posts/a-story-of-regret-and-retiring-a-library-from-hex/

[giscus[bot]]

posts/a-story-of-regret-and-retiring-a-library-from-hex/

I'm retiring a library of mine from Hex. I want to explain why.

https://andrealeopardi.com/posts/a-story-of-regret-and-retiring-a-library-from-hex/

[pmarreck]

Just discovering this after Gleam implemented this feature (and being aware of its existence in ES6, Nix, Go, Python, Rust, etc.). Much of your arguments here seem to be more against the use of the sigil "workaround" than against any syntax change itself. I also don't see why it would be necessary to support string keys, since keyword lists (at least in their usual form where all first tuple values are atoms) and map.key forms already don't support string keys. Lastly, since this shorthand seems to be growing in popularity in other languages, I have to wonder if the idea was simply too new at the time for most developer brains to easily process/model what was actually going on, leading to the feeling of "too much magic" (i.e., the "mental parsing" cost was too high/potentially error-prone).

Also just noticed that https://github.com/kzemek/es6_maps exists, gets you exactly what the ES6 version does (with the caveat that you cannot upgrade Elixir without a compatible version of this... at least they included formatters to go to and from that syntax) and was pushed 3 months ago...

[cgrothaus]

FTR: there is even another library aiming at the same goal - https://github.com/andrewtimberlake/shorthand.