Fork me on GitHub
#cider
<
2015-07-07
>
rui.yang01:07:58

clj-refactor also depends on end v1.1.1, but only v1.0 available on melpa stable.

seancorfield04:07:26

FYI, following up on my earlier issues with paredit, refactor, and Emacs Live: Emacs Live was using an old version of paredit — Sam updated the dev packs again. We’re all good!!

bozhidar07:07:56

@rui.yang: report this issue (one of the clj-refactor’s devs is also the author of edn.el)

rui.yang07:07:46

@bozhidar: reported to edn.el github. end.el also has a lot dependencies, not sure if they are all at melpa stable or not, hopefully yes.

bozhidar07:07:39

I’m not involved there at all

bozhidar07:07:51

edn.el is on my radar for future improvements of cider

bozhidar07:07:08

but right now I don’t need it and I don’t have time to investigate it further

rui.yang07:07:55

@bozhidar: oh, I didn't mean to ask you for the answer or investigation. It's just sometimes missing dependencies in melpa stable is a bit annoying.

bozhidar07:07:26

the problem is that most developers actually use melpa unstable

bozhidar07:07:33

and don’t notice problems like this at all

bozhidar07:07:45

and you don’t get errors from melpa stable

bozhidar07:07:55

if you try to publish something broken there

rui.yang07:07:51

I totally understand. maybe I should change my emacs config install from melpa

bozhidar07:07:59

depends on how adventurous you are simple_smile

malabarba09:07:07

@rui.yang: Honestly, it's nice to have people using melpa-stable and reporting this kind of stuff.

malabarba09:07:12

Melpa-stable is pretty new, and with the system it uses it can be hard for the developers to keep track of all of this stuff. You're being very helpful to the developers by finding these issues and reporting them. simple_smile

bozhidar10:07:27

on the other hand it’d nice if all recipes had some maintainer email attached to them

bozhidar10:07:40

and the system did a couple of sanity checks when building packages

bozhidar10:07:53

notifying maintainers automatically in case of such trivial issues

malabarba10:07:54

Oh, certainly. There's an issue requesting that on Melpa

malabarba10:07:21

It might be easier for them to implement this with 25.1, because package.el now detects and reports packages with unavailable dependencies.

malabarba10:07:41

So it's just a matter of adding that to their server logic.

bozhidar10:07:45

ah, that’s nice

jrychter12:07:57

I see clj-refactor is being discussed. I just tried upgrading to 1.1.0, discovered that my emacs HEAD package manager is broken, fixed the breakage, discovered that clj-refactor 1.1.0 depends on peg-20130320.158, and gave up for the day.

tsdh13:07:09

@jrychter: Really? I use the version 20150706.1342 from MELPA which is only slightly ahead, and that doesn't have a dependency on peg

tsdh13:07:32

@jrychter: Ah, you're right. It's an indirect dependency. clj-refactor depends on edn which in turn depends on peg.

malabarba13:07:37

@tsdh @jrychter Thanks for noticing. I've filed an issue for that. https://github.com/expez/edn.el/issues/4

jrychter13:07:54

@malabarba: well, since we're on that subject (which is offtopic, I know) — I tried 21.5 (HEAD) and package.el didn't work for me. It seems that found-something needs to be moved into an outer let form in package-compute-transaction, otherwise one gets a Symbol's value as variable is void error. I'm puzzled why it worked for anyone, actually. I'm back to 24.5.1 simple_smile

jrychter13:07:53

@malabarba: and thanks for reporting the peg dependency issue. I wouldn't know where to report it.

malabarba14:07:19

@jrychter: Hey! Thanks for catching that. The reason noone else caught it is that it only happens when you try to install a package with a missing dependency (as was the case for you with clj-refactor).

jrychter14:07:40

@malabarba: well, that's me. I always break things. Literally every piece of software breaks for me, because there is always something atypical in my setups. Companies should hire me to be their tester.

malabarba14:07:43

You should launch a unit testing web service that simply runs people's tests on your configs

malabarba14:07:16

Anyway, it's fixed now. Thanks for diping your toes, that was helpful. simple_smile

jrychter14:07:28

Well, it's more annoying if the problems happen to be with my OS, or large applications. Or with the Xbox one login procedure (which I managed to break recently). Or with a top-of-the-line Logitech mouse. Sigh. Anyway, I'm glad I could help!

pbostrom20:07:10

I'm trying to update cider from 0.8.2 -> 0.9.1, when I try to jack-in, it starts the nREPL server but doesn't seem to initialize the cider client buffer. I don't see errors in the Messages buffer:

Starting nREPL server via lein repl :headless...
nREPL server started on 51759
nREPL: Establishing direct connection to localhost:51759 ...
nREPL: Direct connection established
Connected.  May the Source be with you!

nullptr20:07:34

pbostrom: i had the same issue, try deleting the package completely and installing fresh

pbostrom21:07:10

do folks prefer boot vs lein for cider? After I upgraded I noticed cider complaining about lein's version of tools.nrepl 0.2.6, which is several versions behind. Last official lein release is 2.5.1 from January it seems

arrdem21:07:36

there's a profile.clj workaround for that

arrdem21:07:40

but boot is quite nice 😛

pbostrom21:07:57

yeah, I guess I'm wondering out loud if lein is falling behind

arrdem21:07:00

which I'm happy to paste if you don't have it yet

arrdem21:07:08

which is fine

roberto21:07:17

i like lein 😞

gtrak21:07:19

I add cider-nrepl to my java project's dependencies via a maven profile simple_smile

arrdem21:07:28

you are also a literal monster

gtrak21:07:48

dynamic requires in the loader code

arrdem21:07:11

lein is fine. it does its job acceptably. it was never good at composing lots of tasks, and now something designed for that has rolled around.

arrdem21:07:17

but with other tradeoffs.

arrdem21:07:46

frankly I don't think lein or boot is sufficient standalone... make is probably called for to plaster over either

roberto21:07:51

yeah, I haven’t had the need to compose lots of tasks yet. That is why I’m still happy with lein.

arrdem21:07:51

but that's my $0.02

roberto21:07:56

as long as we don’t get like the JS ecosystem.

arrdem21:07:43

I mean the bottom line is that both are thin wrappers around Maven right

arrdem21:07:50

and you can always fall back to "just use Maven"

roberto21:07:32

yeah, I still need to give boot a spin. Didn’t want to get distracted by tooling before I felt comfortable with clojure and emacs.

roberto21:07:43

which I can’t say about JS at this moment.

arrdem21:07:57

Definitely check boot out, I enjoyed playing with it although I'm not using it for anything yet.

roberto21:07:17

cool, yeah, will do.

arrdem21:07:55

also the #C053K90BR folks are pretty helpful 😛

aengelberg21:07:30

@arrdem what is the workaround for out of date nrepl?

arrdem21:07:56

@aengelberg: lemme dig it out for you one sec

arrdem21:07:08

{:user {:plugins             [[cider/cider-nrepl "[0.9.0,)"]
                              [refactor-nrepl "1.1.0-SNAPSHOT"]]
        :dependencies        [[org.clojure/tools.nrepl "0.2.10"]]
        :deploy-repositories [["clojars" {:url "" :creds :gpg}]]}}

arrdem21:07:34

a nearly minimal CIDER profile.clj

arrdem21:07:59

@aengelberg: ^

aengelberg21:07:32

Thanks. I think I'm already doing something like that but I recall there being pain.

aengelberg21:07:50

Now I remember, the pain is that my project uses :pedantic :abort

aengelberg22:07:27

And I don't think there's a way to :exclude the nrepl dependency in leiningen internals.

arrdem22:07:02

*not that I know of

aengelberg22:07:03

Is boot better in that regard?

arrdem22:07:39

it also packages an nREPL version for their boot repl task...

aengelberg22:07:35

Anyway, a bit of a tangent for this channel simple_smile

bozhidar22:07:09

at least they update their dep all the time

bozhidar22:07:31

lein’s been having an identity crisis ever since Phil went to Asia

rui.yang23:07:09

cool, seems that I just fired a duplicate issue around that peg dependency. https://github.com/clojure-emacs/clj-refactor.el/issues/187