mirror of
https://github.com/carbon-language/carbon-lang.git
synced 2026-09-25 11:40:10 +01:00
Jon Ross-Perkins
7c7afc9e32
This leaves a single state for each in the expr loop. I was trying to think through ways to have per-token states, but they felt sort of bulky. Note this is more verbose: but I think the long-term is going to be that when we start wanting to add handlers, we're going to need to switch to different names based on the token found. As a consequence, the parse state logic will end up diverging a little, and we'll just want to align towards boilerplate handlers. Short-term, this opens up a path for saying that each parse node corresponds to precisely one token in success states, and separates out what were becoming big handler functions in check.
24 lines
1.1 KiB
Plaintext
24 lines
1.1 KiB
Plaintext
// Part of the Carbon Language project, under the Apache License v2.0 with LLVM
|
|
// Exceptions. See /LICENSE for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
//
|
|
// AUTOUPDATE
|
|
|
|
fn foo(addr a: i32*);
|
|
|
|
// CHECK:STDOUT: - filename: addr.carbon
|
|
// CHECK:STDOUT: parse_tree: [
|
|
// CHECK:STDOUT: {kind: 'FileStart', text: ''},
|
|
// CHECK:STDOUT: {kind: 'FunctionIntroducer', text: 'fn'},
|
|
// CHECK:STDOUT: {kind: 'IdentifierName', text: 'foo'},
|
|
// CHECK:STDOUT: {kind: 'TuplePatternStart', text: '('},
|
|
// CHECK:STDOUT: {kind: 'IdentifierName', text: 'a'},
|
|
// CHECK:STDOUT: {kind: 'IntTypeLiteral', text: 'i32'},
|
|
// CHECK:STDOUT: {kind: 'PostfixOperatorStar', text: '*', subtree_size: 2},
|
|
// CHECK:STDOUT: {kind: 'BindingPattern', text: ':', subtree_size: 4},
|
|
// CHECK:STDOUT: {kind: 'Address', text: 'addr', subtree_size: 5},
|
|
// CHECK:STDOUT: {kind: 'TuplePattern', text: ')', subtree_size: 7},
|
|
// CHECK:STDOUT: {kind: 'FunctionDecl', text: ';', subtree_size: 10},
|
|
// CHECK:STDOUT: {kind: 'FileEnd', text: ''},
|
|
// CHECK:STDOUT: ]
|