This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2016-02-25
Channels
- # aatree (10)
- # beginners (59)
- # boot (314)
- # braveandtrue (4)
- # cider (50)
- # cljs-dev (12)
- # cljsrn (6)
- # clojure (206)
- # clojure-austin (2)
- # clojure-gamedev (90)
- # clojure-japan (1)
- # clojure-poland (12)
- # clojure-russia (10)
- # clojure-sg (1)
- # clojurescript (86)
- # core-async (2)
- # core-matrix (3)
- # cursive (40)
- # datomic (2)
- # dirac (13)
- # editors (25)
- # emacs (7)
- # hoplon (2)
- # immutant (10)
- # jobs (24)
- # jobs-discuss (1)
- # ldnclj (8)
- # lein-figwheel (19)
- # leiningen (1)
- # mount (7)
- # off-topic (34)
- # om (147)
- # onyx (11)
- # parinfer (151)
- # pedestal (2)
- # re-frame (31)
- # reagent (13)
- # ring-swagger (7)
- # spacemacs (1)
- # yada (11)
@dominicm: re: virtualenv with Neovim: https://github.com/neovim/neovim/issues/3396
@dominicm: I'll do PR soon with fixes and improvements
@juhoteperi: I figured something like that would be possible. I think npm's model works best. Thanks! I'm using it in a project today, so I'll get a few things too I hope. My python-fu isn't as good as it used to be. I need to come up with safer key lookup systems. But it's just so many if/else's that I didn't want to worry about.
@dominicm: I'll fix key lookups and add support for additional data for candidates (arglist, type etc.)
@juhoteperi: Awesome, thank you! I hadn't even started thinking about arglist/type stuff. I look forward to seeing that.
@juhoteperi: I should apologize for some of my bad code in there, I was getting really confused about things that were breaking, and then magically fixed when I changed to doing weird things (like that r variable..). I'm not sure at all what did fix those things in the end. 😛
This is what I call "I have a train coming in 5 minutes, do whatever gets the code to run" debugging.
@dominicm: I think source shouldn't set min_pattern_length
as it looks like that will override user auto_completion_start_length
(or could be bug in deoplete)
but I think other sources don't set it?
fixed in the PR
@juhoteperi: No, most don't.. ah, that was it, to do with the regex value whilst I was debugging. Awesome, thank you.
Looks like fireplace could just embed this deoplete source and it would get used when deoplete is available
clang_complete does that, and it should also work with regular Vim
It's a possibility. I have no idea if that aligns with the pope's view of fireplace though. We can bring it to his attention & see what he thinks though.
@juhoteperi: No idea if you're using your changes, but they enable the docstring coming up whilst you're typing. That's awesome, I didn't think we'd be able to achieve feature parity with the omnicompletion (I hadn't checked out deoplete's features yet). One minor annoyance; keywords put 0 into the scratch buffer. But I'll take a look at the best solution for that a little later.
I have docstring preview disabled (I don't like the preview window - not sure if Neovim has some alternatives)
What is scratch buffer?
I don't like it much either. But I'm glad that we have it, just for the sake of matching. scratch buffer = where docstring preview shows.
Aha. Fixed it, val.get("doc")
returned None
. Using empty string as default works.
I had already pushed the fix so it was merged