Clojure Q&A - Recent questions and answers in Multimethods

Answered: fn? and multimethods

Mon, 07 Oct 2024 22:44:06 +0000

Almost always, you really want ifn? - is that what you want here?

Answered: Please add dispatch-fn to clojure.core

Mon, 29 Nov 2021 14:51:22 +0000

What is the use case for this in ClojureScript and/or why do you want it in Clojure?

Answered: Memory leak using the default method of a multimethod

Wed, 28 Apr 2021 14:36:47 +0000

I don't know if this is a leak so much as a design assumption that the number of dispatch values is bounded.

At one time we did not cache the dispatch values resulting in :default case and this made the :default very slow (this was the primary reason multimethods were considered "slow" for a long time). In 1.8, http://dev.clojure.org/jira/browse/CLJ-1429 changed this and it was a substantial performance boost for a lot of real programs.

Maybe there is some middle ground, not sure.

Answered: Self-Like Pattern Using Multi-Methods

Mon, 12 Apr 2021 23:31:12 +0000

You could also look at building something on top of clojure.core.match. Do not feel boxed-in by multimethods! Or you might be able to get closer to the ideal by writing your own macros to do the dispatching. One thing to watch out for, using Clojure's built-in multimethods dispatching by keys, is that message maps must not have any "extra" keys. For the most part, Clojure patterns embrace the idea that you can stuff extra keys into a map without breaking contracts.

Answered: Best way to test for an integer (including java.lang.Integer) in a multimethod dispatch?

Mon, 28 Dec 2020 02:08:59 +0000

You can give Long and Integer a common ancestor in a custom hierarchy that you tie to the multimethod.

(def h (-> (make-hierarchy)
           (derive Long :num)
           (derive Integer :num)))
(defmulti yy type :hierarchy #'h)
(defmethod yy :num [x] (format "%o is a num" x))
(defmethod yy :default [x] "I don't know")
(yy 7)
(yy (Integer. 42))

Answered: Multimethods resolutions when they have the same dispatch value

Wed, 02 Dec 2020 10:53:59 +0000

Each "defmulti" knows at-most-one method with any particular dispatch value. If a second "defmethod" comes along, it replaces the first. This is a cornerstone of REPL-based development.

On the other hand, there is some interesting work behind-the-scenes to dispatch a multimethod call in cases when the dispatch function returns a value that is related (by the relevant hierarchy) to more than one method's dispatch value. In that case, see prefer-method.

Answered: Implementing a type or record to support defmethod in CLJ

Fri, 17 Jan 2020 19:50:07 +0000

No, this is not a feature I would expect to see in Clojure. I don't understand why it's even useful in ClojureScript?

Answered: behavior of deftests after adding assert-expr methods

Tue, 24 Sep 2019 01:14:42 +0000

Since is is a macro that uses the assert-expr multi-method to expand the code at compile time, changing assert-expr after running deftest will not change the behavior of that test function since it has already been expanded and compiled.

That is why you don't see the new assert-expr functionality until you define new tests (or redefine existing ones).

Answered: MultiFn.prefers() ignores the multimethod's internal hierarchy

Wed, 26 Jun 2019 12:00:00 +0000

Reference: https://clojure.atlassian.net/browse/CLJ-2234 (reported by palisades-lakes)

Answered: Supporting `add-method` on multi fns to install a function instead of a function tail

Wed, 26 Jun 2019 12:00:00 +0000

Reference: https://clojure.atlassian.net/browse/CLJ-2514 (reported by admin)