xo-reader2 stack: parenthesized expressions [WIP]

This commit is contained in:
Roland Conybeare 2026-02-10 15:14:40 -05:00
commit d65eb54102
24 changed files with 1095 additions and 35 deletions

View file

@ -37,6 +37,10 @@ set(SELF_SRCS
ISyntaxStateMachine_DLambdaSsm.cpp
IPrintable_DLambdaSsm.cpp
DParenSsm.cpp
ISyntaxStateMachine_DParenSsm.cpp
IPrintable_DParenSsm.cpp
DExpectFormalArglistSsm.cpp
ISyntaxStateMachine_DExpectFormalArglistSsm.cpp
IPrintable_DExpectFormalArglistSsm.cpp

View file

@ -555,7 +555,6 @@ namespace xo {
case tokentype::tk_end:
case tokentype::N:
break;
return;
}
Super::on_token(tk, p_psm);

View file

@ -8,6 +8,7 @@
#include "SyntaxStateMachine.hpp"
#include "ssm/ISyntaxStateMachine_DProgressSsm.hpp"
#include "DSequenceSsm.hpp"
#include "LambdaSsm.hpp"
#include "syntaxstatetype.hpp"
#include <xo/expression2/Variable.hpp>
#include <xo/expression2/Constant.hpp>
@ -138,6 +139,10 @@ namespace xo {
this->on_bool_token(tk, p_psm);
return;
case tokentype::tk_lambda:
this->on_lambda_token(tk, p_psm);
return;
// all the not-yet handled cases
case tokentype::tk_invalid:
case tokentype::tk_if:
@ -165,7 +170,6 @@ namespace xo {
case tokentype::tk_cmpeq:
case tokentype::tk_cmpne:
case tokentype::tk_type:
case tokentype::tk_lambda:
case tokentype::tk_then:
case tokentype::tk_else:
case tokentype::tk_let:
@ -379,6 +383,15 @@ namespace xo {
p_psm);
}
void
DExpectExprSsm::on_lambda_token(const Token & tk,
ParserStateMachine * p_psm)
{
(void)tk;
DLambdaSsm::start(p_psm);
}
void
DExpectExprSsm::on_parsed_expression(obj<AExpression> expr,
ParserStateMachine * p_psm)
@ -412,17 +425,6 @@ namespace xo {
}
#ifdef NOT_YET
void
expect_expr_xs::on_lambda_token(const token_type & /*tk*/,
parserstatemachine * p_psm)
{
scope log(XO_DEBUG(p_psm->debug_flag()));
//constexpr const char * self_name = "exprstate::on_leftparen";
lambda_xs::start(p_psm);
}
void
expect_expr_xs::on_if_token(const token_type & /*tk*/,
parserstatemachine * p_psm)

View file

@ -1,5 +1,5 @@
/** @file DExprSeqState.cpp
*
*
* @author Roland Conybeare, Jan 2026
**/
@ -9,6 +9,8 @@
#include "DLambdaSsm.hpp"
#include "DProgressSsm.hpp"
#include "DIfElseSsm.hpp"
#include "ParenSsm.hpp"
#include "ExpectExprSsm.hpp"
#include <xo/expression2/DConstant.hpp>
#include <xo/expression2/detail/IExpression_DConstant.hpp>
@ -152,9 +154,12 @@ namespace xo {
this->on_bool_token(tk, p_psm);
return;
case tokentype::tk_leftparen:
this->on_leftparen_token(tk, p_psm);
return;
// all the not-yet handled cases
case tokentype::tk_invalid:
case tokentype::tk_leftparen:
case tokentype::tk_rightparen:
case tokentype::tk_leftbracket:
case tokentype::tk_rightbracket:
@ -389,6 +394,27 @@ namespace xo {
Super::on_token(tk, p_psm);
}
void
DExprSeqState::on_leftparen_token(const Token & tk,
ParserStateMachine * p_psm)
{
switch (seqtype_) {
case exprseqtype::toplevel_interactive: {
DParenSsm::start(p_psm);
p_psm->on_token(Token::leftparen_token());
return;
}
case exprseqtype::toplevel_batch:
break;
case exprseqtype::N:
assert(false); // unreachable
break;
}
Super::on_token(tk, p_psm);
}
void
DExprSeqState::on_parsed_expression(obj<AExpression> expr,
ParserStateMachine * p_psm)

View file

@ -136,6 +136,10 @@ namespace xo {
this->on_yields_token(tk, p_psm);
return;
case tokentype::tk_leftbrace:
this->on_leftbrace_token(tk, p_psm);
return;
// all the not-yet-handled cases
case tokentype::tk_def:
case tokentype::tk_if:
@ -152,7 +156,6 @@ namespace xo {
case tokentype::tk_rightparen:
case tokentype::tk_leftbracket:
case tokentype::tk_rightbracket:
case tokentype::tk_leftbrace:
case tokentype::tk_rightbrace:
case tokentype::tk_leftangle:
case tokentype::tk_rightangle:
@ -179,11 +182,6 @@ namespace xo {
}
Super::on_token(tk, p_psm);
#ifdef OBSOLETE
p_psm->illegal_input_on_token("DLambdaSsm::on_token",
tk,
this->get_expect_str());
#endif
}
void
@ -199,11 +197,6 @@ namespace xo {
}
Super::on_token(tk, p_psm);
#ifdef OBSOLETE
p_psm->illegal_input_on_token("DLambdaSsm::on_lambda_token",
tk,
this->get_expect_str());
#endif
}
void
@ -220,13 +213,28 @@ namespace xo {
}
Super::on_token(tk, p_psm);
#ifdef OBSOLETE
p_psm->illegal_input_on_token("DLambdaSsm::on_yields_token",
tk,
this->get_expect_str());
#endif
}
void
DLambdaSsm::on_leftbrace_token(const Token & tk,
ParserStateMachine * p_psm)
{
if (lmstate_ == lambdastatetype::lm_2) {
// control here when leftbrace immediately follows
// formal param list.
// Otherwise leftbrace arrives in DExpectExprSsm
// pushed via on_parsed_typedescr()
this->lmstate_ = lambdastatetype::lm_4;
DExpectExprSsm::start(p_psm);
// precharge ssm for body with leftbrace
p_psm->on_token(Token::leftbrace_token());
return;
}
Super::on_token(tk, p_psm);
}
void
DLambdaSsm::on_parsed_typedescr(TypeDescr td,

382
src/reader2/DParenSsm.cpp Normal file
View file

@ -0,0 +1,382 @@
/** @file DParenSsm.cpp
*
* @author Roland Conybeare, Feb 2026
**/
#include "ParenSsm.hpp"
#include "syntaxstatetype.hpp"
#include <string_view>
namespace xo {
using xo::facet::with_facet;
using xo::facet::typeseq;
namespace scm {
extern const char *
parenexprstatetype_descr(parenexprstatetype x)
{
switch(x) {
case parenexprstatetype::invalid: return "invalid";
case parenexprstatetype::lparen_0: return "lparen_0";
case parenexprstatetype::lparen_1: return "lparen_1";
case parenexprstatetype::lparen_2: return "lparen_2";
case parenexprstatetype::N: break;
}
return "???parenexprstatetype";
}
std::ostream &
operator<<(std::ostream & os, parenexprstatetype x) {
os << parenexprstatetype_descr(x);
return os;
}
DParenSsm::DParenSsm()
: parenstate_(parenexprstatetype::lparen_0),
expr_{}
{}
DParenSsm *
DParenSsm::make(DArena & mm)
{
void * mem = mm.alloc(typeseq::id<DParenSsm>(),
sizeof(DParenSsm));
return new (mem) DParenSsm();
}
void
DParenSsm::start(ParserStateMachine * p_psm)
{
DParenSsm * paren_ssm = DParenSsm::make(p_psm->parser_alloc());
auto ssm = with_facet<ASyntaxStateMachine>::mkobj(paren_ssm);
p_psm->push_ssm(ssm);
}
syntaxstatetype
DParenSsm::ssm_type() const noexcept
{
return syntaxstatetype::paren;
}
std::string_view
DParenSsm::get_expect_str() const noexcept
{
switch (this->parenstate_) {
case parenexprstatetype::invalid:
case parenexprstatetype::N:
break;
case parenexprstatetype::lparen_0: return "lparen_0";
case parenexprstatetype::lparen_1: return "lparen_1";
case parenexprstatetype::lparen_2: return "lparen_2";
}
return "???parenexprstatetype";
}
void
DParenSsm::on_token(const Token & tk,
ParserStateMachine * p_psm)
{
switch (tk.tk_type()) {
case tokentype::tk_leftparen:
this->on_leftparen_token(tk, p_psm);
return;
// all the not-yet handled cases
case tokentype::tk_symbol:
case tokentype::tk_def:
case tokentype::tk_colon:
case tokentype::tk_singleassign:
case tokentype::tk_semicolon:
case tokentype::tk_invalid:
case tokentype::tk_string:
case tokentype::tk_f64:
case tokentype::tk_i64:
case tokentype::tk_bool:
case tokentype::tk_if:
case tokentype::tk_rightparen:
case tokentype::tk_leftbracket:
case tokentype::tk_rightbracket:
case tokentype::tk_leftbrace:
case tokentype::tk_rightbrace:
case tokentype::tk_leftangle:
case tokentype::tk_rightangle:
case tokentype::tk_lessequal:
case tokentype::tk_greatequal:
case tokentype::tk_dot:
case tokentype::tk_comma:
case tokentype::tk_doublecolon:
case tokentype::tk_assign:
case tokentype::tk_yields:
case tokentype::tk_plus:
case tokentype::tk_minus:
case tokentype::tk_star:
case tokentype::tk_slash:
case tokentype::tk_cmpeq:
case tokentype::tk_cmpne:
case tokentype::tk_type:
case tokentype::tk_lambda:
case tokentype::tk_then:
case tokentype::tk_else:
case tokentype::tk_let:
case tokentype::tk_in:
case tokentype::tk_end:
case tokentype::N:
break;
}
Super::on_token(tk, p_psm);
}
void
DParenSsm::on_leftparen_token(const Token & tk,
ParserStateMachine * p_psm)
{
Super::on_token(tk, p_psm);
}
#ifdef OBSOLETE
void
paren_xs::start(parserstatemachine * p_psm)
{
p_psm->push_exprstate(paren_xs::make());
expect_expr_xs::start(p_psm);
}
bool
paren_xs::admits_rightparen() const {
switch (parenxs_type_) {
case parenexprstatetype::lparen_0:
/* unreachable */
assert(false);
return false;
case parenexprstatetype::lparen_1:
return true;
case parenexprstatetype::invalid:
case parenexprstatetype::n_parenexprstatetype:
/* unreachable */
assert(false);
return false;
}
return false;
}
bool
paren_xs::admits_f64() const {
switch (parenxs_type_) {
case parenexprstatetype::lparen_0:
return true;
case parenexprstatetype::lparen_1:
return false;
case parenexprstatetype::invalid:
case parenexprstatetype::n_parenexprstatetype:
/* unreachable */
assert(false);
return false;
}
return false;
}
void
paren_xs::on_def_token(const token_type & tk,
parserstatemachine * /*p_psm*/)
{
constexpr const char * c_self_name = "paren_xs::on_def";
this->illegal_input_error(c_self_name, tk);
}
void
paren_xs::on_symbol_token(const token_type & /*tk*/,
parserstatemachine * p_psm)
{
constexpr bool c_debug_flag = true;
scope log(XO_DEBUG(c_debug_flag));
log && log(xtag("exstype", p_psm->top_exprstate().exs_type()));
//constexpr const char * self_name = "paren_xs::on_symbol";
/* TODO: lparen_0: treat as variable reference */
assert(false);
}
void
paren_xs::on_typedescr(TypeDescr /*td*/,
parserstatemachine * /*p_psm*/)
{
assert(false);
return;
}
void
paren_xs::on_colon_token(const token_type & tk,
parserstatemachine * /*p_psm*/)
{
constexpr const char * c_self_name = "paren_xs::on_colon";
this->illegal_input_error(c_self_name, tk);
}
void
paren_xs::on_semicolon_token(const token_type & tk,
parserstatemachine * /*p_psm*/)
{
constexpr const char * c_self_name = "paren_xs::on_semicolon";
this->illegal_input_error(c_self_name, tk);
}
void
paren_xs::on_singleassign_token(const token_type & tk,
parserstatemachine * /*p_psm*/)
{
constexpr const char * c_self_name = "paren_xs::on_singleassign";
this->illegal_input_error(c_self_name, tk);
}
void
paren_xs::on_leftparen_token(const token_type & tk,
parserstatemachine * /*p_psm*/)
{
constexpr const char * c_self_name = "paren_xs::on_leftparen";
this->illegal_input_error(c_self_name, tk);
}
void
paren_xs::on_rightparen_token(const token_type & tk,
parserstatemachine * p_psm)
{
constexpr bool c_debug_flag = true;
scope log(XO_DEBUG(c_debug_flag));
constexpr const char * c_self_name = "paren_xs::on_rightparen";
if (!this->admits_rightparen())
{
this->illegal_input_error(c_self_name, tk);
}
if (this->parenxs_type_ == parenexprstatetype::lparen_1) {
rp<Expression> expr = this->gen_expr_;
std::unique_ptr<exprstate> self = p_psm->pop_exprstate();
p_psm->top_exprstate().on_expr(expr, p_psm);
}
}
void
paren_xs::on_i64_token(const token_type & tk,
parserstatemachine * /*p_psm*/)
{
constexpr bool c_debug_flag = true;
scope log(XO_DEBUG(c_debug_flag));
constexpr const char * c_self_name = "paren_xs::on_i64";
this->illegal_input_error(c_self_name, tk);
}
void
paren_xs::on_f64_token(const token_type & tk,
parserstatemachine * /*p_psm*/)
{
constexpr bool c_debug_flag = true;
scope log(XO_DEBUG(c_debug_flag));
constexpr const char * c_self_name = "paren_xs::on_f64";
this->illegal_input_error(c_self_name, tk);
}
void
paren_xs::on_expr(bp<Expression> expr,
parserstatemachine * p_psm)
{
constexpr bool c_debug_flag = true;
scope log(XO_DEBUG(c_debug_flag));
log && log(xtag("exstype", this->exs_type_),
xtag("expr", expr));
switch (this->parenxs_type_) {
case parenexprstatetype::lparen_0: {
this->parenxs_type_ = parenexprstatetype::lparen_1; /* wants on_rightparen */
progress_xs::start(expr.promote(), p_psm);
return;
}
case parenexprstatetype::lparen_1: {
this->gen_expr_ = expr.promote();
/* expect immediate incoming call, this time to on_rightparen() */
return;
}
default:
/* unreachable */
assert(false);
return;
}
} /*on_expr*/
void
paren_xs::on_symbol(const std::string & /*symbol_name*/,
parserstatemachine * /*p_psm*/)
{
switch(this->parenxs_type_) {
case parenexprstatetype::lparen_0:
case parenexprstatetype::lparen_1:
/* NOT IMPLEMENTED */
assert(false);
return;
default:
/* unreachable */
assert(false);
return;
}
}
void
paren_xs::print(std::ostream & os) const {
os << "<paren_xs"
<< xtag("this", (void*)this)
//<< xtag("type", exs_type_);
<< xtag("parenxs_type", parenxs_type_);
if (gen_expr_)
os << xtag("gen_expr", gen_expr_);
os << ">";
}
#endif
bool
DParenSsm::pretty(const ppindentinfo & ppii) const
{
return ppii.pps()->pretty_struct(ppii,
"DParenSsm",
refrtag("parenstate", parenstate_),
refrtag("expect", this->get_expect_str()));
}
} /*namespace scm*/
} /*namespace xo*/
/* end DParenSsm.cpp */

View file

@ -0,0 +1,28 @@
/** @file IPrintable_DParenSsm.cpp
*
* Generated automagically from ingredients:
* 1. code generator:
* [xo-facet/codegen/genfacet]
* arguments:
* --input [idl/IPrintable_DParenSsm.json5]
* 2. jinja2 template for abstract facet .hpp file:
* [iface_facet_any.hpp.j2]
* 3. idl for facet methods
* [idl/IPrintable_DParenSsm.json5]
**/
#include "ssm/IPrintable_DParenSsm.hpp"
namespace xo {
namespace scm {
auto
IPrintable_DParenSsm::pretty(const DParenSsm & self, const ppindentinfo & ppii) -> bool
{
return self.pretty(ppii);
}
} /*namespace scm*/
} /*namespace xo*/
/* end IPrintable_DParenSsm.cpp */

View file

@ -0,0 +1,69 @@
/** @file ISyntaxStateMachine_DParenSsm.cpp
*
* Generated automagically from ingredients:
* 1. code generator:
* [xo-facet/codegen/genfacet]
* arguments:
* --input [idl/ISyntaxStateMachine_DParenSsm.json5]
* 2. jinja2 template for abstract facet .hpp file:
* [iface_facet_any.hpp.j2]
* 3. idl for facet methods
* [idl/ISyntaxStateMachine_DParenSsm.json5]
**/
#include "ssm/ISyntaxStateMachine_DParenSsm.hpp"
namespace xo {
namespace scm {
auto
ISyntaxStateMachine_DParenSsm::ssm_type(const DParenSsm & self) noexcept -> syntaxstatetype
{
return self.ssm_type();
}
auto
ISyntaxStateMachine_DParenSsm::get_expect_str(const DParenSsm & self) noexcept -> std::string_view
{
return self.get_expect_str();
}
auto
ISyntaxStateMachine_DParenSsm::on_token(DParenSsm & self, const Token & tk, ParserStateMachine * p_psm) -> void
{
self.on_token(tk, p_psm);
}
auto
ISyntaxStateMachine_DParenSsm::on_parsed_symbol(DParenSsm & self, std::string_view sym, ParserStateMachine * p_psm) -> void
{
self.on_parsed_symbol(sym, p_psm);
}
auto
ISyntaxStateMachine_DParenSsm::on_parsed_typedescr(DParenSsm & self, TypeDescr td, ParserStateMachine * p_psm) -> void
{
self.on_parsed_typedescr(td, p_psm);
}
auto
ISyntaxStateMachine_DParenSsm::on_parsed_formal(DParenSsm & self, const DUniqueString * param_name, TypeDescr param_type, ParserStateMachine * p_psm) -> void
{
self.on_parsed_formal(param_name, param_type, p_psm);
}
auto
ISyntaxStateMachine_DParenSsm::on_parsed_formal_arglist(DParenSsm & self, DArray * arglist, ParserStateMachine * p_psm) -> void
{
self.on_parsed_formal_arglist(arglist, p_psm);
}
auto
ISyntaxStateMachine_DParenSsm::on_parsed_expression(DParenSsm & self, obj<AExpression> expr, ParserStateMachine * p_psm) -> void
{
self.on_parsed_expression(expr, p_psm);
}
auto
ISyntaxStateMachine_DParenSsm::on_parsed_expression_with_token(DParenSsm & self, obj<AExpression> expr, const Token & tk, ParserStateMachine * p_psm) -> void
{
self.on_parsed_expression_with_token(expr, tk, p_psm);
}
} /*namespace scm*/
} /*namespace xo*/
/* end ISyntaxStateMachine_DParenSsm.cpp */

View file

@ -19,6 +19,8 @@
#include <xo/reader2/ssm/IPrintable_DSequenceSsm.hpp>
#include "ParenSsm.hpp"
#include <xo/reader2/ssm/ISyntaxStateMachine_DExpectFormalArglistSsm.hpp>
#include <xo/reader2/ssm/IPrintable_DExpectFormalArglistSsm.hpp>
@ -85,6 +87,9 @@ namespace xo {
FacetRegistry::register_impl<ASyntaxStateMachine, DProgressSsm>();
FacetRegistry::register_impl<APrintable, DProgressSsm>();
FacetRegistry::register_impl<ASyntaxStateMachine, DParenSsm>();
FacetRegistry::register_impl<APrintable, DParenSsm>();
log && log(xtag("DExprSeqState.tseq", typeseq::id<DExprSeqState>()));
log && log(xtag("DDefineSsm.tseq", typeseq::id<DDefineSsm>()));
log && log(xtag("DLambdaSsm.tseq", typeseq::id<DLambdaSsm>()));
@ -95,6 +100,7 @@ namespace xo {
log && log(xtag("DExpectTypeSsm.tseq", typeseq::id<DExpectTypeSsm>()));
log && log(xtag("DExpectExprSsm.tseq", typeseq::id<DExpectExprSsm>()));
log && log(xtag("DProgressSsm.tseq", typeseq::id<DProgressSsm>()));
log && log(xtag("DParenSsm.tseq", typeseq::id<DParenSsm>()));
log && log(xtag("ASyntaxStateMachine.tseq", typeseq::id<ASyntaxStateMachine>()));
return true;

View file

@ -23,6 +23,8 @@ namespace xo {
return "sequence";
case syntaxstatetype::progress:
return "progress";
case syntaxstatetype::paren:
return "paren";
case syntaxstatetype::expect_toplevel_expression_sequence:
return "expect-toplevel-expression-sequence";
case syntaxstatetype::expect_formal_arglist: