RFC: Add isa as an infix operator

[ararslan]

This implements @yuyichao's suggestion in #19524 of an infix operator version of isa. Thus, with this PR, we have

julia> 1 isa Integer
true

[ararslan]

Any comments? Is this something we want to have?

[quinnj]

I'm a fan 👍

[nalimilan]

I always fail to remember the order of arguments to isa; having it infix would fix that.

[StefanKarpinski]

If this gets the @JeffBezanson go-ahead, we should probably change the usages as well.

[JeffBezanson]

I won't object --- but it's starting to seem like only a matter of time before we allow infix for any function.

[ararslan]

it's starting to seem like only a matter of time before we allow infix for any function.

Actually I was initially against this change for exactly that reason. But upon further reflection, this just seems to me like a natural infix operator, like <: except for values instead of types.

[JeffBezanson]

One observation is that both in and isa should obviously have comparison precedence, but for arbitrary functions the precedence is less clear.

[ararslan]

Would this give you any less pause if there were some kind of symbolic infix operator that lowered to isa rather than having isa itself be infix?

[JeffBezanson]

No, this is fine. I can't think of an operator that would improve on just using isa.

[ararslan]

Okay, cool. Thanks everyone!

[ararslan]

If this is good to merge, does someone want to do the deed?

[oscardssmith]

Can we also add isan in case the type starts with a vowel?

[JeffBezanson]

No. English is too weird :)

[tkelman]

This should go into news since it can change how existing code is parsed

[ararslan]

Okay, I can submit a PR with a news item.

Edit: Done.

[ranjanan]

@tkelman are you planning to include this in 0.6.0-rc2? It's not included in 0.6.0-rc1

[tkelman]

merged in december, what makes you think it isn't in rc1?

[ranjanan]

My bad, I was using 0.5 instead. Forget I asked.