rust/tests/ui/macros/stringify.rs

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

804 lines
25 KiB
Rust
Raw Normal View History

2021-12-23 15:02:31 -06:00
//@ run-pass
//@ edition:2021
//@ compile-flags: --test
#![allow(incomplete_features)]
2021-12-23 15:02:31 -06:00
#![feature(async_closure)]
2023-10-03 14:06:17 -05:00
#![feature(auto_traits)]
#![feature(box_patterns)]
2021-12-23 15:02:31 -06:00
#![feature(const_trait_impl)]
2023-10-19 16:46:28 -05:00
#![feature(coroutines)]
#![feature(decl_macro)]
#![feature(explicit_tail_calls)]
#![feature(if_let_guard)]
#![feature(let_chains)]
2021-12-23 15:02:31 -06:00
#![feature(more_qualified_paths)]
2023-11-21 19:30:43 -06:00
#![feature(never_patterns)]
#![feature(trait_alias)]
#![feature(try_blocks)]
#![feature(type_ascription)]
#![feature(yeet_expr)]
2021-12-23 15:02:31 -06:00
#![deny(unused_macros)]
// These macros force the use of AST pretty-printing by converting the input to
// a particular fragment specifier.
macro_rules! block { ($block:block) => { stringify!($block) }; }
macro_rules! expr { ($expr:expr) => { stringify!($expr) }; }
macro_rules! item { ($item:item) => { stringify!($item) }; }
macro_rules! meta { ($meta:meta) => { stringify!($meta) }; }
macro_rules! pat { ($pat:pat) => { stringify!($pat) }; }
macro_rules! path { ($path:path) => { stringify!($path) }; }
macro_rules! stmt { ($stmt:stmt) => { stringify!($stmt) }; }
macro_rules! ty { ($ty:ty) => { stringify!($ty) }; }
macro_rules! vis { ($vis:vis) => { stringify!($vis) }; }
macro_rules! c1 {
($frag:ident, [$($tt:tt)*], $s:literal) => {
// Prior to #125174:
// - the first of these two lines created a `TokenKind::Interpolated`
// that was printed by the AST pretty printer;
// - the second of these two lines created a token stream that was
// printed by the TokenStream pretty printer.
//
// Now they are both printed by the TokenStream pretty printer. But it
// doesn't hurt to keep both assertions to ensure this remains true.
//
// (This also explains the name `c1`. There used to be a `c2` macro for
// cases where the two pretty printers produced different output.)
assert_eq!($frag!($($tt)*), $s);
assert_eq!(stringify!($($tt)*), $s);
2021-12-23 15:02:31 -06:00
};
}
#[test]
fn test_block() {
c1!(block, [ {} ], "{}");
c1!(block, [ { true } ], "{ true }");
c1!(block, [ { return } ], "{ return }");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(block, [ { return; } ], "{ return; }");
c1!(block,
[ {
2021-12-23 15:02:31 -06:00
let _;
true
} ],
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
"{ let _; true }"
2021-12-23 15:02:31 -06:00
);
// Attributes are not allowed on vanilla blocks.
2021-12-23 15:02:31 -06:00
}
#[test]
fn test_expr() {
// ExprKind::Array
c1!(expr, [ [] ], "[]");
c1!(expr, [ [true] ], "[true]");
c1!(expr, [ [true,] ], "[true,]");
c1!(expr, [ [true, true] ], "[true, true]");
2021-12-23 15:02:31 -06:00
// ExprKind::ConstBlock
// FIXME: todo
2021-12-23 15:02:31 -06:00
// ExprKind::Call
c1!(expr, [ f() ], "f()");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ f::<u8>() ], "f::<u8>()");
c1!(expr, [ f :: < u8>( ) ], "f :: < u8>()");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ f::<1>() ], "f::<1>()");
c1!(expr, [ f::<'a, u8, 1>() ], "f::<'a, u8, 1>()");
c1!(expr, [ f(true) ], "f(true)");
c1!(expr, [ f(true,) ], "f(true,)");
c1!(expr, [ ()() ], "()()");
2021-12-23 15:02:31 -06:00
// ExprKind::MethodCall
c1!(expr, [ x.f() ], "x.f()");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ x.f::<u8>() ], "x.f::<u8>()");
c1!(expr, [ x.collect::<Vec<_>>() ], "x.collect::<Vec<_>>()");
2021-12-23 15:02:31 -06:00
// ExprKind::Tup
c1!(expr, [ () ], "()");
c1!(expr, [ (true,) ], "(true,)");
c1!(expr, [ (true, false) ], "(true, false)");
c1!(expr, [ (true, false,) ], "(true, false,)");
2021-12-23 15:02:31 -06:00
// ExprKind::Binary
c1!(expr, [ true || false ], "true || false");
c1!(expr, [ true || false && false ], "true || false && false");
c1!(expr, [ a < 1 && 2 < b && c > 3 && 4 > d ], "a < 1 && 2 < b && c > 3 && 4 > d");
c1!(expr, [ a & b & !c ], "a & b & !c");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ a + b * c - d + -1 * -2 - -3], "a + b * c - d + -1 * -2 - -3");
c1!(expr, [ x = !y ], "x = !y");
2021-12-23 15:02:31 -06:00
// ExprKind::Unary
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ *expr ], "*expr");
c1!(expr, [ !expr ], "!expr");
c1!(expr, [ -expr ], "-expr");
2021-12-23 15:02:31 -06:00
// ExprKind::Lit
c1!(expr, [ 'x' ], "'x'");
c1!(expr, [ 1_000_i8 ], "1_000_i8");
c1!(expr, [ 1.00000000000000001 ], "1.00000000000000001");
2021-12-23 15:02:31 -06:00
// ExprKind::Cast
c1!(expr, [ expr as T ], "expr as T");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ expr as T<u8> ], "expr as T<u8>");
2021-12-23 15:02:31 -06:00
// ExprKind::Type: there is no syntax for type ascription.
// ExprKind::Let
c1!(expr, [ if let Some(a) = b { c } else { d } ], "if let Some(a) = b { c } else { d }");
c1!(expr, [ if let _ = true && false {} ], "if let _ = true && false {}");
c1!(expr, [ if let _ = (true && false) {} ], "if let _ = (true && false) {}");
c1!(expr,
[ match () { _ if let _ = Struct {} => {} } ],
"match () { _ if let _ = Struct {} => {} }"
);
2021-12-23 15:02:31 -06:00
// ExprKind::If
c1!(expr, [ if true {} ], "if true {}");
c1!(expr, [ if !true {} ], "if !true {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ if ::std::blah() { } else { } ], "if ::std::blah() {} else {}");
c1!(expr, [ if let true = true {} else {} ], "if let true = true {} else {}");
c1!(expr,
[ if true {
2021-12-23 15:02:31 -06:00
} else if false {
} ],
"if true {} else if false {}"
2021-12-23 15:02:31 -06:00
);
c1!(expr,
[ if true {
2021-12-23 15:02:31 -06:00
} else if false {
} else {
} ],
"if true {} else if false {} else {}"
2021-12-23 15:02:31 -06:00
);
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr,
[ if true {
2021-12-23 15:02:31 -06:00
return;
} else if false {
0
} else {
0
} ],
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
"if true { return; } else if false { 0 } else { 0 }"
2021-12-23 15:02:31 -06:00
);
// ExprKind::While
c1!(expr, [ while true {} ], "while true {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ 'a: while true {} ], "'a: while true {}");
c1!(expr, [ while let true = true {} ], "while let true = true {}");
2021-12-23 15:02:31 -06:00
// ExprKind::ForLoop
c1!(expr, [ for _ in x {} ], "for _ in x {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ 'a: for _ in x {} ], "'a: for _ in x {}");
2021-12-23 15:02:31 -06:00
// ExprKind::Loop
c1!(expr, [ loop {} ], "loop {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ 'a: loop {} ], "'a: loop {}");
2021-12-23 15:02:31 -06:00
// ExprKind::Match
c1!(expr, [ match self {} ], "match self {}");
c1!(expr,
[ match self {
2021-12-23 15:02:31 -06:00
Ok => 1,
} ],
"match self { Ok => 1, }"
2021-12-23 15:02:31 -06:00
);
c1!(expr,
[ match self {
2021-12-23 15:02:31 -06:00
Ok => 1,
Err => 0,
} ],
"match self { Ok => 1, Err => 0, }"
2021-12-23 15:02:31 -06:00
);
// ExprKind::Closure
c1!(expr, [ || {} ], "|| {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ |x| {} ], "|x| {}");
c1!(expr, [ |x: u8| {} ], "|x: u8| {}");
c1!(expr, [ || () ], "|| ()");
c1!(expr, [ move || self ], "move || self");
c1!(expr, [ async || self ], "async || self");
c1!(expr, [ async move || self ], "async move || self");
c1!(expr, [ static || self ], "static || self");
c1!(expr, [ static move || self ], "static move || self");
c1!(expr, [ static async || self ], "static async || self");
c1!(expr, [ static async move || self ], "static async move || self");
c1!(expr, [ || -> u8 { self } ], "|| -> u8 { self }");
c1!(expr, [ 1 + || {} ], "1 + || {}");
2021-12-23 15:02:31 -06:00
// ExprKind::Block
c1!(expr, [ {} ], "{}");
c1!(expr, [ unsafe {} ], "unsafe {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ 'a: {} ], "'a: {}");
c1!(expr, [ #[attr] {} ], "#[attr] {}");
c1!(expr,
[
2021-12-23 15:02:31 -06:00
{
#![attr]
}
],
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
"{ #![attr] }"
2021-12-23 15:02:31 -06:00
);
// ExprKind::Async
c1!(expr, [ async {} ], "async {}");
c1!(expr, [ async move {} ], "async move {}");
2021-12-23 15:02:31 -06:00
// ExprKind::Await
c1!(expr, [ expr.await ], "expr.await");
2021-12-23 15:02:31 -06:00
// ExprKind::TryBlock
c1!(expr, [ try {} ], "try {}");
2021-12-23 15:02:31 -06:00
// ExprKind::Assign
c1!(expr, [ expr = true ], "expr = true");
2021-12-23 15:02:31 -06:00
// ExprKind::AssignOp
c1!(expr, [ expr += true ], "expr += true");
2021-12-23 15:02:31 -06:00
// ExprKind::Field
c1!(expr, [ expr.field ], "expr.field");
c1!(expr, [ expr.0 ], "expr.0");
2021-12-23 15:02:31 -06:00
// ExprKind::Index
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ expr[true] ], "expr[true]");
2021-12-23 15:02:31 -06:00
// ExprKind::Range
c1!(expr, [ .. ], "..");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ ..hi ], "..hi");
c1!(expr, [ lo.. ], "lo..");
c1!(expr, [ lo..hi ], "lo..hi");
c1!(expr, [ lo .. hi ], "lo .. hi");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ ..=hi ], "..=hi");
c1!(expr, [ lo..=hi ], "lo..=hi");
c1!(expr, [ -2..=-1 ], "-2..=-1");
2021-12-23 15:02:31 -06:00
// ExprKind::Underscore
// FIXME: todo
2021-12-23 15:02:31 -06:00
// ExprKind::Path
c1!(expr, [ thing ], "thing");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ m::thing ], "m::thing");
c1!(expr, [ self::thing ], "self::thing");
c1!(expr, [ crate::thing ], "crate::thing");
c1!(expr, [ Self::thing ], "Self::thing");
c1!(expr, [ <Self as T>::thing ], "<Self as T>::thing");
c1!(expr, [ Self::<'static> ], "Self::<'static>");
2021-12-23 15:02:31 -06:00
// ExprKind::AddrOf
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ &expr ], "&expr");
c1!(expr, [ &mut expr ], "&mut expr");
c1!(expr, [ &raw const expr ], "&raw const expr");
c1!(expr, [ &raw mut expr ], "&raw mut expr");
2021-12-23 15:02:31 -06:00
// ExprKind::Break
c1!(expr, [ break ], "break");
c1!(expr, [ break 'a ], "break 'a");
c1!(expr, [ break true ], "break true");
c1!(expr, [ break 'a true ], "break 'a true");
2021-12-23 15:02:31 -06:00
// ExprKind::Continue
c1!(expr, [ continue ], "continue");
c1!(expr, [ continue 'a ], "continue 'a");
2021-12-23 15:02:31 -06:00
// ExprKind::Ret
c1!(expr, [ return ], "return");
c1!(expr, [ return true ], "return true");
2021-12-23 15:02:31 -06:00
// ExprKind::InlineAsm: untestable because this test works pre-expansion.
// ExprKind::OffsetOf: untestable because this test works pre-expansion.
2021-12-23 15:02:31 -06:00
// ExprKind::MacCall
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ mac!(...) ], "mac!(...)");
c1!(expr, [ mac![...] ], "mac![...]");
c1!(expr, [ mac! { ... } ], "mac! { ... }");
2021-12-23 15:02:31 -06:00
// ExprKind::Struct
c1!(expr, [ Struct {} ], "Struct {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ <Struct as Trait>::Type {} ], "<Struct as Trait>::Type {}");
c1!(expr, [ Struct { .. } ], "Struct { .. }");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ Struct { ..base } ], "Struct { ..base }");
c1!(expr, [ Struct { x } ], "Struct { x }");
c1!(expr, [ Struct { x, .. } ], "Struct { x, .. }");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ Struct { x, ..base } ], "Struct { x, ..base }");
c1!(expr, [ Struct { x: true } ], "Struct { x: true }");
c1!(expr, [ Struct { x: true, .. } ], "Struct { x: true, .. }");
c1!(expr, [ Struct { x: true, ..base } ], "Struct { x: true, ..base }");
2021-12-23 15:02:31 -06:00
// ExprKind::Repeat
c1!(expr, [ [(); 0] ], "[(); 0]");
2021-12-23 15:02:31 -06:00
// ExprKind::Paren
c1!(expr, [ (expr) ], "(expr)");
2021-12-23 15:02:31 -06:00
// ExprKind::Try
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(expr, [ expr? ], "expr?");
2021-12-23 15:02:31 -06:00
// ExprKind::Yield
c1!(expr, [ yield ], "yield");
c1!(expr, [ yield true ], "yield true");
// ExprKind::Yeet
c1!(expr, [ do yeet ], "do yeet");
c1!(expr, [ do yeet 0 ], "do yeet 0");
// ExprKind::Become
// FIXME: todo
// ExprKind::IncludedBytes
// FIXME: todo
// ExprKind::FormatArgs: untestable because this test works pre-expansion.
// ExprKind::Err: untestable.
// Ones involving attributes.
c1!(expr, [ #[aa] 1 ], "#[aa] 1");
c1!(expr, [ #[aa] #[bb] x ], "#[aa] #[bb] x");
c1!(expr, [ #[aa] 1 + 2 ], "#[aa] 1 + 2");
c1!(expr, [ #[aa] x + 2 ], "#[aa] x + 2");
c1!(expr, [ #[aa] 1 / #[bb] 2 ], "#[aa] 1 / #[bb] 2");
c1!(expr, [ #[aa] x / #[bb] 2 ], "#[aa] x / #[bb] 2");
c1!(expr, [ 1 << #[bb] 2 ], "1 << #[bb] 2");
c1!(expr, [ x << #[bb] 2 ], "x << #[bb] 2");
c1!(expr, [ #[aa] (1 + 2) ], "#[aa] (1 + 2)");
c1!(expr, [ #[aa] #[bb] (x + 2) ], "#[aa] #[bb] (x + 2)");
c1!(expr, [ #[aa] x[0].p ], "#[aa] x[0].p");
c1!(expr, [ #[aa] { #![bb] 0 } ], "#[aa] { #![bb] 0 }");
2021-12-23 15:02:31 -06:00
}
#[test]
fn test_item() {
// ItemKind::ExternCrate
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(item, [ extern crate std; ], "extern crate std;");
c1!(item, [ pub extern crate self as std; ], "pub extern crate self as std;");
2021-12-23 15:02:31 -06:00
// ItemKind::Use
c1!(item, [ pub use crate::{a, b::c}; ], "pub use crate::{ a, b::c };"); // FIXME
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(item, [ pub use A::*; ], "pub use A::*;");
2021-12-23 15:02:31 -06:00
// ItemKind::Static
c1!(item, [ pub static S: () = {}; ], "pub static S: () = {};");
c1!(item, [ static mut S: () = {}; ], "static mut S: () = {};");
c1!(item, [ static S: (); ], "static S: ();");
c1!(item, [ static mut S: (); ], "static mut S: ();");
2021-12-23 15:02:31 -06:00
// ItemKind::Const
c1!(item, [ pub const S: () = {}; ], "pub const S: () = {};");
c1!(item, [ const S: (); ], "const S: ();");
2021-12-23 15:02:31 -06:00
// ItemKind::Fn
c1!(item,
[ pub default const async unsafe extern "C" fn f() {} ],
"pub default const async unsafe extern \"C\" fn f() {}"
2021-12-23 15:02:31 -06:00
);
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(item, [ fn g<T>(t: Vec<Vec<Vec<T>>>) {} ], "fn g<T>(t: Vec<Vec<Vec<T>>>) {}");
c1!(item,
[ fn h<'a>(t: &'a Vec<Cell<dyn D>>) {} ],
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
"fn h<'a>(t: &'a Vec<Cell<dyn D>>) {}"
);
2021-12-23 15:02:31 -06:00
// ItemKind::Mod
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(item, [ pub mod m; ], "pub mod m;");
c1!(item, [ mod m {} ], "mod m {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(item, [ unsafe mod m; ], "unsafe mod m;");
c1!(item, [ unsafe mod m {} ], "unsafe mod m {}");
2021-12-23 15:02:31 -06:00
// ItemKind::ForeignMod
c1!(item, [ extern "C" {} ], "extern \"C\" {}");
c1!(item, [ pub extern "C" {} ], "pub extern \"C\" {}");
c1!(item, [ unsafe extern "C++" {} ], "unsafe extern \"C++\" {}");
2021-12-23 15:02:31 -06:00
// ItemKind::GlobalAsm: untestable because this test works pre-expansion.
2021-12-23 15:02:31 -06:00
// ItemKind::TyAlias
c1!(item,
[
2021-12-23 15:02:31 -06:00
pub default type Type<'a>: Bound
where
Self: 'a,
= T;
],
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
"pub default type Type<'a>: Bound where Self: 'a, = T;"
2021-12-23 15:02:31 -06:00
);
// ItemKind::Enum
c1!(item, [ pub enum Void {} ], "pub enum Void {}");
c1!(item,
[
2021-12-23 15:02:31 -06:00
enum Empty {
Unit,
Tuple(),
Struct {},
}
],
"enum Empty { Unit, Tuple(), Struct {}, }"
2021-12-23 15:02:31 -06:00
);
c1!(item,
[
enum Enum<T>
where
T: 'a,
{
2021-12-23 15:02:31 -06:00
Unit,
Tuple(T),
Struct { t: T },
}
],
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
"enum Enum<T> where T: 'a, { Unit, Tuple(T), Struct { t: T }, }"
2021-12-23 15:02:31 -06:00
);
// ItemKind::Struct
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(item, [ pub struct Unit; ], "pub struct Unit;");
c1!(item, [ struct Tuple(); ], "struct Tuple();");
c1!(item, [ struct Tuple(T); ], "struct Tuple(T);");
c1!(item, [ struct Struct {} ], "struct Struct {}");
c1!(item,
[
2021-12-23 15:02:31 -06:00
struct Struct<T>
where
T: 'a,
{
t: T,
}
],
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
"struct Struct<T> where T: 'a, { t: T, }"
2021-12-23 15:02:31 -06:00
);
// ItemKind::Union
c1!(item, [ pub union Union {} ], "pub union Union {}");
c1!(item,
[
2021-12-23 15:02:31 -06:00
union Union<T> where T: 'a {
t: T,
}
],
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
"union Union<T> where T: 'a { t: T, }"
2021-12-23 15:02:31 -06:00
);
// ItemKind::Trait
c1!(item, [ pub unsafe auto trait Send {} ], "pub unsafe auto trait Send {}");
c1!(item,
[
2021-12-23 15:02:31 -06:00
trait Trait<'a>: Sized
where
Self: 'a,
{
}
],
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
"trait Trait<'a>: Sized where Self: 'a, {}"
2021-12-23 15:02:31 -06:00
);
// ItemKind::TraitAlias
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(item,
[ pub trait Trait<T> = Sized where T: 'a; ],
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
"pub trait Trait<T> = Sized where T: 'a;"
2021-12-23 15:02:31 -06:00
);
// ItemKind::Impl
c1!(item, [ pub impl Struct {} ], "pub impl Struct {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(item, [ impl<T> Struct<T> {} ], "impl<T> Struct<T> {}");
c1!(item, [ pub impl Trait for Struct {} ], "pub impl Trait for Struct {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(item, [ impl<T> const Trait for T {} ], "impl<T> const Trait for T {}");
c1!(item, [ impl ~const Struct {} ], "impl ~const Struct {}");
2021-12-23 15:02:31 -06:00
// ItemKind::MacCall
c1!(item, [ mac!(...); ], "mac!(...);");
c1!(item, [ mac![...]; ], "mac![...];");
c1!(item, [ mac! { ... } ], "mac! { ... }");
2021-12-23 15:02:31 -06:00
// ItemKind::MacroDef
c1!(item,
[
2021-12-23 15:02:31 -06:00
macro_rules! stringify {
() => {};
}
],
"macro_rules! stringify { () => {}; }"
2021-12-23 15:02:31 -06:00
);
c1!(item, [ pub macro stringify() {} ], "pub macro stringify() {}");
// Ones involving attributes.
c1!(item, [ #[aa] mod m; ], "#[aa] mod m;");
c1!(item, [ mod m { #![bb] } ], "mod m { #![bb] }");
c1!(item, [ #[aa] mod m { #![bb] } ], "#[aa] mod m { #![bb] }");
2021-12-23 15:02:31 -06:00
}
#[test]
fn test_meta() {
c1!(meta, [ k ], "k");
c1!(meta, [ k = "v" ], "k = \"v\"");
c1!(meta, [ list(k1, k2 = "v") ], "list(k1, k2 = \"v\")");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(meta, [ serde::k ], "serde::k");
// Attributes are not allowed on metas.
2021-12-23 15:02:31 -06:00
}
#[test]
fn test_pat() {
// PatKind::Wild
c1!(pat, [ _ ], "_");
2021-12-23 15:02:31 -06:00
// PatKind::Ident
c1!(pat, [ _x ], "_x");
c1!(pat, [ ref _x ], "ref _x");
c1!(pat, [ mut _x ], "mut _x");
c1!(pat, [ ref mut _x ], "ref mut _x");
c1!(pat, [ ref mut _x @ _ ], "ref mut _x @ _");
2021-12-23 15:02:31 -06:00
// PatKind::Struct
c1!(pat, [ Struct {} ], "Struct {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(pat, [ Struct::<u8> {} ], "Struct::<u8> {}");
c1!(pat, [ Struct ::< u8 > {} ], "Struct ::< u8 > {}");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(pat, [ Struct::<'static> {} ], "Struct::<'static> {}");
c1!(pat, [ Struct { x } ], "Struct { x }");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(pat, [ Struct { x: _x } ], "Struct { x: _x }");
c1!(pat, [ Struct { .. } ], "Struct { .. }");
c1!(pat, [ Struct { x, .. } ], "Struct { x, .. }");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(pat, [ Struct { x: _x, .. } ], "Struct { x: _x, .. }");
c1!(pat, [ <Struct as Trait>::Type {} ], "<Struct as Trait>::Type {}");
2021-12-23 15:02:31 -06:00
// PatKind::TupleStruct
c1!(pat, [ Tuple() ], "Tuple()");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(pat, [ Tuple::<u8>() ], "Tuple::<u8>()");
c1!(pat, [ Tuple::<'static>() ], "Tuple::<'static>()");
c1!(pat, [ Tuple(x) ], "Tuple(x)");
c1!(pat, [ Tuple(..) ], "Tuple(..)");
c1!(pat, [ Tuple(x, ..) ], "Tuple(x, ..)");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(pat, [ <Struct as Trait>::Type() ], "<Struct as Trait>::Type()");
2021-12-23 15:02:31 -06:00
// PatKind::Or
c1!(pat, [ true | false ], "true | false");
c1!(pat, [ | true ], "| true");
c1!(pat, [ |true| false ], "|true| false");
2021-12-23 15:02:31 -06:00
// PatKind::Path
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(pat, [ crate::Path ], "crate::Path");
c1!(pat, [ Path::<u8> ], "Path::<u8>");
c1!(pat, [ Path::<'static> ], "Path::<'static>");
c1!(pat, [ <Struct as Trait>::Type ], "<Struct as Trait>::Type");
2021-12-23 15:02:31 -06:00
// PatKind::Tuple
c1!(pat, [ () ], "()");
c1!(pat, [ (true,) ], "(true,)");
c1!(pat, [ (true, false) ], "(true, false)");
2021-12-23 15:02:31 -06:00
// PatKind::Box
c1!(pat, [ box pat ], "box pat");
2021-12-23 15:02:31 -06:00
// PatKind::Ref
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(pat, [ &pat ], "&pat");
c1!(pat, [ &mut pat ], "&mut pat");
2021-12-23 15:02:31 -06:00
// PatKind::Lit
c1!(pat, [ 1_000_i8 ], "1_000_i8");
2021-12-23 15:02:31 -06:00
// PatKind::Range
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(pat, [ ..1 ], "..1");
c1!(pat, [ 0.. ], "0..");
c1!(pat, [ 0..1 ], "0..1");
c1!(pat, [ 0..=1 ], "0..=1");
c1!(pat, [ -2..=-1 ], "-2..=-1");
2021-12-23 15:02:31 -06:00
// PatKind::Slice
c1!(pat, [ [] ], "[]");
c1!(pat, [ [true] ], "[true]");
c1!(pat, [ [true,] ], "[true,]");
c1!(pat, [ [true, false] ], "[true, false]");
2021-12-23 15:02:31 -06:00
// PatKind::Rest
c1!(pat, [ .. ], "..");
2021-12-23 15:02:31 -06:00
2023-11-21 19:30:43 -06:00
// PatKind::Never
c1!(pat, [ Some(!) ], "Some(!)");
c1!(pat, [ None | Some(!) ], "None | Some(!)");
2021-12-23 15:02:31 -06:00
// PatKind::Paren
c1!(pat, [ (pat) ], "(pat)");
2021-12-23 15:02:31 -06:00
// PatKind::MacCall
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(pat, [ mac!(...) ], "mac!(...)");
c1!(pat, [ mac![...] ], "mac![...]");
c1!(pat, [ mac! { ... } ], "mac! { ... }");
// Attributes are not allowed on patterns.
2021-12-23 15:02:31 -06:00
}
#[test]
fn test_path() {
c1!(path, [ thing ], "thing");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(path, [ m::thing ], "m::thing");
c1!(path, [ self::thing ], "self::thing");
c1!(path, [ crate::thing ], "crate::thing");
c1!(path, [ Self::thing ], "Self::thing");
c1!(path, [ Self<'static> ], "Self<'static>");
c1!(path, [ Self::<'static> ], "Self::<'static>");
c1!(path, [ Self() ], "Self()");
c1!(path, [ Self() -> () ], "Self() -> ()");
// Attributes are not allowed on paths.
2021-12-23 15:02:31 -06:00
}
#[test]
fn test_stmt() {
// StmtKind::Local
c1!(stmt, [ let _ ], "let _");
c1!(stmt, [ let x = true ], "let x = true");
c1!(stmt, [ let x: bool = true ], "let x: bool = true");
c1!(stmt, [ let (a, b) = (1, 2) ], "let (a, b) = (1, 2)");
c1!(stmt, [ let (a, b): (u32, u32) = (1, 2) ], "let (a, b): (u32, u32) = (1, 2)");
c1!(stmt, [ let _ = f() else { return; } ], "let _ = f() else { return; }");
2021-12-23 15:02:31 -06:00
// StmtKind::Item
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(stmt, [ struct S; ], "struct S;");
c1!(stmt, [ struct S {} ], "struct S {}");
2021-12-23 15:02:31 -06:00
// StmtKind::Expr
c1!(stmt, [ loop {} ], "loop {}");
2021-12-23 15:02:31 -06:00
// StmtKind::Semi
c1!(stmt, [ 1 + 1 ], "1 + 1");
2021-12-23 15:02:31 -06:00
// StmtKind::Empty
c1!(stmt, [ ; ], ";");
2021-12-23 15:02:31 -06:00
// StmtKind::MacCall
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(stmt, [ mac!(...) ], "mac!(...)");
c1!(stmt, [ mac![...] ], "mac![...]");
c1!(stmt, [ mac! { ... } ], "mac! { ... }");
// Ones involving attributes.
c1!(stmt, [ #[aa] 1 ], "#[aa] 1");
c1!(stmt, [ #[aa] #[bb] x ], "#[aa] #[bb] x");
c1!(stmt, [ #[aa] 1 as u32 ], "#[aa] 1 as u32");
c1!(stmt, [ #[aa] x as u32 ], "#[aa] x as u32");
c1!(stmt, [ #[aa] 1 .. #[bb] 2 ], "#[aa] 1 .. #[bb] 2");
c1!(stmt, [ #[aa] x .. #[bb] 2 ], "#[aa] x .. #[bb] 2");
c1!(stmt, [ 1 || #[bb] 2 ], "1 || #[bb] 2");
c1!(stmt, [ x || #[bb] 2 ], "x || #[bb] 2");
c1!(stmt, [ #[aa] (1 + 2) ], "#[aa] (1 + 2)");
c1!(stmt, [ #[aa] #[bb] (x + 2) ], "#[aa] #[bb] (x + 2)");
c1!(stmt, [ #[aa] x[0].p ], "#[aa] x[0].p");
c1!(stmt, [ #[aa] { #![bb] 0 } ], "#[aa] { #![bb] 0 }");
2021-12-23 15:02:31 -06:00
}
#[test]
fn test_ty() {
// TyKind::Slice
c1!(ty, [ [T] ], "[T]");
2021-12-23 15:02:31 -06:00
// TyKind::Array
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(ty, [ [T; 0] ], "[T; 0]");
2021-12-23 15:02:31 -06:00
// TyKind::Ptr
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(ty, [ *const T ], "*const T");
c1!(ty, [ *mut T ], "*mut T");
2021-12-23 15:02:31 -06:00
// TyKind::Ref
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(ty, [ &T ], "&T");
c1!(ty, [ &mut T ], "&mut T");
c1!(ty, [ &'a T ], "&'a T");
c1!(ty, [ &'a mut [T] ], "&'a mut [T]");
c1!(ty, [ &A<B<C<D<E>>>> ], "&A<B<C<D<E>>>>");
c1!(ty, [ &A<B<C<D<E> > > > ], "&A<B<C<D<E> > > >");
2021-12-23 15:02:31 -06:00
// TyKind::BareFn
c1!(ty, [ fn() ], "fn()");
c1!(ty, [ fn() -> () ], "fn() -> ()");
c1!(ty, [ fn(u8) ], "fn(u8)");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(ty, [ fn(x: u8) ], "fn(x: u8)");
c1!(ty, [ for<> fn() ], "for<> fn()");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(ty, [ for<'a> fn() ], "for<'a> fn()");
2021-12-23 15:02:31 -06:00
// TyKind::Never
c1!(ty, [ ! ], "!");
2021-12-23 15:02:31 -06:00
// TyKind::Tup
c1!(ty, [ () ], "()");
c1!(ty, [ (T,) ], "(T,)");
c1!(ty, [ (T, U) ], "(T, U)");
2021-12-23 15:02:31 -06:00
// TyKind::AnonStruct: untestable in isolation.
// TyKind::AnonUnion: untestable in isolation.
2021-12-23 15:02:31 -06:00
// TyKind::Path
c1!(ty, [ T ], "T");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(ty, [ Ref<'a> ], "Ref<'a>");
c1!(ty, [ PhantomData<T> ], "PhantomData<T>");
c1!(ty, [ PhantomData::<T> ], "PhantomData::<T>");
c1!(ty, [ Fn() -> ! ], "Fn() -> !");
c1!(ty, [ Fn(u8) -> ! ], "Fn(u8) -> !");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(ty, [ <Struct as Trait>::Type ], "<Struct as Trait>::Type");
2021-12-23 15:02:31 -06:00
// TyKind::TraitObject
c1!(ty, [ dyn Send ], "dyn Send");
c1!(ty, [ dyn Send + 'a ], "dyn Send + 'a");
c1!(ty, [ dyn 'a + Send ], "dyn 'a + Send");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(ty, [ dyn ?Sized ], "dyn ?Sized");
c1!(ty, [ dyn ~const Clone ], "dyn ~const Clone");
c1!(ty, [ dyn for<'a> Send ], "dyn for<'a> Send");
2021-12-23 15:02:31 -06:00
// TyKind::ImplTrait
c1!(ty, [ impl Send ], "impl Send");
c1!(ty, [ impl Send + 'a ], "impl Send + 'a");
c1!(ty, [ impl 'a + Send ], "impl 'a + Send");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(ty, [ impl ?Sized ], "impl ?Sized");
c1!(ty, [ impl ~const Clone ], "impl ~const Clone");
c1!(ty, [ impl for<'a> Send ], "impl for<'a> Send");
2021-12-23 15:02:31 -06:00
// TyKind::Paren
c1!(ty, [ (T) ], "(T)");
2021-12-23 15:02:31 -06:00
// TyKind::Typeof: unused for now.
2021-12-23 15:02:31 -06:00
// TyKind::Infer
c1!(ty, [ _ ], "_");
// TyKind::ImplicitSelf: there is no syntax for this.
2021-12-23 15:02:31 -06:00
// TyKind::MacCall
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
c1!(ty, [ mac!(...) ], "mac!(...)");
c1!(ty, [ mac![...] ], "mac![...]");
c1!(ty, [ mac! { ... } ], "mac! { ... }");
// TyKind::Err: untestable.
// TyKind::CVarArgs
// FIXME: todo
// Attributes are not allowed on types.
2021-12-23 15:02:31 -06:00
}
#[test]
fn test_vis() {
// VisibilityKind::Public
c1!(vis, [ pub ], "pub");
2021-12-23 15:02:31 -06:00
// VisibilityKind::Restricted
c1!(vis, [ pub(crate) ], "pub(crate)");
c1!(vis, [ pub(self) ], "pub(self)");
c1!(vis, [ pub(super) ], "pub(super)");
c1!(vis, [ pub(in crate) ], "pub(in crate)");
c1!(vis, [ pub(in self) ], "pub(in self)");
c1!(vis, [ pub(in super) ], "pub(in super)");
c1!(vis, [ pub(in path::to) ], "pub(in path::to)");
c1!(vis, [ pub(in ::path::to) ], "pub(in ::path::to)");
c1!(vis, [ pub(in self::path::to) ], "pub(in self::path::to)");
c1!(vis, [ pub(in super::path::to) ], "pub(in super::path::to)");
2021-12-23 15:02:31 -06:00
// VisibilityKind::Inherited
// This one is different because directly calling `vis!` does not work.
macro_rules! inherited_vis { ($vis:vis struct) => { vis!($vis) }; }
assert_eq!(inherited_vis!(struct), "");
assert_eq!(stringify!(), "");
// Attributes are not allowed on visibilities.
2021-12-23 15:02:31 -06:00
}
macro_rules! p {
([$($tt:tt)*], $s:literal) => {
assert_eq!(stringify!($($tt)*), $s);
};
}
#[test]
fn test_punct() {
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
// For all these cases, we should preserve spaces between the tokens.
// Otherwise, any old proc macro that parses pretty-printed code might glue
// together tokens that shouldn't be glued.
p!([ = = < < <= <= == == != != >= >= > > ], "= = < < <= <= == == != != >= >= > >");
p!([ && && & & || || | | ! ! ], "&& && & & || || | | ! !");
p!([ ~ ~ @ @ # # ], "~ ~ @ @ # #");
p!([ . . .. .. ... ... ..= ..=], ". . .. .. ... ... ..= ..=");
p!([ , , ; ; : : :: :: ], ", , ; ; : : :: ::");
p!([ -> -> <- <- => =>], "-> -> <- <- => =>");
p!([ $ $ ? ? ' ' ], "$ $ ? ? ' '");
p!([ + + += += - - -= -= * * *= *= / / /= /= ], "+ + += += - - -= -= * * *= *= / / /= /=");
p!([ % % %= %= ^ ^ ^= ^= << << <<= <<= >> >> >>= >>= ],
"% % %= %= ^ ^ ^= ^= << << <<= <<= >> >> >>= >>=");
p!([ +! ?= |> >>@ --> <-- $$ =====> ], "+! ?= |> >>@ --> <-- $$ =====>");
p!([ ,; ;, ** @@ $+$ >< <> ?? +== ], ",; ;, ** @@ $+$ >< <> ?? +==");
Improve `print_tts` by changing `tokenstream::Spacing`. `tokenstream::Spacing` appears on all `TokenTree::Token` instances, both punct and non-punct. Its current usage: - `Joint` means "can join with the next token *and* that token is a punct". - `Alone` means "cannot join with the next token *or* can join with the next token but that token is not a punct". The fact that `Alone` is used for two different cases is awkward. This commit augments `tokenstream::Spacing` with a new variant `JointHidden`, resulting in: - `Joint` means "can join with the next token *and* that token is a punct". - `JointHidden` means "can join with the next token *and* that token is a not a punct". - `Alone` means "cannot join with the next token". This *drastically* improves the output of `print_tts`. For example, this: ``` stringify!(let a: Vec<u32> = vec![];) ``` currently produces this string: ``` let a : Vec < u32 > = vec! [] ; ``` With this PR, it now produces this string: ``` let a: Vec<u32> = vec![] ; ``` (The space after the `]` is because `TokenTree::Delimited` currently doesn't have spacing information. The subsequent commit fixes this.) The new `print_tts` doesn't replicate original code perfectly. E.g. multiple space characters will be condensed into a single space character. But it's much improved. `print_tts` still produces the old, uglier output for code produced by proc macros. Because we have to translate the generated code from `proc_macro::Spacing` to the more expressive `token::Spacing`, which results in too much `proc_macro::Along` usage and no `proc_macro::JointHidden` usage. So `space_between` still exists and is used by `print_tts` in conjunction with the `Spacing` field. This change will also help with the removal of `Token::Interpolated`. Currently interpolated tokens are pretty-printed nicely via AST pretty printing. `Token::Interpolated` removal will mean they get printed with `print_tts`. Without this change, that would result in much uglier output for code produced by decl macro expansions. With this change, AST pretty printing and `print_tts` produce similar results. The commit also tweaks the comments on `proc_macro::Spacing`. In particular, it refers to "compound tokens" rather than "multi-char operators" because lifetimes aren't operators.
2023-08-07 20:43:44 -05:00
p!([ :#!@|$=&*,+;*~? ], ":#!@|$=&*,+;*~?");
}