ternary requires even more space now

[JayXon]

a ?b :c used to work, with #1498, it no longer parse, have to do a ?b : c now

[bbrk24]

Is this necessarily a bug? I guess it is technically a regression but I don't know that it was ever officially supported.

[JayXon]

it's documented here

Civet/civet.dev/comparison.md

Line 108 in bec932c

* Ternary operator `x ?y :z` needs space before both `?` and `:`

[STRd6]

We could add a compiler flag to disable symbol shorthand.

[JayXon]

The example I gave is not ambiguous because it currently doesn't parse, and there's really only one way to parse it.

But if we support parsing that, then we need to decide whether a ?b :c :d means a ? b(Symbol.for("c")) : d; or a ? b : c(Symbol.for("d"));

[edemaine]

Thanks for finding this!

Given the greedy nature of the parser, it's pretty annoying to fix this, unless we

• add another flag, something like "don't process symbols because I'm waiting for a colon"
• forbid symbols to start implicit function calls (maybe it's rare we want to pass symbols into functions? probably not)
• forbid implicit function calls in ternary (seems u likely)
• define this to be the new behavior and a breaking change, in which case we'd want a compiler flag to disable symbols

Probably worth seeing how much JS code has a space before but not after a colon in a ternary.

[STRd6]

Another thing we could try is maybe add snug ternaries as well like a?b:c

But it may require another parser flag to correctly handle complex cases even still.

[JayXon]

by parser flag, do you mean like a config or parser internal state?

[edemaine]

• add another flag, something like "don't process symbols because I'm waiting for a colon"

In this case, I meant "parser internal state". Sorry for the ambiguity.

It sounds like there might be interest in pursuing this. Snug ternaries would be nice to allow...