Improve performance of UseConsistentIndentation even more > another 10% speedup for formatter #1461
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Outdated suggestions cannot be applied.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.
PR Summary
This now reduced the total CPU time of this rule to only from 12% to 2% on a run of a warm run Invoke-Formatter (compared to current master with the 3 recent speed improvements). Improvements will be smaller for smaller scripts as this was tested against PowerShell's build.psm1 3000+ line module (same as with previous PRs), therefore it rather ensures performance scales much better.
Basically, because the rule goes over every token, any looping, inside it, even when it is only performed for each line as in this case, it has a measurable impact (both in the profiler but also actual timings). Therfore making sure we iterate over fewer items in the
pipelineAsts
list, by having a minimum index, which increases with each line and breaking the loop early if the line number is bigger than the current line. This is possible because the queried functionPositionIsEqual
only returns pipeline asts on the same line.PR Checklist
.cs
,.ps1
and.psm1
files have the correct copyright headerWIP:
to the beginning of the title and remove the prefix when the PR is ready.