Fork me on GitHub
#clr
<
2023-01-31
>
bobcalco18:01:42

@dmiller is there a way to influence the code search path for ClojureCLR? Must it always run in the root of a namespace or can it be told where code/binaries are while staying in the "current directory" (asking for a friend who is building a tool that will need other artifacts relative to current directory)?

dmiller19:01:47

Try environment variable CLOJURE_LOAD_PATH. Like any windows path, semicolon-separated list of directory names.

bobcalco19:01:05

ahhhh. Will give that a spin. thanks.

bobcalco20:01:33

if that works I can undo something I was undoing that made me take lots of showers. not that showers are bad. just that one only needs so many in a day.

😆 4
bobcalco17:02:33

@dmiller will that also affect normal clj if I make it system wide?

dmiller18:02:08

Yes. Load is given just the filename, not the extensions. It checks the standard places, then CLOJURE_LOAD_PATH. (should be the other way around?) Looks for cljr, cljc, clj i that order in each directory it touches.

bobcalco15:02:57

works great - I compute it in cljr on startup based on any current value in that env var plus what I find in deps.edn, set it for the current process, and it works.

bobcalco15:02:42

I did find one minor thing - the user.clj file can't be user.cljr or user.cljc so it seems that use case slipped through the cracks. It is desirable to support extensions consistently in that case too.

bobcalco02:02:03

@dmiller ping about the my last comment above in case you missed it

dmiller02:02:39

I did miss it. I'll create an issue for it and get it fixed soon.

👍 2
dmiller16:02:06

Master branch has fix. I look for all three versions, in order.