Just return integer result code from Lexer.find_longest_match()

This commit is contained in:
Josh Holtrop 2023-07-05 17:07:48 -04:00
parent d56cc2deeb
commit 0d0da49cd5

View File

@ -310,159 +310,88 @@ class <%= @classname %>
result.row = m_input_row; result.row = m_input_row;
result.col = m_input_col; result.col = m_input_col;
result.token = _TOKEN_COUNT; result.token = _TOKEN_COUNT;
auto match_result = find_longest_match(); MatchInfo match_info;
if (match_result.is_eof()) size_t unexpected_input_length;
switch (find_longest_match(match_info, unexpected_input_length))
{ {
case FindLongestMatchResult.FOUND_MATCH:
uint token_to_accept = match_info.accepting_state.token;
if (match_info.accepting_state.code_id.is_valid())
{
Token user_code_token = user_code(match_info.accepting_state.code_id, m_input[m_input_position..(m_input_position + match_info.length)], &result);
/* An invalid Token from user_code() means that the user
* code did not explicitly return a token. So only override
* the token to return if the user code does explicitly
* return a token. */
if (user_code_token.is_valid())
{
token_to_accept = user_code_token.token;
}
}
/* Update the input position tracking. */
m_input_position += match_info.length;
m_input_row += match_info.delta_row;
if (match_info.delta_row != 0u)
{
m_input_col = match_info.delta_col;
}
else
{
m_input_col += match_info.delta_col;
}
result.token = token_to_accept;
result.length = match_info.length;
if (match_info.accepting_state.drop)
{
result.type = Result.Type.DROP;
}
else
{
result.type = Result.Type.TOKEN;
}
return result;
case FindLongestMatchResult.DECODE_ERROR:
result.type = Result.Type.DECODE_ERROR;
return result;
case FindLongestMatchResult.EOF:
result.type = Result.Type.TOKEN; result.type = Result.Type.TOKEN;
result.token = TOKEN___EOF; result.token = TOKEN___EOF;
return result; return result;
}
else if (match_result.is_decode_error()) case FindLongestMatchResult.UNEXPECTED_INPUT:
{
result.type = Result.Type.DECODE_ERROR;
return result;
}
else if (match_result.is_unexpected_input())
{
result.type = Result.Type.UNEXPECTED_INPUT; result.type = Result.Type.UNEXPECTED_INPUT;
return result; return result;
}
uint token_to_accept = match_result.accepting_state.token;
if (match_result.accepting_state.code_id.is_valid())
{
Token user_code_token = user_code(match_result.accepting_state.code_id, m_input[m_input_position..(m_input_position + match_result.length)], &result);
/* An invalid Token from user_code() means that the user
* code did not explicitly return a token. So only override
* the token to return if the user code does explicitly
* return a token. */
if (user_code_token.is_valid())
{
token_to_accept = user_code_token.token;
}
}
/* Update the input position tracking. */ default:
m_input_position += match_result.length; assert(false);
m_input_row += match_result.delta_row;
if (match_result.delta_row != 0u)
{
m_input_col = match_result.delta_col;
}
else
{
m_input_col += match_result.delta_col;
}
result.token = token_to_accept;
result.length = match_result.length;
if (match_result.accepting_state.drop)
{
result.type = Result.Type.DROP;
}
else
{
result.type = Result.Type.TOKEN;
}
return result;
}
/**
* Result type for find_longest_match().
*
* Alternatives:
* - decode_error
* - eof
* - found_match(length, delta_row, delta_col, accepting_state)
* - unexpected_input(unexpected_input_length)
*/
struct FindLongestMatchResult
{
enum : ubyte
{
FOUND_MATCH,
DECODE_ERROR,
EOF,
UNEXPECTED_INPUT,
}
ubyte type;
alias type this;
union
{
struct
{
size_t length;
size_t delta_row;
size_t delta_col;
const(State) * accepting_state;
}
size_t unexpected_input_length;
}
this(ubyte type)
{
this.type = type;
}
this(ubyte type, size_t unexpected_input_length)
{
this.type = type;
this.unexpected_input_length = unexpected_input_length;
}
this(ubyte type, size_t length, size_t delta_row, size_t delta_col, const(State) * accepting_state)
{
this.type = type;
this.length = length;
this.delta_row = delta_row;
this.delta_col = delta_col;
this.accepting_state = accepting_state;
}
static FindLongestMatchResult found_match(size_t length, size_t delta_row, size_t delta_col, const(State) * accepting_state)
{
return FindLongestMatchResult(FOUND_MATCH, length, delta_row, delta_col, accepting_state);
}
static FindLongestMatchResult decode_error()
{
return FindLongestMatchResult(DECODE_ERROR);
}
static FindLongestMatchResult eof()
{
return FindLongestMatchResult(EOF);
}
static FindLongestMatchResult unexpected_input(size_t unexpected_input_length)
{
return FindLongestMatchResult(UNEXPECTED_INPUT, unexpected_input_length);
}
bool is_found_match()
{
return type == FOUND_MATCH;
}
bool is_decode_error()
{
return type == DECODE_ERROR;
}
bool is_eof()
{
return type == EOF;
}
bool is_unexpected_input()
{
return type == UNEXPECTED_INPUT;
} }
} }
private FindLongestMatchResult find_longest_match() enum FindLongestMatchResult
{ {
FindLongestMatchResult longest_match = FindLongestMatchResult.found_match(0, 0, 0, null); FOUND_MATCH,
FindLongestMatchResult attempt_match = longest_match; DECODE_ERROR,
EOF,
UNEXPECTED_INPUT,
}
struct MatchInfo
{
size_t length;
size_t delta_row;
size_t delta_col;
const(State) * accepting_state;
}
private FindLongestMatchResult find_longest_match(
ref MatchInfo out_match_info,
ref size_t out_unexpected_input_length)
{
MatchInfo longest_match;
MatchInfo attempt_match;
uint current_state = modes[m_mode].state_table_offset; uint current_state = modes[m_mode].state_table_offset;
for (;;) for (;;)
{ {
@ -494,11 +423,13 @@ class <%= @classname %>
} }
else if (longest_match.length > 0) else if (longest_match.length > 0)
{ {
return longest_match; out_match_info = longest_match;
return FindLongestMatchResult.FOUND_MATCH;
} }
else else
{ {
return FindLongestMatchResult.unexpected_input(attempt_match.length + code_point_length); out_unexpected_input_length = attempt_match.length + code_point_length;
return FindLongestMatchResult.UNEXPECTED_INPUT;
} }
break; break;
@ -507,22 +438,24 @@ class <%= @classname %>
if (longest_match.length > 0) if (longest_match.length > 0)
{ {
/* We have a match, so use it. */ /* We have a match, so use it. */
return longest_match; out_match_info = longest_match;
return FindLongestMatchResult.FOUND_MATCH;
} }
else if (attempt_match.length != 0) else if (attempt_match.length != 0)
{ {
/* There is a partial match - error! */ /* There is a partial match - error! */
return FindLongestMatchResult.unexpected_input(attempt_match.length); out_unexpected_input_length = attempt_match.length;
return FindLongestMatchResult.UNEXPECTED_INPUT;
} }
else else
{ {
/* Valid EOF return. */ /* Valid EOF return. */
return FindLongestMatchResult.eof(); return FindLongestMatchResult.EOF;
} }
break; break;
case Decoder.Result.DECODE_ERROR: case Decoder.Result.DECODE_ERROR:
return FindLongestMatchResult.decode_error(); return FindLongestMatchResult.DECODE_ERROR;
default: default:
assert(false); assert(false);