Open atomontage opened 7 years ago
Just ran into this too and voting for it too. I never knew that Emacs sets those.
Also I think a better approach re the dependencies would be push a feature flag for it, so that loading prove will only conditionally add the dependencies for colours. Since other systems will only want to load the generic prove system, having two ASDF system wouldn't really work, would it?
(Lastly, another common check is whether the output is connected to a terminal, isatty basically - at least that's a case where you probably do want colours. But I imagine that's not available portably, or in UIOP yet; while osicat has it, getting the right FD would be another nightmare. Nevermind then.)
ANSI escapes will mess up the SLIME REPL buffer in Emacs (I think this is known already).
Emacs 25.1 here does not set the EMACS environment variable on subshells but it does set INSIDE_EMACS e.g.
CL-USER> (sb-unix::posix-getenv "INSIDE_EMACS") "25.1.1,comint"
CL-USER> (sb-unix::posix-getenv "EMACS") NIL
Ideally you should check both EMACS and INSIDE_EMACS.
Finally, to be super-duper pedantic about this, it'd be amazing if the entire color part was made optional (via separate ASDF system) so that those of us who despise colors of any sort in our test output don't have to install dependencies we'll never use.