This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2016-04-22
Channels
- # admin-announcements (7)
- # beginners (56)
- # boot (69)
- # cider (168)
- # cljs-dev (2)
- # clojure (170)
- # clojure-austin (25)
- # clojure-beijing (3)
- # clojure-belgium (2)
- # clojure-france (3)
- # clojure-poland (17)
- # clojure-russia (115)
- # clojure-uk (40)
- # clojurebridge (3)
- # clojurescript (87)
- # cursive (9)
- # datomic (30)
- # dirac (18)
- # editors (3)
- # emacs (14)
- # hoplon (195)
- # immutant (14)
- # jobs (3)
- # jobs-discuss (4)
- # leiningen (11)
- # melbourne (5)
- # mount (42)
- # off-topic (5)
- # om (24)
- # onyx (48)
- # parinfer (53)
- # proton (1)
- # protorepl (2)
- # re-frame (3)
- # reactive (2)
- # reagent (29)
- # rum (5)
- # spacemacs (4)
- # untangled (91)
- # yada (1)
OK, I almost have cider 12 more or less working - it's astonishing how much effort this appears to require (moving from 9 to 12). I'm not sure, but I think the only odd thing left is cider-connect no longer has cider mode enabled by default in the cider-nrepl buffer. Other than adding hooks to try to force this, is there anything 'obvious' that I may be missing?
OK, maybe there was something else in operation that gave function args previously?
Yes, but the old hook for that cider-turn-on-eldoc-mode, or whatever, was removed (previously deprecated) and supposedly should be replaced by eldoc-mode, but that appears to not have any effect
Or just (global-eldoc-mode 1)
anywhere (though that will affect more than just cider)
Worth a try, this: (add-hook 'cider-mode-hook 'cider-turn-on-eldoc-mode)`` used to work
Also, why in the world, when giving function signature, does the new cider stuff (I have no ideas what combination is doing this), print the fully qualified name space for the function?? Is that configurable?
Well, that was the only place that eldoc mode occurred in my setup. It's possible that adding the hook you suggest won't work either - I need to try that...
(dolist (fn '(paredit-mode
subword-mode
eldoc-mode))
(when (fboundp fn)
(add-hook 'clojure-mode-hook fn)
(add-hook 'cider-repl-mode-hook fn)))
that's a nice generalized way to do this stuff
OK, your hook does enable this again, without having full cider mode.
But the odd thing now is when connecting the final result is a weird 'Project directory: <directory of file cider-connect launched from> in the mini buffer. What is that? And why is it there?
Actually, that's really annoying as it isn't the actual project directory!!
You get the random quote if you remove the hook
OK, I'm wrong about that - I removed that hook and for some reason I no longer get the random message - could this be related to using a second editor session to connect up to the server? The first connection (which I didn't shutdown) still gives the random message
Yes, two
I had a session running and that one didn't have your hook. But I was trying out your hook in a totally separate editor session
So, two totally different emacs instances running against the same server
OK, I shut everything down and am putting the hook back and will see what happens here
The thing is that cider-jack-in
knows the project its connecting to, but cider-connect
doesn't. So we need you to confirm the directory.
That's odd. The first emacs didn't ask for that and just gave the random quote. I now shutdown both emacs and fired up a new one with the hook but didn't stop the server. Still get the weird project dir msg and no random quote. Will now stop server as well and retry
I mean clojure server. And yes, shutting that down and the emacs (all of them) down and running a new emacs with your hook now: starts cider-nrepl, gives the random quote with no weird project directory mini buffer request, and does provide function signatures.
I have no idea what toxic interactions were happening
cider-connect. And just tried a second emacs into the new clojure server and - yes, now in that second emacs when issuing cider-connect, upon connection, I get no random message and the weird Project directory mini buffer request.
I'm not sure I ever had two emacs going into the same clojure server using cider and nrepl. I know this worked seamlessly with the old swank/slime combo.
How could it not find one?
You mean running the emacs outside the directory ?
Let me try that
If call cider-connect
while actually inside a clojure project, cider-connect will not ask you anything
Confirmed - if I start it from a file that is nested in the src of a project no project directory request and random msg is given
I think the better way of saying that is if you issue cider-connect in a buffer on a file that is in the src directory of a project it won't ask anything
Actually, why does cider care about this??? That seems like a server side issue only
does it just look up the directory tree from the files position until/unless it finds a project.clj?
OK, but again, why does cider care about this at all??
I don't get that - the connections are on different sockets, right? And the servers would have the different name spaces for the projects involved
So, cider should not need to know anything about projects
That seems like a strictly server side concern
Ah, ok, that does make sense
Well, many thanks malabarba for wasting some cycles on this to get things a bit clearer on my side
I hope/think I now have a new working cider 12 setup
@bozhidar: re: “ is ConcurrentList imported in your namespace?” - I was trying to trigger the completion within the :import
namespace section: (:import [ConcurrentLi…])
. Should completion work there?
@nooga: there's a variable that controls whether overlays use font locking, you can find that if you browse the customize menu. Just make sure it's t.
But I don't see anything wrong with the image you posted. Try evaluating something that actually has font locking, like a var or a map with keywords.
Hi, will cider look for upgrade when it load? My cider startup very slowly now. Is it possible a network issue?
It's slow when elisp eval (use-package cider)
or (require 'cider)
. It takes about 2 minutes.
Debugger entered--Lisp error: (quit) call-process("ssh" nil t nil "-o" "ControlPath=%C" "host.does.not.exist")
You can also try (setq tramp-ssh-controlmaster-options "-o ControlMaster=auto -o ControlPath='tramp.%%[email protected]%%h:%%p' -o ControlPersist=no")
@malabarba: I don’t like the yellow border ;D
@malabarba: any idea where to look for that face? I’m quite a newb in .el stuff
What if cider-repl-set-ns
managed a ring of namespaces (ala emacs' yank ring)? Then it would be easy to loop through recent namespaces and pop back and forth between a couple.
If you're looping a lot between namespaces, maybe you should be using files instead. :-)
so you can easily loop through those and then C-u C-c M-z
brings up the current ns in the repl
did the name of this cider--region-for-defun-at-point change post version 0.9? I'm getting a 'Symbol's definition is void ...' for this when trying to compile (via C-c C-c). I'm also using package cider-eval-sexp-fu for brief highlighting and am thinking this is where the problem is coming from
@jsa-aerial: I had to update cider-eval-sexp-fu
, I don't remember if it was for that particular function but I now have 0.12
and latest master
for cider-eval-sexp-fu
yes, I see in the 'fu' code a change on just this point
Ah - is that available via package update or does it require manual stuff
I always do manual, maybe ping the maintainer if you don't see the updated package
yes, the code was updated on git, but package-list-packages doesn't see an update available for the package...
There's something odd here - the MELPA site actually has the version with the fix, but I can't get emacs package system to realize it is a package with an update. I did get it to work by hacking the .el and 'installing from buffer', but that seems like a recipe for errors
Yes, but it apparently is the 'old' one
Hey @malabarba, what is the magic switch to turn off the 'ghostly' in buffer result Cider gives upon form evaluation?
that's it! Thanks!