And #community-development is still there to discuss alternatives to Slack...

is there a way to force syntax-quote to not resolve a symbol? something like

`(defui ~name static om/IQuery (query [this] '[]))

so that static doesn't become my.ns/static

one way to do it is ~'static

that seems to work, thanks!

why when I

(->> ["a" "b" "c"]
   (fn [coll] (nth coll 2)))

(->> ["a" "b" "c"]
   (clojure.string/join))

I run into this quite often when I want to third some collection in a this thread last macro.

I guess it has to do with how the ->> macro resolves the fn. if I defn the third function it works.

yeah, you've gotta wrap it in an extra parens

(->> ["a" "b" "c"]
   ((fn [coll] (nth coll 2))))

->> just threads stuff into the last position of the forms. So with the second example, it just turns into (clojure.string/join ["a" "b" "c"]). If you think about the first one then, all its doing is adding more to a (fn ...) form.

ah yes, that works. And makes perfect sense @chrisjd 🙂 thanks

@baritonehands 4clojure also uses clojail and it has very fine grain restrictions for educational purposes

given a vector like this

(def l [1 {:class "first"}
                             [2 {:class "second"}]])

I would like to be able to add more lists inside the second nested vector

and get something like this [1 {:class "first"} [2 {:class "second"} [["one"] ["two"]]]]

this actually works (assoc-in l [2 2] [["one"] ["two"]])

but I wonder if there is a better way that doesn't need to hard code indices?

@seancorfield the bot is this one: https://github.com/verma/clj-slackbot. It does use Clojail

@andrea.crotti with specter it's (setval [LAST AFTER-ELEM] [["one"] ["two"]] l)

will perform ~3x faster compared to assoc-in version

hello 🙂 i read on cascalog introduction that you can Query anything – Query HDFS data, database data, and/or local data

that means you can query a relational table with cascalog or any database?

@takis_ you do that by implementing cascading's "Tap" interface

depending on the db an implementation may already exist

i am so new that i don't understand , all i was thinking was using cascalog instead of sql, thank you

cascalog is for batch processing so isn't a drop-in replacement for sql

you could implement a special planner which incorporates the fine-grained query capabilities necessary, but afaik no one's done that yet

thanks alot for your time and cascalog 🙂 i will read to fully get it

is it possible to get an exhaustive list of generated values by spec? similar to excercise, but give you all possible generated values. This would make sense for example for a spec that is #{:a 😛 :c}

that is just about the only case I can think about where it is possible

it is impossible for general fn specs

for set specs you can use (s/form ::your-spec) to get all the values

I don't quite get what's the desired shape for :prep-tasks in the :uberjar configuration

if I look at here for example https://github.com/technomancy/leiningen/blob/master/sample.project.clj#L257

it's a bit weird :prep-tasks [["protobuf" "compile"] "javac" "compile"]

at the moment I have

:prep-tasks ["compile"
                          ["cljsbuild" "once" "min"]]

and can't find how I'm supposed to write it

@schmee thanks!

btw is it possible to share spec between clojurescript and clojure? for clojure I'm using the backported library with 1.8

dunno off-hand, try asking in #clojure-spec 🙂

@schmee is there a channel for everything? 😄

btw answering myself this seems to work

:prep-tasks ["compile"
                          ["garden" "once"]
                          ["cljsbuild" "once" "min"]]

still don't quite get what's the format it wants

and it's quite different from that example in the lein docs, but well as long as it works 😄

@andrea.crotti it accepts either a string (for a command that has no args) or a vector of strings

and the whole thing should be in a vector

so :prep-tasks [["compile"] ["garden" "once"] ["cljsbuild" "once" "min"]] would behave the same as what you have now

Ah ok I get it now thanks

is there an accepted way of testing if a function/macro in the core namespace is side effecting?

there's no reliable way to know

not programmatically, that is

many but not all things that are not safe in transactions have names ending in !

and the thing that makes them transaction unsafe is some side effect that might be repeated (or a risk of bad state coordination)

how to turn a collection into nil if it's empty, but return if it's not empty?

@pwrflx not-empty

or seq, if you want to get a sequence back

@stuartsierra thx!

you're welcome!

is there some way of using spec's valid?, explain, conform, etc. with a map that contains the keys without namespace? The spec itself I've defined to accept let's say {:mypec/a "a" :myspec/b "b"}, but i have a map that has simply {:a "a" and 😛 "b"} and would like to know if it is valid or not..

is that good practice at all?

however, using the namespaced variants is a lot of extra typing 🙂 also, when deserializing a map from JSON, that is without a namespace by default..

I see 1.9 has a convenience feature for this: https://dev.clojure.org/jira/browse/CLJ-1910

@pwrflx there's also :req-un

(s/keys :req-un [::a ::b])

Stands for require unqualified

is it possible to customize the way clojure printout the exceptions data ?

Fir spec, there's explain-out

But I'm not sure for normal exceptions

@didibus thx!