Decide whether to keep pointer sized integers as the default

[brson]

I know there are other issues related to this, but I want to make a firm decision one way or the other on whether we want to commit to int and uint being pointer sized. Whatever the decision add it to the design FAQ (#4047). Nominating.

[huonw]

I don't see how this is different to #6023 (removing the default to int) and #9940 (changing the int name).

Also, if we keep a type called int, and it's non-pointer-sized, then it's either fixed size (i.e. it's i32 or something, in which case there is no point in keeping it) or it's arbitrary precision (but then I'd argue that it should be called Int).

[brson]

They are both related, and motivated by the notion that pointer sized ints are wrong: #6023 just suggests removing inference and #9940 is a proposal to rename int and uint. There other potential solutions if int and uint should not be pointer sized.

[pnkfelix]

Accepted for 1.0, P-backcompat-lang

[nikomatsakis]

A summary of what I see as the related issues in this discussion:

• what does uint mean, if anything
  • nothing, just use uintptr
  • always use 32 bit
  • use bigint
• do we want a default fallback when we can't infer a specific integer type
• what types do we accept for indexing and shifting (currently we're kind of permissive)

[brson]

Nominating for closing, along with #6023 and #9940. Unless we do this right now, I don't see it happening.

[thestinger]

I think it's fine to have int and uint as pointer-size integers, and the bug I opened about it was not much more than a bikeshed. I do think it's really important to remove the fallback to int to prevent bugs caused by relying on type inference.

[pnkfelix]

closing.