xo-reader2: + DDefineSsm + utest

This commit is contained in:
Roland Conybeare 2026-01-19 00:38:43 -05:00
commit d3066ef88d
13 changed files with 762 additions and 3 deletions

View file

@ -20,7 +20,7 @@ add_definitions(${PROJECT_CXX_FLAGS})
# ---------------------------------------------------------------- # ----------------------------------------------------------------
# output targets # output targets
#add_subdirectory(utest) add_subdirectory(utest)
# note: manual target; generated code committed to git # note: manual target; generated code committed to git
xo_add_genfacet( xo_add_genfacet(

View file

@ -0,0 +1,13 @@
{
mode: "implementation",
includes: [ "\"SyntaxStateMachine.hpp\"",
"\"ssm/ISyntaxStateMachine_Xfer.hpp\"" ],
local_types: [ ],
namespace1: "xo",
namespace2: "scm",
facet_idl: "idl/SyntaxStateMachine.json5",
brief: "provide ASyntaxStateMachine interface for DDefineSsm",
using_doxygen: true,
repr: "DDefineSsm",
doc: [ "implement ASyntaxStateMachine for DDefineSsm" ],
}

View file

@ -0,0 +1,97 @@
/** @file DDefineSsm.hpp
*
* @author Roland Conybeare, Jan 2026
**/
#include "ParserStateMachine.hpp"
#include "SyntaxStateMachine.hpp"
#include "syntaxstatetype.hpp"
#include <xo/facet/obj.hpp>
namespace xo {
namespace scm {
/**
* @pre
*
* def foo : f64 = 1 ;
* ^ ^ ^ ^ ^ ^ ^ ^
* | | | | | | | (done)
* | | | | | | def_6:expect_rhs_expression:expr_progress
* | | | | | def_5:expect_rhs_expression
* | | | | def_4
* | | | def_3:expect_type
* | | def_2
* | def_1:expect_symbol
* def_0
* expect_toplevel_expression_sequence
*
* def_0 --on_def_token()--> def_1
* def_1 --on_symbol()--> def_2
* def_2 --on_colon_token()--> def_3
* --on_singleassign_token()--> def_5
* def_3 --on_typedescr()--> def_4
* def_4 --on_singleassign_token()--> def_5
* def_5 --on_expr()--> def_6
* def_6 --on_semicolon_token()--> (done)
*
* def_1:expect_symbol: got 'def' keyword, symbol to follow
* def_1: got symbol name
* def_3:expect_symbol got (optional) colon, type name to follow
* def_4: got symbol type
* def_6:expect_rhs_expression got (optional) equal sign, value to follow
* (done): definition complete, pop exprstate from stack
*
* @endpre
**/
enum class defexprstatetype {
invalid = -1,
def_0,
def_1,
def_2,
def_3,
def_4,
def_5,
def_6,
n_defexprstatetype,
};
extern const char * defexprstatetype_descr(defexprstatetype x);
std::ostream &
operator<<(std::ostream & os, defexprstatetype x);
/** @class DDefineSsm
* @brief state machine for parsing a define expression
**/
class DDefineSsm {
public:
public:
/** @defgroup scm-define-ssm-facet syntaxstatemachine facet methods **/
///@{
/** identifies the ssm implemented here **/
syntaxstatetype ssm_type() const noexcept;
/** text describing expected/allowed input to this ssm in current state.
* Intended to drive error mesages
**/
std::string_view get_expect_str() const noexcept;
/** update state for this syntax on incoming token @p tk,
* overall parser state in @p p_psm
**/
void on_if_token(const Token & tk, ParserStateMachine * p_psm);
///@}
private:
/** identify define-expression state **/
defexprstatetype defstate_;
};
} /*namespace scm*/
} /*namespace xo*/
/* end DDefineSsm.hpp */

View file

@ -198,7 +198,7 @@ namespace xo {
* @return parsed expression, if @p tk completes an expression. * @return parsed expression, if @p tk completes an expression.
* otherwise nullptr * otherwise nullptr
**/ **/
const ParserResult & include_token(const token_type & tk); const ParserResult & on_token(const token_type & tk);
/** reset parsed result expression; use using return value from /** reset parsed result expression; use using return value from
* @ref include_token. Complicating api here to avoid copying ParserResult * @ref include_token. Complicating api here to avoid copying ParserResult

View file

@ -0,0 +1,66 @@
/** @file ISyntaxStateMachine_DDefineSsm.hpp
*
* Generated automagically from ingredients:
* 1. code generator:
* [/Users/roland/proj/xo-umbrella2/xo-facet/codegen/genfacet]
* arguments:
* --input [idl/ISyntaxStateMachine_DDefineSsm.json5]
* 2. jinja2 template for abstract facet .hpp file:
* [iface_facet_repr.hpp.j2]
* 3. idl for facet methods
* [idl/ISyntaxStateMachine_DDefineSsm.json5]
**/
#pragma once
#include "SyntaxStateMachine.hpp"
#include "SyntaxStateMachine.hpp"
#include "ssm/ISyntaxStateMachine_Xfer.hpp"
#include "DDefineSsm.hpp"
namespace xo { namespace scm { class ISyntaxStateMachine_DDefineSsm; } }
namespace xo {
namespace facet {
template <>
struct FacetImplementation<xo::scm::ASyntaxStateMachine,
xo::scm::DDefineSsm>
{
using ImplType = xo::scm::ISyntaxStateMachine_Xfer
<xo::scm::DDefineSsm,
xo::scm::ISyntaxStateMachine_DDefineSsm>;
};
}
}
namespace xo {
namespace scm {
/** @class ISyntaxStateMachine_DDefineSsm
**/
class ISyntaxStateMachine_DDefineSsm {
public:
/** @defgroup scm-syntaxstatemachine-ddefinessm-type-traits **/
///@{
using Copaque = xo::scm::ASyntaxStateMachine::Copaque;
using Opaque = xo::scm::ASyntaxStateMachine::Opaque;
///@}
/** @defgroup scm-syntaxstatemachine-ddefinessm-methods **/
///@{
// const methods
/** identify a type of syntax state machine **/
static syntaxstatetype ssm_type(const DDefineSsm & self) noexcept;
/** text describing expected/allowed input to this ssm in current state **/
static std::string_view get_expect_str(const DDefineSsm & self) noexcept;
// non-const methods
/** update state machine for incoming define-keyworkd-token @p tk **/
static void on_def_token(DDefineSsm & self, const Token & tk, ParserStateMachine * ps_psm);
/** update state machine for incoming if-keyword-token @p tk **/
static void on_if_token(DDefineSsm & self, const Token & tk, ParserStateMachine * p_psm);
///@}
};
} /*namespace scm*/
} /*namespace xo*/
/* end */

View file

@ -21,6 +21,9 @@ namespace xo {
/** toplevel of some translation unit. See @ref DExprSeqState **/ /** toplevel of some translation unit. See @ref DExprSeqState **/
expect_toplevel_expression_sequence, expect_toplevel_expression_sequence,
/** handle define-expression. See @ref DDefineSsm **/
defexpr,
/** comes lasts, counts number of valid enums **/ /** comes lasts, counts number of valid enums **/
N N
}; };

View file

@ -14,6 +14,8 @@ set(SELF_SRCS
DExprSeqState.cpp DExprSeqState.cpp
ISyntaxStateMachine_DExprSeqState.cpp ISyntaxStateMachine_DExprSeqState.cpp
DDefineSsm.cpp
reader2_register_facets.cpp reader2_register_facets.cpp
reader2_register_types.cpp reader2_register_types.cpp
) )

View file

@ -0,0 +1,395 @@
/** @file DDefineSsm.cpp
*
* @author Roland Conybeare, Jan 2026
**/
#include "DDefineSsm.hpp"
#ifdef NOT_YET
#include "parserstatemachine.hpp"
#include "expect_symbol_xs.hpp"
#include "expect_expr_xs.hpp"
#include "expect_type_xs.hpp"
#include "pretty_expression.hpp"
#endif
namespace xo {
namespace scm {
// ----- defexprstatetype -----
const char *
defexprstatetype_descr(defexprstatetype x) {
switch (x) {
case defexprstatetype::invalid: return "invalid";
case defexprstatetype::def_0: return "def_0";
case defexprstatetype::def_1: return "def_1";
case defexprstatetype::def_2: return "def_2";
case defexprstatetype::def_3: return "def_3";
case defexprstatetype::def_4: return "def_4";
case defexprstatetype::def_5: return "def_5";
case defexprstatetype::def_6: return "def_6";
case defexprstatetype::n_defexprstatetype: break;
}
return "???defexprstatetype";
}
std::ostream &
operator<<(std::ostream & os, defexprstatetype x) {
os << defexprstatetype_descr(x);
return os;
}
// ----- define_xs -----
#ifdef NOT_YET
std::unique_ptr<define_xs>
define_xs::make() {
return std::make_unique<define_xs>(define_xs(DefineExprAccess::make_empty()));
}
void
define_xs::start(parserstatemachine * p_psm)
{
scope log(XO_DEBUG(p_psm->debug_flag()));
p_psm->push_exprstate(define_xs::make());
p_psm->top_exprstate().on_def_token(token_type::def(), p_psm);
}
define_xs::define_xs(rp<DefineExprAccess> def_expr)
: exprstate(exprstatetype::defexpr),
defxs_type_{defexprstatetype::def_0},
def_expr_{std::move(def_expr)}
{}
// const char *
// define_xs::get_expect_str() const { ... }
void
define_xs::on_expr(bp<Expression> expr,
parserstatemachine * p_psm)
{
scope log(XO_DEBUG(p_psm->debug_flag()));
log && log(xtag("defxs_type", defxs_type_));
if (this->defxs_type_ == defexprstatetype::def_5) {
/* have all the ingredients to create an expression
* representing a definition
*
* 1. if ir_type is a symbol, interpret as variable name.
* Need to be able to locate variable by type
* 2. if ir_type is an expression, adopt as rhs
*/
rp<Expression> rhs_value = expr.promote();
if (this->cvt_expr_) {
this->cvt_expr_->assign_arg(rhs_value);
} else {
/* note: establishes .def_expr_ valuetype */
this->def_expr_->assign_rhs(rhs_value);
}
rp<Expression> def_expr = this->def_expr_;
this->defxs_type_ = defexprstatetype::def_6;
return;
}
constexpr const char * c_self_name = "define_xs::on_expr";
const char * exp = get_expect_str();
this->illegal_input_on_expr(c_self_name, expr, exp, p_psm);
}
void
define_xs::on_expr_with_semicolon(bp<Expression> expr,
parserstatemachine * p_psm)
{
scope log(XO_DEBUG(p_psm->debug_flag()));
log && log(xtag("defxs_type", defxs_type_));
this->on_expr(expr, p_psm);
/* semicolon is allowed to terminate def expr */
this->on_semicolon_token(token_type::semicolon(), p_psm);
}
void
define_xs::on_symbol(const std::string & symbol_name,
parserstatemachine * p_psm)
{
scope log(XO_DEBUG(p_psm->debug_flag()));
log && log(xtag("defxs_type", defxs_type_), xtag("env_stack_size", p_psm->env_stack_size()));
if (this->defxs_type_ == defexprstatetype::def_1) {
this->defxs_type_ = defexprstatetype::def_2;
this->def_expr_->assign_lhs_name(symbol_name);
// if this is a genuine top-level define (i.e. nesting level = 0),
// then we need to upsert so we can refer to rhs later.
//
// In other contexts (e.g. body-of-lambda) will be rewriting
// {
// def y = foo(x,x);
// bar(y,y);
// }
// into something like
// {
// (lambda (y123) bar(y123,y123))(foo(x,x));
// }
//
// This works in the body of lambda, because we don't evaluate anything
// until lambda definition is complete.
//
// For interactive top-level defs we want to evaluate as we go,
// so need incremental bindings.
if (p_psm->env_stack_size() == 1) {
/* remember variable binding in lexical context,
* so we can refer to it later
*/
p_psm->upsert_var(this->def_expr_->lhs_variable());
}
return;
}
constexpr const char * c_self_name = "define_xs::on_symbol";
const char * exp = this->get_expect_str();
this->illegal_input_on_symbol(c_self_name, symbol_name, exp, p_psm);
}
void
define_xs::on_typedescr(TypeDescr td,
parserstatemachine * p_psm)
{
scope log(XO_DEBUG(p_psm->debug_flag()));
log && log("defxs_type", defxs_type_);
if (this->defxs_type_ == defexprstatetype::def_3) {
this->defxs_type_ = defexprstatetype::def_4;
this->cvt_expr_ = ConvertExprAccess::make(td /*dest_type*/,
nullptr /*source_expr*/);
/* note: establishes .def_expr_ valuetype */
this->def_expr_->assign_rhs(this->cvt_expr_);
return;
}
constexpr const char * c_self_name = "define_xs::on_symbol";
const char * exp = this->get_expect_str();
this->illegal_input_on_type(c_self_name, td, exp, p_psm);
}
void
define_xs::on_def_token(const token_type & tk,
parserstatemachine * p_psm)
{
constexpr bool c_debug_flag = true;
scope log(XO_DEBUG(c_debug_flag));
log && log("defxs_type", defxs_type_);
if (this->defxs_type_ == defexprstatetype::def_0) {
this->defxs_type_ = defexprstatetype::def_1;
expect_symbol_xs::start(p_psm);
return;
}
constexpr const char * c_self_name = "define_xs::on_def_token";
const char * exp = this->get_expect_str();
this->illegal_input_on_token(c_self_name, tk, exp, p_psm);
}
void
define_xs::on_colon_token(const token_type & tk,
parserstatemachine * p_psm)
{
constexpr bool c_debug_flag = true;
scope log(XO_DEBUG(c_debug_flag));
log && log("defxs_type", defxs_type_);
if (this->defxs_type_ == defexprstatetype::def_2) {
this->defxs_type_ = defexprstatetype::def_3;
expect_type_xs::start(p_psm);
return;
}
constexpr const char * c_self_name = "define_xs::on_symbol";
const char * exp = this->get_expect_str();
this->illegal_input_on_token(c_self_name, tk, exp, p_psm);
}
void
define_xs::on_semicolon_token(const token_type & tk,
parserstatemachine * p_psm)
{
/* def expr consumes semicolon */
scope log(XO_DEBUG(p_psm->debug_flag()));
log && log("defxs_type", defxs_type_);
if (this->defxs_type_ == defexprstatetype::def_6) {
rp<DefineExprAccess> def_expr = this->def_expr_;
std::unique_ptr<exprstate> self = p_psm->pop_exprstate();
p_psm->top_exprstate().on_expr(def_expr, p_psm);
return;
}
constexpr const char * c_self_name = "define_xs::on_symbol";
const char * exp = this->get_expect_str();
this->illegal_input_on_token(c_self_name, tk, exp, p_psm);
}
void
define_xs::on_singleassign_token(const token_type & tk,
parserstatemachine * p_psm)
{
constexpr bool c_debug_flag = true;
scope log(XO_DEBUG(c_debug_flag));
log && log("defxs_type", defxs_type_);
if ((this->defxs_type_ == defexprstatetype::def_2)
|| (this->defxs_type_ == defexprstatetype::def_4))
{
this->defxs_type_ = defexprstatetype::def_5;
expect_expr_xs::start(p_psm);
return;
}
constexpr const char * c_self_name = "define_xs::on_singleassign_token";
const char * exp = get_expect_str();
this->illegal_input_on_token(c_self_name, tk, exp, p_psm);
}
void
define_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 = "define_xs::on_rightparen";
const char * exp = get_expect_str();
this->illegal_input_on_token(c_self_name, tk, exp, p_psm);
}
void
define_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 = "define_xs::on_i64";
const char * exp = get_expect_str();
this->illegal_input_on_token(c_self_name, tk, exp, p_psm);
}
void
define_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 = "define_xs::on_f64";
const char * exp = get_expect_str();
this->illegal_input_on_token(c_self_name, tk, exp, p_psm);
}
void
define_xs::print(std::ostream & os) const {
os << "<define_xs"
<< xtag("defxs_type", defxs_type_);
os << ">";
}
bool
define_xs::pretty_print(const xo::print::ppindentinfo & ppii) const
{
return ppii.pps()->pretty_struct(ppii, "define_xs",
refrtag("defxs_type", defxs_type_));
}
#endif
////////////////////////////////////////////////////////////////
syntaxstatetype
DDefineSsm::ssm_type() const noexcept
{
return syntaxstatetype::defexpr;
}
std::string_view
DDefineSsm::get_expect_str() const noexcept
{
/*
* def foo = 1 ;
* def foo : f64 = 1 ;
* ^ ^ ^ ^ ^ ^ ^ ^
* | | | | | | | (done)
* | | | | | | def_6
* | | | | | def_5:expect_rhs_expression
* | | | | def_4
* | | | def_3:expect_type
* | | def_2
* | def_1:expect_symbol
* expect_toplevel_expression_sequence
*
* note that we skip from def_2 -> def_5 if '=' instead of ':'
*/
switch (this->defstate_) {
case defexprstatetype::invalid:
case defexprstatetype::def_0:
case defexprstatetype::n_defexprstatetype:
assert(false); // impossible
return nullptr;
case defexprstatetype::def_1:
return "symbol";
case defexprstatetype::def_2:
return "singleassign|colon";
case defexprstatetype::def_4:
return "singleassign";
case defexprstatetype::def_3:
return "type";
case defexprstatetype::def_5:
return "expression";
case defexprstatetype::def_6:
return "semicolon";
}
return "?expect";
}
void
DDefineSsm::on_if_token(const Token & tk,
ParserStateMachine * p_psm)
{
p_psm->illegal_input_on_token("DDefineSsm::on_if_token",
tk,
this->get_expect_str());
}
} /*namespace scm*/
} /*namespace xo*/
/* end DDefineSsm.cpp */

View file

@ -0,0 +1,50 @@
/** @file ISyntaxStateMachine_DDefineSsm.cpp
*
* Generated automagically from ingredients:
* 1. code generator:
* [/Users/roland/proj/xo-umbrella2/xo-facet/codegen/genfacet]
* arguments:
* --input [idl/ISyntaxStateMachine_DDefineSsm.json5]
* 2. jinja2 template for abstract facet .hpp file:
* [iface_facet_any.hpp.j2]
* 3. idl for facet methods
* [idl/ISyntaxStateMachine_DDefineSsm.json5]
**/
#include "ssm/ISyntaxStateMachine_DDefineSsm.hpp"
namespace xo {
namespace scm {
auto
ISyntaxStateMachine_DDefineSsm::ssm_type(const DDefineSsm & self) noexcept
-> syntaxstatetype
{
return self.ssm_type();
}
auto
ISyntaxStateMachine_DDefineSsm::get_expect_str(const DDefineSsm & self) noexcept
-> std::string_view
{
return self.get_expect_str();
}
auto
ISyntaxStateMachine_DDefineSsm::on_def_token(DDefineSsm & self,
const Token & tk,
ParserStateMachine * ps_psm) -> void
{
self.on_def_token(tk, ps_psm);
}
auto
ISyntaxStateMachine_DDefineSsm::on_if_token(DDefineSsm & self,
const Token & tk,
ParserStateMachine * p_psm) -> void
{
self.on_if_token(tk, p_psm);
}
} /*namespace scm*/
} /*namespace xo*/
/* end ISyntaxStateMachine_DDefineSsm.cpp */

View file

@ -48,7 +48,7 @@ namespace xo {
} }
const ParserResult & const ParserResult &
SchematikaParser::include_token(const token_type & tk) SchematikaParser::on_token(const token_type & tk)
{ {
scope log(XO_DEBUG(debug_flag_), xtag("tk", tk)); scope log(XO_DEBUG(debug_flag_), xtag("tk", tk));

View file

@ -0,0 +1,11 @@
# build unittest xo-reader2/utest
set(UTEST_EXE utest.reader2)
set(UTEST_SRCS
reader2_utest_main.cpp
SchematikaParser.test.cpp
)
xo_add_utest_executable(${UTEST_EXE} ${UTEST_SRCS})
xo_self_dependency(${UTEST_EXE} xo_reader2)
xo_external_target_dependency(${UTEST_EXE} Catch2 Catch2::Catch2)

View file

@ -0,0 +1,95 @@
/** @file SchematikaParser.test.cpp
*
* @author Roland Conybeare, Jan 2026
**/
#include <xo/reader2/SchematikaParser.hpp>
#include <xo/alloc2/arena/IAllocator_DArena.hpp>
#include <catch2/catch.hpp>
namespace xo {
using xo::scm::SchematikaParser;
using xo::scm::ParserResult;
using xo::scm::parser_result_type;
using xo::scm::Token;
using xo::mm::ArenaConfig;
using xo::mm::AAllocator;
using xo::mm::DArena;
using xo::facet::with_facet;
namespace ut {
TEST_CASE("SchematikaParser-ctor", "[reader2][SchematikaParser]")
{
ArenaConfig config;
config.name_ = "test-arena";
config.size_ = 16 * 1024;
DArena expr_arena = DArena::map(config);
obj<AAllocator> expr_alloc = with_facet<AAllocator>::mkobj(&expr_arena);
SchematikaParser parser(config, &expr_alloc, false /*debug_flag*/);
REQUIRE(parser.debug_flag() == false);
REQUIRE(parser.is_at_toplevel() == true);
}
TEST_CASE("SchematikaParser-begin-interactive", "[reader2][SchematikaParser]")
{
ArenaConfig config;
config.name_ = "test-arena";
config.size_ = 16 * 1024;
DArena expr_arena = DArena::map(config);
obj<AAllocator> expr_alloc = with_facet<AAllocator>::mkobj(&expr_arena);
SchematikaParser parser(config, &expr_alloc, false /*debug_flag*/);
parser.begin_interactive_session();
// after begin_interactive_session, parser has toplevel exprseq
// but is still "at toplevel" in the sense of ready for input
REQUIRE(parser.has_incomplete_expr() == false);
}
TEST_CASE("SchematikaParser-begin-batch", "[reader2][SchematikaParser]")
{
ArenaConfig config;
config.name_ = "test-arena";
config.size_ = 16 * 1024;
DArena expr_arena = DArena::map(config);
obj<AAllocator> expr_alloc = with_facet<AAllocator>::mkobj(&expr_arena);
SchematikaParser parser(config, &expr_alloc, false /*debug_flag*/);
parser.begin_translation_unit();
// after begin_translation_unit, parser has toplevel exprseq
// but is still "at toplevel" in the sense of ready for input
REQUIRE(parser.has_incomplete_expr() == false);
}
TEST_CASE("SchematikaParser-interactive-if", "[reader2][SchematikaParser]")
{
ArenaConfig config;
config.name_ = "test-arena";
config.size_ = 16 * 1024;
DArena expr_arena = DArena::map(config);
obj<AAllocator> expr_alloc = with_facet<AAllocator>::mkobj(&expr_arena);
SchematikaParser parser(config, &expr_alloc, false /*debug_flag*/);
parser.begin_interactive_session();
parser.on_token(Token::if_token());
// after begin_interactive_session, parser has toplevel exprseq
// but is still "at toplevel" in the sense of ready for input
REQUIRE(parser.has_incomplete_expr() == false);
}
} /*namespace ut*/
} /*namespace xo*/
/* end SchematikaParser.test.cpp */

View file

@ -0,0 +1,27 @@
/** @file reader2_utest_main.cpp
*
* @author Roland Conybeare, Jan 2026
**/
#include <xo/subsys/Subsystem.hpp>
#define CATCH_CONFIG_RUNNER
#include "catch2/catch.hpp"
int
main(int argc, char* argv[])
{
using xo::Subsystem;
// initialize subsystems
Subsystem::initialize_all();
// Run Catch2's test session
int result = Catch::Session().run(argc, argv);
// cleanup here, if any
return result;
}
/* end reader2_utest_main.cpp */