roland/xo-reader2
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

xo-expression2: + DApplyExpr::make

Browse source
This commit is contained in:
Roland Conybeare 2026-01-26 15:45:40 -05:00
commit 774a6ebe22
1 changed files with 25 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

28
src/reader2/DProgressSsm.cpp
View file

@ -9,9 +9,15 @@
#include "DExpectExprSsm.hpp"
#include "ssm/ISyntaxStateMachine_DExpectExprSsm.hpp"
#include <xo/expression2/DApplyExpr.hpp>
#include <xo/expression2/detail/IExpression_DApplyExpr.hpp>
#include <xo/expression2/DConstant.hpp>
#include <xo/expression2/detail/IExpression_DConstant.hpp>
#include <xo/procedure2/init_primitives.hpp> // for xo::scm::Primitives
#include <xo/procedure2/detail/IGCObject_DPrimitive_gco_2_gco_gco.hpp>
#ifdef NOT_YET
#include "DApplySsm.hpp"
#include "ssm/ISyntaxStateMachine_DApplySsm.hpp"
@ -1179,20 +1185,36 @@ namespace xo {
break;
case optype::op_multiply:
#ifdef NOT_YET
{
auto pm_obj = with_facet<AGCObject>::mkobj(&Primitives::s_mul_gco_gco_pm);
auto fn_expr = DConstant::make(p_psm->expr_alloc(), pm_obj);
/* note:
* 1. don't assume we know lhs_ / rhs_ value types yet.
* perhaps have expression like
* f(..) * g(..)
* where f is the function that contains current ssm.
*
* 2. consequence: we need representation for
* polymorphic multiply on unknown numeric arguments.
*
* 3. TypeRef::dwim(..) is a placeholder.
* Plan to later provide abstract interpreter
* (ie compiler pass :) to drive type inference/unification
*
* 4. Alternatively could supply type-annotation syntax
* so human can assist inference; context here is we want
* to automate the boring stuff
*/
DApplyExpr::make2();
TypeRef tref = TypeRef::dwim
(TypeRef::prefix_type::from_chars("_mul_gco"),
nullptr);
return DApplyExpr::make2(p_psm->expr_alloc(),
tref, fn_expr, lhs_, rhs_);
}
#endif
break;
case optype::op_divide: