Calling meta on records should exclude preserved own metadata?

[athos]

version

The current head of the master branch (207e994)

platform

Clojure CLI version 1.11.1.1165
macOS 12.6

problem & repro

When calling the meta function on a record object, you'll get metadata with the record's preserved own metadata merged:

(require '[sci.core :as sci])
(sci/eval-string "(defrecord R []) (def r (->R)) (meta (with-meta (->R) {:val 42}))")
;=> {:val 42,
;    :type user.R,
;    :sci.impl/record true,
;    :sci.impl/constructor nil,
;    :sci.impl/var nil,
;    :sci.impl.record/map-constructor nil}

Is this the intended behavior or one of the limitations in the current record implementation?

expected behavior

Expected (meta (with-meta x m)) = m to hold for any valid x and m.

[borkdude]

Previously this was because record was just a regular map and metadata indicated that it was considered a record in SCI, but this implementation has changed and we may be able to get rid of this.

[athos]

Yeah, I was somewhat guessing the current situation from reading this TODO.

[borkdude]

How critical is this issue for you?

[athos]

Not so critical.
I’m now working on some of my libraries to make them support Babashka and along the way I found some tests (e.g. around here) were failing on Babashka. But some tweaks on the test code will fix the failures, so it’s not a big deal about this case.

[borkdude]

The way I usually write tests like that (especially in clj-kondo test cases) is to assert that the result has at least the expected keys, but can contain more, so that when adding new keys, existing tests don't break. But I will leave this open and see when it can be fixed eventually.

[borkdude]

Should be fixed now. Fix also pushed to babashka.

[athos]

Great! Thanks!!
I'm amazed every time I see how productive you are 😳

[borkdude]

Thanks for sponsoring :)

On Sun, 23 Oct 2022 at 15:38, Shogo Ohta ***@***.***> wrote: Great! Thanks!! I'm amazed every time I see how productive you are 😳 — Reply to this email directly, view it on GitHub <#819 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACFSBVOZTGOIQY6IZBQM5TWEU5VTANCNFSM6AAAAAARLF47T4> . You are receiving this because you modified the open/close state.Message ID: ***@***.***>

-- https://www.michielborkent.nl https://www.eetvoorjeleven.nu