Add a @cubing/icons/js entry point, transpiled via tsc

[lgarron]

For a discussion of tradeoffs, see: #146

[jfly]

LGTM.

Not a blocking comment, just a thought: is there some sort of "end to end" testing we could/should be doing to detect issues like this before our users are affected?

[lgarron]

Not a blocking comment, just a thought: is there some sort of "end to end" testing we could/should be doing to detect issues like this before our users are affected?

I've tried to do some of this in the past, and I'm currently on the side of "this isn't worth the maintenance tradeoff". There are a lot of caveats and complications due to ecosystem "quirks".

In any case, there was no regression here, this PR isn't so much a bug as an enhancement. Since it's using raw tsc and uses both exports and legacy path-based index files, it should be as cross-compatible as possible.

A "not quite end-to-end" test to run the built output might be worthwhile, lemme know if you wanna talk through some options.

[jfly]

In any case, there was no regression here, this PR isn't so much a bug as an enhancement.

Maybe I'm confused about what we're doing here. Didn't the 3.0 release of cubing icons break things for @dmint789? If so, that's a regression.

[lgarron]

In any case, there was no regression here, this PR isn't so much a bug as an enhancement.

Maybe I'm confused about what we're doing here. Didn't the 3.0 release of cubing icons break things for @dmint789? If so, that's a regression.

Ah, nah. @cubing/icons/ts was a new entry point in v3.0.0. I was hoping it would be enough to address #119 for @dmint789, but it wasn't enough.