parser: Ensure that all nonterminals have tokens after parsing

This commit is contained in:
Vadim Petrochenkov 2021-05-06 16:21:40 +03:00
parent 9a576175cc
commit cbdfa1edca
8 changed files with 612 additions and 31 deletions

View file

@ -82,7 +82,8 @@ impl AstLike for crate::token::Nonterminal {
Nonterminal::NtMeta(attr_item) => attr_item.tokens_mut(),
Nonterminal::NtPath(path) => path.tokens_mut(),
Nonterminal::NtVis(vis) => vis.tokens_mut(),
_ => panic!("Called tokens_mut on {:?}", self),
Nonterminal::NtBlock(block) => block.tokens_mut(),
Nonterminal::NtIdent(..) | Nonterminal::NtLifetime(..) | Nonterminal::NtTT(..) => None,
}
}
}

View file

@ -342,16 +342,10 @@ impl<'a> Parser<'a> {
// If we support tokens at all
if let Some(target_tokens) = ret.tokens_mut() {
if let Some(target_tokens) = target_tokens {
assert!(
!self.capture_cfg,
"Encountered existing tokens with capture_cfg set: {:?}",
target_tokens
);
} else {
if target_tokens.is_none() {
// Store se our newly captured tokens into the AST node
*target_tokens = Some(tokens.clone());
};
}
}
let final_attrs = ret.attrs();

View file

@ -94,17 +94,7 @@ impl<'a> Parser<'a> {
/// Parses an expression, forcing tokens to be collected
pub fn parse_expr_force_collect(&mut self) -> PResult<'a, P<Expr>> {
// If we have outer attributes, then the call to `collect_tokens_trailing_token`
// will be made for us.
if matches!(self.token.kind, TokenKind::Pound | TokenKind::DocComment(..)) {
self.parse_expr()
} else {
// If we don't have outer attributes, then we need to ensure
// that collection happens by using `collect_tokens_no_attrs`.
// Expression don't support custom inner attributes, so `parse_expr`
// will never try to collect tokens if we don't have outer attributes.
self.collect_tokens_no_attrs(|this| this.parse_expr())
}
self.collect_tokens_no_attrs(|this| this.parse_expr())
}
pub fn parse_anon_const_expr(&mut self) -> PResult<'a, AnonConst> {

View file

@ -63,6 +63,7 @@ enum BlockMode {
/// Whether or not we should force collection of tokens for an AST node,
/// regardless of whether or not it has attributes
#[derive(Clone, Copy, PartialEq)]
pub enum ForceCollect {
Yes,
No,

View file

@ -1,5 +1,6 @@
use rustc_ast::ptr::P;
use rustc_ast::token::{self, Nonterminal, NonterminalKind, Token};
use rustc_ast::AstLike;
use rustc_ast_pretty::pprust;
use rustc_errors::PResult;
use rustc_span::symbol::{kw, Ident};
@ -102,7 +103,7 @@ impl<'a> Parser<'a> {
// which requires having captured tokens available. Since we cannot determine
// in advance whether or not a proc-macro will be (transitively) invoked,
// we always capture tokens for any `Nonterminal` which needs them.
Ok(match kind {
let mut nt = match kind {
NonterminalKind::Item => match self.parse_item(ForceCollect::Yes)? {
Some(item) => token::NtItem(item),
None => {
@ -169,7 +170,19 @@ impl<'a> Parser<'a> {
return Err(self.struct_span_err(self.token.span, msg));
}
}
})
};
// If tokens are supported at all, they should be collected.
if matches!(nt.tokens_mut(), Some(None)) {
panic!(
"Missing tokens for nt {:?} at {:?}: {:?}",
nt,
nt.span(),
pprust::nonterminal_to_string(&nt)
);
}
Ok(nt)
}
}

View file

@ -73,7 +73,11 @@ impl<'a> Parser<'a> {
// or `auto trait` items. We aim to parse an arbitrary path `a::b` but not something
// that starts like a path (1 token), but it fact not a path.
// Also, we avoid stealing syntax from `parse_item_`.
self.parse_stmt_path_start(lo, attrs, force_collect)?
if force_collect == ForceCollect::Yes {
self.collect_tokens_no_attrs(|this| this.parse_stmt_path_start(lo, attrs))
} else {
self.parse_stmt_path_start(lo, attrs)
}?
} else if let Some(item) =
self.parse_item_common(attrs.clone(), false, true, |_| true, force_collect)?
{
@ -85,7 +89,13 @@ impl<'a> Parser<'a> {
self.mk_stmt(lo, StmtKind::Empty)
} else if self.token != token::CloseDelim(token::Brace) {
// Remainder are line-expr stmts.
let e = self.parse_expr_res(Restrictions::STMT_EXPR, Some(attrs))?;
let e = if force_collect == ForceCollect::Yes {
self.collect_tokens_no_attrs(|this| {
this.parse_expr_res(Restrictions::STMT_EXPR, Some(attrs))
})
} else {
self.parse_expr_res(Restrictions::STMT_EXPR, Some(attrs))
}?;
self.mk_stmt(lo.to(e.span), StmtKind::Expr(e))
} else {
self.error_outer_attrs(&attrs.take_for_recovery());
@ -93,13 +103,8 @@ impl<'a> Parser<'a> {
}))
}
fn parse_stmt_path_start(
&mut self,
lo: Span,
attrs: AttrWrapper,
force_collect: ForceCollect,
) -> PResult<'a, Stmt> {
let stmt = self.collect_tokens_trailing_token(attrs, force_collect, |this, attrs| {
fn parse_stmt_path_start(&mut self, lo: Span, attrs: AttrWrapper) -> PResult<'a, Stmt> {
let stmt = self.collect_tokens_trailing_token(attrs, ForceCollect::No, |this, attrs| {
let path = this.parse_path(PathStyle::Expr)?;
if this.eat(&token::Not) {

View file

@ -0,0 +1,37 @@
// check-pass
// aux-build:test-macros.rs
#![feature(decl_macro)]
#![feature(stmt_expr_attributes)]
#![no_std] // Don't load unnecessary hygiene information from std
extern crate std;
#[macro_use]
extern crate test_macros;
macro mac {
(expr $expr:expr) => {
#[derive(Print)]
enum E {
V = { let _ = $expr; 0 },
}
},
(stmt $stmt:stmt) => {
#[derive(Print)]
enum E {
V = { let _ = { $stmt }; 0 },
}
},
}
const PATH: u8 = 2;
fn main() {
mac!(expr #[allow(warnings)] 0);
mac!(stmt 0);
mac!(stmt {});
mac!(stmt PATH);
mac!(stmt 0 + 1);
mac!(stmt PATH + 1);
}

View file

@ -0,0 +1,540 @@
PRINT-DERIVE INPUT (DISPLAY): enum E { V = { let _ = #[allow(warnings)] 0 ; 0 }, }
PRINT-DERIVE DEEP-RE-COLLECTED (DISPLAY): enum E { V = { let _ = #[allow(warnings)] #[allow(warnings)] 0 ; 0 }, }
PRINT-DERIVE INPUT (DEBUG): TokenStream [
Ident {
ident: "enum",
span: #4 bytes(299..303),
},
Ident {
ident: "E",
span: #4 bytes(304..305),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "V",
span: #4 bytes(320..321),
},
Punct {
ch: '=',
spacing: Alone,
span: #4 bytes(322..323),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "let",
span: #4 bytes(326..329),
},
Ident {
ident: "_",
span: #4 bytes(330..331),
},
Punct {
ch: '=',
spacing: Alone,
span: #4 bytes(332..333),
},
Group {
delimiter: None,
stream: TokenStream [
Punct {
ch: '#',
spacing: Alone,
span: #0 bytes(541..542),
},
Group {
delimiter: Bracket,
stream: TokenStream [
Ident {
ident: "allow",
span: #0 bytes(543..548),
},
Group {
delimiter: Parenthesis,
stream: TokenStream [
Ident {
ident: "warnings",
span: #0 bytes(549..557),
},
],
span: #0 bytes(548..558),
},
],
span: #0 bytes(542..559),
},
Punct {
ch: '#',
spacing: Alone,
span: #0 bytes(541..542),
},
Group {
delimiter: Bracket,
stream: TokenStream [
Ident {
ident: "allow",
span: #0 bytes(543..548),
},
Group {
delimiter: Parenthesis,
stream: TokenStream [
Ident {
ident: "warnings",
span: #0 bytes(549..557),
},
],
span: #0 bytes(548..558),
},
],
span: #0 bytes(542..559),
},
Literal {
kind: Integer,
symbol: "0",
suffix: None,
span: #0 bytes(560..561),
},
],
span: #4 bytes(334..339),
},
Punct {
ch: ';',
spacing: Alone,
span: #4 bytes(339..340),
},
Literal {
kind: Integer,
symbol: "0",
suffix: None,
span: #4 bytes(341..342),
},
],
span: #4 bytes(324..344),
},
Punct {
ch: ',',
spacing: Alone,
span: #4 bytes(344..345),
},
],
span: #4 bytes(306..355),
},
]
PRINT-DERIVE INPUT (DISPLAY): enum E { V = { let _ = { 0; } ; 0 }, }
PRINT-DERIVE DEEP-RE-COLLECTED (DISPLAY): enum E { V = { let _ = { 0 } ; 0 }, }
PRINT-DERIVE INPUT (DEBUG): TokenStream [
Ident {
ident: "enum",
span: #8 bytes(423..427),
},
Ident {
ident: "E",
span: #8 bytes(428..429),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "V",
span: #8 bytes(444..445),
},
Punct {
ch: '=',
spacing: Alone,
span: #8 bytes(446..447),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "let",
span: #8 bytes(450..453),
},
Ident {
ident: "_",
span: #8 bytes(454..455),
},
Punct {
ch: '=',
spacing: Alone,
span: #8 bytes(456..457),
},
Group {
delimiter: Brace,
stream: TokenStream [
Group {
delimiter: None,
stream: TokenStream [
Literal {
kind: Integer,
symbol: "0",
suffix: None,
span: #0 bytes(578..579),
},
],
span: #8 bytes(460..465),
},
],
span: #8 bytes(458..467),
},
Punct {
ch: ';',
spacing: Alone,
span: #8 bytes(467..468),
},
Literal {
kind: Integer,
symbol: "0",
suffix: None,
span: #8 bytes(469..470),
},
],
span: #8 bytes(448..472),
},
Punct {
ch: ',',
spacing: Alone,
span: #8 bytes(472..473),
},
],
span: #8 bytes(430..483),
},
]
PRINT-DERIVE INPUT (DISPLAY): enum E { V = { let _ = { { } } ; 0 }, }
PRINT-DERIVE INPUT (DEBUG): TokenStream [
Ident {
ident: "enum",
span: #12 bytes(423..427),
},
Ident {
ident: "E",
span: #12 bytes(428..429),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "V",
span: #12 bytes(444..445),
},
Punct {
ch: '=',
spacing: Alone,
span: #12 bytes(446..447),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "let",
span: #12 bytes(450..453),
},
Ident {
ident: "_",
span: #12 bytes(454..455),
},
Punct {
ch: '=',
spacing: Alone,
span: #12 bytes(456..457),
},
Group {
delimiter: Brace,
stream: TokenStream [
Group {
delimiter: None,
stream: TokenStream [
Group {
delimiter: Brace,
stream: TokenStream [],
span: #0 bytes(596..598),
},
],
span: #12 bytes(460..465),
},
],
span: #12 bytes(458..467),
},
Punct {
ch: ';',
spacing: Alone,
span: #12 bytes(467..468),
},
Literal {
kind: Integer,
symbol: "0",
suffix: None,
span: #12 bytes(469..470),
},
],
span: #12 bytes(448..472),
},
Punct {
ch: ',',
spacing: Alone,
span: #12 bytes(472..473),
},
],
span: #12 bytes(430..483),
},
]
PRINT-DERIVE INPUT (DISPLAY): enum E { V = { let _ = { PATH; } ; 0 }, }
PRINT-DERIVE DEEP-RE-COLLECTED (DISPLAY): enum E { V = { let _ = { PATH } ; 0 }, }
PRINT-DERIVE INPUT (DEBUG): TokenStream [
Ident {
ident: "enum",
span: #16 bytes(423..427),
},
Ident {
ident: "E",
span: #16 bytes(428..429),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "V",
span: #16 bytes(444..445),
},
Punct {
ch: '=',
spacing: Alone,
span: #16 bytes(446..447),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "let",
span: #16 bytes(450..453),
},
Ident {
ident: "_",
span: #16 bytes(454..455),
},
Punct {
ch: '=',
spacing: Alone,
span: #16 bytes(456..457),
},
Group {
delimiter: Brace,
stream: TokenStream [
Group {
delimiter: None,
stream: TokenStream [
Ident {
ident: "PATH",
span: #0 bytes(615..619),
},
],
span: #16 bytes(460..465),
},
],
span: #16 bytes(458..467),
},
Punct {
ch: ';',
spacing: Alone,
span: #16 bytes(467..468),
},
Literal {
kind: Integer,
symbol: "0",
suffix: None,
span: #16 bytes(469..470),
},
],
span: #16 bytes(448..472),
},
Punct {
ch: ',',
spacing: Alone,
span: #16 bytes(472..473),
},
],
span: #16 bytes(430..483),
},
]
PRINT-DERIVE INPUT (DISPLAY): enum E { V = { let _ = { 0 + 1; } ; 0 }, }
PRINT-DERIVE DEEP-RE-COLLECTED (DISPLAY): enum E { V = { let _ = { 0 + 1 } ; 0 }, }
PRINT-DERIVE INPUT (DEBUG): TokenStream [
Ident {
ident: "enum",
span: #20 bytes(423..427),
},
Ident {
ident: "E",
span: #20 bytes(428..429),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "V",
span: #20 bytes(444..445),
},
Punct {
ch: '=',
spacing: Alone,
span: #20 bytes(446..447),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "let",
span: #20 bytes(450..453),
},
Ident {
ident: "_",
span: #20 bytes(454..455),
},
Punct {
ch: '=',
spacing: Alone,
span: #20 bytes(456..457),
},
Group {
delimiter: Brace,
stream: TokenStream [
Group {
delimiter: None,
stream: TokenStream [
Literal {
kind: Integer,
symbol: "0",
suffix: None,
span: #0 bytes(636..637),
},
Punct {
ch: '+',
spacing: Alone,
span: #0 bytes(638..639),
},
Literal {
kind: Integer,
symbol: "1",
suffix: None,
span: #0 bytes(640..641),
},
],
span: #20 bytes(460..465),
},
],
span: #20 bytes(458..467),
},
Punct {
ch: ';',
spacing: Alone,
span: #20 bytes(467..468),
},
Literal {
kind: Integer,
symbol: "0",
suffix: None,
span: #20 bytes(469..470),
},
],
span: #20 bytes(448..472),
},
Punct {
ch: ',',
spacing: Alone,
span: #20 bytes(472..473),
},
],
span: #20 bytes(430..483),
},
]
PRINT-DERIVE INPUT (DISPLAY): enum E { V = { let _ = { PATH + 1; } ; 0 }, }
PRINT-DERIVE DEEP-RE-COLLECTED (DISPLAY): enum E { V = { let _ = { PATH + 1 } ; 0 }, }
PRINT-DERIVE INPUT (DEBUG): TokenStream [
Ident {
ident: "enum",
span: #24 bytes(423..427),
},
Ident {
ident: "E",
span: #24 bytes(428..429),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "V",
span: #24 bytes(444..445),
},
Punct {
ch: '=',
spacing: Alone,
span: #24 bytes(446..447),
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "let",
span: #24 bytes(450..453),
},
Ident {
ident: "_",
span: #24 bytes(454..455),
},
Punct {
ch: '=',
spacing: Alone,
span: #24 bytes(456..457),
},
Group {
delimiter: Brace,
stream: TokenStream [
Group {
delimiter: None,
stream: TokenStream [
Ident {
ident: "PATH",
span: #0 bytes(658..662),
},
Punct {
ch: '+',
spacing: Alone,
span: #0 bytes(663..664),
},
Literal {
kind: Integer,
symbol: "1",
suffix: None,
span: #0 bytes(665..666),
},
],
span: #24 bytes(460..465),
},
],
span: #24 bytes(458..467),
},
Punct {
ch: ';',
spacing: Alone,
span: #24 bytes(467..468),
},
Literal {
kind: Integer,
symbol: "0",
suffix: None,
span: #24 bytes(469..470),
},
],
span: #24 bytes(448..472),
},
Punct {
ch: ',',
spacing: Alone,
span: #24 bytes(472..473),
},
],
span: #24 bytes(430..483),
},
]