Avoid wcwidth(), wcrtomb() and mbrtowc() on ASCII/ISO8859-1 characters.
ASCII <-> UTF has trivial mappings. Avoid wcrtomb() and mbrtowc(). ISO-8859-1 is all narrow characters, and cheap to test for. It might be possible to cheaply test other popular UTF blocks and/or planes as well. These two changes get 2-3x faster input processing on Linux and FreeBSD. Performance improvement in actual usage is more modest but still significant.
This commit is contained in:
@@ -80,10 +80,15 @@ void Parser::UTF8Parser::input( char c, Actions &ret )
|
||||
{
|
||||
assert( buf_len < BUF_SIZE );
|
||||
|
||||
/* 1-byte UTF-8 character, aka ASCII? Cheat. */
|
||||
if ( buf_len == 0 && static_cast<unsigned char>(c) <= 0x7f ) {
|
||||
parser.input( static_cast<wchar_t>(c), ret );
|
||||
return;
|
||||
}
|
||||
|
||||
buf[ buf_len++ ] = c;
|
||||
|
||||
/* This function will only work in a UTF-8 locale. */
|
||||
|
||||
wchar_t pwc;
|
||||
mbstate_t ps = mbstate_t();
|
||||
|
||||
|
||||
@@ -61,7 +61,13 @@ void Emulator::print( const Parser::Print *act )
|
||||
{
|
||||
assert( act->char_present );
|
||||
|
||||
int chwidth = act->ch == L'\0' ? -1 : wcwidth( act->ch );
|
||||
const wchar_t ch = act->ch;
|
||||
|
||||
/*
|
||||
* Check for printing ISO 8859-1 first, it's a cheap way to detect
|
||||
* some common narrow characters.
|
||||
*/
|
||||
const int chwidth = ch == L'\0' ? -1 : ( Cell::isprint_iso8859_1( ch ) ? 1 : wcwidth( ch ));
|
||||
|
||||
Cell *this_cell = fb.get_mutable_cell();
|
||||
|
||||
@@ -100,7 +106,7 @@ void Emulator::print( const Parser::Print *act )
|
||||
}
|
||||
|
||||
fb.reset_cell( this_cell );
|
||||
this_cell->append( act->ch );
|
||||
this_cell->append( ch );
|
||||
this_cell->width = chwidth;
|
||||
fb.apply_renditions_to_cell( this_cell );
|
||||
|
||||
@@ -134,7 +140,7 @@ void Emulator::print( const Parser::Print *act )
|
||||
}
|
||||
if ( combining_cell->contents.size() < 32 ) {
|
||||
/* seems like a reasonable limit on combining characters */
|
||||
combining_cell->append( act->ch );
|
||||
combining_cell->append( ch );
|
||||
}
|
||||
act->handled = true;
|
||||
}
|
||||
|
||||
@@ -126,8 +126,19 @@ namespace Terminal {
|
||||
|
||||
bool compare( const Cell &other ) const;
|
||||
|
||||
static void append_to_str( std::string &dest, const wchar_t c )
|
||||
// Is this a printing ISO 8859-1 character?
|
||||
static bool isprint_iso8859_1( const wchar_t c )
|
||||
{
|
||||
return ( c <= 0xff && c >= 0xa0 ) || ( c <= 0x7e && c >= 0x20 );
|
||||
}
|
||||
|
||||
static void append_to_str( std::string &dest, const wchar_t c )
|
||||
{
|
||||
/* ASCII? Cheat. */
|
||||
if ( static_cast<uint32_t>(c) <= 0x7f ) {
|
||||
dest.push_back( static_cast<char>(c) );
|
||||
return;
|
||||
}
|
||||
static mbstate_t ps = mbstate_t();
|
||||
char tmp[MB_LEN_MAX];
|
||||
size_t ignore = wcrtomb(NULL, 0, &ps);
|
||||
@@ -138,6 +149,11 @@ namespace Terminal {
|
||||
|
||||
void append( const wchar_t c )
|
||||
{
|
||||
/* ASCII? Cheat. */
|
||||
if ( static_cast<uint32_t>(c) <= 0x7f ) {
|
||||
contents.push_back( static_cast<char>(c) );
|
||||
return;
|
||||
}
|
||||
static mbstate_t ps = mbstate_t();
|
||||
char tmp[MB_LEN_MAX];
|
||||
size_t ignore = wcrtomb(NULL, 0, &ps);
|
||||
|
||||
Reference in New Issue
Block a user