struct and union declarations

[nrc]

I'd like to discuss generics separately (#29). I'd like to include unit and tuple structs here.

[nrc]

Single-line tuple structs are already covered in the guide: https://github.com/rust-lang-nursery/fmt-rfcs/blob/master/guide/guide.md#tuples-and-tuple-structs

[strega-nil]

I prefer:

struct Foo /* <- must be on one line */ { /* <- on the first line if it fits, otherwise pull it down */
  b: B, /* <- indented once, must end with a trailing comma */
  long_name:
    LongName, /* <- if required, pull the type name down to the next line and indent again */
} /* <- on it's own line */

[nrc]

And to clarify - a trailing comma on the last field?

[joshtriplett]

@nrc Yes. That avoids having to add a comma when adding another field, or delete a comma when deleting a field.

[bluss]

Edit: Where clauses are a separate issue, see #38

---

What about structs with trait bounds?

struct Peekable<I>
    where I: Iterator,
          I::Item: Debug, // This is only to illustrate more bounds
{
    peek: Option<I::Item>,
    iter: I,
}

• If the where clause is treated as a list, it too should always have a trailing comma; new additions go on a new line.

[Screwtapello]

I'm not a fan of visual identing, and it seems like the tone in other issues is against it. I'd rather have:

struct Peekable<I> where // 'where' on the same line, following the example of '{'
    I: Iterator,         // one block indent, just like struct fields or any other list item
    I::Item: Debug,      // trailing comma, of course
{ // visually ends the 'where' items and begins the struct items
    peek: Option<I::Item>,
    iter: I,
}

[bluss]

A hanging where works well for functions and methods too (they often have longer initial lines), so taken together I prefer it for all of them for consistency. A where at the end of a line can be far removed from the actual where clauses, so I don't like reading it that way. It's also been the prevailing style inside the rust project itself, what I can see.

[nrc]

I'd like to also discuss where clauses separately, since they equally affect functions, traits etc., I created #38 for that.

[nrc]

I think this is ready for FCP, summarised in #30 (comment).

[johnthagen]

@nrc I would like to take a crack at this one.

[nrc]

assigning to myself for mentoring and so we can see @johnthagen is working on it

[nrc]

Closed by #53