-
-
Notifications
You must be signed in to change notification settings - Fork 406
Closed
Description
Given we are trying to optimise things, it makes sense we have benchmarks to measure whether things actually got faster or slower. My best guess would be to synthesise a large number of files (say 500?), use lsp-test, and then fire a representative set of requests/changes at them (e.g. a boat load of hover because that's what emacs does, and a little bit of everything else). The result should be useful to figure out why things like forkOn give performance improvements, as per https://gitlab.haskell.org/ghc/ghc/issues/18224#note_275367. CC @bgamari, @wz1000 and ndmitchell/shake#751
Metadata
Metadata
Assignees
Labels
No labels