Clojurians Log v2

To expound a bit, let’s say I’m trying to match a map pattern:

(let [expr {:present "yes"
              "notkeyword" "missing"}]
    (m/find expr
      {:present "yes"
       :missing "no"
       (m/pred keyword? ?k) "missing"}
      true
      _
      false))

{(m/pred keyword? ?k) "missing"
   :missing "no"}

For now, I’m mapping each mapentry of the map pattern into individual “patterns”, and then seeing which work with m/find, and which don’t. If they don’t, I merge them back into a pattern to display the “missing” map entries.

Yeah, that would definitely be cool to have. Not really sure we can make something like that happen given the current setup. We've talked about build some separate explainer for why a pattern didn't match.

👍

I just wanted to make sure there wasn’t some idiomatic obvious thing I was missing 🙂

The error refers to the fact that once you have !v once ?v

probably

Thanks, I thought meander kept those separate but I'll give it a try

(let [m {:a "a", :b "b'"} v [:a :b :c]]
  (m/rewrite [m v]
    [?m (m/gather (m/pred #(contains? ?m %) !xs))]
    [!xs ...]))
;; => [:a :b]

it works ; )

😄 thanks!

np

btw, I really like what you're doing with doxa. I have been working through https://github.com/fulcro-community/fulcro-exercises and doxa makes handling the normalized database much nicer

thx!

I'm quietly working on a major update, but I have many conceptual dilemmas

Such as? (I don't mean to pry but I'm very curious)

I don't know how much compatibility with datascript I should keep

I can only speak for myself: very little 🙂

the current use of vector as ident, as in fulcro, has many advantages but is not without disadvantages

I'm exclusively interested in doxa as a convenience for fulcro. I'm sure there are other use cases but, again, I can only speak for myself

I don't really need it either, but a couple of people have asked about schema, or the possibility that q is not a macro and can take a query on runetime

I can see the value of taking a query at runtime.

It seems to me that schema can (and should?) be handled as a layer above doxa. seems pretty straightforward given that with is available

you know, you speak for yourself, but all the people who use doxa can sit at the same table

🙂

Is there a #doxa channel?

no, not yet

maybe it will happen if there is a need and doxa has an official release

for now, it's just thinktering

sounds good to me. in the meantime, I'll keep enjoying it and looking forward to your new release

https://github.com/ribelo/doxa/tree/memoized-query

for now, everything is happening in a separate branch

If you have any questions, have any ideas or find a bug, please don't hesitate to tell me

it is much better to work knowing that someone is watching

🙂

Yes, I agree and I'm happy to provide feedback

🙃

regarding the meander example you provided, what build of meander are you using? I tried your solution in 602 and 643. It didn't work in either

huh

{:mvn/version "0.0.643"}

clj or cljs?

shouldn't matter, but maybe it does

clj

Double checking that this is what works for you:

(let [m {:a "a", :b "b'"} v [:a :b :c]]
    (m/rewrite [m v]
      [?m (m/gather (m/pred #(contains? ?m %) !xs))]
      [!xs ...]))

Ah, it works in 643 but not in 602

I screwed up earlier when testing against 643

😉

The worst thing about the meander is that it's a moving target

True but it's so nice that these minor hiccups are worth it 🙂

noprompt is doing a gigantic job on its own and it is so complicated that it is difficult to help

definitely

I currently use meander for anything more complicated than get/get-in