256 lines
4.9 KiB
Hare
256 lines
4.9 KiB
Hare
use ascii; // TODO: maybe use unicode?
|
|
use bufio;
|
|
use encoding::utf8;
|
|
use fmt;
|
|
use io;
|
|
use memio;
|
|
use os;
|
|
use strings;
|
|
|
|
// my cod prob sux :(
|
|
|
|
export type lexer = struct {
|
|
in: io::handle,
|
|
strbuf: memio::stream,
|
|
path: str,
|
|
loc: (uint, uint),
|
|
prevloc: (uint, uint),
|
|
unread: (rune | void),
|
|
};
|
|
|
|
export fn newlexer(in: io::handle, path: str) lexer = {
|
|
return lexer {
|
|
in = in,
|
|
strbuf = memio::dynamic(),
|
|
path = path,
|
|
loc = (1, 0),
|
|
unread = void,
|
|
...
|
|
};
|
|
};
|
|
|
|
export fn close(lex: *lexer) void = {
|
|
io::close(&lex.strbuf)!;
|
|
};
|
|
|
|
export fn next(lex: *lexer) (token | io::EOF | error) = {
|
|
const rn = match (nextrunews(lex)?) {
|
|
case io::EOF =>
|
|
return io::EOF;
|
|
case let rn: rune =>
|
|
yield rn;
|
|
};
|
|
|
|
switch (rn) {
|
|
case '(' =>
|
|
return punctuation::LEFT_PAREN: token;
|
|
case ')' =>
|
|
return punctuation::RIGHT_PAREN: token;
|
|
case '[' =>
|
|
return punctuation::LEFT_SQUARE_BRACKET: token;
|
|
case ']' =>
|
|
return punctuation::RIGHT_SQUARE_BRACKET: token;
|
|
case '{' =>
|
|
return punctuation::LEFT_CURLY_BRACKET: token;
|
|
case '}' =>
|
|
return punctuation::RIGHT_CURLY_BRACKET: token;
|
|
case '\\' =>
|
|
return punctuation::BACKSLASH: token;
|
|
case ':' =>
|
|
return punctuation::COLON: token;
|
|
case '"' =>
|
|
match (scanstr(lex)?) {
|
|
case let s: str =>
|
|
return s;
|
|
case io::EOF =>
|
|
return io::EOF;
|
|
};
|
|
case =>
|
|
yield;
|
|
};
|
|
|
|
unget(lex, rn);
|
|
return word{ v = scanword(lex)? };
|
|
};
|
|
|
|
fn nextrune(lex: *lexer) (rune | io::EOF | error) = {
|
|
match (lex.unread) {
|
|
case let rn: rune =>
|
|
lex.prevloc = lex.loc;
|
|
lex.unread = void;
|
|
if (rn == '\n') {
|
|
lex.loc = (lex.loc.0 + 1, 0);
|
|
} else {
|
|
lex.loc.1 += 1;
|
|
};
|
|
return rn;
|
|
case void =>
|
|
yield;
|
|
};
|
|
|
|
match (bufio::read_rune(lex.in)) {
|
|
case let rn: rune =>
|
|
lex.prevloc = lex.loc;
|
|
if (rn == '\n') {
|
|
lex.loc = (lex.loc.0 + 1, 0);
|
|
} else {
|
|
lex.loc.1 += 1;
|
|
};
|
|
return rn;
|
|
case io::EOF =>
|
|
return io::EOF;
|
|
case let err: io::error =>
|
|
return err;
|
|
case utf8::invalid =>
|
|
return lex.loc: invalid;
|
|
};
|
|
};
|
|
|
|
fn nextrunews(lex: *lexer) (rune | io::EOF | error) = {
|
|
for (true) {
|
|
match (nextrune(lex)?) {
|
|
case let rn: rune =>
|
|
if (ascii::isspace(rn)) {
|
|
continue;
|
|
};
|
|
return rn;
|
|
case io::EOF =>
|
|
return io::EOF;
|
|
};
|
|
};
|
|
};
|
|
|
|
fn unget(lex: *lexer, rn: rune) void = {
|
|
assert(lex.unread is void);
|
|
lex.unread = rn;
|
|
lex.loc = lex.prevloc;
|
|
};
|
|
|
|
fn scanword(lex: *lexer) (str | error) = {
|
|
memio::reset(&lex.strbuf);
|
|
for (true) {
|
|
const rn = match (nextrune(lex)?) {
|
|
case let rn: rune =>
|
|
yield rn;
|
|
case io::EOF =>
|
|
break;
|
|
};
|
|
if (ascii::isspace(rn)) {
|
|
unget(lex, rn);
|
|
break;
|
|
};
|
|
memio::appendrune(&lex.strbuf, rn)!;
|
|
};
|
|
return memio::string(&lex.strbuf)!;
|
|
};
|
|
|
|
fn scanstr(lex: *lexer) (str | io::EOF | error) = {
|
|
memio::reset(&lex.strbuf);
|
|
for (true) {
|
|
const rn = match (nextrune(lex)?) {
|
|
case let rn: rune =>
|
|
yield rn;
|
|
case io::EOF =>
|
|
return lex.loc: invalid;
|
|
};
|
|
|
|
switch (rn) {
|
|
case '"' => break;
|
|
case '\\' =>
|
|
memio::appendrune(&lex.strbuf, scanescape(lex)?)!;
|
|
case =>
|
|
memio::appendrune(&lex.strbuf, rn)!;
|
|
};
|
|
};
|
|
return memio::string(&lex.strbuf)!;
|
|
};
|
|
|
|
fn scanescape(lex: *lexer) (rune | error) = {
|
|
const rn = match (nextrune(lex)?) {
|
|
case let rn: rune =>
|
|
yield rn;
|
|
case io::EOF =>
|
|
return lex.loc: invalid;
|
|
};
|
|
|
|
switch (rn) {
|
|
case '"' =>
|
|
return '"';
|
|
case '\\' =>
|
|
return '\\';
|
|
case '\n' =>
|
|
return '\n';
|
|
case =>
|
|
return lex.loc: invalid;
|
|
};
|
|
};
|
|
|
|
// Tests! :)
|
|
|
|
fn tnext(lex: *lexer) token = {
|
|
match (next(lex)!) {
|
|
case let t: token =>
|
|
return t;
|
|
case =>
|
|
assert(false);
|
|
return word { v = "" };
|
|
};
|
|
};
|
|
|
|
@test fn test_next() void = {
|
|
let lex = newlexer(&memio::fixed(
|
|
strings::toutf8("\"hello\" \\greeting def")),
|
|
"<string>");
|
|
defer close(&lex);
|
|
|
|
let tk = tnext(&lex);
|
|
assert(tk is str && tk: str == "hello");
|
|
let tk = tnext(&lex);
|
|
assert(tk is punctuation && tk: punctuation == punctuation::BACKSLASH);
|
|
let tk = tnext(&lex);
|
|
assert(tk is word && (tk: word).v == "greeting");
|
|
let tk = tnext(&lex);
|
|
assert(tk is word && (tk: word).v == "def");
|
|
};
|
|
|
|
@test fn test_nextrune() void = {
|
|
let lex = newlexer(&memio::fixed(strings::toutf8("a\nb")),
|
|
"<string>");
|
|
defer close(&lex);
|
|
|
|
assert(nextrune(&lex)! == 'a');
|
|
assert(nextrune(&lex)! == '\n');
|
|
assert(nextrune(&lex)! == 'b');
|
|
assert(lex.loc.0 == 2u && lex.loc.1 == 1u);
|
|
};
|
|
|
|
@test fn test_nextrunews() void = {
|
|
let lex = newlexer(&memio::fixed(strings::toutf8("\n a")),
|
|
"<string>");
|
|
defer close(&lex);
|
|
|
|
assert(nextrunews(&lex)! == 'a');
|
|
assert(lex.loc.0 == 2u && lex.loc.1 == 2u);
|
|
};
|
|
|
|
@test fn test_scanword() void = {
|
|
let lex = newlexer(&memio::fixed(strings::toutf8("string->number .")),
|
|
"<string>");
|
|
defer close(&lex);
|
|
|
|
assert(scanword(&lex)! == "string->number");
|
|
};
|
|
|
|
@test fn test_scanstr() void = {
|
|
let lex = newlexer(&memio::fixed(strings::toutf8("\"\\\\hello\\\"world!\\\n\"")),
|
|
"<string>");
|
|
defer close(&lex);
|
|
|
|
assert(nextrune(&lex)! == '"');
|
|
match (scanstr(&lex)!) {
|
|
case io::EOF =>
|
|
assert(false);
|
|
case let s: str =>
|
|
assert(s == "\\hello\"world!\n");
|
|
};
|
|
};
|