work has already been done in #1371 to reduce the startup impact of click. further work could be done to profile the application's initialisation route and remove further overhead.
eventually, an implementation of the language server protocol is desirable, removing the need for IDEs to run proselint's startup routine more than once. an intermediary implementation similar to the work done in eslint_d may be beneficial, and would only require allowing stdin mode to lint continuously.
this is a follow up of #1363.
work has already been done in #1371 to reduce the startup impact of click. further work could be done to profile the application's initialisation route and remove further overhead.
eventually, an implementation of the language server protocol is desirable, removing the need for IDEs to run proselint's startup routine more than once. an intermediary implementation similar to the work done in
eslint_dmay be beneficial, and would only require allowing stdin mode to lint continuously.