Files
carbon-lang/explorer/syntax/lexer.lpp
T
2023-06-15 08:42:14 -07:00

452 lines
18 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
*/
%{
#include <cstdlib>
#include "common/check.h"
#include "common/error.h"
#include "explorer/syntax/lex_helper.h"
#include "explorer/syntax/lex_scan_helper.h"
#include "explorer/syntax/parse_and_lex_context.h"
#include "explorer/syntax/parser.h"
#include "llvm/ADT/StringExtras.h"
#include "llvm/Support/FormatVariadic.h"
%}
/* Turn off legacy bits we don't need. */
%option noyywrap nounput nodefault
%option reentrant
/* Lexing a token immediately after consuming some whitespace. */
%s AFTER_WHITESPACE
/*
* Lexing a token immediately after consuming an operand-ending token:
* a closing bracket, identifier, or literal.
*/
%s AFTER_OPERAND
/* table-begin */
ABSTRACT "abstract"
ADDR "addr"
ALIAS "alias"
AMPERSAND "&"
AMPERSAND_EQUAL "&="
AND "and"
API "api"
ARROW "->"
AS "as"
AUTO "auto"
BASE "base"
BOOL "bool"
BREAK "break"
CARET "^"
CARET_EQUAL "^="
CASE "case"
CHOICE "choice"
CLASS "class"
COLON ":"
COLON_BANG ":!"
COMMA ","
CONSTRAINT "constraint"
CONTINUE "continue"
DEFAULT "default"
DESTRUCTOR "destructor"
DOUBLE_ARROW "=>"
ELSE "else"
EQUAL "="
EQUAL_EQUAL "=="
EXTEND "extend"
FALSE "false"
FN "fn"
FN_TYPE "__Fn"
FOR "for"
FORALL "forall"
GREATER ">"
GREATER_EQUAL ">="
GREATER_GREATER ">>"
GREATER_GREATER_EQUAL ">>="
IF "if"
IMPL "impl"
IMPLS "impls"
IMPORT "import"
IN "in"
INTERFACE "interface"
LEFT_CURLY_BRACE "{"
LEFT_PARENTHESIS "("
LEFT_SQUARE_BRACKET "["
LESS "<"
LESS_EQUAL "<="
LESS_LESS "<<"
LESS_LESS_EQUAL "<<="
LET "let"
LIBRARY "library"
MATCH "match"
MATCH_FIRST "__match_first"
MINUS "-"
MINUS_EQUAL "-="
MINUS_MINUS "--"
MIX "__mix"
MIXIN "__mixin"
NAMESPACE "namespace"
NOT "not"
NOT_EQUAL "!="
OR "or"
PACKAGE "package"
PERCENT "%"
PERCENT_EQUAL "%="
PERIOD "."
PIPE "|"
PIPE_EQUAL "|="
PLUS "+"
PLUS_EQUAL "+="
PLUS_PLUS "++"
REQUIRE "require"
RETURN "return"
RETURNED "returned"
RIGHT_CURLY_BRACE "}"
RIGHT_PARENTHESIS ")"
RIGHT_SQUARE_BRACKET "]"
SELF "Self"
SEMICOLON ";"
SLASH "/"
SLASH_EQUAL "/="
STAR_EQUAL "*="
STRING "String"
TEMPLATE "template"
THEN "then"
TRUE "true"
TYPE "type"
UNDERSCORE "_"
UNIMPL_EXAMPLE "__unimplemented_example_infix"
VAR "var"
VIRTUAL "virtual"
WHERE "where"
WHILE "while"
/* table-end */
/* This should be kept table-like, but isn't automatic due to spaces. */
identifier [A-Za-z_][A-Za-z0-9_]*
/* TODO: Remove Print special casing once we have variadics or overloads. */
intrinsic_identifier (Print|__intrinsic_[A-Za-z0-9_]*)
sized_type_literal [iuf][1-9][0-9]*
integer_literal [0-9]+
horizontal_whitespace [ \t\r]
whitespace [ \t\r\n]
one_line_comment \/\/[^\n]*\n
operand_start [(A-Za-z0-9_\"]
%%
%{
// Code run each time yylex is called.
// Begin with an empty token span starting where its previous end was.
context.current_token_position.step();
%}
/* table-begin */
{ABSTRACT} { return CARBON_SIMPLE_TOKEN(ABSTRACT); }
{ADDR} { return CARBON_SIMPLE_TOKEN(ADDR); }
{ALIAS} { return CARBON_SIMPLE_TOKEN(ALIAS); }
{AMPERSAND_EQUAL} { return CARBON_SIMPLE_TOKEN(AMPERSAND_EQUAL); }
{AMPERSAND} { return CARBON_SIMPLE_TOKEN(AMPERSAND); }
{AND} { return CARBON_SIMPLE_TOKEN(AND); }
{API} { return CARBON_SIMPLE_TOKEN(API); }
{ARROW} { return CARBON_SIMPLE_TOKEN(ARROW); }
{AS} { return CARBON_SIMPLE_TOKEN(AS); }
{AUTO} { return CARBON_SIMPLE_TOKEN(AUTO); }
{BASE} { return CARBON_SIMPLE_TOKEN(BASE); }
{BOOL} { return CARBON_SIMPLE_TOKEN(BOOL); }
{BREAK} { return CARBON_SIMPLE_TOKEN(BREAK); }
{CARET_EQUAL} { return CARBON_SIMPLE_TOKEN(CARET_EQUAL); }
{CARET} { return CARBON_SIMPLE_TOKEN(CARET); }
{CASE} { return CARBON_SIMPLE_TOKEN(CASE); }
{CHOICE} { return CARBON_SIMPLE_TOKEN(CHOICE); }
{CLASS} { return CARBON_SIMPLE_TOKEN(CLASS); }
{COLON_BANG} { return CARBON_SIMPLE_TOKEN(COLON_BANG); }
{COLON} { return CARBON_SIMPLE_TOKEN(COLON); }
{COMMA} { return CARBON_SIMPLE_TOKEN(COMMA); }
{CONSTRAINT} { return CARBON_SIMPLE_TOKEN(CONSTRAINT); }
{CONTINUE} { return CARBON_SIMPLE_TOKEN(CONTINUE); }
{DEFAULT} { return CARBON_SIMPLE_TOKEN(DEFAULT); }
{DESTRUCTOR} { return CARBON_SIMPLE_TOKEN(DESTRUCTOR); }
{DOUBLE_ARROW} { return CARBON_SIMPLE_TOKEN(DOUBLE_ARROW); }
{ELSE} { return CARBON_SIMPLE_TOKEN(ELSE); }
{EQUAL_EQUAL} { return CARBON_SIMPLE_TOKEN(EQUAL_EQUAL); }
{EQUAL} { return CARBON_SIMPLE_TOKEN(EQUAL); }
{EXTEND} { return CARBON_SIMPLE_TOKEN(EXTEND); }
{FALSE} { return CARBON_SIMPLE_TOKEN(FALSE); }
{FN_TYPE} { return CARBON_SIMPLE_TOKEN(FN_TYPE); }
{FN} { return CARBON_SIMPLE_TOKEN(FN); }
{FORALL} { return CARBON_SIMPLE_TOKEN(FORALL); }
{FOR} { return CARBON_SIMPLE_TOKEN(FOR); }
{GREATER_EQUAL} { return CARBON_SIMPLE_TOKEN(GREATER_EQUAL); }
{GREATER_GREATER_EQUAL} { return CARBON_SIMPLE_TOKEN(GREATER_GREATER_EQUAL); }
{GREATER_GREATER} { return CARBON_SIMPLE_TOKEN(GREATER_GREATER); }
{GREATER} { return CARBON_SIMPLE_TOKEN(GREATER); }
{IF} { return CARBON_SIMPLE_TOKEN(IF); }
{IMPLS} { return CARBON_SIMPLE_TOKEN(IMPLS); }
{IMPL} { return CARBON_SIMPLE_TOKEN(IMPL); }
{IMPORT} { return CARBON_SIMPLE_TOKEN(IMPORT); }
{INTERFACE} { return CARBON_SIMPLE_TOKEN(INTERFACE); }
{IN} { return CARBON_SIMPLE_TOKEN(IN); }
{LEFT_CURLY_BRACE} { return CARBON_SIMPLE_TOKEN(LEFT_CURLY_BRACE); }
{LEFT_PARENTHESIS} { return CARBON_SIMPLE_TOKEN(LEFT_PARENTHESIS); }
{LEFT_SQUARE_BRACKET} { return CARBON_SIMPLE_TOKEN(LEFT_SQUARE_BRACKET); }
{LESS_EQUAL} { return CARBON_SIMPLE_TOKEN(LESS_EQUAL); }
{LESS_LESS_EQUAL} { return CARBON_SIMPLE_TOKEN(LESS_LESS_EQUAL); }
{LESS_LESS} { return CARBON_SIMPLE_TOKEN(LESS_LESS); }
{LESS} { return CARBON_SIMPLE_TOKEN(LESS); }
{LET} { return CARBON_SIMPLE_TOKEN(LET); }
{LIBRARY} { return CARBON_SIMPLE_TOKEN(LIBRARY); }
{MATCH_FIRST} { return CARBON_SIMPLE_TOKEN(MATCH_FIRST); }
{MATCH} { return CARBON_SIMPLE_TOKEN(MATCH); }
{MINUS_EQUAL} { return CARBON_SIMPLE_TOKEN(MINUS_EQUAL); }
{MINUS_MINUS} { return CARBON_SIMPLE_TOKEN(MINUS_MINUS); }
{MINUS} { return CARBON_SIMPLE_TOKEN(MINUS); }
{MIXIN} { return CARBON_SIMPLE_TOKEN(MIXIN); }
{MIX} { return CARBON_SIMPLE_TOKEN(MIX); }
{NAMESPACE} { return CARBON_SIMPLE_TOKEN(NAMESPACE); }
{NOT_EQUAL} { return CARBON_SIMPLE_TOKEN(NOT_EQUAL); }
{NOT} { return CARBON_SIMPLE_TOKEN(NOT); }
{OR} { return CARBON_SIMPLE_TOKEN(OR); }
{PACKAGE} { return CARBON_SIMPLE_TOKEN(PACKAGE); }
{PERCENT_EQUAL} { return CARBON_SIMPLE_TOKEN(PERCENT_EQUAL); }
{PERCENT} { return CARBON_SIMPLE_TOKEN(PERCENT); }
{PERIOD} { return CARBON_SIMPLE_TOKEN(PERIOD); }
{PIPE_EQUAL} { return CARBON_SIMPLE_TOKEN(PIPE_EQUAL); }
{PIPE} { return CARBON_SIMPLE_TOKEN(PIPE); }
{PLUS_EQUAL} { return CARBON_SIMPLE_TOKEN(PLUS_EQUAL); }
{PLUS_PLUS} { return CARBON_SIMPLE_TOKEN(PLUS_PLUS); }
{PLUS} { return CARBON_SIMPLE_TOKEN(PLUS); }
{REQUIRE} { return CARBON_SIMPLE_TOKEN(REQUIRE); }
{RETURNED} { return CARBON_SIMPLE_TOKEN(RETURNED); }
{RETURN} { return CARBON_SIMPLE_TOKEN(RETURN); }
{SELF} { return CARBON_SIMPLE_TOKEN(SELF); }
{SEMICOLON} { return CARBON_SIMPLE_TOKEN(SEMICOLON); }
{SLASH_EQUAL} { return CARBON_SIMPLE_TOKEN(SLASH_EQUAL); }
{SLASH} { return CARBON_SIMPLE_TOKEN(SLASH); }
{STAR_EQUAL} { return CARBON_SIMPLE_TOKEN(STAR_EQUAL); }
{STRING} { return CARBON_SIMPLE_TOKEN(STRING); }
{TEMPLATE} { return CARBON_SIMPLE_TOKEN(TEMPLATE); }
{THEN} { return CARBON_SIMPLE_TOKEN(THEN); }
{TRUE} { return CARBON_SIMPLE_TOKEN(TRUE); }
{TYPE} { return CARBON_SIMPLE_TOKEN(TYPE); }
{UNDERSCORE} { return CARBON_SIMPLE_TOKEN(UNDERSCORE); }
{UNIMPL_EXAMPLE} { return CARBON_SIMPLE_TOKEN(UNIMPL_EXAMPLE); }
{VAR} { return CARBON_SIMPLE_TOKEN(VAR); }
{VIRTUAL} { return CARBON_SIMPLE_TOKEN(VIRTUAL); }
{WHERE} { return CARBON_SIMPLE_TOKEN(WHERE); }
{WHILE} { return CARBON_SIMPLE_TOKEN(WHILE); }
/* table-end */
/* More modern Bisons provide make_EOF. */
<<EOF>> { return CARBON_SIMPLE_TOKEN(END_OF_FILE); }
{RIGHT_PARENTHESIS} {
BEGIN(AFTER_OPERAND);
return CARBON_SIMPLE_TOKEN(RIGHT_PARENTHESIS);
}
{RIGHT_CURLY_BRACE} {
BEGIN(AFTER_OPERAND);
return CARBON_SIMPLE_TOKEN(RIGHT_CURLY_BRACE);
}
{RIGHT_SQUARE_BRACKET} {
BEGIN(AFTER_OPERAND);
return CARBON_SIMPLE_TOKEN(RIGHT_SQUARE_BRACKET);
}
/*
* For a `*` operator, we look at whitespace and local context to determine the
* arity and fixity. There are two ways to write a binary operator:
*
* 1) Whitespace on both sides.
* 2) Whitespace on neither side, and the previous token is considered to be
* the end of an operand, and the next token is considered to be the start
* of an operand.
*
* Otherwise, the operator is unary, but we also check for whitespace to help
* the parser enforce the rule that whitespace is not permitted between the
* operator and its operand, leading to three more cases:
*
* 3) Whitespace before (but implicitly not after, because that would give a
* longer match and hit case 1): this can only be a prefix operator.
* 4) Whitespace after and not before: this can only be a postfix operator.
* 5) No whitespace on either side (otherwise the longest match would take us
* to case 4): this is a unary operator and could be either prefix or
* postfix.
*/
/* `*` operator case 1: */
<AFTER_WHITESPACE>"*"{whitespace}+ {
BEGIN(AFTER_WHITESPACE);
return CARBON_SIMPLE_TOKEN(BINARY_STAR);
}
/* `*` operator case 2: */
<AFTER_OPERAND>"*"/{operand_start} { return CARBON_SIMPLE_TOKEN(BINARY_STAR); }
/* `*` operator case 3: */
<AFTER_WHITESPACE>"*" { return CARBON_SIMPLE_TOKEN(PREFIX_STAR); }
/* `*` operator case 4: */
<INITIAL,AFTER_OPERAND>"*"{whitespace}+ {
BEGIN(AFTER_WHITESPACE);
return CARBON_SIMPLE_TOKEN(POSTFIX_STAR);
}
/* `*` operator case 5: */
<INITIAL,AFTER_OPERAND>"*" { return CARBON_SIMPLE_TOKEN(UNARY_STAR); }
{sized_type_literal} {
BEGIN(AFTER_OPERAND);
return CARBON_ARG_TOKEN(sized_type_literal, yytext);
}
{intrinsic_identifier} {
BEGIN(AFTER_OPERAND);
Carbon::ErrorOr<Carbon::IntrinsicExpression::Intrinsic> intrinsic =
Carbon::IntrinsicExpression::FindIntrinsic(yytext, context.source_loc());
if (intrinsic.ok()) {
return CARBON_ARG_TOKEN(intrinsic_identifier, *intrinsic);
} else {
return context.RecordSyntaxError(std::move(intrinsic).error());
}
}
{identifier} {
BEGIN(AFTER_OPERAND);
return CARBON_ARG_TOKEN(identifier, yytext);
}
{integer_literal} {
BEGIN(AFTER_OPERAND);
int val = 0;
if (!llvm::to_integer(yytext, val)) {
return context.RecordSyntaxError(
llvm::formatv("Invalid integer literal: {0}", yytext));
}
return CARBON_ARG_TOKEN(integer_literal, val);
}
#*\" {
// Found a double quote character.
Carbon::StringLexHelper str_lex_helper(yytext, yyscanner, context);
const std::string& s = str_lex_helper.str();
const int hashtag_num = s.find_first_of('"');
if (!str_lex_helper.Advance()) {
return CARBON_SIMPLE_TOKEN(END_OF_FILE);
}
while (!str_lex_helper.is_eof()) {
switch (str_lex_helper.last_char()) {
case '\n': // Fall through.
case '\v': // Fall through.
case '\f': // Fall through.
case '\r':
return context.RecordSyntaxError(
llvm::formatv("missing closing quote in single-line string: {0}",
str_lex_helper.str()));
break;
case '"':
if (Carbon::ReadHashTags(str_lex_helper, hashtag_num)) {
// Reach closing quotes, break out of the loop.
return Carbon::ProcessSingleLineString(str_lex_helper.str(), context,
hashtag_num);
}
break;
case '\\':
if (Carbon::ReadHashTags(str_lex_helper, hashtag_num)) {
// Read the escaped char.
if (!str_lex_helper.Advance()) {
continue;
}
// Read the next char.
str_lex_helper.Advance();
}
break;
default:
str_lex_helper.Advance();
}
}
return CARBON_SIMPLE_TOKEN(END_OF_FILE);
}
#*\'\'\' {
// Multi-line string literal.
Carbon::StringLexHelper str_lex_helper(yytext, yyscanner, context);
const std::string& s = str_lex_helper.str();
const int hashtag_num = s.find_first_of('\'');
if (!str_lex_helper.Advance()) {
return CARBON_SIMPLE_TOKEN(END_OF_FILE);
}
while (!str_lex_helper.is_eof()) {
switch (str_lex_helper.last_char()) {
case '\'':
// Check for 2 more '\''s on block string.
if (!(str_lex_helper.Advance() && str_lex_helper.last_char() == '\'')) {
continue;
}
if (!(str_lex_helper.Advance() && str_lex_helper.last_char() == '\'')) {
continue;
}
// Now we are at the last ' of '''.
if (Carbon::ReadHashTags(str_lex_helper, hashtag_num)) {
// Reach closing quotes, break out of the loop.
return Carbon::ProcessMultiLineString(str_lex_helper.str(), context,
hashtag_num);
}
break;
case '\\':
if (Carbon::ReadHashTags(str_lex_helper, hashtag_num)) {
// Read the escaped char.
if (!str_lex_helper.Advance()) {
continue;
}
// Read the next char.
str_lex_helper.Advance();
}
break;
default:
str_lex_helper.Advance();
}
}
return CARBON_SIMPLE_TOKEN(END_OF_FILE);
}
{one_line_comment} {
// Advance end by 1 line, resetting the column to zero.
context.current_token_position.lines(1);
// Make the span empty by setting start to end.
context.current_token_position.step();
}
{horizontal_whitespace}+ {
// Make the span empty by setting start to end.
context.current_token_position.step();
BEGIN(AFTER_WHITESPACE);
}
\n+ {
// Advance end by yyleng lines, resetting the column to zero.
context.current_token_position.lines(yyleng);
// Make the span empty by setting start to end.
context.current_token_position.step();
BEGIN(AFTER_WHITESPACE);
}
. {
return context.RecordSyntaxError(
llvm::formatv("invalid character '\\x{0}' in source file.",
llvm::toHex(llvm::StringRef(yytext, 1))));
}
%%
auto YyinputWrapper(yyscan_t yyscanner) -> int { return yyinput(yyscanner); }