Chiebot-Mirror
/
protobuf
mirror of https://github.com/protocolbuffers/protobuf.git
8
0
Fork 0
Code Issues Projects Releases Wiki Activity

Reduce stack usage on recursion by making all error generation lazy on

Browse Source 
out-of-line functions. This avoids local std::string temporaries and alike.

It also speeds up code because we delay constructing the errors via StrCat
until an error has actually occurred.

PiperOrigin-RevId: 579269516
pull/14638/head
Protobuf Team Bot 1 year ago committed by Copybara-Service
parent af75cb31f4
commit 70cc55aeb2
2 changed files with 117 additions and 68 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 123
      src/google/protobuf/compiler/parser.cc
  2. 62
      src/google/protobuf/compiler/parser.h

123
src/google/protobuf/compiler/parser.cc
Unescape View File

@ -201,7 +201,7 @@ bool Parser::TryConsume(absl::string_view text) {
}
}
bool Parser::Consume(absl::string_view text, absl::string_view error) {
bool Parser::Consume(absl::string_view text, ErrorMaker error) {
if (TryConsume(text)) {
return true;
} else {
@ -211,10 +211,11 @@ bool Parser::Consume(absl::string_view text, absl::string_view error) {
}
bool Parser::Consume(absl::string_view text) {
return Consume(text, absl::StrCat("Expected \"", text, "\"."));
return Consume(text,
[&] { return absl::StrCat("Expected \"", text, "\"."); });
}
bool Parser::ConsumeIdentifier(std::string* output, absl::string_view error) {
bool Parser::ConsumeIdentifier(std::string* output, ErrorMaker error) {
if (LookingAtType(io::Tokenizer::TYPE_IDENTIFIER)) {
*output = input_->current().text;
input_->Next();
@ -225,7 +226,7 @@ bool Parser::ConsumeIdentifier(std::string* output, absl::string_view error) {
}
}
bool Parser::ConsumeInteger(int* output, absl::string_view error) {
bool Parser::ConsumeInteger(int* output, ErrorMaker error) {
if (LookingAtType(io::Tokenizer::TYPE_INTEGER)) {
uint64_t value = 0;
if (!io::Tokenizer::ParseInteger(input_->current().text,
@ -243,7 +244,7 @@ bool Parser::ConsumeInteger(int* output, absl::string_view error) {
}
}
bool Parser::ConsumeSignedInteger(int* output, absl::string_view error) {
bool Parser::ConsumeSignedInteger(int* output, ErrorMaker error) {
bool is_negative = false;
uint64_t max_value = std::numeric_limits<int32_t>::max();
if (TryConsume("-")) {
@ -258,7 +259,7 @@ bool Parser::ConsumeSignedInteger(int* output, absl::string_view error) {
}
bool Parser::ConsumeInteger64(uint64_t max_value, uint64_t* output,
absl::string_view error) {
ErrorMaker error) {
if (LookingAtType(io::Tokenizer::TYPE_INTEGER)) {
if (!io::Tokenizer::ParseInteger(input_->current().text, max_value,
output)) {
@ -283,7 +284,7 @@ bool Parser::TryConsumeInteger64(uint64_t max_value, uint64_t* output) {
return false;
}
bool Parser::ConsumeNumber(double* output, absl::string_view error) {
bool Parser::ConsumeNumber(double* output, ErrorMaker error) {
if (LookingAtType(io::Tokenizer::TYPE_FLOAT)) {
*output = io::Tokenizer::ParseFloat(input_->current().text);
input_->Next();
@ -320,7 +321,7 @@ bool Parser::ConsumeNumber(double* output, absl::string_view error) {
}
}
bool Parser::ConsumeString(std::string* output, absl::string_view error) {
bool Parser::ConsumeString(std::string* output, ErrorMaker error) {
if (LookingAtType(io::Tokenizer::TYPE_STRING)) {
io::Tokenizer::ParseString(input_->current().text, output);
input_->Next();
@ -372,32 +373,34 @@ bool Parser::ConsumeEndOfDeclaration(absl::string_view text,
if (TryConsumeEndOfDeclaration(text, location)) {
return true;
} else {
RecordError(absl::StrCat("Expected \"", text, "\"."));
RecordError([&] { return absl::StrCat("Expected \"", text, "\"."); });
return false;
}
}
// -------------------------------------------------------------------
void Parser::RecordError(int line, int column, absl::string_view error) {
void Parser::RecordError(int line, int column, ErrorMaker error) {
if (error_collector_ != nullptr) {
error_collector_->RecordError(line, column, error);
error_collector_->RecordError(line, column, error.get());
}
had_errors_ = true;
}
void Parser::RecordError(absl::string_view error) {
void Parser::RecordError(ErrorMaker error) {
RecordError(input_->current().line, input_->current().column, error);
}
void Parser::RecordWarning(int line, int column, absl::string_view warning) {
void Parser::RecordWarning(int line, int column, ErrorMaker error) {
if (error_collector_ != nullptr) {
error_collector_->RecordWarning(line, column, warning);
error_collector_->RecordWarning(line, column, error.get());
}
}
void Parser::RecordWarning(absl::string_view warning) {
RecordWarning(input_->current().line, input_->current().column, warning);
// Invokes error_collector_->RecordWarning() with the line and column number
// of the current token.
void Parser::RecordWarning(ErrorMaker error) {
RecordWarning(input_->current().line, input_->current().column, error);
}
// -------------------------------------------------------------------
@ -590,12 +593,13 @@ bool Parser::ValidateEnum(const EnumDescriptorProto* proto) {
}
if (has_allow_alias && !allow_alias) {
std::string error = absl::StrCat(
"\"", proto->name(),
"\" declares 'option allow_alias = false;' which has no effect. "
"Please remove the declaration.");
// This needlessly clutters declarations with nops.
RecordError(error);
RecordError([=] {
return absl::StrCat(
"\"", proto->name(),
"\" declares 'option allow_alias = false;' which has no effect. "
"Please remove the declaration.");
});
return false;
}
@ -611,14 +615,15 @@ bool Parser::ValidateEnum(const EnumDescriptorProto* proto) {
}
}
if (allow_alias && !has_duplicates) {
std::string error = absl::StrCat(
"\"", proto->name(),
"\" declares support for enum aliases but no enum values share field "
"numbers. Please remove the unnecessary 'option allow_alias = true;' "
"declaration.");
// Generate an error if an enum declares support for duplicate enum values
// and does not use it protect future authors.
RecordError(error);
RecordError([=] {
return absl::StrCat(
"\"", proto->name(),
"\" declares support for enum aliases but no enum values share field "
"numbers. Please remove the unnecessary 'option allow_alias = true;' "
"declaration.");
});
return false;
}
@ -627,9 +632,13 @@ bool Parser::ValidateEnum(const EnumDescriptorProto* proto) {
if (!allow_alias) {
for (const auto& enum_value : proto->value()) {
if (!IsUpperUnderscore(enum_value.name())) {
RecordWarning(absl::StrCat(
"Enum constant should be in UPPER_CASE. Found: ", enum_value.name(),
". See https://developers.google.com/protocol-buffers/docs/style"));
RecordWarning([&] {
return absl::StrCat(
"Enum constant should be in UPPER_CASE. Found: ",
enum_value.name(),
". See "
"https://developers.google.com/protocol-buffers/docs/style");
});
}
}
}
@ -732,8 +741,9 @@ bool Parser::ParseSyntaxIdentifier(const FileDescriptorProto* file,
if (has_edition) {
if (!Edition_Parse(absl::StrCat("EDITION_", syntax), &edition_) ||
edition_ < Edition::EDITION_2023) {
RecordError(syntax_token.line, syntax_token.column,
absl::StrCat("Unknown edition \"", syntax, "\"."));
RecordError(syntax_token.line, syntax_token.column, [&] {
return absl::StrCat("Unknown edition \"", syntax, "\".");
});
return false;
}
syntax_identifier_ = "editions";
@ -743,10 +753,11 @@ bool Parser::ParseSyntaxIdentifier(const FileDescriptorProto* file,
syntax_identifier_ = syntax;
if (syntax != "proto2" && syntax != "proto3" &&
!stop_after_syntax_identifier_) {
RecordError(syntax_token.line, syntax_token.column,
absl::StrCat("Unrecognized syntax identifier \"", syntax,
"\". This parser "
"only recognizes \"proto2\" and \"proto3\"."));
RecordError(syntax_token.line, syntax_token.column, [&] {
return absl::StrCat("Unrecognized syntax identifier \"", syntax,
"\". This parser "
"only recognizes \"proto2\" and \"proto3\".");
});
return false;
}
@ -847,9 +858,12 @@ bool Parser::ParseMessageDefinition(
DescriptorPool::ErrorCollector::NAME);
DO(ConsumeIdentifier(message->mutable_name(), "Expected message name."));
if (!IsUpperCamelCase(message->name())) {
RecordWarning(absl::StrCat(
"Message name should be in UpperCamelCase. Found: ", message->name(),
". See https://developers.google.com/protocol-buffers/docs/style"));
RecordWarning([=] {
return absl::StrCat(
"Message name should be in UpperCamelCase. Found: ",
message->name(),
". See https://developers.google.com/protocol-buffers/docs/style");
});
}
}
DO(ParseMessageBlock(message, message_location, containing_file));
@ -1080,15 +1094,19 @@ bool Parser::ParseMessageFieldNoLabel(
DO(ConsumeIdentifier(field->mutable_name(), "Expected field name."));
if (!IsLowerUnderscore(field->name())) {
RecordWarning(absl::StrCat(
"Field name should be lowercase. Found: ", field->name(),
". See: https://developers.google.com/protocol-buffers/docs/style"));
RecordWarning([=] {
return absl::StrCat(
"Field name should be lowercase. Found: ", field->name(),
". See: https://developers.google.com/protocol-buffers/docs/style");
});
}
if (IsNumberFollowUnderscore(field->name())) {
RecordWarning(absl::StrCat(
"Number should not come right after an underscore. Found: ",
field->name(),
". See: https://developers.google.com/protocol-buffers/docs/style"));
RecordWarning([=] {
return absl::StrCat(
"Number should not come right after an underscore. Found: ",
field->name(),
". See: https://developers.google.com/protocol-buffers/docs/style");
});
}
}
DO(Consume("=", "Missing field number."));
@ -1819,18 +1837,17 @@ bool Parser::ParseReserved(DescriptorProto* message,
}
}
bool Parser::ParseReservedName(std::string* name,
absl::string_view error_message) {
bool Parser::ParseReservedName(std::string* name, ErrorMaker error_message) {
// Capture the position of the token, in case we have to report an
// error after it is consumed.
int line = input_->current().line;
int col = input_->current().column;
DO(ConsumeString(name, error_message));
if (!io::Tokenizer::IsIdentifier(*name)) {
RecordWarning(
line, col,
absl::StrFormat("Reserved name \"%s\" is not a valid identifier.",
*name));
RecordWarning(line, col, [=] {
return absl::StrFormat("Reserved name \"%s\" is not a valid identifier.",
*name);
});
}
return true;
}
@ -1847,7 +1864,7 @@ bool Parser::ParseReservedNames(DescriptorProto* message,
}
bool Parser::ParseReservedIdentifier(std::string* name,
absl::string_view error_message) {
ErrorMaker error_message) {
DO(ConsumeIdentifier(name, error_message));
return true;
}

62
src/google/protobuf/compiler/parser.h
Unescape View File

@ -16,6 +16,7 @@
#include <cstdint>
#include <string>
#include <type_traits>
#include <utility>
#include "absl/container/flat_hash_map.h"
@ -139,33 +140,65 @@ class PROTOBUF_EXPORT Parser {
// true. Otherwise, return false without logging an error.
bool TryConsume(absl::string_view text);
// In the following functions the error is passed as a lazily evaluated
// callable to reduce stack usage and delay the actual execution of the error
// statement.
// Super simple type erasure interface. Similar to absl::FunctionRef but takes
// the callable by value. Optimized for lambdas with at most a single pointer
// as payload.
class ErrorMaker {
using StorageT = void*;
public:
template <typename F,
typename = std::enable_if_t<std::is_same<
std::string, decltype(std::declval<F>()())>::value>>
ErrorMaker(F f) {
static_assert(sizeof(F) <= sizeof(StorageT), "");
static_assert(alignof(F) <= alignof(StorageT), "");
static_assert(std::is_trivially_destructible<F>::value, "");
::new (static_cast<void*>(storage_)) F(f);
func_ = [](const void* p) { return (*reinterpret_cast<const F*>(p))(); };
}
// This overload helps callers that just want to pass a literal string.
ErrorMaker(const char* error) : error_(error), func_(nullptr) {}
std::string get() const { return func_ ? func_(storage_) : error_; }
private:
union {
alignas(StorageT) char storage_[sizeof(StorageT)];
const char* error_;
};
std::string (*func_)(const void*);
};
// These attempt to read some kind of token from the input. If successful,
// they return true. Otherwise they return false and add the given error
// to the error list.
// Consume a token with the exact text given.
bool Consume(absl::string_view text, absl::string_view error);
bool Consume(absl::string_view text, ErrorMaker error);
// Same as above, but automatically generates the error "Expected \"text\".",
// where "text" is the expected token text.
bool Consume(absl::string_view text);
// Consume a token of type IDENTIFIER and store its text in "output".
bool ConsumeIdentifier(std::string* output, absl::string_view error);
bool ConsumeIdentifier(std::string* output, ErrorMaker error);
// Consume an integer and store its value in "output".
bool ConsumeInteger(int* output, absl::string_view error);
bool ConsumeInteger(int* output, ErrorMaker error);
// Consume a signed integer and store its value in "output".
bool ConsumeSignedInteger(int* output, absl::string_view error);
bool ConsumeSignedInteger(int* output, ErrorMaker error);
// Consume a 64-bit integer and store its value in "output". If the value
// is greater than max_value, an error will be reported.
bool ConsumeInteger64(uint64_t max_value, uint64_t* output,
absl::string_view error);
bool ConsumeInteger64(uint64_t max_value, uint64_t* output, ErrorMaker error);
// Try to consume a 64-bit integer and store its value in "output". No
// error is reported on failure, allowing caller to consume token another way.
bool TryConsumeInteger64(uint64_t max_value, uint64_t* output);
// Consume a number and store its value in "output". This will accept
// tokens of either INTEGER or FLOAT type.
bool ConsumeNumber(double* output, absl::string_view error);
bool ConsumeNumber(double* output, ErrorMaker error);
// Consume a string literal and store its (unescaped) value in "output".
bool ConsumeString(std::string* output, absl::string_view error);
bool ConsumeString(std::string* output, ErrorMaker error);
// Consume a token representing the end of the statement. Comments between
// this token and the next will be harvested for documentation. The given
@ -188,18 +221,18 @@ class PROTOBUF_EXPORT Parser {
// Error logging helpers
// Invokes error_collector_->RecordError(), if error_collector_ is not NULL.
void RecordError(int line, int column, absl::string_view error);
PROTOBUF_NOINLINE void RecordError(int line, int column, ErrorMaker error);
// Invokes error_collector_->RecordError() with the line and column number
// of the current token.
void RecordError(absl::string_view error);
PROTOBUF_NOINLINE void RecordError(ErrorMaker error);
// Invokes error_collector_->RecordWarning(), if error_collector_ is not NULL.
void RecordWarning(int line, int column, absl::string_view warning);
PROTOBUF_NOINLINE void RecordWarning(int line, int column, ErrorMaker error);
// Invokes error_collector_->RecordWarning() with the line and column number
// of the current token.
void RecordWarning(absl::string_view warning);
PROTOBUF_NOINLINE void RecordWarning(ErrorMaker error);
// Records a location in the SourceCodeInfo.location table (see
// descriptor.proto). We use RAII to ensure that the start and end locations
@ -379,11 +412,10 @@ class PROTOBUF_EXPORT Parser {
const LocationRecorder& message_location);
bool ParseReservedNames(DescriptorProto* message,
const LocationRecorder& parent_location);
bool ParseReservedName(std::string* name, absl::string_view error_message);
bool ParseReservedName(std::string* name, ErrorMaker error_message);
bool ParseReservedIdentifiers(DescriptorProto* message,
const LocationRecorder& parent_location);
bool ParseReservedIdentifier(std::string* name,
absl::string_view error_message);
bool ParseReservedIdentifier(std::string* name, ErrorMaker error_message);
bool ParseReservedNumbers(DescriptorProto* message,
const LocationRecorder& parent_location);
bool ParseReserved(EnumDescriptorProto* message,

Write Preview
Loading…
Cancel
Save