i'm mostly concerned about divide by zero exceptions, which it looks like ratios do not prevent. maybe i should use doubles - you can divide a double by zero, looks like.

@michael740 is this what you're trying to do?:

(/ (+ 20 2) (+ 5 2))
=> 22/7

that would do it! i do think ratios are not the right tool for me, after all.

Also see quot which is regular division (not a ratio).

It rounds to the nearest integer.

@michael740, it sounds like you’re trying to do algebraic manipulations. maybe something like https://github.com/clojure-numerics/expresso ?

I have a cyclical data structure built with Refs that I would like to print. At some points I'd like the print the hex part of ref.toString() without printing the contents. What's the best to access this part Ref?

hey folks. I’m trying to parse the XML from a wordpress export to move into blog using cryogen or perun, some nice clojure static site generators. I think I need some pointers in parsing the xml though, as i’m able to follow examples on clojure doc and clojure’s own test to see how to get a single map from a post using the queries in the gist below. However, I’m unable to figure out how to iterate through an xml data structure to do this for the other 150-odd posts available to me - when I try to get the next I get the whole blob of XML, which is no good fo me. I’ve spent a good few hours reading up on how clojure zippers work, and trying various kinds of recursion, loops, and other macros but alas - no dice. I’m defeated by something I could do in pretty quickly in other languages :( You can see my code in the gist below, along with the source xml - how can I work with each post in the xml, to get a map back to write to a text files, or dump in a database? https://gist.github.com/mrchrisadams/cb9b3af65757b9fb9253b0c3cfa1d6cd

What is the correct way to test for a thrown ex-info using clojure.test? At the call site of ex-info I can specify data in a hashmap. With the (is (thrown? ...)) I would like to test the information in that hash map to make sure I'm getting the correct exception information.

You can wrap it in a try/catch, assign the exception to a local variable, use ex-data to get the data out, and check whatever part of its contents you wish, potentially using one or more (is (= x y)) expressions.

And if that proves a bit too much boiler-plate, you can hide it away by extending clojure.test with custom assertions by adding methods to the assert-expr multimethod - see thrown? and thrown-with-msg? themselves for examples.

@andy.fingerhut can you help me construct a try/catch will only catch the exception I designated in the (throw (ex-info ...)) ? And importantly will not accidentally catch any other.

@kelvended, How do I find the code for thrown? ? When I press M-. in emacs/cider, it cannot find the source code.

Are you concerned that some other exception other than the one you are testing will occur? Every JVM exception has a class, and there are many such classes. You can give a class name in the catch expression, and such a catch will only catch exceptions with that class, or one of its subclasses.

yes, so which class name should I give for exceptions generated by ex-info ?

clojure.lang.ExceptionInfo <<--

You can find out in a REPL by doing a catch on the most general class for all JVM exceptions, Throwable, and then in the exception handler print out the (class x) of the exception x

(doc ex-info)
-------------------------
clojure.core/ex-info
([msg map] [msg map cause])
  Create an instance of ExceptionInfo, a RuntimeException subclass
   that carries a map of additional data.

@gerritjvv yes reading your reply as we speak. Thanks for the helpful example.

Suddenly clojure -A:new stopped working though I have it defined in ~/.clojure/deps.edn. It instead tries to look for a deps.edn in the current directory. Used to work.

Lots of functions take a predicate such as remove. I find myself often writing the following ideom.

(remove (fn [x] x=something) my-sequence)

(remove (eql something) my-sequence)

it's idiomatic to use sets in this case: (remove #{something} sequence)

or you could always use (partial = something) or #(= % something) , no need for a specialized function

ah because a singleton set acts like an equivalence predicate for its singular content. clever

Yup, unless the set contains false or nil. Sets when called as fns return the arg if present and nil otherwise, both are falsy in this case and contains? would be more appropriate.

Does clojure.java.shell work under Windows? Can't seem to make it work.

Once I've written a clojure program, are there any ways to decrease the startup time? It seems rather large. Can I compile to binary or something?

Is there a cousin of map which does not accumulate return values? for example simply returns the sequence it iterated over?

doseq?

in an immutable language that's kinda built in. (do (map f coll) coll)

oh wait you want to return the original seq

but most likely you want (do (doseq ...) coll) since its for side effects

We might be interpreting the question differently. Map returns a lazy sequence. Whether you realize it all, is up to you.

(take 5 (map (partial * 10) (range)))
  ;; => (0 10 20 30 40)

Actually I don't care what the return value is, I just want to call something like map and not accumulate the return values.

(doseq [x items]
   (do-something-with x))

@jimka.issy this is what run! is for

user=> (run! println [:a :b :c])
:a
:b
:c
nil

Can someone please help me understand this error message?

clojure-rte.core> (let [a '((1 2 3) (2 1 3) (1 3 2) (3 1 2) (2 3 1) (3 2 1))
                        b  #{(1 2 3)
                             (1 3 2)
                             (2 1 3)
                             (2 3 1)
                             (3 1 2)
                             (3 2 1)}]
                    (= (set a) b))
Execution error (ClassCastException) at clojure-rte.core/eval20206 (form-init4386036104585295471.clj:15176).
class java.lang.Long cannot be cast to class clojure.lang.IFn (java.lang.Long is in module java.base of loader 'bootstrap'; clojure.lang.IFn is in unnamed module of loader 'app')
clojure-rte.core> 

IFn is the interface for things you can call

(f) in clojure means "call f"

(1) means "call 1"

1 isn't an IFn, you can't call it

Oh, is #{(1 2 3)} trying to call the function 1 ? I just need to quote the items.

or use [] instead of () to get a sequential collection

you can quote the entirety '#{(1 2 3) ...}

or use vectors #{[1 2 3] ...}

both would work here

so [] interprets its contents as literals, and #{} evaluates them.

[] builds a collection and takes no further action

() builds a collection then executes

' prevents the execution step if any, and also prevents resolving symbols to their current value

this might help see some of the differences:

ins)user=> (def a 1)
#'user/a
(ins)user=> (= [a] '(1))
true
(ins)user=> '(a)
(a)
(ins)user=> [a]
[1]

I’m having trouble identifying the source of a bug. Isn’t there a way to pause execution at specific points and evaluate what’s going on in that environment? e.g. in the R language you can use the browser() function to stop evaluation and give you back control at thatpoint

there are "debug repl" implementations that allow that, but in practice since our data types are immutable, it suffices to capture data in scope (a simple def inside your function can do that while debugging)

or the new tap> facility

CIDER has a nice debugger if you are using Emacs.

I use Cursive and so do have the ability to set breakpoints but I actually find myself using https://github.com/dgrnbrg/spyscope more.

nice thanks everyone I’ll check those points

Is there a cousin of let which lets the n'th binding rever to bindings 0 through n-1 ?

In Common Lisp it is called let*

in scheme it is called letseq

I think normal let gives you that ability.

equivalent of:

(let [a 1]
  (let [b (+ 1 a)]
    (let [c (+ a b)] ...

user=> (let [x 5 y (inc x)] y)
6

Yes, this is just normal let in Clojure

Are you sure? experiments show otherwise.

clojure-rte.core> (let [a 100]
                    (let [a a]
                      a))
100
clojure-rte.core> 

ahhh I see, each expression on the right hand side can use any variable above on the left hand side. yes it seems you're right.

So we cant swap a and b with

(let [a b
      b a] ...)

True, but you can do (let [[a b] [b a]])

I think…

Yeah

(let [a 1 b 2
      [a b] [b a]]
  [a b])
=> [2 1]

And apprently I can bind the same value multiple times. That's cool!

(let [a (f)
      a (remove-duplicates a)
      a (sort a)]
  ...)

if its a series of transforms youre doing using one of the threading macros may be useful: (-> (f) remove-duplicates sort) https://clojure.org/guides/threading_macros

is lein install the preferred way to get local dependencies into your classpath in a lein project?

just a pedagogical example. the general case might be more than just a series of transforms of course

If I have a sequence of objects, and from each object I can extract a key and a value, how can I best build the hashmap from that?

clojure-rte.core> (rte-to-dfa '(:cat (:or ::Cat ::Fox) :sigma ::Lion ::Wolf))
([0 [[0 :clojure-rte.core/Fox 1] [0 :clojure-rte.core/Cat 1]]]
 [1 [[1 :sigma 3]]]
 [2 [[2 :clojure-rte.core/Wolf 4]]]
 [3 [[3 :clojure-rte.core/Lion 2]]]
 [4 nil])
clojure-rte.core> 

[ { :clojure-rte.core/Fox 1
                      :clojure-rte.core/Cat 1}
                   {:sigma 3}
                   {:clojure-rte.core/Wolf 4}
                   {:clojure-rte.core/Lion 2}
                   {}]

This represents a finite-automaton. where arr[n] is a map of transitions, the key is the transition label, and the value is another index in the array.

including arr[4] = {} as there are no transitions from state 4

@jimka.issy hmm interesting puzzle, what is rte-to-dfa (..something something finite automata?) and is this the ideal place to do this conversion... if so i can think of a few hacky ways to get that output from that input, but nothing especially elegant comes to mind... maybe someone has better ideas on changing a sequence of vectors into a hash map reliably while dropping a bunch of index vals

People asserts that a Clojure codebase has more parenthesis than those in other languages, is it true though? I think it could be true for the density of parens, but I believe that we have fewer lines and parens in absolute number for the same number of feature than other languages?

@neo2551 We probably have more ( ) in arithmetic expressions but fewer overall than C-family languages.

(and, yes, the productivity factor can mean a substantial reduction in lines and therefore "punctuation" too)

The domain expired for http://www.4clojure.com

Thanks a lot for your answer! @seancorfield

suppose my data structure is a two dimensional vector. is there a more idiomatic approach to getting values from that matrix than (nth (nth...))? i'm seeing clojure style guides advise against index based collection access.

Hey! I while ago someone shared a website that, given the inputs of the function, returned the functions that could be called. It was specific for Clojure, and it was pretty nice because you could find functions you didn't know existed. I'm trying to look for it on Google, but I can't. Does someone have it?

https://borkdude.github.io/re-find.web/

Hello expertes, i'm trying to do some basic java interop but it's not working, what i'm doing wrong here?

(import  java.security.SecureRandom
         java.math.BigInteger)

user> (.probablePrime java.math.BigInteger 1024 (SecureRandom.))
Execution error (IllegalArgumentException) at user/eval14128 (form-init5284286588528896621.clj:155).
No matching method probablePrime found taking 2 args for class java.lang.Class

apparently i'm not casting the method on the right class