I ran into surprising behaviour where a symbol does not get metadata when using the ^:some-key syntax

(def map-meta-from-reader ^:foo {})
(def map-meta-explicit (with-meta {} {:foo true}))

(def symbol-meta-from-reader ^:foo 'a)
(def symbol-meta-explicit (with-meta 'a {:foo true}))

(meta symbol-meta-from-reader) ;; => nil
(meta symbol-meta-explicit) ;; => {:foo 1}
(meta map-meta-from-reader) ;; => {:foo true}
(meta map-meta-explicit) ;; => {:foo true}

I would expect that (meta ^:foo 'a) returns {:foo true}, but that is not the case

Is this a known issue?

@chrisblom ^:foo 'a is ^:foo (quote a), the eval of which is a. this is expected behavior. note that '^:foo a does what you're expecting above.

ok, that seems to work indeed

what @arrdem said, note that ^:foo 'a isn't losing any metadata, ,it's just attaching metadata to the (quote a) form which is only accessible at compile time

ow wow, that is new to me, i’ve never seen the quoted metadata reader syntax '^:some-key before

that's not syntax

it's just the composition of ' and ^:

do the expansion in your head

that's equivalent to (quote ^:foo a)

ok, that makes sense

ok, i was trying to use ^:marker some-symbol for a config format

luckily (meta (read-string "^:foo a")) returns the expected result

The library coding standards can be made gender neutral. There are 2 occurrences of he in https://dev.clojure.org/display/community/Library+Coding+Standards : One should only extend a protocol to a type if he owns either the type or the protocol. Replace he with either they, one or she/he, or similar?

@petterik Alex is offline, but I would imagine that any rework on that article would start by republishing it on http://clojure.org.

I think that's a small enough and reasonable enough change that I went ahead and made it.

It's hard to say whether that page belongs on http://clojure.org ...

That’s just one of the confluence pages - anyone can edit them.

And Sean just did it.