standardize on -Vprint:... (still support -Xprint:... as alias)

[SethTisue]

(I think we should try to make progress on observing the -V, -W, -X, -Y distinctions internally and in docs, etc)

(prompted by participating in the compiler spree today and noticing that the contributing guide was still recommending -Xprint)

[soronpo]

I don't know if it was ever discussed before, but "-Y" and "-V" look too similar and it's always confusing to me.

[som-snytt]

It's -V for "verbose" to aid discoverability. Similarly -W for "warning".

Nominally, -X options may change behavior, such as -Xmax-inlines or -Xkind-projector. -Y for "private" also represented "forking" behaviors (now "experimental" as a feature), with the mnemonic that it looks like a fork in the road.

There are fine distinctions: -Xmigration could be -Wmigration, for example. I liked -Xlint for legacy reasons (and the pun).

[soronpo]

-Y for "private"

I wish this was -P, so it will look different

[SethTisue]

this ship has sailed, I think

fwiw, I have not personally ever noticed any difficulty distinguishing V from Y

[som-snytt]

-P is for plugin.

But it's not too late to take this back to the committee on world alphabets. So long as we're working on upsilon or the Greek y, let's clean up eta and epsilon. Really no reason to have two symbols when a diacritic will do. We don't blame the ancient Greeks enough. Also democracy, they could have nailed that down better.

[som-snytt]

I'm reminded that there are usability improvements at #17563

where -Vprint defaults to typer as in Scala 2, and -Vphases:elimByName highlights the phase in the huge list.

Those features are useful to the casual user or contributor, not necessarily the PhD wrapping up their research. So I agree that the sense of what is useful or confusing depends on context.

(Also -Z was reserved for options to the typelevel fork, or possibly the zio. Missed opportunity!)

[som-snytt]

standardise