a0226f9bff
G++ 10.1 emits inappropriate "use of uninitialized data" warnings when
compiling f18. The warnings stem from two sites in templatized code
whose multiple instantiations magnified the number of warnings.
These changes dodge those warnings by making some innocuous changes to
the code. In the parser, the idiom defaulted(cut >> x), which yields a
parser that always succeeds, has been replaced with a new equivalent
pass<T>() parser that returns a default-constructed value T{} in an
arguably more readable fashion. This idiom was the only attestation of
the basic parser cut, so it has been removed and the remaining code
simplified. In Evaluate/traverse.h, a return {}; was replaced with a
return of a default-constructed member.
Differential Revision: https://reviews.llvm.org/D81747
101 lines
4.1 KiB
C++
101 lines
4.1 KiB
C++
//===-- lib/Parser/stmt-parser.h --------------------------------*- C++ -*-===//
|
|
//
|
|
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
|
|
// See https://llvm.org/LICENSE.txt for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#ifndef FORTRAN_PARSER_STMT_PARSER_H_
|
|
#define FORTRAN_PARSER_STMT_PARSER_H_
|
|
|
|
// Basic parsing of statements.
|
|
|
|
#include "basic-parsers.h"
|
|
#include "token-parsers.h"
|
|
|
|
namespace Fortran::parser {
|
|
|
|
// statement(p) parses Statement<P> for some statement type P that is the
|
|
// result type of the argument parser p, while also handling labels and
|
|
// end-of-statement markers.
|
|
|
|
// R611 label -> digit [digit]...
|
|
constexpr auto label{space >> digitString64 / spaceCheck};
|
|
|
|
template <typename PA>
|
|
inline constexpr auto unterminatedStatement(const PA &p) {
|
|
return skipStuffBeforeStatement >>
|
|
sourced(construct<Statement<typename PA::resultType>>(
|
|
maybe(label), space >> p));
|
|
}
|
|
|
|
constexpr auto endOfLine{
|
|
"\n"_ch >> ok || fail("expected end of line"_err_en_US)};
|
|
|
|
constexpr auto semicolons{";"_ch >> skipMany(";"_tok) / space / maybe("\n"_ch)};
|
|
constexpr auto endOfStmt{
|
|
space >> withMessage("expected end of statement"_err_en_US,
|
|
semicolons || endOfLine)};
|
|
constexpr auto forceEndOfStmt{recovery(endOfStmt, SkipPast<'\n'>{})};
|
|
|
|
template <typename PA> inline constexpr auto statement(const PA &p) {
|
|
return unterminatedStatement(p) / endOfStmt;
|
|
}
|
|
|
|
// unlabeledStatement() is basically statement() for those few situations
|
|
// in Fortran where a statement cannot have a label.
|
|
template <typename PA> inline constexpr auto unlabeledStatement(const PA &p) {
|
|
return space >>
|
|
sourced(construct<UnlabeledStatement<typename PA::resultType>>(p));
|
|
}
|
|
|
|
// This unambiguousStatement() variant of statement() provides better error
|
|
// recovery for contexts containing statements that might have trailing
|
|
// garbage, but it must be used only when no instance of the statement in
|
|
// question could also be a legal prefix of some other statement that might
|
|
// be valid at that point. It only makes sense to use this within "some()"
|
|
// or "many()" so as to not end the list of statements.
|
|
template <typename PA> inline constexpr auto unambiguousStatement(const PA &p) {
|
|
return unterminatedStatement(p) / forceEndOfStmt;
|
|
}
|
|
|
|
constexpr auto ignoredStatementPrefix{
|
|
skipStuffBeforeStatement >> maybe(label) >> maybe(name / ":") >> space};
|
|
|
|
// Error recovery within a statement() call: skip *to* the end of the line,
|
|
// unless at an END or CONTAINS statement.
|
|
constexpr auto inStmtErrorRecovery{!"END"_tok >> !"CONTAINS"_tok >>
|
|
SkipTo<'\n'>{} >> construct<ErrorRecovery>()};
|
|
|
|
// Error recovery within statement sequences: skip *past* the end of the line,
|
|
// but not over an END or CONTAINS statement.
|
|
constexpr auto skipStmtErrorRecovery{!"END"_tok >> !"CONTAINS"_tok >>
|
|
SkipPast<'\n'>{} >> construct<ErrorRecovery>()};
|
|
|
|
// Error recovery across statements: skip the line, unless it looks
|
|
// like it might end the containing construct.
|
|
constexpr auto stmtErrorRecoveryStart{ignoredStatementPrefix};
|
|
constexpr auto skipBadLine{SkipPast<'\n'>{} >> construct<ErrorRecovery>()};
|
|
constexpr auto executionPartErrorRecovery{stmtErrorRecoveryStart >>
|
|
!"END"_tok >> !"CONTAINS"_tok >> !"ELSE"_tok >> !"CASE"_tok >>
|
|
!"TYPE IS"_tok >> !"CLASS"_tok >> !"RANK"_tok >>
|
|
!("!$OMP "_sptok >> ("END"_tok || "SECTION"_id)) >> skipBadLine};
|
|
|
|
// END statement error recovery
|
|
constexpr auto missingOptionalName{pure<std::optional<Name>>()};
|
|
constexpr auto noNameEnd{"END" >> missingOptionalName};
|
|
constexpr auto atEndOfStmt{space >>
|
|
withMessage("expected end of statement"_err_en_US, lookAhead(";\n"_ch))};
|
|
constexpr auto bareEnd{noNameEnd / recovery(atEndOfStmt, SkipTo<'\n'>{})};
|
|
|
|
constexpr auto endStmtErrorRecovery{
|
|
("END"_tok >> SkipTo<'\n'>{} || ok) >> missingOptionalName};
|
|
|
|
constexpr auto progUnitEndStmtErrorRecovery{
|
|
(many(!"END"_tok >> SkipPast<'\n'>{}) >>
|
|
("END"_tok >> SkipTo<'\n'>{} || consumedAllInput)) >>
|
|
missingOptionalName};
|
|
} // namespace Fortran::parser
|
|
#endif // FORTRAN_PARSER_STMT_PARSER_H_
|