Operator expressions grammar #141
Conversation
| @@ -1,5 +1,18 @@ | |||
| # Operator expressions | |||
|
|
|||
| > **<sup>Syntax</sup>** | |||
| > _OperatorExpression_ : | |||
There was a problem hiding this comment.
Grouped expression will be located directly on the Expression production. I think it would be out of place if put together with the OperatorExpressions -- even though it is in the same section. I put it in the operator expression section mainly because I think it is too small to be on a separate section by itself.
What do you think?
There was a problem hiding this comment.
I think grammar should be associated with its prose. GroupedExpression is on this page, so it should show up here. If it feels out of place in the grammar, it should feel out of place on the prose and moved to its own page, albeit the shortness is concerning. Although some examples could help with that.
| @@ -36,6 +53,11 @@ assert_eq!(y, 20); | |||
|
|
|||
| ## Borrow operators | |||
|
|
|||
| > **<sup>Syntax</sup>** | |||
| > _BorrowExpression_ : | |||
| > (`&`|`&&`) [_Expression_] | |||
There was a problem hiding this comment.
Because the lexer will give us && when it finds two contiguous ampersands, since && is a valid symbol. The same problem occurs with << and >> in generics/paths, and || in closures. Example:
let & & mut a = & & mut 10;
let && mut a = && mut 10;I'm currently very busy, but I'll put some explanation in the next days, since it isn't really obvious why && is accepted.
brauliobz
force-pushed
the
grammar_operator_expr
branch
from
e5568bb
to
9c4b7a1
Compare
src/expressions/operator-expr.md
Outdated
| @@ -60,8 +82,25 @@ let mut array = [-2, 3, 9]; | |||
| } | |||
| ``` | |||
|
|
|||
| For ease of typing, consecutive borrow expressions can be expressed by the | |||
There was a problem hiding this comment.
Ease of typing compared to what? To most people reading the reference, it would seem obvious that && is allowed and can be chained just like e.g. --1 works (though is useless). Though I'm guessing that the x && y operator giving an && token meaning is to blame here?
There was a problem hiding this comment.
Yeah, indeed the compound && being a single token is to blame here. I've changed the text to only refer to this fact.
|
@brauliobz ping! what's the status? |
|
@mrhota Hey, I'm back! I'm rebasing my branches and after that I'll deal with this PR. |
brauliobz
force-pushed
the
grammar_operator_expr
branch
from
9c4b7a1
to
e79d1ef
Compare
|
💟 Thanks! |
Part of #84.