we will now see that if theiafter the alternation doesn't match and we backtrack into the alternation, we won't need to backtrack into thed*orh*, as it can't give back anything that would allow theito match.Best reviewed with whitespace disabled: https://github.com/dotnet/runtime/pull/1769/files?w=1
Contributes to #1349
cc: @danmosemsft, @pgovind, @eerhardt, @ViktorHofer, @lpereira