976 lines
29 KiB
C++
976 lines
29 KiB
C++
#include "nasal_new_parse.h"
|
|
|
|
const error& parse::compile(const lexer& lexer) {
|
|
toks=lexer.result().data();
|
|
ptr=in_func=in_loop=0;
|
|
|
|
root={toks[0].loc, ast_root};
|
|
while(!lookahead(tok::eof)) {
|
|
root.add(expr());
|
|
if (lookahead(tok::semi)) {
|
|
match(tok::semi);
|
|
} else if (need_semi_check(root.child().back()) && !lookahead(tok::eof)) {
|
|
// the last expression can be recognized without semi
|
|
die(prevspan, "expected \";\"");
|
|
}
|
|
}
|
|
root.update_span();
|
|
return err;
|
|
}
|
|
|
|
void parse::easter_egg() const {
|
|
std::clog
|
|
<< " _,,,_ \n"
|
|
<< " .' `'. \n"
|
|
<< " / ____ \\ Fucking Nasal Parser \n"
|
|
<< " | .-'_ _\\/ / \n"
|
|
<< " \\_/ a a| / \n"
|
|
<< " (,` \\ | .----. \n"
|
|
<< " | -' | /| '--. \n"
|
|
<< " \\ '= / || ]| `-. \n"
|
|
<< " /`-.__.' || ]| ::| \n"
|
|
<< " .-'`-.__ \\__ || ]| ::| \n"
|
|
<< " / `` `. || ]| ::| \n"
|
|
<< " _ | \\ \\ \\ \\| ]| .-' \n"
|
|
<< " / \\| \\ | \\ L.__ .--'( \n"
|
|
<< " | |\\ `. | \\ ,---|_ \\---------, \n"
|
|
<< " | | '. './\\ \\/ .--._|=- |_ /| \n"
|
|
<< " | \\ '. `'.'. /`\\/ .-' '. / | \n"
|
|
<< " | | `'. `;-:-;`)| |-./ | \n"
|
|
<< " | /_ `'--./_ ` )/'-------------')/) | \n"
|
|
<< " \\ | `\"\"\"\"----\"`\\//`\"\"`/,===..'`````````/ ( |\n"
|
|
<< " | | / `---` `===' / ) | \n"
|
|
<< " / \\ / / ( | \n"
|
|
<< " | '------. |'--------------------'| ) | \n"
|
|
<< " \\ `-| | / | \n"
|
|
<< " `--...,______| | ( | \n"
|
|
<< " | | | | ) ,| \n"
|
|
<< " | | | | ( /|| \n"
|
|
<< " | | | | )/ `\" \n"
|
|
<< " / \\ | | (/ \n"
|
|
<< " .' /I\\ '.| | /) \n"
|
|
<< " .-'_.'/ \\'. | | / \n"
|
|
<< " ``` `\"\"\"` `| .-------------------.|| \n"
|
|
<< " `\"` `\"` \n";
|
|
}
|
|
|
|
void parse::die(const span& loc, string info) {
|
|
err.err("parse", loc, info);
|
|
}
|
|
|
|
void parse::match(tok type,const char* info) {
|
|
if (!lookahead(type)) {
|
|
if (info) {
|
|
die(thisspan, info);
|
|
return;
|
|
}
|
|
switch(type) {
|
|
case tok::num:die(thisspan, "expected number"); break;
|
|
case tok::str:die(thisspan, "expected string"); break;
|
|
case tok::id: die(thisspan, "expected identifier");break;
|
|
default: die(thisspan, "expected '"+tokname.at(type)+"'"); break;
|
|
}
|
|
return;
|
|
}
|
|
if (lookahead(tok::eof)) {
|
|
return;
|
|
}
|
|
next();
|
|
}
|
|
|
|
bool parse::lookahead(tok type) {
|
|
return toks[ptr].type==type;
|
|
}
|
|
|
|
bool parse::is_call(tok type) {
|
|
return type==tok::lcurve || type==tok::lbracket || type==tok::dot;
|
|
}
|
|
|
|
bool parse::check_comma(const tok* panic_set) {
|
|
for(u32 i=0;panic_set[i]!=tok::null;++i) {
|
|
if (lookahead(panic_set[i])) {
|
|
die(prevspan, "expected ',' between scalars");
|
|
return true;
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
bool parse::check_tuple() {
|
|
u32 check_ptr=ptr, curve=1, bracket=0, brace=0;
|
|
while(toks[++check_ptr].type!=tok::eof && curve) {
|
|
switch(toks[check_ptr].type) {
|
|
case tok::lcurve: ++curve; break;
|
|
case tok::lbracket: ++bracket; break;
|
|
case tok::lbrace: ++brace; break;
|
|
case tok::rcurve: --curve; break;
|
|
case tok::rbracket: --bracket; break;
|
|
case tok::rbrace: --brace; break;
|
|
default: break;
|
|
}
|
|
if (curve==1 && !bracket && !brace && toks[check_ptr].type==tok::comma) {
|
|
return true;
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
bool parse::check_func_end(const ast& node) {
|
|
u32 type=node.type();
|
|
if (type==ast_func) {
|
|
return true;
|
|
} else if (type==ast_num || type==ast_id || type==ast_str || type==ast_nil || type==ast_vec || type==ast_hash) {
|
|
return false;
|
|
}
|
|
if (node.child().empty() || (
|
|
type!=ast_def &&
|
|
type!=ast_equal &&
|
|
type!=ast_addeq &&
|
|
type!=ast_subeq &&
|
|
type!=ast_multeq &&
|
|
type!=ast_diveq &&
|
|
type!=ast_lnkeq
|
|
)
|
|
) {
|
|
return false;
|
|
} else {
|
|
return check_func_end(node.child().back());
|
|
}
|
|
return false;
|
|
}
|
|
|
|
bool parse::check_special_call() {
|
|
// special call means like this: function_name(a:1,b:2,c:3);
|
|
u32 check_ptr=ptr, curve=1, bracket=0, brace=0;
|
|
while(toks[++check_ptr].type!=tok::eof && curve) {
|
|
switch(toks[check_ptr].type) {
|
|
case tok::lcurve: ++curve; break;
|
|
case tok::lbracket: ++bracket;break;
|
|
case tok::lbrace: ++brace; break;
|
|
case tok::rcurve: --curve; break;
|
|
case tok::rbracket: --bracket;break;
|
|
case tok::rbrace: --brace; break;
|
|
default: break;
|
|
}
|
|
// m?1:0 will be recognized as normal parameter
|
|
if (curve==1 && !bracket && !brace && toks[check_ptr].type==tok::quesmark) {
|
|
return false;
|
|
}
|
|
if (curve==1 && !bracket && !brace && toks[check_ptr].type==tok::colon) {
|
|
return true;
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
bool parse::need_semi_check(const ast& node) {
|
|
u32 type=node.type();
|
|
if (type==ast_for || type==ast_foreach || type==ast_forindex || type==ast_while || type==ast_cond) {
|
|
return false;
|
|
}
|
|
return !check_func_end(node);
|
|
}
|
|
|
|
ast parse::null() {
|
|
return {toks[ptr].loc, ast_null};
|
|
}
|
|
|
|
ast parse::nil() {
|
|
return {toks[ptr].loc, ast_nil};
|
|
}
|
|
|
|
ast parse::num() {
|
|
ast node(toks[ptr].loc, ast_num);
|
|
node.set_num(str2num(toks[ptr].str.c_str()));
|
|
match(tok::num);
|
|
return node;
|
|
}
|
|
|
|
ast parse::str() {
|
|
ast node(toks[ptr].loc, ast_str);
|
|
node.set_str(toks[ptr].str);
|
|
match(tok::str);
|
|
return node;
|
|
}
|
|
|
|
ast parse::id() {
|
|
ast node(toks[ptr].loc, ast_id);
|
|
node.set_str(toks[ptr].str);
|
|
match(tok::id);
|
|
return node;
|
|
}
|
|
|
|
ast parse::bools() {
|
|
ast node(toks[ptr].loc, ast_bool);
|
|
node.set_str(toks[ptr].str);
|
|
if (lookahead(tok::tktrue)) {
|
|
match(tok::tktrue);
|
|
} else {
|
|
match(tok::tkfalse);
|
|
}
|
|
return node;
|
|
}
|
|
|
|
ast parse::vec() {
|
|
// panic set for this token is not ','
|
|
// this is the FIRST set of calculation
|
|
// array end with tok::null=0
|
|
const tok panic[]={
|
|
tok::id,tok::str,tok::num,tok::tktrue,
|
|
tok::tkfalse,tok::opnot,tok::sub,tok::tknil,
|
|
tok::func,tok::var,tok::lcurve,tok::floater,
|
|
tok::lbrace,tok::lbracket,tok::null
|
|
};
|
|
ast node(toks[ptr].loc, ast_vec);
|
|
match(tok::lbracket);
|
|
while(!lookahead(tok::rbracket)) {
|
|
node.add(calc());
|
|
if (lookahead(tok::comma)) {
|
|
match(tok::comma);
|
|
} else if (lookahead(tok::eof)) {
|
|
break;
|
|
} else if (!lookahead(tok::rbracket) && !check_comma(panic)) {
|
|
break;
|
|
}
|
|
}
|
|
node.update_span(thisspan);
|
|
match(tok::rbracket, "expected ']' when generating vector");
|
|
return node;
|
|
}
|
|
|
|
ast parse::hash() {
|
|
ast node(toks[ptr].loc, ast_hash);
|
|
match(tok::lbrace);
|
|
while(!lookahead(tok::rbrace)) {
|
|
node.add(pair());
|
|
if (lookahead(tok::comma)) {
|
|
match(tok::comma);
|
|
} else if (lookahead(tok::id) || lookahead(tok::str)) { // first set of hashmember
|
|
die(prevspan, "expected ',' between hash members");
|
|
} else {
|
|
break;
|
|
}
|
|
}
|
|
node.update_span(thisspan);
|
|
match(tok::rbrace, "expected '}' when generating hash");
|
|
return node;
|
|
}
|
|
|
|
ast parse::pair() {
|
|
ast node(toks[ptr].loc, ast_pair);
|
|
if (lookahead(tok::id)) {
|
|
node.add(id());
|
|
} else if (lookahead(tok::str)) {
|
|
node.add(str());
|
|
} else {
|
|
match(tok::id, "expected hashmap key");
|
|
}
|
|
match(tok::colon);
|
|
node.add(calc());
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::func() {
|
|
++in_func;
|
|
ast node(toks[ptr].loc, ast_func);
|
|
match(tok::func);
|
|
if (lookahead(tok::lcurve)) {
|
|
node.add(params());
|
|
} else {
|
|
node.add(null());
|
|
}
|
|
node.add(exprs());
|
|
--in_func;
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::params() {
|
|
ast node(toks[ptr].loc, ast_params);
|
|
match(tok::lcurve);
|
|
while(!lookahead(tok::rcurve)) {
|
|
ast tmp=id();
|
|
if (lookahead(tok::eq) || lookahead(tok::ellipsis)) {
|
|
ast special_arg(toks[ptr].loc, ast_null);
|
|
if (lookahead(tok::eq)) {
|
|
match(tok::eq);
|
|
special_arg=std::move(tmp);
|
|
special_arg.set_type(ast_default);
|
|
special_arg.add(calc());
|
|
} else {
|
|
match(tok::ellipsis);
|
|
special_arg=std::move(tmp);
|
|
special_arg.set_type(ast_dynamic);
|
|
}
|
|
node.add(std::move(special_arg));
|
|
} else {
|
|
node.add(std::move(tmp));
|
|
}
|
|
if (lookahead(tok::comma)) {
|
|
match(tok::comma);
|
|
} else if (lookahead(tok::id)) { // first set of identifier
|
|
die(prevspan, "expected ',' between identifiers");
|
|
} else {
|
|
break;
|
|
}
|
|
}
|
|
node.update_span(thisspan);
|
|
match(tok::rcurve, "expected ')' after parameter list");
|
|
return node;
|
|
}
|
|
|
|
ast parse::lcurve_expr() {
|
|
if (toks[ptr+1].type==tok::var)
|
|
return definition();
|
|
return check_tuple()?multi_assgin():calc();
|
|
}
|
|
|
|
ast parse::expr() {
|
|
tok type=toks[ptr].type;
|
|
if ((type==tok::brk || type==tok::cont) && !in_loop) {
|
|
die(thisspan, "must use break/continue in loops");
|
|
}
|
|
if (type==tok::ret && !in_func) {
|
|
die(thisspan, "must use return in functions");
|
|
}
|
|
switch(type) {
|
|
case tok::tknil:
|
|
case tok::num:
|
|
case tok::str:
|
|
case tok::id:
|
|
case tok::tktrue:
|
|
case tok::tkfalse:
|
|
case tok::func:
|
|
case tok::lbracket:
|
|
case tok::lbrace:
|
|
case tok::sub:
|
|
case tok::floater:
|
|
case tok::opnot: return calc(); break;
|
|
case tok::var: return definition(); break;
|
|
case tok::lcurve: return lcurve_expr(); break;
|
|
case tok::rfor:
|
|
case tok::forindex:
|
|
case tok::foreach:
|
|
case tok::rwhile: return loop(); break;
|
|
case tok::rif: return cond(); break;
|
|
case tok::cont: return continue_expr(); break;
|
|
case tok::brk: return break_expr(); break;
|
|
case tok::ret: return ret_expr(); break;
|
|
case tok::semi: break;
|
|
default:
|
|
die(thisspan, "incorrect token <"+toks[ptr].str+">");
|
|
next();
|
|
break;
|
|
}
|
|
|
|
// unreachable
|
|
return {toks[ptr].loc, ast_null};
|
|
}
|
|
|
|
ast parse::exprs() {
|
|
if (lookahead(tok::eof)) {
|
|
die(thisspan, "expected expression block");
|
|
return null();
|
|
}
|
|
ast node(toks[ptr].loc, ast_block);
|
|
if (lookahead(tok::lbrace)) {
|
|
match(tok::lbrace);
|
|
while(!lookahead(tok::rbrace) && !lookahead(tok::eof)) {
|
|
node.add(expr());
|
|
if (lookahead(tok::semi)) {
|
|
match(tok::semi);
|
|
} else if (need_semi_check(node.child().back()) && !lookahead(tok::rbrace)) {
|
|
// the last expression can be recognized without semi
|
|
die(prevspan, "expected ';'");
|
|
}
|
|
}
|
|
match(tok::rbrace, "expected '}' when generating expressions");
|
|
} else {
|
|
node.add(expr());
|
|
if (lookahead(tok::semi))
|
|
match(tok::semi);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::calc() {
|
|
ast node=bitwise_or();
|
|
if (lookahead(tok::quesmark)) {
|
|
// trinocular calculation
|
|
ast tmp(toks[ptr].loc, ast_trino);
|
|
match(tok::quesmark);
|
|
tmp.add(std::move(node));
|
|
tmp.add(calc());
|
|
match(tok::colon);
|
|
tmp.add(calc());
|
|
node=std::move(tmp);
|
|
} else if (tok::eq<=toks[ptr].type && toks[ptr].type<=tok::lnkeq) {
|
|
// tok::eq~tok::lnkeq is 37 to 42,ast_equal~ast_lnkeq is 21~26
|
|
ast tmp(toks[ptr].loc, (u32)toks[ptr].type-(u32)tok::eq+ast_equal);
|
|
tmp.add(std::move(node));
|
|
match(toks[ptr].type);
|
|
tmp.add(calc());
|
|
node=std::move(tmp);
|
|
} else if (toks[ptr].type==tok::btandeq || toks[ptr].type==tok::btoreq || toks[ptr].type==tok::btxoreq) {
|
|
ast tmp(toks[ptr].loc, (u32)toks[ptr].type-(u32)tok::btandeq+ast_btandeq);
|
|
tmp.add(std::move(node));
|
|
match(toks[ptr].type);
|
|
tmp.add(calc());
|
|
node=std::move(tmp);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::bitwise_or() {
|
|
ast node=bitwise_xor();
|
|
while(lookahead(tok::btor)) {
|
|
ast tmp(toks[ptr].loc, ast_bitor);
|
|
tmp.add(std::move(node));
|
|
match(tok::btor);
|
|
tmp.add(bitwise_xor());
|
|
tmp.update_span();
|
|
node=std::move(tmp);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::bitwise_xor() {
|
|
ast node=bitwise_and();
|
|
while(lookahead(tok::btxor)) {
|
|
ast tmp(toks[ptr].loc, ast_bitxor);
|
|
tmp.add(std::move(node));
|
|
match(tok::btxor);
|
|
tmp.add(bitwise_and());
|
|
tmp.update_span();
|
|
node=std::move(tmp);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::bitwise_and() {
|
|
ast node=or_expr();
|
|
while(lookahead(tok::btand)) {
|
|
ast tmp(toks[ptr].loc, ast_bitand);
|
|
tmp.add(std::move(node));
|
|
match(tok::btand);
|
|
tmp.add(or_expr());
|
|
tmp.update_span();
|
|
node=std::move(tmp);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::or_expr() {
|
|
ast node=and_expr();
|
|
while(lookahead(tok::opor)) {
|
|
ast tmp(toks[ptr].loc, ast_or);
|
|
tmp.add(std::move(node));
|
|
match(tok::opor);
|
|
tmp.add(and_expr());
|
|
tmp.update_span();
|
|
node=std::move(tmp);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::and_expr() {
|
|
ast node=cmp_expr();
|
|
while(lookahead(tok::opand)) {
|
|
ast tmp(toks[ptr].loc, ast_and);
|
|
tmp.add(std::move(node));
|
|
match(tok::opand);
|
|
tmp.add(cmp_expr());
|
|
tmp.update_span();
|
|
node=std::move(tmp);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::cmp_expr() {
|
|
ast node=additive_expr();
|
|
while(tok::cmpeq<=toks[ptr].type && toks[ptr].type<=tok::geq) {
|
|
// tok::cmpeq~tok::geq is 43~48,ast_cmpeq~ast_geq is 27~32
|
|
ast tmp(toks[ptr].loc, (u32)toks[ptr].type-(u32)tok::cmpeq+ast_cmpeq);
|
|
tmp.add(std::move(node));
|
|
match(toks[ptr].type);
|
|
tmp.add(additive_expr());
|
|
tmp.update_span();
|
|
node=std::move(tmp);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::additive_expr() {
|
|
ast node=multive_expr();
|
|
while(lookahead(tok::add) || lookahead(tok::sub) || lookahead(tok::floater)) {
|
|
ast tmp(toks[ptr].loc, ast_null);
|
|
switch(toks[ptr].type) {
|
|
case tok::add: tmp.set_type(ast_add); break;
|
|
case tok::sub: tmp.set_type(ast_sub); break;
|
|
case tok::floater: tmp.set_type(ast_link); break;
|
|
default: break;
|
|
}
|
|
tmp.add(std::move(node));
|
|
match(toks[ptr].type);
|
|
tmp.add(multive_expr());
|
|
tmp.update_span();
|
|
node=std::move(tmp);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::multive_expr() {
|
|
ast node=(lookahead(tok::sub) || lookahead(tok::opnot) || lookahead(tok::floater))?unary():scalar();
|
|
while(lookahead(tok::mult) || lookahead(tok::div)) {
|
|
ast tmp(toks[ptr].loc, (u32)toks[ptr].type-(u32)tok::mult+ast_mult);
|
|
tmp.add(std::move(node));
|
|
match(toks[ptr].type);
|
|
tmp.add((lookahead(tok::sub) || lookahead(tok::opnot) || lookahead(tok::floater))?unary():scalar());
|
|
tmp.update_span();
|
|
node=std::move(tmp);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::unary() {
|
|
ast node(toks[ptr].loc,ast_null);
|
|
switch(toks[ptr].type) {
|
|
case tok::sub: node.set_type(ast_neg);match(tok::sub);break;
|
|
case tok::opnot: node.set_type(ast_lnot);match(tok::opnot);break;
|
|
case tok::floater: node.set_type(ast_bnot);match(tok::floater);break;
|
|
default: break;
|
|
}
|
|
node.add((lookahead(tok::sub) || lookahead(tok::opnot) || lookahead(tok::floater))?unary():scalar());
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::scalar() {
|
|
ast node(toks[ptr].loc, ast_null);
|
|
if (lookahead(tok::tknil)) {
|
|
node=nil();
|
|
match(tok::tknil);
|
|
} else if (lookahead(tok::num)) {
|
|
node=num();
|
|
} else if (lookahead(tok::str)) {
|
|
node=str();
|
|
} else if (lookahead(tok::id)) {
|
|
node=id();
|
|
} else if (lookahead(tok::tktrue) || lookahead(tok::tkfalse)) {
|
|
node=bools();
|
|
} else if (lookahead(tok::func)) {
|
|
node=func();
|
|
} else if (lookahead(tok::lbracket)) {
|
|
node=vec();
|
|
} else if (lookahead(tok::lbrace)) {
|
|
node=hash();
|
|
} else if (lookahead(tok::lcurve)) {
|
|
const auto& loc=toks[ptr].loc;
|
|
match(tok::lcurve);
|
|
node=calc();
|
|
node.set_begin(loc.begin_line, loc.begin_column);
|
|
node.update_span(thisspan);
|
|
match(tok::rcurve);
|
|
} else if (lookahead(tok::var)) {
|
|
match(tok::var);
|
|
node.set_type(ast_def);
|
|
node.add(id());
|
|
match(tok::eq);
|
|
node.add(calc());
|
|
} else {
|
|
die(thisspan, "expected scalar");
|
|
return node;
|
|
}
|
|
// check call and avoid ambiguous syntax
|
|
if (is_call(toks[ptr].type) && !(lookahead(tok::lcurve) && toks[ptr+1].type==tok::var)) {
|
|
ast tmp=std::move(node);
|
|
node={toks[ptr].loc, ast_call};
|
|
node.add(std::move(tmp));
|
|
while(is_call(toks[ptr].type)) {
|
|
node.add(call_scalar());
|
|
}
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::call_scalar() {
|
|
switch(toks[ptr].type) {
|
|
case tok::lcurve: return callf(); break;
|
|
case tok::lbracket: return callv(); break;
|
|
case tok::dot: return callh(); break;
|
|
default: break;
|
|
}
|
|
// unreachable
|
|
return {toks[ptr].loc, ast_nil};
|
|
}
|
|
|
|
ast parse::callh() {
|
|
ast node(toks[ptr].loc, ast_callh);
|
|
match(tok::dot);
|
|
node.set_str(toks[ptr].str);
|
|
node.set_end(toks[ptr].loc.end_line, toks[ptr].loc.end_column);
|
|
match(tok::id, "expected hashmap key"); // get key
|
|
return node;
|
|
}
|
|
|
|
ast parse::callv() {
|
|
// panic set for this token is not ','
|
|
// this is the FIRST set of subvec
|
|
// array end with tok::null=0
|
|
const tok panic[]={
|
|
tok::id,tok::str,tok::num,tok::tktrue,
|
|
tok::tkfalse,tok::opnot,tok::sub,tok::tknil,
|
|
tok::func,tok::var,tok::lcurve,tok::floater,
|
|
tok::lbrace,tok::lbracket,tok::colon,tok::null
|
|
};
|
|
ast node(toks[ptr].loc, ast_callv);
|
|
match(tok::lbracket);
|
|
while(!lookahead(tok::rbracket)) {
|
|
node.add(subvec());
|
|
if (lookahead(tok::comma)) {
|
|
match(tok::comma);
|
|
} else if (lookahead(tok::eof)) {
|
|
break;
|
|
} else if (!lookahead(tok::rbracket) && !check_comma(panic)) {
|
|
break;
|
|
}
|
|
}
|
|
if (node.size()==0) {
|
|
die(thisspan, "expected index value");
|
|
}
|
|
node.update_span(thisspan);
|
|
match(tok::rbracket, "expected ']' when calling vector");
|
|
return node;
|
|
}
|
|
|
|
ast parse::callf() {
|
|
// panic set for this token is not ','
|
|
// this is the FIRST set of calculation/hashmember
|
|
// array end with tok::null=0
|
|
const tok panic[]={
|
|
tok::id,tok::str,tok::num,tok::tktrue,
|
|
tok::tkfalse,tok::opnot,tok::sub,tok::tknil,
|
|
tok::func,tok::var,tok::lcurve,tok::floater,
|
|
tok::lbrace,tok::lbracket,tok::null
|
|
};
|
|
ast node(toks[ptr].loc, ast_callf);
|
|
bool special_call=check_special_call();
|
|
match(tok::lcurve);
|
|
while(!lookahead(tok::rcurve)) {
|
|
node.add(special_call?pair():calc());
|
|
if (lookahead(tok::comma))
|
|
match(tok::comma);
|
|
else if (lookahead(tok::eof))
|
|
break;
|
|
else if (!lookahead(tok::rcurve) && !check_comma(panic))
|
|
break;
|
|
}
|
|
node.update_span(thisspan);
|
|
match(tok::rcurve, "expected ')' when calling function");
|
|
return node;
|
|
}
|
|
|
|
ast parse::subvec() {
|
|
ast node=lookahead(tok::colon)?nil():calc();
|
|
if (lookahead(tok::colon)) {
|
|
ast tmp(toks[ptr].loc, ast_subvec);
|
|
match(tok::colon);
|
|
tmp.add(std::move(node));
|
|
tmp.add((lookahead(tok::comma) || lookahead(tok::rbracket))?nil():calc());
|
|
node=std::move(tmp);
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::definition() {
|
|
ast node(toks[ptr].loc, ast_def);
|
|
if (lookahead(tok::var)) {
|
|
match(tok::var);
|
|
switch(toks[ptr].type) {
|
|
case tok::id: node.add(id());break;
|
|
case tok::lcurve: node.add(outcurve_def());break;
|
|
default: die(thisspan, "expected identifier");break;
|
|
}
|
|
} else if (lookahead(tok::lcurve)) {
|
|
node.add(incurve_def());
|
|
}
|
|
match(tok::eq);
|
|
if (lookahead(tok::lcurve)) {
|
|
node.add(check_tuple()?multi_scalar():calc());
|
|
} else {
|
|
node.add(calc());
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::incurve_def() {
|
|
const auto& loc=toks[ptr].loc;
|
|
match(tok::lcurve);
|
|
match(tok::var);
|
|
ast node=multi_id();
|
|
node.update_span(thisspan);
|
|
match(tok::rcurve);
|
|
node.set_begin(loc.begin_line, loc.begin_column);
|
|
return node;
|
|
}
|
|
|
|
ast parse::outcurve_def() {
|
|
const auto& loc=toks[ptr].loc;
|
|
match(tok::lcurve);
|
|
ast node=multi_id();
|
|
node.update_span(thisspan);
|
|
match(tok::rcurve);
|
|
node.set_begin(loc.begin_line, loc.begin_column);
|
|
return node;
|
|
}
|
|
|
|
ast parse::multi_id() {
|
|
ast node(toks[ptr].loc, ast_multi_id);
|
|
while(!lookahead(tok::eof)) {
|
|
// only identifier is allowed here
|
|
// but we check it at codegen stage
|
|
node.add(calc());
|
|
if (lookahead(tok::comma)) {
|
|
match(tok::comma);
|
|
} else if (lookahead(tok::id)) { // first set of identifier
|
|
die(prevspan, "expected ',' between identifiers");
|
|
} else {
|
|
break;
|
|
}
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::multi_scalar() {
|
|
// if check_call_memory is true,we will check if value called here can reach a memory space
|
|
const tok panic[]={
|
|
tok::id,tok::str,tok::num,tok::tktrue,
|
|
tok::tkfalse,tok::opnot,tok::sub,tok::tknil,
|
|
tok::func,tok::var,tok::lcurve,tok::floater,
|
|
tok::lbrace,tok::lbracket,tok::null
|
|
};
|
|
ast node(toks[ptr].loc, ast_tuple);
|
|
match(tok::lcurve);
|
|
while(!lookahead(tok::rcurve)) {
|
|
node.add(calc());
|
|
if (lookahead(tok::comma)) {
|
|
match(tok::comma);
|
|
} else if (lookahead(tok::eof)) {
|
|
break;
|
|
} else if (!lookahead(tok::rcurve) && !check_comma(panic)) {
|
|
break;
|
|
}
|
|
}
|
|
node.update_span(thisspan);
|
|
match(tok::rcurve, "expected ')' after multi-scalar");
|
|
return node;
|
|
}
|
|
|
|
ast parse::multi_assgin() {
|
|
ast node(toks[ptr].loc, ast_multi_assign);
|
|
node.add(multi_scalar());
|
|
match(tok::eq);
|
|
if (lookahead(tok::eof)) {
|
|
die(thisspan, "expected value list");
|
|
return node;
|
|
}
|
|
if (lookahead(tok::lcurve)) {
|
|
node.add(check_tuple()?multi_scalar():calc());
|
|
} else {
|
|
node.add(calc());
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::loop() {
|
|
++in_loop;
|
|
ast node(toks[ptr].loc, ast_null);
|
|
switch(toks[ptr].type) {
|
|
case tok::rwhile: node=while_loop(); break;
|
|
case tok::rfor: node=for_loop(); break;
|
|
case tok::forindex:
|
|
case tok::foreach: node=forei_loop(); break;
|
|
default: break;
|
|
}
|
|
--in_loop;
|
|
return node;
|
|
}
|
|
|
|
ast parse::while_loop() {
|
|
ast node(toks[ptr].loc, ast_while);
|
|
match(tok::rwhile);
|
|
match(tok::lcurve);
|
|
node.add(calc());
|
|
match(tok::rcurve);
|
|
node.add(exprs());
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::for_loop() {
|
|
ast node(toks[ptr].loc, ast_for);
|
|
match(tok::rfor);
|
|
match(tok::lcurve);
|
|
// first expression
|
|
if (lookahead(tok::eof)) {
|
|
die(thisspan, "expected definition");
|
|
}
|
|
if (lookahead(tok::semi)) {
|
|
node.add(null());
|
|
} else if (lookahead(tok::var)) {
|
|
node.add(definition());
|
|
} else if (lookahead(tok::lcurve)) {
|
|
node.add(lcurve_expr());
|
|
} else {
|
|
node.add(calc());
|
|
}
|
|
match(tok::semi, "expected ';' in for(;;)");
|
|
// conditional expression
|
|
if (lookahead(tok::eof)) {
|
|
die(thisspan, "expected conditional expr");
|
|
}
|
|
if (lookahead(tok::semi)) {
|
|
node.add(null());
|
|
} else {
|
|
node.add(calc());
|
|
}
|
|
match(tok::semi, "expected ';' in for(;;)");
|
|
//after loop expression
|
|
if (lookahead(tok::eof)) {
|
|
die(thisspan, "expected calculation");
|
|
}
|
|
if (lookahead(tok::rcurve)) {
|
|
node.add(null());
|
|
} else {
|
|
node.add(calc());
|
|
}
|
|
match(tok::rcurve);
|
|
node.add(exprs());
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::forei_loop() {
|
|
ast node(toks[ptr].loc, ast_null);
|
|
switch(toks[ptr].type) {
|
|
case tok::forindex:node.set_type(ast_forindex);match(tok::forindex);break;
|
|
case tok::foreach: node.set_type(ast_foreach); match(tok::foreach); break;
|
|
default: break;
|
|
}
|
|
match(tok::lcurve);
|
|
// first expression
|
|
// foreach/forindex must have an iterator to loop through
|
|
if (!lookahead(tok::var) && !lookahead(tok::id)) {
|
|
die(thisspan, "expected iterator");
|
|
}
|
|
node.add(iter_gen());
|
|
match(tok::semi, "expected ';' in foreach/forindex(iter;vector)");
|
|
if (lookahead(tok::eof)) {
|
|
die(thisspan, "expected vector");
|
|
}
|
|
node.add(calc());
|
|
match(tok::rcurve);
|
|
node.add(exprs());
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::iter_gen() {
|
|
ast node(toks[ptr].loc, ast_null);
|
|
if (lookahead(tok::var)) {
|
|
match(tok::var);
|
|
node.set_type(ast_iter);
|
|
node.add(id());
|
|
} else {
|
|
node.set_type(ast_call);
|
|
node.add(id());
|
|
while(is_call(toks[ptr].type)) {
|
|
node.add(call_scalar());
|
|
}
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::cond() {
|
|
ast node(toks[ptr].loc, ast_cond);
|
|
|
|
// generate if
|
|
ast ifnode(toks[ptr].loc, ast_if);
|
|
match(tok::rif);
|
|
match(tok::lcurve);
|
|
ifnode.add(calc());
|
|
match(tok::rcurve);
|
|
ifnode.add(exprs());
|
|
ifnode.update_span();
|
|
node.add(std::move(ifnode));
|
|
|
|
// generate elsif
|
|
while(lookahead(tok::elsif)) {
|
|
ast elsifnode(toks[ptr].loc, ast_elsif);
|
|
match(tok::elsif);
|
|
match(tok::lcurve);
|
|
elsifnode.add(calc());
|
|
match(tok::rcurve);
|
|
elsifnode.add(exprs());
|
|
elsifnode.update_span();
|
|
node.add(std::move(elsifnode));
|
|
}
|
|
|
|
// generate else
|
|
if (lookahead(tok::relse)) {
|
|
ast elsenode(toks[ptr].loc, ast_else);
|
|
match(tok::relse);
|
|
elsenode.add(exprs());
|
|
elsenode.update_span();
|
|
node.add(std::move(elsenode));
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|
|
|
|
ast parse::continue_expr() {
|
|
ast node(toks[ptr].loc, ast_continue);
|
|
match(tok::cont);
|
|
return node;
|
|
}
|
|
|
|
ast parse::break_expr() {
|
|
ast node(toks[ptr].loc, ast_break);
|
|
match(tok::brk);
|
|
return node;
|
|
}
|
|
|
|
ast parse::ret_expr() {
|
|
ast node(toks[ptr].loc, ast_ret);
|
|
match(tok::ret);
|
|
tok type=toks[ptr].type;
|
|
if (type==tok::tknil || type==tok::num ||
|
|
type==tok::str || type==tok::id ||
|
|
type==tok::func || type==tok::sub ||
|
|
type==tok::opnot || type==tok::lcurve ||
|
|
type==tok::lbracket || type==tok::lbrace ||
|
|
type==tok::tktrue || type==tok::tkfalse
|
|
) {
|
|
node.add(calc());
|
|
}
|
|
node.update_span();
|
|
return node;
|
|
}
|