Hey! I’m reading Programming Clojure and there’s this snippet: “Clojure is high signal, low noise…” I’m having a hard time to understand this. Where can I be familiar with these terms? What’s the definition of a high signal and low noise language?

It is a figure of speech meaning the amount of code it takes to perform some action is very small with very little outside what is minimally required

It is an analogy to https://en.m.wikipedia.org/wiki/Signal-to-noise_ratio where for example in Java to print something out, the "signal" might be System.out.println("foo") and the "noise" would be the class and method around it

Oh, I get it. Thanks a lot @hiredman

anyone have an upsert function for mysql on their shelf by any chance?

@mzavarella It's going to depend on how you identify your primary key, but what I've seen is to run an update! based on the PK and if that returns zero rows updated, do an insert! with the PK assoc'd into the data -- but you really need to wrap the insert! in try/`catch` in case another thread also attempts the insert (and this is assuming you have a unique constraint on your PK so a second insert will fail), and if the insert fails, because of that constraint, you need to retry the upsert.

awesome, thanks Sean!

Didn't even think to use try/`catch`

Reading around, it looks like you might be able to just do INSERT ... ON DUPLICATE KEY UPDATE but I'm not sure how that would work with auto-generated PKs...

See https://www.xaprb.com/blog/2006/02/21/flexible-insert-and-update-in-mysql/ for more options.

(non-standard SQL caveats, not available in all versions of MySQL caveats, etc, etc)

If you have non-primary (keys) only, you can still do the update-or-insert approach (assuming auto-gen'd PK).

Anyone played with claypoole? I have a pool of processes doing http document uploads. They're all using a token stored in an atom, when the token expires I'd really like one Thread to know that it has the job of updating the token, I guess it won't do any really harm for them all to update the atom but it just doesn't seem right, should I bracket the token check/update in some other way?

@zugnush Perhaps you could push the check/update into an agent so that piece of work would be single threaded?

thanks Sean, I think might be what I need, I had it in my head that i needed to check the thread id somehow but that sounds much better. Ta. Loving jdbc-next btw.

Cool. Best channel to ask Qs about next.jdbc is #sql to make sure I see it. Always happy to hear feedback on it (and on its docs).

Quick spec question…

(require '[clojure.spec.alpha :as s])
(require '[clojure.test.check.generators :as gen])
(s/def ::non-map #(not (map? %)))
(gen/generate (s/gen ::non-map))

1. Unhandled clojure.lang.ExceptionInfo
   Spec assertion failed.

         Spec: nil
        Value: nil

     Problems: 

                 alpha.clj:  282  clojure.spec.alpha/gensub

I am taking a stab in the dark here, but I suspect that the built in generators do not handle the case where you have a function you wrote (`#(not (map? %))` in this case) as the spec predicate. Probably you would need to write your own generator function for such a spec.

I think that is correct: (s/gen #(not (map? %))) says that its unable to construct a generator 🙂

@alexmiller FYI, I find a tool who already do the job https://github.com/hagmonk/find-deps

Hi, I learned some (very little) clojure before, and am in the progress of relearning it. I was comparing its expressiveness with some other languages (python and go), but so far I don't find it that more expressive as people claimed. Let's take 4clojure problem 41 for example:

;; Drop Every Nth Item - Easy
;; Write a function which drops every Nth item from a sequence.

Here is my clojure version

(defn to-vector [& rest]
  rest)

(defn keep-index? [n [index _]]
  (not
   (zero?
    (mod (inc index) n))))

(defn my-drop-every [coll n]
  ;; Use transduers to transform the collection
  (into []
        (comp (map-indexed to-vector)
              (filter (partial keep-index? n))
              (map second))
        coll
        ))

but it's much longer than python version:

def my_drop_every(coll, n):
    return [x for idx, x in enumerate(coll)
        if (idx + 1) % n != 0]

and even the go version is not so bad

func MyDropEvery(coll []int, n int) []int {
	newcoll := make([]int, len(coll))
	for index, i := range(coll) {
		if index % n != 0 {
			newcoll = append(newcoll, i)
		}
	}
	return newcoll;
}

am I not getting the idiomatic way to do it in clojure? Thx!

(for [[i el] (map list (rest (range)) cool) :when (not (zero? (mod i n)))] el)

also possible, (apply concat (partition-all (dec n) n coll))

(defn drop-every [n coll]
  (mapcat (fn [x i]
            (when (not= (mod x n) 0)
              [x]))
          coll
          (range)))

the partition version is neat, both are idiomatic approaches imo

also possible, (keep-indexed (fn [i x] (when-not (zero? (mod (inc i) n)) x)) coll)

One thing I’d say is that for this type of thing, there’s not really that much to choose Clojure over any other language. The python solution here is definitely more readable than most of our suggestions. I think the advantages of Clojure imo becomes more apparent once you get into exploring and working with data from databases, apis etc. When most things are data, and all libraries are built around the assumption that you’ll be working with data.

I think the keep-index version by @michael.gaare is on the same level as the python version. > I think the advantages of Clojure imo becomes more apparent once you get into exploring and working with data from databases, apis etc. When most things are data, and all libraries are built around the assumption that you’ll be working with data. That's the part I don't quite get yet. I heard so many of people saying things like this on reddit or in the clojure conference videos. Can you give me a concrete example?

@danieleneal ☝️

Hey all, new to clojure, coming from node 👋. I am trying to use vscode with some goodies (autocompletion, linting, etc.) and then run that file with repl. On each save, re-run this file. This way I get quick feedback loop and can work in familiar environment instead of in terminal. Does this make sense? If not, what are some alternative methods?

Have you heard of REPL-Driven Development (RDD)? https://github.com/stuarthalloway/presentations/wiki/REPL-Driven-Development

@manutter51 yes, I have, but I would like to read more about this. Being that I have several layers of new (syntax, language, environment, etc.) trying to have some familiarity so I can make progress learning the language

There’s a link to a good video on that github page, worth a look if you haven’t seen it before.

I suggest using Calva (since I am maintaining that and since you use vscode) and use it's in-editor REPL for getting quick feedback on your code. So, write code inside (comment ...) forms and use Calva's commands for evaluating current form and current top level form to examine the results of stuff inside those comment forms. It's in the Calva README: Something to try first.

Also if you want to get the feel of the repl, you can check out (night-code)[https://sekao.net/nightcode/], but I would just go with calva instead

I love nightcode! It is perfect for a beginner looking to toy around with Clojure. The Insta-REPL feature is awesome and the lack of functionality minimizes distractions with keyboard shortcuts. If nightcode had git functionality, similar to VS Code, and a shortcut to comment out lines of text, I’d probably use it exclusively. Since it doesn’t, I’m using Calva on VS Code for my main text editor.

@here i am having 2 years of experience in clojure & clojurescript, can anyone help me in getting remote work or full time work?

I’m putting together a simple use case in a demo web app I’m using to learn clojure. I have a req handler for account registration, when this handler runs I want to insert a new row in the database and then send a welcome email. So far I have this https://gist.github.com/hendore/2d8686fddcec2342688372b300d811e6#file-registration-clj however it’s left me with so many questions, I’m certain there are obvious answers to those that have had more exposure to clojure but I feel a bit lost on where to go next to solve a few things. 1: If the send-welcome-email! function was implemented and actually sent an email, I wouldn’t want this to happen before returning a response, in other languages I would dispatch an event / add a job to a queue to run later to send the email, what is the clojure way for doing this sort of thing? 2: It’s possible that the insert-account! function will throw PSQLException what’s the standard practice for handling this and returning a nicer error back to the request handler function to be given to the user, try/catch maybe? If so where would be the best place to catch the exception and how would I handle that in the request handler to return a nice message? 3: I would like to hash the plain text password before inserting the account, it makes sense for me to do that in insert-account! should I use something like the threading macro to assoc or update password in the account map with a hashed password

(let [account (-> account-info
                  (select-keys [:name :email :phone :password])
                  (update :password hash-plaintext-password))]
  (…))

There's a lot to unpack there and the answers to most of those questions is "it depends".

Totally

just looking for some pointers / links to go do some further investigation more than a step by step, but I understand there’s a lot being asked here 🙈

1. You want send-welcome-email! to run after successful completion of the registration "work". It doesn't really matter if it happens before the handler returns or not, as long as it happens after successful registration. I would probably have something like

(when success ; whatever that is
  (future (send-welcome-email! ,,,)))

But you could do it via an agent or a queue or whatever you're used to.

Perfect, just brought up the future and agent docs this is great :thumbsup:

2. Yes, try/`catch` is the right choice here -- exceptions are idiomatic in Clojure because they are idiomatic in the host language.

agents look very similar to atoms :thinking_face:

As for the try/catch would it be best to place that in the function that could potentially throw the exception insert-account! and handle it there, maybe return [:error "Email is already in use"] or would you let that exception bubble up and handle it in the register-account! or registration-req-handler functions?

3. looks fine (but you probably want to seed the hash, perhaps with the email or something)

4. There are lots of options for validation. Too complicated for me to type out right now.

No worries, thanks again for your input, you’ve been really helpful the last couple of days I really appreciate the time you’ve taken to help

Re: returning errors -- consider keywords at the lower level (`:ok`, :email-in-use) and turn them into user-facing strings closer to the user.

Ah, that makes better sense

Is this the idomatic way in clojure for returning success / errors ?

It depends 🙂

😄

How much information do you need to convey to callers

You can use keywords, or tuples, or maps, depending on how much information you need to convey back to the caller.