Add macro declaration syntax.

[gilbens-starkware]

No description provided.

[reviewable-StarkWare]

This change is 

[orizi]

Reviewed 8 of 14 files at r1, 1 of 2 files at r2, all commit messages.
Reviewable status: 9 of 15 files reviewed, 2 unresolved discussions (waiting on @gilbens-starkware)

---

crates/cairo-lang-parser/src/parser.rs line 1487 at r2 (raw file):

            SyntaxKind::TerminalDiv => self.take::<TerminalDiv>().into(),
            SyntaxKind::TerminalDivEq => self.take::<TerminalDivEq>().into(),
            SyntaxKind::TerminalDollar => self.take::<TerminalDollar>().into(),

should dollar be not fully part of the identifier?

Code quote:

            SyntaxKind::TerminalDollar => self.take::<TerminalDollar>().into(),

---

crates/cairo-lang-parser/src/parser.rs line 1861 at r2 (raw file):

    /// Parses an expr block, while allowing placeholder expressions. Restores the previous
    /// placeholder expression setting after parsing.
    fn parse_block_with_placeholders(&mut self) -> ExprBlockGreen {

doesn't this have exactly a single caller?

Code quote:

 parse_block_with_placeholders(

[gilbens-starkware]

Reviewable status: 9 of 15 files reviewed, 2 unresolved discussions (waiting on @orizi)

---

crates/cairo-lang-parser/src/parser.rs line 1487 at r2 (raw file):

Previously, orizi wrote…

should dollar be not fully part of the identifier?

You mean removing dollar as a token, and allowing it as an identifier character (only in the first position) in the lexer level?
That was my first intention, but then how is $false is lexed? (or any other keyword). I thought of two options, either as an identifier, and we'll need to check for keywords in another place, or as $ and false (similar to what I did, but with $ as an identifier I guess), but I thought it will only add more edge cases, so I opted to always split them.

---

crates/cairo-lang-parser/src/parser.rs line 1861 at r2 (raw file):

Previously, orizi wrote…

doesn't this have exactly a single caller?

There is, however, it felt safer to wrap it in a function (to prevent turning the flag on and forgetting to turn it off).

[orizi]

Reviewable status: 9 of 15 files reviewed, 1 unresolved discussion (waiting on @gilbens-starkware)

---

crates/cairo-lang-parser/src/parser.rs line 1487 at r2 (raw file):

Previously, gilbens-starkware (Gil Ben-Shachar) wrote…

You mean removing dollar as a token, and allowing it as an identifier character (only in the first position) in the lexer level?
That was my first intention, but then how is $false is lexed? (or any other keyword). I thought of two options, either as an identifier, and we'll need to check for keywords in another place, or as $ and false (similar to what I did, but with $ as an identifier I guess), but I thought it will only add more edge cases, so I opted to always split them.

i mostly mean that there shouldn't be a space allowed between the $ and the identifier.

[gilbens-starkware]

Reviewable status: 9 of 15 files reviewed, 1 unresolved discussion (waiting on @orizi)

---

crates/cairo-lang-parser/src/parser.rs line 1487 at r2 (raw file):

Previously, orizi wrote…

i mostly mean that there shouldn't be a space allowed between the $ and the identifier.

Asserting that the trailing trivia is empty? It can be done, but it deviates from any other thing that we do with whitespaces.
In rust it works btw, also a trailing comment after the $ works.

[orizi]

Reviewable status: 9 of 15 files reviewed, all discussions resolved

---

crates/cairo-lang-parser/src/parser.rs line 1487 at r2 (raw file):

Previously, gilbens-starkware (Gil Ben-Shachar) wrote…

Asserting that the trailing trivia is empty? It can be done, but it deviates from any other thing that we do with whitespaces.
In rust it works btw, also a trailing comment after the $ works.

ok - so just change the formatter at some point to make this non-confusing.

[orizi]

Reviewed 5 of 14 files at r1, 1 of 2 files at r2.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @gilbens-starkware)