Please can I make a plea for fixing the create namespace workflow?

If I accidentally create a .clj file when I wanted a cljs/cljc one more time I think I shall go mad! 🙂

@sandbags You mean https://github.com/cursive-ide/cursive/issues/904, right?

One thing that isn’t mentioned in that issue is what to do about cljc files, since there’s no way to autodetect when to use those. https://github.com/cursive-ide/cursive/issues/893 also suggests a way to convert between the file types - does that sound useful?

@cfleming did you think about a timetable for boot integration ? I know how to generate a lein project-file from boot and that of course works for now, but it seems like direct integration could offer so much more

@pseud I don’t have a timeline, but it’s definitely something I want to do soon. It’s one of the most requested items.

And if you would then drop lein support so as to send a message to the unbelievers … 🤓

Umm, that might be a bit extreme 🙂

@cfleming could you look at files already in the namespace?

typically in my experience if i have a cljs file in a namespace (folder) the next i create will also be a cljc, etc...

@sandbags: Should that read: “if i have a cljc file in a namespace”?

I’m guessing yes. So yeah, perhaps either that, or just remember what the user last chose.

@pseud this is what i use to make intellij and boot projects friends: https://clojars.org/onetom/boot-lein-generate and here is a full example where u can quickly try it too as a reference: https://github.com/onetom/hoplon-layouts

@onetom - interesting 🙂 Will have to read this over to fully grok, but any little step toward making the process of synchronization more transparent is a good one.

@cfleming i guess i was addressing that you might have the same namespace in both a CLJ and CLJS

@cfleming they are different folders ususally in Cursive

@sandbags Right, but you would normally have a clj and a cljs, and cljc files normally go under clj, right?

i’m assuming that you can tell which is which

@cfleming exactly

So if you create a file in clj, I can’t tell whether you want clj or cljc

oh, i usually have a cljc folder

Except by remembering what you did last time, or something similar.

except when i don’t have clj & cljc

Ok, I can check that.

can you look what else is in that namespace?

I’m not sure yet, probably.

i.e. the first time ask, subsequently default to whatever is already in that namespace?

And if there’s a mix, use the last one, or something similar.

i think that sounds reasonable

typically i cannot imagine i would mix clj and cljc in the same folder

but even if you do i can’t think of a better approach

Ok, sounds good.

@bfabry That’s fixed in the latest EAP: https://github.com/cursive-ide/cursive/issues/1560

mm that's what I'm on

Version: 1.4.0-eap3-2016.2

No, I mean the one I’m going to release shortly 🙂

haha

awesome thanks

I should have said “next EAP"

How does one connect to a remote socket server repl (not nrepl) using cursive?

when speccing functions, is it normal to put the (s/fdef my-fn) above the (defn my-fn) line? i feel like i’ve seen this in one or two official bits of clojure documentation

if so, will cursive support this? currently i see “my-fn cannot be resolved” when i try to do this

i realize it would be helpful if i could link to a specific piece of documentation that does what i said on that first line, trying to remember..

oh i think i saw dnolen do this in the video of his talk from last week

@joshg Currently there’s no native support for this. I’m planning to add this soon as part of adding support for CLJS REPLs.

@cfleming Got it. Thanks!

@joshg: As a workaround, you can use https://github.com/mfikes/tubular

@jrheard Currently there’s no native support for that either 🙂

hee

cool, thanks

But if it should be supported, let me know.

will do! i’ll keep an eye out for actual cognitect documentation that says that this is an official pattern

@jrheard that's not a pattern

fdef can appear anywhere

this necessarily needs to be the case

if you want to spec someone else’s thing

makes sense

@dnolen So it’s valid to spec a fn ahead of the fn definition in the same ns?

exactly

Ok, I’ll fix that.

❤️ thanks for clearing that up david!

@jrheard Could you file an issue for that?

you got it

Thanks

it also simplifies stuff like optional type checking

you don’t want to enforce an ordering

Ok, no problem.

@jrheard That said, it’s also valid to put the spec after the fn, so that will work as a workaround in the meantime.

yup, i’ll continue doing that in the meantime 🙂

thanks!

@denik 1. you are trying to make code look (very subjectively) good by downgrading semantics, which actually (I am pretty sure "objectively") matter more: aligning :content of one element with :margin-top of another gives you nothing other than "looks cool, I don't even need to read it!". 2. your examples are rather naive if you compare it with any just-a-tiny-bit-non-trivial piece of code using any CSS frameforks, where you have 3 divs-deep structures with 50-chars-long class names combos just to align a button. Those rarely fit within 120-chars-wide editor tabs, let alone having 2 of them viewed at once side by side. 3. most of the line changes with subsequent re-formatting will generate tens of lines in commit diff. That's just a waste of people's attention. 4. that's a pain in the ass to implement for such a tiny subjective benefit of "looking cool"

please be nice 🙂

@jrheard I was asked for detailed explanation : )