Files
carbon-lang/toolchain/parser/parser_state.def
T
Jon Ross-PerkinsandChandler Carruth 84deb62aef Work on ParseTree structure to use more bracketed structures. (#2416)
This works on multiple statements to make them better for the bracketing model. Stub nodes are added in more cases of invalid syntax, simply so that the semantics has reliably structured input. Comments in parse_node_kind.def now try to show the expected parse tree structure in postorder form.

This labels If, While, and For a little differently in parse nodes so that at the start of the postorder traversal, it'll already be available to semantics which structure is being processed. I need to do a little more with If in particular, but this felt like a reasonable stopping point.

While this makes significant parser changes, the changes to parser_state.def are minimal, mostly naming-related. The actual flow isn't substantively changed, just a couple minor names and the new As(If|While) state which allows distinguishing IfCondition and WhileCondition.

Co-authored-by: Chandler Carruth <chandlerc@gmail.com>
2022-11-28 15:02:59 -08:00

561 lines
14 KiB
Modula-2

// 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
//
// Note that this is an X-macro header.
//
// It does not use `#include` guards, and instead is designed to be `#include`ed
// after the x-macro is defined in order for its inclusion to expand to the
// desired output. The x-macro for this header is `CARBON_PARSE_NODE_KIND`. The
// definition provided will be removed at the end of this file to clean up.
#ifndef CARBON_PARSER_STATE
#error "Must define the x-macro to use this file."
#endif
// Handles the `{` of a brace expression.
//
// If `CloseCurlyBrace`:
// 1. BraceExpressionFinishAsUnknown
// Else:
// 1. BraceExpressionParameterAsUnknown
// 2. BraceExpressionFinishAsUnknown
CARBON_PARSER_STATE(BraceExpression)
// Handles a brace expression parameter. Note this will always start as unknown,
// but should be known after the first valid parameter. All later inconsistent
// parameters are invalid.
//
// If valid:
// 1. DesignatorExpressionAsStruct
// 2. BraceExpressionParameterAfterDesignatorAs(Type|Value|Unknown)
// Else:
// 1. BraceExpressionParameterFinishAs(Type|Value|Unknown)
CARBON_PARSER_STATE(BraceExpressionParameterAsType)
CARBON_PARSER_STATE(BraceExpressionParameterAsValue)
CARBON_PARSER_STATE(BraceExpressionParameterAsUnknown)
// Handles a brace expression parameter after the initial designator. This
// should be at a `:` or `=`, depending on whether it's a type or value literal.
//
// If valid:
// 1. Expression
// 2. BraceExpressionParameterFinishAs(Type|Value|Unknown)
// Else:
// 1. BraceExpressionParameterFinishAs(Type|Value|Unknown)
CARBON_PARSER_STATE(BraceExpressionParameterAfterDesignatorAsType)
CARBON_PARSER_STATE(BraceExpressionParameterAfterDesignatorAsValue)
CARBON_PARSER_STATE(BraceExpressionParameterAfterDesignatorAsUnknown)
// Handles the end of a brace expression parameter.
//
// If `Comma`:
// 1. BraceExpressionParameterAsUnknown
// Else:
// (state done)
CARBON_PARSER_STATE(BraceExpressionParameterFinishAsType)
CARBON_PARSER_STATE(BraceExpressionParameterFinishAsValue)
CARBON_PARSER_STATE(BraceExpressionParameterFinishAsUnknown)
// Handles the `}` of a brace expression.
//
// Always:
// (state done)
CARBON_PARSER_STATE(BraceExpressionFinishAsType)
CARBON_PARSER_STATE(BraceExpressionFinishAsValue)
CARBON_PARSER_STATE(BraceExpressionFinishAsUnknown)
// Handles a call expression `(...)`.
//
// If `CloseParen`:
// 1. CallExpressionFinish
// Else:
// 1. Expression
// 2. CallExpressionParameterFinish
// 3. CallExpressionFinish
CARBON_PARSER_STATE(CallExpression)
// Handles the `,` or `)` after a call parameter.
//
// If `Comma`:
// 1. Expression
// 2. CallExpressionParameterFinish
// Else:
// (state done)
CARBON_PARSER_STATE(CallExpressionParameterFinish)
// Handles finishing the call expression.
//
// Always:
// (state done)
CARBON_PARSER_STATE(CallExpressionFinish)
// Handles processing at the `{` on a typical code block.
//
// If `OpenCurlyBrace`:
// 1. StatementScopeLoop
// 2. CodeBlockFinish
// Else:
// 1. Statement
// 2. CodeBlockFinish
CARBON_PARSER_STATE(CodeBlock)
// Handles processing at the `}` on a typical code block, after a statement
// scope is done.
//
// Always:
// (state done)
CARBON_PARSER_STATE(CodeBlockFinish)
// Handles processing of a declaration scope. Things like fn, class, interface,
// and so on.
//
// If `EndOfFile`:
// (state done)
// If `Fn`:
// 1. FunctionIntroducer
// 2. DeclarationLoop
// If `Package`:
// 1. Package
// 2. DeclarationLoop
// If `Semi`:
// 1. DeclarationLoop
// If `Var`:
// 1. Var
// 2. DeclarationLoop
// If `interface`:
// 1. InterfaceIntroducer
// 2. DeclarationLoop
// Else:
// 1. DeclarationLoop
CARBON_PARSER_STATE(DeclarationLoop)
// Handles a designator expression, such as `.z` in `x.(y.z)`.
//
// Always:
// (state done)
CARBON_PARSER_STATE(DesignatorAsExpression)
CARBON_PARSER_STATE(DesignatorAsStruct)
// Handles processing of an expression.
//
// If valid prefix operator:
// 1. Expression
// 2. ExpressionLoopForPrefix
// Else:
// 1. ExpressionInPostfix
// 2. ExpressionLoop
CARBON_PARSER_STATE(Expression)
// Handles the initial part of postfix expressions, such as an identifier or
// literal value, then proceeds to the loop.
//
// If `Identifier` or literal (including type literals):
// 1. ExpressionInPostfixLoop
// If `OpenCurlyBrace`:
// 1. BraceExpression
// 2. ExpressionInPostfixLoop
// If `OpenParen`:
// 1. ParenExpression
// 2. ExpressionInPostfixLoop
// Else:
// (state done)
CARBON_PARSER_STATE(ExpressionInPostfix)
// Handles looping through elements following the initial postfix expression,
// such as designators or parenthesized parameters.
//
// If `Period`:
// 1. DesignatorAsExpression
// 2. ExpressionInPostfixLoop
// If `OpenParen`:
// 1. CallExpression
// 2. ExpressionInPostfixLoop
// Else:
// (state done)
CARBON_PARSER_STATE(ExpressionInPostfixLoop)
// Handles processing of an expression.
//
// If binary operator:
// 1. Expression
// 2. ExpressionLoopForBinary
// If postfix operator:
// 1. ExpressionLoop
// Else:
// (state done)
CARBON_PARSER_STATE(ExpressionLoop)
// Completes an ExpressionLoop pass by adding an infix operator, then goes back
// to ExpressionLoop.
//
// Always:
// 1. ExpressionLoop
CARBON_PARSER_STATE(ExpressionLoopForBinary)
// Completes an ExpressionLoop pass by adding a prefix operator, then goes back
// to ExpressionLoop.
//
// Always:
// 1. ExpressionLoop
CARBON_PARSER_STATE(ExpressionLoopForPrefix)
// Handles the `;` for an expression statement, which is different from most
// keyword statements.
//
// Always:
// (state done)
CARBON_PARSER_STATE(ExpressionStatementFinish)
// Handles processing of a function's `fn <name>(`, and enqueues parameter list
// handling.
//
// If invalid:
// (state done)
// If parenthesized parameters:
// 1. PatternAsFunctionParameter
// 2. FunctionParameterListFinish
// 3. FunctionAfterParameterList
// Else:
// 1. FunctionParameterListFinish
// 2. FunctionAfterParameterList
CARBON_PARSER_STATE(FunctionIntroducer)
// Starts function parameter processing.
//
// Always:
// 1. PatternAsFunctionParameter
// 2. FunctionParameterFinish
CARBON_PARSER_STATE(FunctionParameter)
// Finishes function parameter processing, including `,`. If there are more
// parameters, enqueues another parameter processing state.
//
// If `Comma` without `CloseParen`:
// 1. FunctionParameter
// Else:
// (state done)
CARBON_PARSER_STATE(FunctionParameterFinish)
// Handles processing of a function's parameter list `)`.
//
// Always:
// (state done)
CARBON_PARSER_STATE(FunctionParameterListFinish)
// Handles processing of a function's syntax after `)`, primarily the
// possibility a `->` return type is there. Always enqueues signature finish
// handling.
//
// If `MinusGreater`:
// 1. Expression
// 2. FunctionReturnTypeFinish
// 3. FunctionSignatureFinish
// Else:
// 1. FunctionSignatureFinish
CARBON_PARSER_STATE(FunctionAfterParameterList)
// Finishes a function return type.
//
// Always:
// (state done)
CARBON_PARSER_STATE(FunctionReturnTypeFinish)
// Finishes a function signature. If it's a declaration, the function is done;
// otherwise, this also starts definition processing.
//
// If `Semi`:
// (state done)
// If `OpenCurlyBrace`:
// 1. StatementScopeLoop
// 2. FunctionDefinitionFinish
// Else:
// (state done)
CARBON_PARSER_STATE(FunctionSignatureFinish)
// Finishes a function definition.
//
// Always:
// (state done)
CARBON_PARSER_STATE(FunctionDefinitionFinish)
// Finishes an interface definition.
//
// Always:
// (state done)
CARBON_PARSER_STATE(InterfaceDefinitionFinish)
// Handles parsing the body of an interface.
//
// If `}`:
// (state done)
// Else:
// 1. InterfaceDefinitionLoop
CARBON_PARSER_STATE(InterfaceDefinitionLoop)
// Handles processing of a intefaces's `interface <name> {`.
//
// If invalid:
// 1. InterfaceDefinitionFinish
// If `{` is missing:
// 1. InterfaceDefinitionFinish
// Else:
// 1. InterfaceDefinitionLoop
// 2. InterfaceDefinitionFinish
CARBON_PARSER_STATE(InterfaceIntroducer)
// Handles `package`.
//
// Always:
// (state done)
CARBON_PARSER_STATE(Package)
// Handles the processing of a `(condition)` up through the expression.
//
// Always:
// 1. Expression
// 2. ParenConditionAs(If|While)Finish
CARBON_PARSER_STATE(ParenConditionAsIf)
CARBON_PARSER_STATE(ParenConditionAsWhile)
// Finishes the processing of a `(condition)` after the expression.
//
// Always:
// (state done)
CARBON_PARSER_STATE(ParenConditionFinishAsIf)
CARBON_PARSER_STATE(ParenConditionFinishAsWhile)
// Handles the `(` of a parenthesized expression.
//
// If `CloseParen`:
// 1. ParenExpressionFinishAsTuple
// Else:
// 1. Expression
// 2. ParenExpressionParameterFinishAsUnknown
// 3. ParenExpressionFinish
CARBON_PARSER_STATE(ParenExpression)
// Handles the end of a parenthesized expression's parameter. This will start as
// AsUnknown on the first parameter; if there are more, it switches to AsTuple
// processing.
//
// If `Comma` without `CloseParen`:
// 1. Expression
// 2. ParenExpressionParameterFinishAsTuple
// SPECIAL: Parent becomes ParenExpressionFinishAsTuple
// If `Comma` with `CloseParen`:
// (state done)
// SPECIAL: Parent becomes ParenExpressionFinishAsTuple
// Else `CloseParen`:
// (state done)
CARBON_PARSER_STATE(ParenExpressionParameterFinishAsUnknown)
CARBON_PARSER_STATE(ParenExpressionParameterFinishAsTuple)
// Handles the `)` of a parenthesized expression.
//
// Always:
// (state done)
CARBON_PARSER_STATE(ParenExpressionFinish)
CARBON_PARSER_STATE(ParenExpressionFinishAsTuple)
// Handles pattern parsing for a pattern, enqueuing type expression processing.
// This covers function parameter and `var` support.
//
// If valid:
// 1. Expression
// 2. PatternFinish
// Else:
// 1. PatternFinish
CARBON_PARSER_STATE(PatternAsFunctionParameter)
CARBON_PARSER_STATE(PatternAsVariable)
// Finishes pattern processing.
//
// Always:
// (state done)
CARBON_PARSER_STATE(PatternFinish)
// Handles a single statement. While typically within a statement block, this
// can also be used for error recovery where we expect a statement block and
// are missing braces.
//
// If `Break`:
// 1. StatementBreakFinish
// (state done)
// If `Continue`:
// 1. StatementContinueFinish
// (state done)
// If `For`:
// 1. StatementForHeader
// 2. StatementForFinish
// If `If`:
// 1. StatementIf
// If `Return`:
// 1. StatementReturn
// If `Var`:
// 1. VarAsSemicolon
// If `While`:
// 1. StatementWhile
// Else:
// 1. Expression
// 2. ExpressionStatementFinish
CARBON_PARSER_STATE(Statement)
// Handles `break` processing at the `;`.
//
// Always:
// (state done)
CARBON_PARSER_STATE(StatementBreakFinish)
// Handles `continue` processing at the `;`.
//
// Always:
// (state done)
CARBON_PARSER_STATE(StatementContinueFinish)
// Handles `for` processing of `(var`, proceeding to a pattern before
// continuing.
//
// If no `OpenParen`:
// 1. CodeBlock
// If `Var`:
// 1. VarAsFor
// 2. StatementForHeaderIn
// Else:
// 1. StatementForHeaderIn
CARBON_PARSER_STATE(StatementForHeader)
// Handles `for` procesisng of `in`, proceeding to an expression before
// continuing.
//
// If `In` or `Colon`:
// 1. Expression
// 2. StatementForHeaderFinish
// Else:
// 1. StatementForHeaderFinish
CARBON_PARSER_STATE(StatementForHeaderIn)
// Handles `for` processing of `)`, proceeding to the statement block.
//
// Always:
// 1. CodeBlock
CARBON_PARSER_STATE(StatementForHeaderFinish)
// Handles `for` processing at the final `}`.
//
// Always:
// (state done)
CARBON_PARSER_STATE(StatementForFinish)
// Handles `if` processing at the start.
//
// Always:
// 1. ParenConditionAsIf
// 2. StatementIfConditionFinish
CARBON_PARSER_STATE(StatementIf)
// Handles `if` processing between the condition and start of the first code
// block.
//
// Always:
// 1. CodeBlock
// 2. StatementIfThenBlockFinish
CARBON_PARSER_STATE(StatementIfConditionFinish)
// Handles `if` processing after the end of the first code block, with the
// optional `else`.
//
// If `Else` then `If`:
// 1. CodeBlock
// 2. StatementIfElseBlockFinish
// If `Else`:
// 1. StatementIf
// 2. StatementIfElseBlockFinish
// Else:
// (state done)
CARBON_PARSER_STATE(StatementIfThenBlockFinish)
// Handles `if` processing after a provided `else` code block.
//
// Always:
// (state done)
CARBON_PARSER_STATE(StatementIfElseBlockFinish)
// Handles `return` processing.
//
// If `Semi`:
// 1. StatementReturnFinish
// Else:
// 1. Expression
// 2. StatementReturnFinish
CARBON_PARSER_STATE(StatementReturn)
// Handles `return` processing at the `;` when there's an expression.
//
// Always:
// (state done)
CARBON_PARSER_STATE(StatementReturnFinish)
// Handles processing of statements within a scope.
//
// If `CloseCurlyBrace`:
// (state done)
// Else:
// 1. Statement
// 2. StatementScopeLoop
CARBON_PARSER_STATE(StatementScopeLoop)
// Handles `while` processing.
//
// Always:
// 1. ParenConditionAsWhile
// 2. StatementWhileConditionFinish
CARBON_PARSER_STATE(StatementWhile)
// Handles `while` processing between the condition and start of the code block.
//
// Always:
// 1. CodeBlock
// 2. StatementWhileBlockFinish
CARBON_PARSER_STATE(StatementWhileConditionFinish)
// Handles `while` processing after the end of the code block.
//
// Always:
// (state done)
CARBON_PARSER_STATE(StatementWhileBlockFinish)
// Handles the start of a `var`.
//
// Always:
// 1. PatternAsVariable
// 2. VarAfterPattern
// 3. VarFinishAs(Semicolon|For)
CARBON_PARSER_STATE(VarAsSemicolon)
CARBON_PARSER_STATE(VarAsFor)
// Handles `var` after the pattern, either followed by an initializer or the
// semicolon.
//
// If `Equal`:
// 1. Expression
// 2. VarAfterInitializer
// Else:
// (state done)
CARBON_PARSER_STATE(VarAfterPattern)
// Handles `var` after the initializer, wrapping up its subtree.
//
// Always:
// (state done)
CARBON_PARSER_STATE(VarAfterInitializer)
// Handles `var` parsing at the end.
//
// Always:
// (state done)
CARBON_PARSER_STATE(VarFinishAsSemicolon)
CARBON_PARSER_STATE(VarFinishAsFor)
#undef CARBON_PARSER_STATE