Rich Harris

I think that must be a Java-only option, it doesn't seem to have any effect on the JS output. Perhaps we should try and benchmark the Java version as well...

Wow. I'd assumed that since closure-compiler-js was compiled from the Java source, it would yield the same results (compression-wise, not perf-wise). Needs a 'don't actually use this!' notice on the...

Ha, yep — GNU C++ Compiler !== Google Closure Compiler 😆

Thinking about this, the low-hanging fruit would be any ExpressionStatement nodes whose expression is anything except a CallExpression, NewExpression, AssignmentExpression or UpdateExpression... does that sound about right? There's also assignments...

Incidentally, disabling mangling for classes and functions is relatively easy — just need to change [this function](https://github.com/Rich-Harris/butternut/blob/61283e96d58827bcef355a1772666d5956cb9793/src/program/Scope.js#L173-L184): ```diff Object.keys( this.declarations ).forEach( name => { const declaration = this.declarations[ name ];...

On the flip side, when mangling *is* enabled, we should remove function names for all FunctionExpression nodes whose names are not referenced inside the function (which should be easy to...

(one fairly major wrinkle that was just pointed out to me — maintaining full opacity for the bottom element doesn't really work if the top element's non-opaque areas don't correspond...

Came here via https://github.com/sveltejs/kit/issues/5215 — thanks for all the research everyone. Long term the solution is definitely to switch to the Web Animations API, but I wonder if we can't...

Was pointed to this set of checks as well, which may contain some additional stuff we can implement https://github.com/dequelabs/axe-core/blob/develop/doc/rule-descriptions.md

I've transferred this to the Svelte repo, as that's where the bug lies (rather than in SvelteKit). Note that the repro can be simplified — it's not necessary to have...