perf: do not use concat_string if it is not needed

At the earlier stage of the development, I thought that it is useful if
lexer-combinators generate error messages, because by doing this,
parser would not need to generate an error message. But now it turned
out that to show an appropriate error message, parser need to generate
according to the context. And almost all the messages from lexer are
discarded. So I added another parameter to lexer-combinator to suppress
error message generation. In the future, we may want to remove messages
completely from lexers, but currently I will keep it. Removing those
unused message generation makes the parsing process faster.
This commit is contained in:
ToruNiina
2019-04-16 21:09:59 +09:00
parent b3917aaadf
commit 91966a6917

View File

@@ -56,7 +56,8 @@ struct character
static constexpr char target = C; static constexpr char target = C;
template<typename Cont> template<typename Cont>
static result<region<Cont>, std::string> invoke(location<Cont>& loc) static result<region<Cont>, std::string>
invoke(location<Cont>& loc, const bool msg = false)
{ {
static_assert(std::is_same<char, typename Cont::value_type>::value, static_assert(std::is_same<char, typename Cont::value_type>::value,
"internal error: container::value_type should be `char`."); "internal error: container::value_type should be `char`.");
@@ -66,10 +67,14 @@ struct character
const char c = *(loc.iter()); const char c = *(loc.iter());
if(c != target) if(c != target)
{
if(msg)
{ {
return err(concat_to_string("expected '", show_char(target), return err(concat_to_string("expected '", show_char(target),
"' but got '", show_char(c), "'.")); "' but got '", show_char(c), "'."));
} }
return err("");
}
loc.advance(); // update location loc.advance(); // update location
return ok(region<Cont>(loc, first, loc.iter())); return ok(region<Cont>(loc, first, loc.iter()));
@@ -91,7 +96,8 @@ struct in_range
static constexpr char lower = Low; static constexpr char lower = Low;
template<typename Cont> template<typename Cont>
static result<region<Cont>, std::string> invoke(location<Cont>& loc) static result<region<Cont>, std::string>
invoke(location<Cont>& loc, const bool msg = false)
{ {
static_assert(std::is_same<char, typename Cont::value_type>::value, static_assert(std::is_same<char, typename Cont::value_type>::value,
"internal error: container::value_type should be `char`."); "internal error: container::value_type should be `char`.");
@@ -101,11 +107,15 @@ struct in_range
const char c = *(loc.iter()); const char c = *(loc.iter());
if(c < lower || upper < c) if(c < lower || upper < c)
{
if(msg)
{ {
return err(concat_to_string("expected character in range " return err(concat_to_string("expected character in range "
"[", show_char(lower), ", ", show_char(upper), "] but got ", "[", show_char(lower), ", ", show_char(upper), "] but got ",
"'", show_char(c), "'.")); "'", show_char(c), "'."));
} }
return err("");
}
loc.advance(); loc.advance();
return ok(region<Cont>(loc, first, loc.iter())); return ok(region<Cont>(loc, first, loc.iter()));
@@ -125,7 +135,8 @@ template<typename Combinator>
struct exclude struct exclude
{ {
template<typename Cont> template<typename Cont>
static result<region<Cont>, std::string> invoke(location<Cont>& loc) static result<region<Cont>, std::string>
invoke(location<Cont>& loc, const bool msg = false)
{ {
static_assert(std::is_same<char, typename Cont::value_type>::value, static_assert(std::is_same<char, typename Cont::value_type>::value,
"internal error: container::value_type should be `char`."); "internal error: container::value_type should be `char`.");
@@ -133,13 +144,16 @@ struct exclude
if(loc.iter() == loc.end()) {return err("not sufficient characters");} if(loc.iter() == loc.end()) {return err("not sufficient characters");}
auto first = loc.iter(); auto first = loc.iter();
auto rslt = Combinator::invoke(loc); auto rslt = Combinator::invoke(loc, msg);
if(rslt.is_ok()) if(rslt.is_ok())
{ {
loc.reset(first); loc.reset(first);
return err(concat_to_string( if(msg)
"invalid pattern (", Combinator::pattern(), ") appeared ", {
rslt.unwrap().str())); return err(concat_to_string("invalid pattern (",
Combinator::pattern(), ") appeared ", rslt.unwrap().str()));
}
return err("");
} }
loc.reset(std::next(first)); // XXX maybe loc.advance() is okay but... loc.reset(std::next(first)); // XXX maybe loc.advance() is okay but...
return ok(region<Cont>(loc, first, loc.iter())); return ok(region<Cont>(loc, first, loc.iter()));
@@ -156,12 +170,13 @@ template<typename Combinator>
struct maybe struct maybe
{ {
template<typename Cont> template<typename Cont>
static result<region<Cont>, std::string> invoke(location<Cont>& loc) static result<region<Cont>, std::string>
invoke(location<Cont>& loc, const bool msg = false)
{ {
static_assert(std::is_same<char, typename Cont::value_type>::value, static_assert(std::is_same<char, typename Cont::value_type>::value,
"internal error: container::value_type should be `char`."); "internal error: container::value_type should be `char`.");
const auto rslt = Combinator::invoke(loc); const auto rslt = Combinator::invoke(loc, msg);
if(rslt.is_ok()) if(rslt.is_ok())
{ {
return rslt; return rslt;
@@ -182,34 +197,36 @@ template<typename Head, typename ... Tail>
struct sequence<Head, Tail...> struct sequence<Head, Tail...>
{ {
template<typename Cont> template<typename Cont>
static result<region<Cont>, std::string> invoke(location<Cont>& loc) static result<region<Cont>, std::string>
invoke(location<Cont>& loc, const bool msg = false)
{ {
static_assert(std::is_same<char, typename Cont::value_type>::value, static_assert(std::is_same<char, typename Cont::value_type>::value,
"internal error: container::value_type should be `char`."); "internal error: container::value_type should be `char`.");
const auto first = loc.iter(); const auto first = loc.iter();
const auto rslt = Head::invoke(loc); const auto rslt = Head::invoke(loc, msg);
if(rslt.is_err()) if(rslt.is_err())
{ {
loc.reset(first); loc.reset(first);
return err(rslt.unwrap_err()); return err(rslt.unwrap_err());
} }
return sequence<Tail...>::invoke(loc, std::move(rslt.unwrap()), first); return sequence<Tail...>::invoke(loc, std::move(rslt.unwrap()), first, msg);
} }
// called from the above function only, recursively. // called from the above function only, recursively.
template<typename Cont, typename Iterator> template<typename Cont, typename Iterator>
static result<region<Cont>, std::string> static result<region<Cont>, std::string>
invoke(location<Cont>& loc, region<Cont> reg, Iterator first) invoke(location<Cont>& loc, region<Cont> reg, Iterator first,
const bool msg = false)
{ {
const auto rslt = Head::invoke(loc); const auto rslt = Head::invoke(loc, msg);
if(rslt.is_err()) if(rslt.is_err())
{ {
loc.reset(first); loc.reset(first);
return err(rslt.unwrap_err()); return err(rslt.unwrap_err());
} }
reg += rslt.unwrap(); // concat regions reg += rslt.unwrap(); // concat regions
return sequence<Tail...>::invoke(loc, std::move(reg), first); return sequence<Tail...>::invoke(loc, std::move(reg), first, msg);
} }
static std::string pattern() static std::string pattern()
@@ -224,9 +241,10 @@ struct sequence<Head>
// would be called from sequence<T ...>::invoke only. // would be called from sequence<T ...>::invoke only.
template<typename Cont, typename Iterator> template<typename Cont, typename Iterator>
static result<region<Cont>, std::string> static result<region<Cont>, std::string>
invoke(location<Cont>& loc, region<Cont> reg, Iterator first) invoke(location<Cont>& loc, region<Cont> reg, Iterator first,
const bool msg = false)
{ {
const auto rslt = Head::invoke(loc); const auto rslt = Head::invoke(loc, msg);
if(rslt.is_err()) if(rslt.is_err())
{ {
loc.reset(first); loc.reset(first);
@@ -245,14 +263,15 @@ template<typename Head, typename ... Tail>
struct either<Head, Tail...> struct either<Head, Tail...>
{ {
template<typename Cont> template<typename Cont>
static result<region<Cont>, std::string> invoke(location<Cont>& loc) static result<region<Cont>, std::string>
invoke(location<Cont>& loc, const bool msg = false)
{ {
static_assert(std::is_same<char, typename Cont::value_type>::value, static_assert(std::is_same<char, typename Cont::value_type>::value,
"internal error: container::value_type should be `char`."); "internal error: container::value_type should be `char`.");
const auto rslt = Head::invoke(loc); const auto rslt = Head::invoke(loc, msg);
if(rslt.is_ok()) {return rslt;} if(rslt.is_ok()) {return rslt;}
return either<Tail...>::invoke(loc); return either<Tail...>::invoke(loc, msg);
} }
static std::string pattern() static std::string pattern()
@@ -264,11 +283,12 @@ template<typename Head>
struct either<Head> struct either<Head>
{ {
template<typename Cont> template<typename Cont>
static result<region<Cont>, std::string> invoke(location<Cont>& loc) static result<region<Cont>, std::string>
invoke(location<Cont>& loc, const bool msg = false)
{ {
static_assert(std::is_same<char, typename Cont::value_type>::value, static_assert(std::is_same<char, typename Cont::value_type>::value,
"internal error: container::value_type should be `char`."); "internal error: container::value_type should be `char`.");
return Head::invoke(loc); return Head::invoke(loc, msg);
} }
static std::string pattern() static std::string pattern()
{ {
@@ -287,13 +307,14 @@ template<typename T, std::size_t N>
struct repeat<T, exactly<N>> struct repeat<T, exactly<N>>
{ {
template<typename Cont> template<typename Cont>
static result<region<Cont>, std::string> invoke(location<Cont>& loc) static result<region<Cont>, std::string>
invoke(location<Cont>& loc, const bool msg = false)
{ {
region<Cont> retval(loc); region<Cont> retval(loc);
const auto first = loc.iter(); const auto first = loc.iter();
for(std::size_t i=0; i<N; ++i) for(std::size_t i=0; i<N; ++i)
{ {
auto rslt = T::invoke(loc); auto rslt = T::invoke(loc, msg);
if(rslt.is_err()) if(rslt.is_err())
{ {
loc.reset(first); loc.reset(first);
@@ -313,14 +334,15 @@ template<typename T, std::size_t N>
struct repeat<T, at_least<N>> struct repeat<T, at_least<N>>
{ {
template<typename Cont> template<typename Cont>
static result<region<Cont>, std::string> invoke(location<Cont>& loc) static result<region<Cont>, std::string>
invoke(location<Cont>& loc, const bool msg = false)
{ {
region<Cont> retval(loc); region<Cont> retval(loc);
const auto first = loc.iter(); const auto first = loc.iter();
for(std::size_t i=0; i<N; ++i) for(std::size_t i=0; i<N; ++i)
{ {
auto rslt = T::invoke(loc); auto rslt = T::invoke(loc, msg);
if(rslt.is_err()) if(rslt.is_err())
{ {
loc.reset(first); loc.reset(first);
@@ -330,7 +352,7 @@ struct repeat<T, at_least<N>>
} }
while(true) while(true)
{ {
auto rslt = T::invoke(loc); auto rslt = T::invoke(loc, msg);
if(rslt.is_err()) if(rslt.is_err())
{ {
return ok(std::move(retval)); return ok(std::move(retval));
@@ -348,12 +370,13 @@ template<typename T>
struct repeat<T, unlimited> struct repeat<T, unlimited>
{ {
template<typename Cont> template<typename Cont>
static result<region<Cont>, std::string> invoke(location<Cont>& loc) static result<region<Cont>, std::string>
invoke(location<Cont>& loc, const bool msg = false)
{ {
region<Cont> retval(loc); region<Cont> retval(loc);
while(true) while(true)
{ {
auto rslt = T::invoke(loc); auto rslt = T::invoke(loc, msg);
if(rslt.is_err()) if(rslt.is_err())
{ {
return ok(std::move(retval)); return ok(std::move(retval));