We're updating the issue view to help you get more done. 

Namespaced record fields in defrecord

Description

Currently, Clojure records—the preferred Clojure solution to single-dispatch polymorphic data—support only bare, non-namespaced field names. In contrast, the new clojure.spec standard library opinionatedly focuses on fields identified by namespaced keywords.

  • "spec will allow (only) namespace-qualified keywords and symbols to name specs."

  • "Encourage and support attribute-granularity specs of namespaced keyword to value-spec."

  • "People using namespaced keys for their informational maps" is "a practice we'd like to see grow."

The spec guide notes that "unqualified keys can also be used to validate record attributes", using the :req-un and :opt-un options in spec/keys.

In order for records to leverage clojure.spec fully, however, it may be worth somehow adding support namespaced record fields in defrecord.

One example of how this might be done is something like (defrecord Record [x/a y/b] ...). One disadvantage is that it is not clear how to specify that a field belongs to the current namespace. Allowing keywords would allow double-colon :: syntax to be used (defrecord Record [::a] ...), but this may be confusing. (Alternatively, a syntax for namespacing symbols in the current namespace, similarly to double-colon keywords, might instead be implemented, but that would be out of scope of this issue.)

(Also out of scope of this issue, though also related, would be whether CLJ-1910 namespaced maps could somehow be applied to record map literals (e.g., #foo.Record{:a 2}.)

Environment

None

Status

Assignee

Unassigned

Reporter

import

Approval

None

Patch

None

Priority

Major