(comp forget past)

#bliss

great view

That's really beautiful :melting_face:

Bergen is quite breathtaking too.

wonderful shot. Kudos 🌹

Thx very much 😊

I like the way you made it monochrome. very dramatic

Thanks @U052852ES

It works as implemented 👍

there's also a clj-kondo rule called "used-underscored-binding" which is off by default

why is it off by default? I've never seen it used for realz

because only critical things are on by default (and those that are not critical were on by default in the early days)

ok, no biggie. I think any code review would ❓this code

Task failed successfully

I’ve heard that code uses too many CPU cycles, and this seems like an excellent way to shave off some bytes! 😁

hahahahaha

Out of curisosity, what were you expecting to happen? I see a couple of options • compiler says can’t have multiple params with same name. • compiler chooses to assign to the first paramter • @U064X3EF3 showing up on your door, hit you across the head and says “Don’t do that!”

I thought the semantic of is ignore, so I was assuming that the compiler would tell me that I can’t refer to an ignored value. But of course it’s just a convention. Reflecting on it, I think the best case is that the compiler could reject it on the basis that it can’t disambiguate the . Alex did actually pop up on the feed for apropos after we talked about that :)

There is no semantic. Reused function parameter shadow previous. If you don’t like that, don’t do it. :)

As Stu once said, Clojure is optimized for correct programs ™️

I thought it was a joke but :man-shrugging::skin-tone-3:

People frequently use _ for multiple ignored args - making this an error would break all those programs

I only mean for the use not the [_ _ _]

I think there's two things here:

(defn foo [x x] x)

^^^ permitting the use of _ is odd? Well, imho at least

clojure never had any semantic for _ - it's just a user convention (in many programming languages)

Yeah, it's just a name 🙂

semantics and conventions are cousins ... and sometime their activities are borderline illegal

clojure doesn't prevent you from doing anything illegal, that's why we have linters ;)

In fact, Clojure is very gray about what "illegal" is, in my view! Lots of "interesting" things you can do.

Like ^:private being less of a locked door and more like a "please don't enter" sign

I like that very much about clojure ;)

er, some things are illegal ... otherwise what was the point of spec? It depends on your definition of correct vs legal 🙂

the underscore stuff isn't part of the fn spec, so there's your answer ;)

Sure, sure. Like, you can easily write code that the Reader won't read. In that sense, there's an endless amount of locked doors.

Missing parens and so on

The fun thing is that clojure lets us write some tooling. What else are you going to waste your life on? ;) https://github.com/borkdude/edamame?tab=readme-ov-file#fix-incomplete-expressions

this thread 😛

Some things are invalid, some things are undefined but with observable behaviors. The example thread is not any of those: _ is a valid name, and repeating param names shadows previous (same as let). Everything else flows from that.

btw just found this related issue in the clj-kondo backlog: https://github.com/clj-kondo/clj-kondo/issues/1732

perhaps a new rule :same-named-param or so would be good

(better name suggestions welcome)

fwiw I think returning or otherwise using _ or for that matter _x in the body should be warned cos it's likely a mistake, and kondo could highlight it without disturbing the compiler's flow.

this is already a linting rule called :used-underscored-binding - these are two separate things

I'm just referring to using same named function arguments in general (not related to underscores)

ok ... I understand the distinction, I wasn't aware of the existing rule