| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
…peedup for formatter
There was a problem hiding this comment.
Really awesome! Honestly I think removing LINQ use generally from PSSA could provide performance boosts across the board
Sorry, something went wrong.
Co-Authored-By: Robert Holt <rjmholt@gmail.com>
Now, Linq consumers 18.5% of the CPU (but this include everything including SMA, therefore it is hard to say how much of that stems from PSSA only), so there is some potential but it's not as big as one would think. |
Sorry, something went wrong.
| Back | FazBrowse Home | New Git URL |
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 function PositionIsEqual only returns pipeline asts on the same line.
PR Checklist