Evaluate block does not work for top-most block

[hovsater]

First of all, thanks for all work you've put into Chlorine, it's a wonderful package. ❤️

I'm not sure this is a bug or not. If it's not, feel free to close this issue. 🙂

I thought the difference between Chlorine: Evaluate Block and Chlorine: Evaluate Top Block is that Chlorine: Evaluate Block will evaluate the current form (relative to the cursor position) while Chlorine: Evaluate Top Block would evaluate the top-most form (relative to the cursor position).

(for [x (range 1 10)] (* x 2))  ; Cursor is immediately after `x`.

The above example would evaluate to (1 2 3 4 5 6 7 8 9) using Chlorine: Evaluate Block while it would evaluate to (2 4 6 8 10 12 14 16 18) using Chlorine: Evaluate Top Block. Of course, I could achieve the same thing by moving my cursor to the start of the line and run Chlorine: Evaluate Block, correctly giving me (2 4 6 8 10 12 14 16 18) as well.

This all make sense to me. Now here's a different example.

Thread$State

The above example would evaluate to java.lang.Thread$State using Chlorine: Evaluate Top Block. But Chlorine: Evaluate Block does not evaluate this example at all.

Is that a bug, or am I just misunderstanding how these features work?

[mauricioszabo]

It's not a bug, nor a misunderstanding - it's what we call a "leaking abstraction" 🤣

In the beginning, both "evaluate top block" and "evaluate block" did not evaluate anything - because it's not a "block", really, just the "current var". But then I changed the way I detect top-blocks so I could re-use the code on VSCode.... and then things became weird. The current approach does detect top-level variables as "top-blocks", and also sometimes detects more things (like @(:some state) will include the @, and #_(+ 1 2) will evaluate the block - this is intentional - but will not evaluate if there are multiple #_ before the top-block).

I could add support for Thread$State, but then I have to check for (:something other-var) to eval the "block", not the "current var". I'm working on a migration to Pathom to solve things for me more easily and then I'll be able to track these inconsistencies (by using Pathom I don't have to re-parse the contents multiple times to detect blocks/top-blocks/vars/namespaces/etc - on some big files like the clojure/core.clj this is a big performance hit).

I will probably try to get the current "block", if it it does not exist, it'll get the current "var". What do you think?

[hovsater]

Thanks again for an awesome explanation about the intricacies of how Chlorine work under the hood. I think what you're suggesting makes sense. I don't think it's worth pursuing a solution to this until you've started the migration. While I haven't used Pathom myself, I can see the benefit of being able to express the parsing more fluently instead of having to do multiple passes over the result.

While I'm still learning Clojure myself (day 3 to be exact 😄), I'd be happy to contribute in financial terms if needed. Do you have any preferred medium for financial contributions?