mirror of
https://github.com/carbon-language/carbon-lang.git
synced 2026-10-06 07:24:42 +01:00
Initial support for expression parsing. (#455)
This commit is contained in:
+53
-19
@@ -58,9 +58,8 @@ class ParseTree::Parser {
|
||||
|
||||
// Start parsing one (or more) subtrees of nodes.
|
||||
//
|
||||
// This returns a marker representing start position. It will also enforce
|
||||
// that at least *some* node is added using this starting position. Multiple
|
||||
// nodes can be added if they share a start position though.
|
||||
// This returns a marker representing start position. Multiple nodes can be
|
||||
// added if they share a start position.
|
||||
auto StartSubtree() -> SubtreeStart;
|
||||
|
||||
// Add a node to the parse tree that potentially has a subtree larger than
|
||||
@@ -69,7 +68,7 @@ class ParseTree::Parser {
|
||||
// Requires a start marker be passed to compute the size of the subtree rooted
|
||||
// at this node.
|
||||
auto AddNode(ParseNodeKind n_kind, TokenizedBuffer::Token t,
|
||||
SubtreeStart& start, bool has_error = false) -> Node;
|
||||
SubtreeStart start, bool has_error = false) -> Node;
|
||||
|
||||
// If the current token is an opening symbol for a matched group, skips
|
||||
// forward to one past the matched closing symbol and returns true. Otherwise,
|
||||
@@ -79,28 +78,33 @@ class ParseTree::Parser {
|
||||
// Skip forward to the token immediately after the given token.
|
||||
auto SkipTo(TokenizedBuffer::Token t) -> void;
|
||||
|
||||
// Skips forward to move past the likely end of a declaration.
|
||||
// Find the next token of the given kind at the current bracketing level.
|
||||
auto FindNext(TokenKind t) -> llvm::Optional<TokenizedBuffer::Token>;
|
||||
|
||||
// Callback used if we find a semicolon when skipping to the end of a
|
||||
// declaration or statement.
|
||||
using SemiHandler = llvm::function_ref<
|
||||
auto(TokenizedBuffer::Token semi)->llvm::Optional<Node>>;
|
||||
|
||||
// Skips forward to move past the likely end of a declaration or statement.
|
||||
//
|
||||
// Looks forward, skipping over any matched symbol groups, to find the next
|
||||
// position that is likely past the end of a declaration. This is a heuristic
|
||||
// and should only be called when skipping past parse errors.
|
||||
// position that is likely past the end of a declaration or statement. This
|
||||
// is a heuristic and should only be called when skipping past parse errors.
|
||||
//
|
||||
// The strategy for recognizing when we have likely passed the end of a
|
||||
// declaration:
|
||||
// - If we get to close curly brace, we likely ended the entire context of
|
||||
// declarations.
|
||||
// - If we get to a semicolon, that should have ended the declaration.
|
||||
// declaration or statement:
|
||||
// - If we get to close curly brace, we likely ended the entire context.
|
||||
// - If we get to a semicolon, that should have ended the declaration or
|
||||
// statement.
|
||||
// - If we get to a new line from the `SkipRoot` token, but with the same or
|
||||
// less indentation, there is likely a missing semicolon. Continued
|
||||
// declarations across multiple lines should be indented.
|
||||
// declarations or statements across multiple lines should be indented.
|
||||
//
|
||||
// If we find a semicolon based on this skipping, we try to build a parse node
|
||||
// to represent it and will return that node. Otherwise we will return an
|
||||
// empty optional. If `IsInsideDeclaration` is true (the default) we build a
|
||||
// node that marks the end of the declaration we are inside. Otherwise we
|
||||
// build an empty declaration node.
|
||||
auto SkipPastLikelyDeclarationEnd(TokenizedBuffer::Token skip_root,
|
||||
bool is_inside_declaration = true)
|
||||
// If we find a semicolon based on this skipping, we call `on_semi_` to try
|
||||
// to build a parse node to represent it, and will return that node.
|
||||
// Otherwise we will return an empty optional.
|
||||
auto SkipPastLikelyEnd(TokenizedBuffer::Token skip_root, SemiHandler on_semi)
|
||||
-> llvm::Optional<Node>;
|
||||
|
||||
// Parses the signature of the function, consisting of a parameter list and an
|
||||
@@ -126,6 +130,36 @@ class ParseTree::Parser {
|
||||
// even when a node is returned.
|
||||
auto ParseDeclaration() -> llvm::Optional<Node>;
|
||||
|
||||
// Parses a parenthesized expression.
|
||||
auto ParseParenExpression() -> llvm::Optional<Node>;
|
||||
|
||||
// Parses a primary expression, which is either a terminal portion of an
|
||||
// expression tree, such as an identifier or literal, or a parenthesized
|
||||
// expression.
|
||||
auto ParsePrimaryExpression() -> llvm::Optional<Node>;
|
||||
|
||||
// Parses a designator expression suffix starting with `.`.
|
||||
auto ParseDesignatorExpression(SubtreeStart start, bool has_errors)
|
||||
-> llvm::Optional<Node>;
|
||||
|
||||
// Parses a call expression suffix starting with `(`.
|
||||
auto ParseCallExpression(SubtreeStart start, bool has_errors)
|
||||
-> llvm::Optional<Node>;
|
||||
|
||||
// Parses a postfix expression, which is a primary expression followed by
|
||||
// zero or more of the following:
|
||||
//
|
||||
// - function applications
|
||||
// - array indexes (TODO)
|
||||
// - designators
|
||||
auto ParsePostfixExpression() -> llvm::Optional<Node>;
|
||||
|
||||
// Parses an expression.
|
||||
auto ParseExpression() -> llvm::Optional<Node>;
|
||||
|
||||
// Parses an expression statement: an expression followed by a semicolon.
|
||||
auto ParseExpressionStatement() -> llvm::Optional<Node>;
|
||||
|
||||
ParseTree& tree;
|
||||
TokenizedBuffer& tokens;
|
||||
TokenDiagnosticEmitter& emitter;
|
||||
|
||||
Reference in New Issue
Block a user