_{tachometer-reporter-action v2 for Benchmarks}

Size Change: +255 B (+0.41%)

Total Size: 62.2 kB

_{compressed-size-action}

coverage: 99.611%. remained the same
when pulling 17fee02 on feat/prop-attr-directives
into 4c20c23 on main.

coverage: 99.611%. remained the same
when pulling 17fee02 on feat/prop-attr-directives
into 4c20c23 on main.

I'll mark this as "ready", though I'm happy to still discuss whether we even want to do this and how it should interact with other things like RTS.

IMO, it is a nice bit of power to give users though it is a bit of a deviation and might confuse some.

Edit: What in the world is up with Coveralls

coverage: 99.611%. remained the same
when pulling 806e605 on feat/prop-attr-directives
into 16aeb9f on main.

coverage: 99.611%. remained the same
when pulling 806e605 on feat/prop-attr-directives
into 16aeb9f on main.

coverage: 99.612%. remained the same
when pulling a9bb51f on feat/prop-attr-directives
into f7f9d9b on main.

@rschristian I like this direction, but I would love to avoid using namespaces for the prefixes. I believe vue uses something like <foo .value="bar"> for properties and <foo @value="bar"> for attributes. JSX doesn't support that though.

Interestingly, JSX does allow for two suffixes: $ and -. The - suffix makes a decent fit for attributes because - at the end of an identifier is invalid so it's an uncommon property name (obj['foo-']):

<div prop$="foo" />;
<div attr-="bar" />;

The other tidbit I wanted to add to this discussion is that we do technically already have a "force to use attribute" syntax, which is to uppercase the name (<div ATTR="bar" />). Unless a Custom Element happens to have an all-upper-case property (which would be extremely weird), it'll always hit setAttribute('NAME') which is case-insensitive and results in attributeChangedCallback('name').

The other tidbit I wanted to add to this discussion is that we do technically already have a "force to use attribute" syntax, which is to uppercase the name (<div ATTR="bar" />).

I'm fully aware, but it's a hard thing to push users toward; as stable as it is, it's an obvious hack that most raise an eyebrow at when it's suggested.

There's a few threads recently in particular w/ users uneasy after I've suggested using capitalization to get around it, hence, this PR.

It's a shame directives require some opt-in, I don't love how unclear $ & - as suffixes are comparatively. Not as easy to spot either.