* [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609]
@ 2024-01-07 0:00 waffl3x
2024-01-09 20:58 ` Jason Merrill
2024-01-10 17:15 ` Patrick Palka
0 siblings, 2 replies; 8+ messages in thread
From: waffl3x @ 2024-01-07 0:00 UTC (permalink / raw)
To: waffl3x; +Cc: Jason Merrill, gcc-patches
[-- Attachment #1: Type: text/plain, Size: 635 bytes --]
Bootstrapped and tested on x86_64-linux with no regressions.
I'm considering this finished, I have CWG2586 working but I have not
included it in this version of the patch. I was not happy with the
amount of work I had done on it. I will try to get it finished before
we get cut off, and I'm pretty sure I can. I just don't want to risk
missing the boat for the whole patch just for that.
There aren't too many changes from v7, it's mostly just cleaned up.
There are a few though, so do take a look, if there's anything severe I
can rush to fix it if necessary.
That's all, hopefully all is good, fingers crossed.
Alex
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-C-23-P0847R7-deducing-this-prerequisite-changes.-PR1.patch --]
[-- Type: text/x-patch; name=0001-C-23-P0847R7-deducing-this-prerequisite-changes.-PR1.patch, Size: 34823 bytes --]
From cd122053dfad741a7d90adcd45929af768ce643f Mon Sep 17 00:00:00 2001
From: Waffl3x <waffl3x@protonmail.com>
Date: Sun, 31 Dec 2023 03:16:36 -0700
Subject: [PATCH 1/4] C++23 P0847R7 (deducing this) - prerequisite changes.
[PR102609]
Adds the xobj_flag member to lang_decl_fn and a corresponding member access
macro and predicate to support the addition of explicit object member
functions. Additionally, since explicit object member functions are also
non-static member functions, we need to change uses of
DECL_NONSTATIC_MEMBER_FUNCTION_P to clarify whether they intend to include or
exclude them.
Many of these alterations are authored by Jason Merril.
PR c++/102609
gcc/cp/ChangeLog:
PR c++/102609
C++23 P0847R7 (deducing this) - prerequisite changes. [PR102609]
* cp-tree.h (struct lang_decl_fn): New data member.
(DECL_NONSTATIC_MEMBER_FUNCTION_P): Poison.
(DECL_IOBJ_MEMBER_FUNCTION_P): Define.
(DECL_FUNCTION_XOBJ_FLAG): Define.
(DECL_XOBJ_MEMBER_FUNCTION_P): Define.
(DECL_OBJECT_MEMBER_FUNCTION_P): Define.
(DECL_FUNCTION_MEMBER_P): Don't use DECL_NONSTATIC_MEMBER_FUNCTION_P.
(DECL_CONST_MEMFUNC_P): Likewise.
(DECL_VOLATILE_MEMFUNC_P): Likewise.
(DECL_NONSTATIC_MEMBER_P): Likewise.
* module.cc (trees_out::lang_decl_bools): Handle xobj_flag.
(trees_in::lang_decl_bools): Handle xobj_flag.
* call.cc (build_this_conversion):
(add_function_candidate):
(add_template_candidate_real):
(add_candidates):
(maybe_warn_class_memaccess):
(cand_parms_match):
(joust):
(do_warn_dangling_reference): Don't use it.
* class.cc (finalize_literal_type_property):
(finish_struct):
(resolve_address_of_overloaded_function):
* constexpr.cc (is_valid_constexpr_fn):
(cxx_bind_parameters_in_call):
* contracts.cc (build_contract_condition_function):
* cp-objcp-common.cc (cp_decl_dwarf_attribute):
* cxx-pretty-print.cc (cxx_pretty_printer::postfix_expression):
(cxx_pretty_printer::declaration_specifiers):
(cxx_pretty_printer::direct_declarator):
* decl.cc (cp_finish_decl):
(grok_special_member_properties):
(start_preparsed_function):
(record_key_method_defined):
* decl2.cc (cp_handle_deprecated_or_unavailable):
* init.cc (find_uninit_fields_r):
(build_offset_ref):
* lambda.cc (lambda_expr_this_capture):
(maybe_generic_this_capture):
(nonlambda_method_basetype):
* mangle.cc (write_nested_name):
* method.cc (early_check_defaulted_comparison):
(skip_artificial_parms_for):
(num_artificial_parms_for):
* module.cc (trees_out::lang_decl_bools):
(trees_in::lang_decl_bools):
* pt.cc (is_specialization_of_friend):
(determine_specialization):
(copy_default_args_to_explicit_spec):
(check_explicit_specialization):
(tsubst_contract_attribute):
(check_non_deducible_conversions):
(more_specialized_fn):
(maybe_instantiate_noexcept):
(register_parameter_specializations):
(value_dependent_expression_p):
* search.cc (shared_member_p):
(lookup_member):
(field_access_p):
* semantics.cc (finish_omp_declare_simd_methods):
* tree.cc (lvalue_kind):
* typeck.cc (invalid_nonstatic_memfn_p): Don't use
DECL_NONSTATIC_MEMBER_FUNCTION_P.
libcc1/ChangeLog:
PR c++/102609
C++23 P0847R7 (deducing this) - prerequisite changes. [PR102609]
* libcp1plugin.cc (plugin_pragma_push_user_expression): Don't use
DECL_NONSTATIC_MEMBER_FUNCTION_P.
Signed-off-by: Waffl3x <waffl3x@protonmail.com>
---
gcc/cp/call.cc | 25 +++++++++++++------------
gcc/cp/class.cc | 6 +++---
gcc/cp/constexpr.cc | 4 ++--
gcc/cp/contracts.cc | 6 +++---
gcc/cp/cp-objcp-common.cc | 4 ++--
gcc/cp/cp-tree.h | 37 ++++++++++++++++++++++++++++++-------
gcc/cp/cxx-pretty-print.cc | 6 +++---
gcc/cp/decl.cc | 8 ++++----
gcc/cp/decl2.cc | 2 +-
gcc/cp/init.cc | 4 ++--
gcc/cp/lambda.cc | 9 +++++----
gcc/cp/mangle.cc | 4 ++--
gcc/cp/method.cc | 8 ++++----
gcc/cp/module.cc | 6 ++++--
gcc/cp/pt.cc | 38 +++++++++++++++++++-------------------
gcc/cp/search.cc | 5 +++--
gcc/cp/semantics.cc | 2 +-
gcc/cp/tree.cc | 2 +-
gcc/cp/typeck.cc | 2 +-
libcc1/libcp1plugin.cc | 2 +-
20 files changed, 104 insertions(+), 76 deletions(-)
diff --git a/gcc/cp/call.cc b/gcc/cp/call.cc
index 6ac87a298b2..46f2ccdfc48 100644
--- a/gcc/cp/call.cc
+++ b/gcc/cp/call.cc
@@ -2309,7 +2309,7 @@ build_this_conversion (tree fn, tree ctype,
tree& parmtype, tree& argtype, tree& arg,
int flags, tsubst_flags_t complain)
{
- gcc_assert (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn)
+ gcc_assert (DECL_IOBJ_MEMBER_FUNCTION_P (fn)
&& !DECL_CONSTRUCTOR_P (fn));
/* The type of the implicit object parameter ('this') for
@@ -2522,7 +2522,7 @@ add_function_candidate (struct z_candidate **candidates,
{
tree parmtype = TREE_VALUE (parmnode);
if (i == 0
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (fn)
+ && DECL_IOBJ_MEMBER_FUNCTION_P (fn)
&& !DECL_CONSTRUCTOR_P (fn))
t = build_this_conversion (fn, ctype, parmtype, argtype, arg,
flags, complain);
@@ -3479,7 +3479,7 @@ add_template_candidate_real (struct z_candidate **candidates, tree tmpl,
/* We don't do deduction on the in-charge parameter, the VTT
parameter or 'this'. */
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (tmpl))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (tmpl))
{
if (first_arg_without_in_chrg != NULL_TREE)
first_arg_without_in_chrg = NULL_TREE;
@@ -3589,7 +3589,7 @@ add_template_candidate_real (struct z_candidate **candidates, tree tmpl,
convs = alloc_conversions (nargs);
if (shortcut_bad_convs
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (tmpl)
+ && DECL_IOBJ_MEMBER_FUNCTION_P (tmpl)
&& !DECL_CONSTRUCTOR_P (tmpl))
{
/* Check the 'this' conversion before proceeding with deduction.
@@ -6617,7 +6617,7 @@ add_candidates (tree fns, tree first_arg, const vec<tree, va_gc> *args,
tree fn_first_arg = NULL_TREE;
const vec<tree, va_gc> *fn_args = args;
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn))
+ if (DECL_OBJECT_MEMBER_FUNCTION_P (fn))
{
/* Figure out where the object arg comes from. If this
function is a non-static member and we didn't get an
@@ -10784,7 +10784,8 @@ maybe_warn_class_memaccess (location_t loc, tree fndecl,
type. If so, and if the class has no non-trivial bases or members,
be more permissive. */
if (current_function_decl
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (current_function_decl)
+ && DECL_OBJECT_MEMBER_FUNCTION_P (current_function_decl)
+ /* ??? is_object_parameter? */
&& is_this_parameter (tree_strip_nop_conversions (dest)))
{
tree ctx = DECL_CONTEXT (current_function_decl);
@@ -12684,8 +12685,8 @@ cand_parms_match (z_candidate *c1, z_candidate *c2)
tree parms2 = TYPE_ARG_TYPES (TREE_TYPE (fn2));
if (DECL_FUNCTION_MEMBER_P (fn1)
&& DECL_FUNCTION_MEMBER_P (fn2)
- && (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn1)
- != DECL_NONSTATIC_MEMBER_FUNCTION_P (fn2)))
+ && (DECL_STATIC_FUNCTION_P (fn1)
+ != DECL_STATIC_FUNCTION_P (fn2)))
{
/* Ignore 'this' when comparing the parameters of a static member
function with those of a non-static one. */
@@ -12870,7 +12871,7 @@ joust (struct z_candidate *cand1, struct z_candidate *cand2, bool warn,
print_z_candidate (input_location,
N_("candidate 2:"), l);
if (w->fn == l->fn
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (w->fn)
+ && DECL_IOBJ_MEMBER_FUNCTION_P (w->fn)
&& (type_memfn_quals (TREE_TYPE (w->fn))
& TYPE_QUAL_CONST) == 0)
{
@@ -14019,7 +14020,7 @@ do_warn_dangling_reference (tree expr, bool arg_p)
because R refers to one of the int elements of V, not to
a temporary object. Member operator* may return a reference
but probably not to one of its arguments. */
- || (DECL_NONSTATIC_MEMBER_FUNCTION_P (fndecl)
+ || (DECL_OBJECT_MEMBER_FUNCTION_P (fndecl)
&& DECL_OVERLOADED_OPERATOR_P (fndecl)
&& DECL_OVERLOADED_OPERATOR_IS (fndecl, INDIRECT_REF)))
return NULL_TREE;
@@ -14045,7 +14046,7 @@ do_warn_dangling_reference (tree expr, bool arg_p)
tree arg = CALL_EXPR_ARG (expr, i);
/* Check that this argument initializes a reference, except for
the argument initializing the object of a member function. */
- if (!DECL_NONSTATIC_MEMBER_FUNCTION_P (fndecl)
+ if (!DECL_IOBJ_MEMBER_FUNCTION_P (fndecl)
&& !TYPE_REF_P (TREE_TYPE (arg)))
continue;
STRIP_NOPS (arg);
@@ -14065,7 +14066,7 @@ do_warn_dangling_reference (tree expr, bool arg_p)
const S& s = S().self();
where 's' dangles. If we've gotten here, the object this function
is invoked on is not a temporary. */
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fndecl))
+ if (DECL_OBJECT_MEMBER_FUNCTION_P (fndecl))
break;
}
return NULL_TREE;
diff --git a/gcc/cp/class.cc b/gcc/cp/class.cc
index 1954e0a5ed3..692b238dd7e 100644
--- a/gcc/cp/class.cc
+++ b/gcc/cp/class.cc
@@ -5968,7 +5968,7 @@ finalize_literal_type_property (tree t)
for (fn = TYPE_FIELDS (t); fn; fn = DECL_CHAIN (fn))
if (TREE_CODE (fn) == FUNCTION_DECL
&& DECL_DECLARED_CONSTEXPR_P (fn)
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (fn)
+ && DECL_IOBJ_MEMBER_FUNCTION_P (fn)
&& !DECL_CONSTRUCTOR_P (fn))
{
DECL_DECLARED_CONSTEXPR_P (fn) = false;
@@ -7932,7 +7932,7 @@ finish_struct (tree t, tree attributes)
if (flag_openmp)
for (tree decl = TYPE_FIELDS (t); decl; decl = DECL_CHAIN (decl))
if (TREE_CODE (decl) == FUNCTION_DECL
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (decl))
+ && DECL_OBJECT_MEMBER_FUNCTION_P (decl))
if (tree attr = lookup_attribute ("omp declare variant base",
DECL_ATTRIBUTES (decl)))
omp_declare_variant_finalize (decl, attr);
@@ -8727,7 +8727,7 @@ resolve_address_of_overloaded_function (tree target_type,
/* Good, exactly one match. Now, convert it to the correct type. */
fn = TREE_PURPOSE (matches);
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn)
+ if (DECL_OBJECT_MEMBER_FUNCTION_P (fn)
&& !(complain & tf_ptrmem_ok) && !flag_ms_extensions)
{
static int explained;
diff --git a/gcc/cp/constexpr.cc b/gcc/cp/constexpr.cc
index eff43a42353..b12473c44af 100644
--- a/gcc/cp/constexpr.cc
+++ b/gcc/cp/constexpr.cc
@@ -291,7 +291,7 @@ is_valid_constexpr_fn (tree fun, bool complain)
/* C++14 DR 1684 removed this restriction. */
if (cxx_dialect < cxx14
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (fun)
+ && DECL_IOBJ_MEMBER_FUNCTION_P (fun)
&& !CLASSTYPE_LITERAL_P (DECL_CONTEXT (fun)))
{
ret = false;
@@ -1886,7 +1886,7 @@ cxx_bind_parameters_in_call (const constexpr_ctx *ctx, tree t, tree fun,
non_constant_p, overflow_p);
/* Check we aren't dereferencing a null pointer when calling a non-static
member function, which is undefined behaviour. */
- if (i == 0 && DECL_NONSTATIC_MEMBER_FUNCTION_P (fun)
+ if (i == 0 && DECL_OBJECT_MEMBER_FUNCTION_P (fun)
&& integer_zerop (arg)
/* But ignore calls from within compiler-generated code, to handle
cases like lambda function pointer conversion operator thunks
diff --git a/gcc/cp/contracts.cc b/gcc/cp/contracts.cc
index 035ca4827e7..c505b67d339 100644
--- a/gcc/cp/contracts.cc
+++ b/gcc/cp/contracts.cc
@@ -1398,7 +1398,7 @@ build_contract_condition_function (tree fndecl, bool pre)
{
if (TREE_TYPE (fndecl) == error_mark_node)
return error_mark_node;
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fndecl)
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (fndecl)
&& !TYPE_METHOD_BASETYPE (TREE_TYPE (fndecl)))
return error_mark_node;
@@ -1421,7 +1421,7 @@ build_contract_condition_function (tree fndecl, bool pre)
arg_type && arg_type != void_list_node;
arg_type = TREE_CHAIN (arg_type))
{
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fndecl)
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (fndecl)
&& TYPE_ARG_TYPES (TREE_TYPE (fn)) == arg_type)
{
class_type = TREE_TYPE (TREE_VALUE (arg_type));
@@ -1451,7 +1451,7 @@ build_contract_condition_function (tree fndecl, bool pre)
}
TREE_TYPE (fn) = build_function_type (value_type, arg_types);
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fndecl))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (fndecl))
TREE_TYPE (fn) = build_method_type (class_type, TREE_TYPE (fn));
DECL_NAME (fn) = copy_node (DECL_NAME (fn));
diff --git a/gcc/cp/cp-objcp-common.cc b/gcc/cp/cp-objcp-common.cc
index ee88df5767b..3970c4cddd5 100644
--- a/gcc/cp/cp-objcp-common.cc
+++ b/gcc/cp/cp-objcp-common.cc
@@ -349,7 +349,7 @@ cp_decl_dwarf_attribute (const_tree decl, int attr)
case DW_AT_reference:
if (TREE_CODE (decl) == FUNCTION_DECL
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (decl)
+ && DECL_IOBJ_MEMBER_FUNCTION_P (decl)
&& FUNCTION_REF_QUALIFIED (TREE_TYPE (decl))
&& !FUNCTION_RVALUE_QUALIFIED (TREE_TYPE (decl)))
return 1;
@@ -357,7 +357,7 @@ cp_decl_dwarf_attribute (const_tree decl, int attr)
case DW_AT_rvalue_reference:
if (TREE_CODE (decl) == FUNCTION_DECL
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (decl)
+ && DECL_IOBJ_MEMBER_FUNCTION_P (decl)
&& FUNCTION_REF_QUALIFIED (TREE_TYPE (decl))
&& FUNCTION_RVALUE_QUALIFIED (TREE_TYPE (decl)))
return 1;
diff --git a/gcc/cp/cp-tree.h b/gcc/cp/cp-tree.h
index 1979572c365..41d480f8e6e 100644
--- a/gcc/cp/cp-tree.h
+++ b/gcc/cp/cp-tree.h
@@ -2962,8 +2962,9 @@ struct GTY(()) lang_decl_fn {
unsigned coroutine_p : 1;
unsigned implicit_constexpr : 1;
unsigned escalated_p : 1;
+ unsigned xobj_func : 1;
- unsigned spare : 8;
+ unsigned spare : 7;
/* 32-bits padding on 64-bit host. */
@@ -3361,33 +3362,55 @@ struct GTY(()) lang_decl {
#define DECL_STATIC_FUNCTION_P(NODE) \
(LANG_DECL_FN_CHECK (NODE)->static_function)
-/* Nonzero for FUNCTION_DECL means that this decl is a non-static
+/* Nonzero for FUNCTION_DECL means that this decl is a non-static member
+ function. C++23 explicit object member functions are also considered
+ non-static, but most former uses of this macro meant implicit object member
+ function. Instead of this macro, use DECL_IOBJ_MEMBER_FUNCTION_P or
+ DECL_OBJECT_MEMBER_FUNCTION_P. */
+#define DECL_NONSTATIC_MEMBER_FUNCTION_P(NODE) did_you_mean_object_or_iobj
+
+/* Nonzero for FUNCTION_DECL means that this decl is an implicit object
member function. */
-#define DECL_NONSTATIC_MEMBER_FUNCTION_P(NODE) \
+#define DECL_IOBJ_MEMBER_FUNCTION_P(NODE) \
(TREE_CODE (TREE_TYPE (NODE)) == METHOD_TYPE)
+/* Simple member access, only valid for FUNCTION_DECL nodes. */
+#define DECL_FUNCTION_XOBJ_FLAG(NODE) \
+ (LANG_DECL_FN_CHECK (NODE)->xobj_func)
+
+/* Nonzero if NODE is an xobj member function,
+ safely evaluates to false for all non FUNCTION_DECL nodes. */
+#define DECL_XOBJ_MEMBER_FUNCTION_P(NODE) \
+ (TREE_CODE (STRIP_TEMPLATE (NODE)) == FUNCTION_DECL \
+ && DECL_FUNCTION_XOBJ_FLAG (NODE) == 1)
+
+/* Nonzero if NODE is a member function with an object argument,
+ in other words, a non-static member function. */
+#define DECL_OBJECT_MEMBER_FUNCTION_P(NODE) \
+ (DECL_IOBJ_MEMBER_FUNCTION_P (NODE) || DECL_XOBJ_MEMBER_FUNCTION_P (NODE))
+
/* Nonzero for FUNCTION_DECL means that this decl is a member function
(static or non-static). */
#define DECL_FUNCTION_MEMBER_P(NODE) \
- (DECL_NONSTATIC_MEMBER_FUNCTION_P (NODE) || DECL_STATIC_FUNCTION_P (NODE))
+ (DECL_OBJECT_MEMBER_FUNCTION_P (NODE) || DECL_STATIC_FUNCTION_P (NODE)) \
/* Nonzero for FUNCTION_DECL means that this member function
has `this' as const X *const. */
#define DECL_CONST_MEMFUNC_P(NODE) \
- (DECL_NONSTATIC_MEMBER_FUNCTION_P (NODE) \
+ (DECL_IOBJ_MEMBER_FUNCTION_P (NODE) \
&& CP_TYPE_CONST_P (TREE_TYPE (TREE_VALUE \
(TYPE_ARG_TYPES (TREE_TYPE (NODE))))))
/* Nonzero for FUNCTION_DECL means that this member function
has `this' as volatile X *const. */
#define DECL_VOLATILE_MEMFUNC_P(NODE) \
- (DECL_NONSTATIC_MEMBER_FUNCTION_P (NODE) \
+ (DECL_IOBJ_MEMBER_FUNCTION_P (NODE) \
&& CP_TYPE_VOLATILE_P (TREE_TYPE (TREE_VALUE \
(TYPE_ARG_TYPES (TREE_TYPE (NODE))))))
/* Nonzero for a DECL means that this member is a non-static member. */
#define DECL_NONSTATIC_MEMBER_P(NODE) \
- (DECL_NONSTATIC_MEMBER_FUNCTION_P (NODE) \
+ (DECL_OBJECT_MEMBER_FUNCTION_P (NODE) \
|| TREE_CODE (NODE) == FIELD_DECL)
/* Nonzero for a FIELD_DECL means that this member object type
diff --git a/gcc/cp/cxx-pretty-print.cc b/gcc/cp/cxx-pretty-print.cc
index 6a82358f370..53c0722d2f3 100644
--- a/gcc/cp/cxx-pretty-print.cc
+++ b/gcc/cp/cxx-pretty-print.cc
@@ -553,7 +553,7 @@ cxx_pretty_printer::postfix_expression (tree t)
instantiation time. */
if (TREE_CODE (fun) != FUNCTION_DECL)
;
- else if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fun))
+ else if (DECL_OBJECT_MEMBER_FUNCTION_P (fun))
{
tree object = (code == AGGR_INIT_EXPR
? (AGGR_INIT_VIA_CTOR_P (t)
@@ -1342,7 +1342,7 @@ cxx_pretty_printer::declaration_specifiers (tree t)
do not have a type-specifier in their return types. */
if (DECL_CONSTRUCTOR_P (t) || DECL_CONV_FN_P (t))
function_specifier (t);
- else if (DECL_NONSTATIC_MEMBER_FUNCTION_P (t))
+ else if (DECL_IOBJ_MEMBER_FUNCTION_P (t))
declaration_specifiers (TREE_TYPE (TREE_TYPE (t)));
else
c_pretty_printer::declaration_specifiers (t);
@@ -1700,7 +1700,7 @@ cxx_pretty_printer::direct_declarator (tree t)
expression (t);
pp_cxx_parameter_declaration_clause (this, t);
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (t))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (t))
{
padding = pp_before;
pp_cxx_cv_qualifier_seq (this, pp_cxx_implicit_parameter_type (t));
diff --git a/gcc/cp/decl.cc b/gcc/cp/decl.cc
index 27f17808934..b5b23b86b62 100644
--- a/gcc/cp/decl.cc
+++ b/gcc/cp/decl.cc
@@ -8526,7 +8526,7 @@ cp_finish_decl (tree decl, tree init, bool init_const_expr_p,
&& TREE_CODE (decl) == FUNCTION_DECL
/* #pragma omp declare variant on methods handled in finish_struct
instead. */
- && (!DECL_NONSTATIC_MEMBER_FUNCTION_P (decl)
+ && (!DECL_OBJECT_MEMBER_FUNCTION_P (decl)
|| COMPLETE_TYPE_P (DECL_CONTEXT (decl))))
if (tree attr = lookup_attribute ("omp declare variant base",
DECL_ATTRIBUTES (decl)))
@@ -15516,7 +15516,7 @@ grok_special_member_properties (tree decl)
tree class_type;
if (TREE_CODE (decl) == USING_DECL
- || !DECL_NONSTATIC_MEMBER_FUNCTION_P (decl))
+ || !DECL_OBJECT_MEMBER_FUNCTION_P (decl))
return;
class_type = DECL_CONTEXT (decl);
@@ -17706,7 +17706,7 @@ start_preparsed_function (tree decl1, tree attrs, int flags)
/* Start the statement-tree, start the tree now. */
DECL_SAVED_TREE (decl1) = push_stmt_list ();
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (decl1))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (decl1))
{
/* We know that this was set up by `grokclassfn'. We do not
wait until `store_parm_decls', since evil parse errors may
@@ -18175,7 +18175,7 @@ outer_curly_brace_block (tree fndecl)
static void
record_key_method_defined (tree fndecl)
{
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fndecl)
+ if (DECL_OBJECT_MEMBER_FUNCTION_P (fndecl)
&& DECL_VIRTUAL_P (fndecl)
&& !processing_template_decl)
{
diff --git a/gcc/cp/decl2.cc b/gcc/cp/decl2.cc
index 0850d3f5bce..c8eddf4cc26 100644
--- a/gcc/cp/decl2.cc
+++ b/gcc/cp/decl2.cc
@@ -5625,7 +5625,7 @@ cp_handle_deprecated_or_unavailable (tree decl, tsubst_flags_t complain)
if (cxx_dialect >= cxx11
&& DECL_P (decl)
&& DECL_ARTIFICIAL (decl)
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (decl)
+ && DECL_IOBJ_MEMBER_FUNCTION_P (decl)
&& copy_fn_p (decl))
{
/* Don't warn if the flag was disabled around the class definition
diff --git a/gcc/cp/init.cc b/gcc/cp/init.cc
index 6444f0a8518..fb5f404805b 100644
--- a/gcc/cp/init.cc
+++ b/gcc/cp/init.cc
@@ -866,7 +866,7 @@ find_uninit_fields_r (tree *tp, int *walk_subtrees, void *data)
else if (code == CALL_EXPR)
{
tree fn = get_callee_fndecl (init);
- if (fn && DECL_NONSTATIC_MEMBER_FUNCTION_P (fn))
+ if (fn && DECL_IOBJ_MEMBER_FUNCTION_P (fn))
{
tree op = CALL_EXPR_ARG (init, 0);
if (TREE_CODE (op) == ADDR_EXPR)
@@ -2477,7 +2477,7 @@ build_offset_ref (tree type, tree member, bool address_p,
-- in a mem-initializer for a constructor for that class or for
a class derived from that class (_class.base.init_). */
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (member))
+ if (DECL_OBJECT_MEMBER_FUNCTION_P (member))
{
/* Build a representation of the qualified name suitable
for use as the operand to "&" -- even though the "&" is
diff --git a/gcc/cp/lambda.cc b/gcc/cp/lambda.cc
index 5990a6de736..fc6a0708b66 100644
--- a/gcc/cp/lambda.cc
+++ b/gcc/cp/lambda.cc
@@ -834,8 +834,9 @@ lambda_expr_this_capture (tree lambda, int add_capture_p)
if (!LAMBDA_FUNCTION_P (containing_function))
{
- /* We found a non-lambda function. */
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (containing_function))
+ /* We found a non-lambda function.
+ There is no this pointer in xobj member functions. */
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (containing_function))
/* First parameter is 'this'. */
init = DECL_ARGUMENTS (containing_function);
break;
@@ -969,7 +970,7 @@ maybe_generic_this_capture (tree object, tree fns)
for (lkp_iterator iter (fns); iter; ++iter)
if (((!id_expr && TREE_CODE (*iter) != USING_DECL)
|| TREE_CODE (*iter) == TEMPLATE_DECL)
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (*iter))
+ && DECL_IOBJ_MEMBER_FUNCTION_P (*iter))
{
/* Found a non-static member. Capture this. */
lambda_expr_this_capture (lam, /*maybe*/-1);
@@ -1012,7 +1013,7 @@ nonlambda_method_basetype (void)
tree fn = TYPE_CONTEXT (type);
if (!fn || TREE_CODE (fn) != FUNCTION_DECL
- || !DECL_NONSTATIC_MEMBER_FUNCTION_P (fn))
+ || !DECL_IOBJ_MEMBER_FUNCTION_P (fn))
/* No enclosing non-lambda method. */
return NULL_TREE;
if (!LAMBDA_FUNCTION_P (fn))
diff --git a/gcc/cp/mangle.cc b/gcc/cp/mangle.cc
index 365d470f46e..99fabbf3587 100644
--- a/gcc/cp/mangle.cc
+++ b/gcc/cp/mangle.cc
@@ -1231,9 +1231,9 @@ write_nested_name (const tree decl)
write_char ('N');
- /* Write CV-qualifiers, if this is a member function. */
+ /* Write CV-qualifiers, if this is an iobj member function. */
if (TREE_CODE (decl) == FUNCTION_DECL
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (decl))
+ && DECL_IOBJ_MEMBER_FUNCTION_P (decl))
{
if (DECL_VOLATILE_MEMFUNC_P (decl))
write_char ('V');
diff --git a/gcc/cp/method.cc b/gcc/cp/method.cc
index f645066077f..d6ad5aa41d6 100644
--- a/gcc/cp/method.cc
+++ b/gcc/cp/method.cc
@@ -1187,7 +1187,7 @@ early_check_defaulted_comparison (tree fn)
ok = false;
}
- bool mem = DECL_NONSTATIC_MEMBER_FUNCTION_P (fn);
+ bool mem = DECL_IOBJ_MEMBER_FUNCTION_P (fn);
if (mem && type_memfn_quals (TREE_TYPE (fn)) != TYPE_QUAL_CONST)
{
error_at (loc, "defaulted %qD must be %<const%>", fn);
@@ -1230,7 +1230,7 @@ early_check_defaulted_comparison (tree fn)
if (saw_bad || (saw_byval && saw_byref))
{
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (fn))
error_at (loc, "defaulted member %qD must have parameter type "
"%<const %T&%>", fn, ctx);
else if (saw_bad)
@@ -3606,7 +3606,7 @@ lazily_declare_fn (special_function_kind sfk, tree type)
tree
skip_artificial_parms_for (const_tree fn, tree list)
{
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (fn))
list = TREE_CHAIN (list);
else
return list;
@@ -3626,7 +3626,7 @@ num_artificial_parms_for (const_tree fn)
{
int count = 0;
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (fn))
count++;
else
return 0;
diff --git a/gcc/cp/module.cc b/gcc/cp/module.cc
index 0bd46414da9..2d0c0a3eaab 100644
--- a/gcc/cp/module.cc
+++ b/gcc/cp/module.cc
@@ -5683,8 +5683,9 @@ trees_out::lang_decl_bools (tree t)
WB (lang->u.fn.has_dependent_explicit_spec_p);
WB (lang->u.fn.immediate_fn_p);
WB (lang->u.fn.maybe_deleted);
- WB (lang->u.fn.escalated_p);
/* We do not stream lang->u.fn.implicit_constexpr. */
+ WB (lang->u.fn.escalated_p);
+ WB (lang->u.fn.xobj_func);
goto lds_min;
case lds_decomp: /* lang_decl_decomp. */
@@ -5753,8 +5754,9 @@ trees_in::lang_decl_bools (tree t)
RB (lang->u.fn.has_dependent_explicit_spec_p);
RB (lang->u.fn.immediate_fn_p);
RB (lang->u.fn.maybe_deleted);
- RB (lang->u.fn.escalated_p);
/* We do not stream lang->u.fn.implicit_constexpr. */
+ RB (lang->u.fn.escalated_p);
+ RB (lang->u.fn.xobj_func);
goto lds_min;
case lds_decomp: /* lang_decl_decomp. */
diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
index 2817657a8bb..72eb124d116 100644
--- a/gcc/cp/pt.cc
+++ b/gcc/cp/pt.cc
@@ -1446,9 +1446,9 @@ is_specialization_of_friend (tree decl, tree friend_decl)
`this' parameter. */
friend_args_type = TYPE_ARG_TYPES (friend_type);
decl_args_type = TYPE_ARG_TYPES (decl_type);
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (friend_decl))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (friend_decl))
friend_args_type = TREE_CHAIN (friend_args_type);
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (decl))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (decl))
decl_args_type = TREE_CHAIN (decl_args_type);
return compparms (decl_args_type, friend_args_type);
@@ -2236,7 +2236,7 @@ determine_specialization (tree template_id,
that the const qualification is the same. Since
get_bindings does not try to merge the "this" parameter,
we must do the comparison explicitly. */
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (fn))
{
if (!same_type_p (TREE_VALUE (fn_arg_types),
TREE_VALUE (decl_arg_types)))
@@ -2359,14 +2359,14 @@ determine_specialization (tree template_id,
/* Adjust the type of DECL in case FN is a static member. */
decl_arg_types = TYPE_ARG_TYPES (TREE_TYPE (decl));
if (DECL_STATIC_FUNCTION_P (fn)
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (decl))
+ && DECL_IOBJ_MEMBER_FUNCTION_P (decl))
decl_arg_types = TREE_CHAIN (decl_arg_types);
if (!compparms (TYPE_ARG_TYPES (TREE_TYPE (fn)),
decl_arg_types))
continue;
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn)
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (fn)
&& (type_memfn_rqual (TREE_TYPE (decl))
!= type_memfn_rqual (TREE_TYPE (fn))))
continue;
@@ -2552,7 +2552,7 @@ copy_default_args_to_explicit_spec (tree decl)
old_type = TREE_TYPE (decl);
spec_types = TYPE_ARG_TYPES (old_type);
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (decl))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (decl))
{
/* Remove the this pointer, but remember the object's type for
CV quals. */
@@ -3137,7 +3137,7 @@ check_explicit_specialization (tree declarator,
make DECL a static member function as well. */
if (DECL_FUNCTION_TEMPLATE_P (tmpl)
&& DECL_STATIC_FUNCTION_P (tmpl)
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (decl))
+ && DECL_IOBJ_MEMBER_FUNCTION_P (decl))
revert_static_member_fn (decl);
/* If this is a specialization of a member template of a
@@ -11814,7 +11814,7 @@ tsubst_contract_attribute (tree decl, tree t, tree args,
/* For member functions, make this available for semantic analysis. */
tree save_ccp = current_class_ptr;
tree save_ccr = current_class_ref;
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (decl))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (decl))
{
tree arg_types = TYPE_ARG_TYPES (TREE_TYPE (decl));
tree this_type = TREE_TYPE (TREE_VALUE (arg_types));
@@ -22087,7 +22087,7 @@ check_non_deducible_conversions (tree parms, const tree *args, unsigned nargs,
{
/* Non-constructor methods need to leave a conversion for 'this', which
isn't included in nargs here. */
- unsigned offset = (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn)
+ unsigned offset = (DECL_IOBJ_MEMBER_FUNCTION_P (fn)
&& !DECL_CONSTRUCTOR_P (fn));
for (unsigned ia = 0;
@@ -25314,16 +25314,16 @@ more_specialized_fn (tree pat1, tree pat2, int len)
I think think the old G++ behavior of just skipping the object
parameter when comparing to a static member function was better, so
let's stick with that for now. This is CWG2834. --jason 2023-12 */
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (decl1)) /* FIXME or explicit */
+ if (DECL_OBJECT_MEMBER_FUNCTION_P (decl1))
{
len--; /* LEN is the number of significant arguments for DECL1 */
args1 = TREE_CHAIN (args1);
}
- else if (DECL_NONSTATIC_MEMBER_FUNCTION_P (decl2)) /* FIXME or explicit */
+ else if (DECL_OBJECT_MEMBER_FUNCTION_P (decl2))
args2 = TREE_CHAIN (args2);
}
- else if (DECL_NONSTATIC_MEMBER_FUNCTION_P (decl1) /* FIXME implicit only */
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (decl2))
+ else if (DECL_IOBJ_MEMBER_FUNCTION_P (decl1)
+ && DECL_IOBJ_MEMBER_FUNCTION_P (decl2))
{
/* Note DR2445 also (IMO wrongly) removed the "only one" above, which
would break e.g. cpp1y/lambda-generic-variadic5.C. */
@@ -25331,12 +25331,12 @@ more_specialized_fn (tree pat1, tree pat2, int len)
args1 = TREE_CHAIN (args1);
args2 = TREE_CHAIN (args2);
}
- else if (DECL_NONSTATIC_MEMBER_FUNCTION_P (decl1) /* FIXME implicit only */
- || DECL_NONSTATIC_MEMBER_FUNCTION_P (decl2))
+ else if (DECL_IOBJ_MEMBER_FUNCTION_P (decl1)
+ || DECL_IOBJ_MEMBER_FUNCTION_P (decl2))
{
/* The other is a non-member or explicit object member function;
rewrite the implicit object parameter to a reference. */
- tree ns = DECL_NONSTATIC_MEMBER_FUNCTION_P (decl2) ? decl2 : decl1;
+ tree ns = DECL_IOBJ_MEMBER_FUNCTION_P (decl2) ? decl2 : decl1;
tree &nsargs = ns == decl2 ? args2 : args1;
tree obtype = TREE_TYPE (TREE_VALUE (nsargs));
@@ -26669,7 +26669,7 @@ maybe_instantiate_noexcept (tree fn, tsubst_flags_t complain)
push_deferring_access_checks (dk_no_deferred);
input_location = DECL_SOURCE_LOCATION (fn);
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (fn)
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (fn)
&& !DECL_LOCAL_DECL_P (fn))
{
/* If needed, set current_class_ptr for the benefit of
@@ -26732,7 +26732,7 @@ register_parameter_specializations (tree pattern, tree inst)
{
tree tmpl_parm = DECL_ARGUMENTS (pattern);
tree spec_parm = DECL_ARGUMENTS (inst);
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (inst))
+ if (DECL_IOBJ_MEMBER_FUNCTION_P (inst))
{
register_local_specialization (spec_parm, tmpl_parm);
spec_parm = skip_artificial_parms_for (inst, spec_parm);
@@ -28075,7 +28075,7 @@ value_dependent_expression_p (tree expression)
cause the call to be considered value-dependent. We also
look through it in potential_constant_expression. */
if (i == 0 && fn && DECL_DECLARED_CONSTEXPR_P (fn)
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (fn)
+ && DECL_IOBJ_MEMBER_FUNCTION_P (fn)
&& TREE_CODE (op) == ADDR_EXPR)
op = TREE_OPERAND (op, 0);
if (value_dependent_expression_p (op))
diff --git a/gcc/cp/search.cc b/gcc/cp/search.cc
index ac79b625b6e..3473d66b88c 100644
--- a/gcc/cp/search.cc
+++ b/gcc/cp/search.cc
@@ -1008,7 +1008,7 @@ shared_member_p (tree t)
/* Conservatively assume a dependent using-declaration
might resolve to a non-static member. */
return false;
- if (DECL_NONSTATIC_MEMBER_FUNCTION_P (decl))
+ if (DECL_OBJECT_MEMBER_FUNCTION_P (decl))
return false;
}
return true;
@@ -1264,7 +1264,7 @@ lookup_member (tree xbasetype, tree name, int protect, bool want_type,
decl = strip_using_decl (decl);
/* A dependent USING_DECL will be checked after tsubsting. */
if (TREE_CODE (decl) != USING_DECL
- && !DECL_NONSTATIC_MEMBER_FUNCTION_P (decl)
+ && !DECL_IOBJ_MEMBER_FUNCTION_P (decl)
&& !perform_or_defer_access_check (basetype_path, decl, decl,
complain, afi))
return error_mark_node;
@@ -1737,6 +1737,7 @@ field_access_p (tree component_ref, tree field_decl, tree field_type)
return false;
tree ptr = STRIP_NOPS (TREE_OPERAND (indirect_ref, 0));
+ /* ??? is_object_parameter? */
if (!is_this_parameter (ptr))
return false;
diff --git a/gcc/cp/semantics.cc b/gcc/cp/semantics.cc
index 64839b1ac87..3bc426fa130 100644
--- a/gcc/cp/semantics.cc
+++ b/gcc/cp/semantics.cc
@@ -6645,7 +6645,7 @@ finish_omp_declare_simd_methods (tree t)
for (tree x = TYPE_FIELDS (t); x; x = DECL_CHAIN (x))
{
if (TREE_CODE (x) == USING_DECL
- || !DECL_NONSTATIC_MEMBER_FUNCTION_P (x))
+ || !DECL_IOBJ_MEMBER_FUNCTION_P (x))
continue;
tree ods = lookup_attribute ("omp declare simd", DECL_ATTRIBUTES (x));
if (!ods || !TREE_VALUE (ods))
diff --git a/gcc/cp/tree.cc b/gcc/cp/tree.cc
index d17b9b34891..80596e3d976 100644
--- a/gcc/cp/tree.cc
+++ b/gcc/cp/tree.cc
@@ -298,7 +298,7 @@ lvalue_kind (const_tree ref)
case FUNCTION_DECL:
/* All functions (except non-static-member functions) are
lvalues. */
- return (DECL_NONSTATIC_MEMBER_FUNCTION_P (ref)
+ return (DECL_IOBJ_MEMBER_FUNCTION_P (ref)
? clk_none : clk_ordinary);
case BASELINK:
diff --git a/gcc/cp/typeck.cc b/gcc/cp/typeck.cc
index a6e2f4ee7da..2abf9f3b905 100644
--- a/gcc/cp/typeck.cc
+++ b/gcc/cp/typeck.cc
@@ -2352,7 +2352,7 @@ invalid_nonstatic_memfn_p (location_t loc, tree expr, tsubst_flags_t complain)
if (is_overloaded_fn (expr) && !really_overloaded_fn (expr))
expr = get_first_fn (expr);
if (TREE_TYPE (expr)
- && DECL_NONSTATIC_MEMBER_FUNCTION_P (expr))
+ && DECL_IOBJ_MEMBER_FUNCTION_P (expr))
{
if (complain & tf_error)
{
diff --git a/libcc1/libcp1plugin.cc b/libcc1/libcp1plugin.cc
index 3c8e9e8b436..108d5786167 100644
--- a/libcc1/libcp1plugin.cc
+++ b/libcc1/libcp1plugin.cc
@@ -468,7 +468,7 @@ plugin_pragma_push_user_expression (cpp_reader *)
}
}
- if (unchanged_cfun || DECL_NONSTATIC_MEMBER_FUNCTION_P (changed_func_decl))
+ if (unchanged_cfun || DECL_OBJECT_MEMBER_FUNCTION_P (changed_func_decl))
{
/* Check whether the oracle supplies us with a "this", and if
so, arrange for data members and this itself to be
--
2.43.0
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609]
2024-01-07 0:00 [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609] waffl3x
@ 2024-01-09 20:58 ` Jason Merrill
2024-01-09 21:26 ` waffl3x
2024-01-09 21:56 ` Jason Merrill
2024-01-10 17:15 ` Patrick Palka
1 sibling, 2 replies; 8+ messages in thread
From: Jason Merrill @ 2024-01-09 20:58 UTC (permalink / raw)
To: waffl3x; +Cc: gcc-patches
On 1/6/24 19:00, waffl3x wrote:
> Bootstrapped and tested on x86_64-linux with no regressions.
>
> I'm considering this finished, I have CWG2586 working but I have not
> included it in this version of the patch. I was not happy with the
> amount of work I had done on it. I will try to get it finished before
> we get cut off, and I'm pretty sure I can. I just don't want to risk
> missing the boat for the whole patch just for that.
>
> There aren't too many changes from v7, it's mostly just cleaned up.
> There are a few though, so do take a look, if there's anything severe I
> can rush to fix it if necessary.
>
> That's all, hopefully all is good, fingers crossed.
Great. Given where we are in the release cycle, I'm thinking to put
these with only minimal changes and then do any further adjustments
afterward.
For this first one, I needed to fix the commit message to wrap at 75
columns so that it fits in 80 columns with the initial padding added by
'git log'. I also needed to adjust the ChangeLog entry to please git
gcc-verify. And I changed the credit note to a Co-authored-by.
The other commit messages only needed wrapping.
Thanks!
Jason
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609]
2024-01-09 20:58 ` Jason Merrill
@ 2024-01-09 21:26 ` waffl3x
2024-01-09 21:56 ` Jason Merrill
1 sibling, 0 replies; 8+ messages in thread
From: waffl3x @ 2024-01-09 21:26 UTC (permalink / raw)
To: Jason Merrill; +Cc: gcc-patches
On Tuesday, January 9th, 2024 at 1:58 PM, Jason Merrill <jason@redhat.com> wrote:
>
>
> On 1/6/24 19:00, waffl3x wrote:
>
> > Bootstrapped and tested on x86_64-linux with no regressions.
> >
> > I'm considering this finished, I have CWG2586 working but I have not
> > included it in this version of the patch. I was not happy with the
> > amount of work I had done on it. I will try to get it finished before
> > we get cut off, and I'm pretty sure I can. I just don't want to risk
> > missing the boat for the whole patch just for that.
> >
> > There aren't too many changes from v7, it's mostly just cleaned up.
> > There are a few though, so do take a look, if there's anything severe I
> > can rush to fix it if necessary.
> >
> > That's all, hopefully all is good, fingers crossed.
>
>
> Great. Given where we are in the release cycle, I'm thinking to put
> these with only minimal changes and then do any further adjustments
> afterward.
>
> For this first one, I needed to fix the commit message to wrap at 75
> columns so that it fits in 80 columns with the initial padding added by
> 'git log'. I also needed to adjust the ChangeLog entry to please git
> gcc-verify. And I changed the credit note to a Co-authored-by.
>
> The other commit messages only needed wrapping.
>
> Thanks!
>
> Jason
Sounds good to me, it's been a pleasure working with you. If there's
anything you would like me to do differently for future patches just
let me know. For now, I think I will take a break.
Alex
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609]
2024-01-09 20:58 ` Jason Merrill
2024-01-09 21:26 ` waffl3x
@ 2024-01-09 21:56 ` Jason Merrill
2024-01-09 22:34 ` waffl3x
1 sibling, 1 reply; 8+ messages in thread
From: Jason Merrill @ 2024-01-09 21:56 UTC (permalink / raw)
To: waffl3x; +Cc: gcc-patches
[-- Attachment #1: Type: text/plain, Size: 1288 bytes --]
On 1/9/24 15:58, Jason Merrill wrote:
> On 1/6/24 19:00, waffl3x wrote:
>> Bootstrapped and tested on x86_64-linux with no regressions.
>>
>> I'm considering this finished, I have CWG2586 working but I have not
>> included it in this version of the patch. I was not happy with the
>> amount of work I had done on it. I will try to get it finished before
>> we get cut off, and I'm pretty sure I can. I just don't want to risk
>> missing the boat for the whole patch just for that.
>>
>> There aren't too many changes from v7, it's mostly just cleaned up.
>> There are a few though, so do take a look, if there's anything severe I
>> can rush to fix it if necessary.
>>
>> That's all, hopefully all is good, fingers crossed.
>
> Great. Given where we are in the release cycle, I'm thinking to put
> these with only minimal changes and then do any further adjustments
> afterward.
>
> For this first one, I needed to fix the commit message to wrap at 75
> columns so that it fits in 80 columns with the initial padding added by
> 'git log'. I also needed to adjust the ChangeLog entry to please git
> gcc-verify. And I changed the credit note to a Co-authored-by.
>
> The other commit messages only needed wrapping.
I just pushed your patches along with these two follow-ons:
[-- Attachment #2: 0001-c-explicit-object-cleanups.patch --]
[-- Type: text/x-patch, Size: 7356 bytes --]
From 5a6d3b1737843aa64d83ffc5d639fa0afa5d8318 Mon Sep 17 00:00:00 2001
From: Jason Merrill <jason@redhat.com>
Date: Tue, 9 Jan 2024 16:00:52 -0500
Subject: [PATCH 1/2] c++: explicit object cleanups
To: gcc-patches@gcc.gnu.org
The FIXME in xobj_iobj_parameters_correspond was due to expecting
TYPE_MAIN_VARIANT to be the same for all equivalent types, which is not the
case. And I adjusted some comments that I disagree with; the iobj parameter
adjustment only applies to overload resolution, we can handle that in
cand_parms_match (and I have WIP for that).
gcc/cp/ChangeLog:
* call.cc (build_over_call): Refactor handle_arg lambda.
* class.cc (xobj_iobj_parameters_correspond): Fix FIXME.
* method.cc (defaulted_late_check): Adjust comments.
---
gcc/cp/call.cc | 24 ++++++++++++------------
gcc/cp/class.cc | 40 ++++++++++++----------------------------
gcc/cp/method.cc | 7 +------
3 files changed, 25 insertions(+), 46 deletions(-)
diff --git a/gcc/cp/call.cc b/gcc/cp/call.cc
index dca8e5090e2..7d3d67600c8 100644
--- a/gcc/cp/call.cc
+++ b/gcc/cp/call.cc
@@ -10187,11 +10187,11 @@ build_over_call (struct z_candidate *cand, int flags, tsubst_flags_t complain)
parm = TREE_CHAIN (parm);
}
- auto handle_arg = [fn, flags, complain](tree type,
- tree arg,
- int const param_index,
- conversion *conv,
- bool const conversion_warning)
+ auto handle_arg = [fn, flags](tree type,
+ tree arg,
+ int const param_index,
+ conversion *conv,
+ tsubst_flags_t const arg_complain)
{
/* Set user_conv_p on the argument conversions, so rvalue/base handling
knows not to allow any more UDCs. This needs to happen after we
@@ -10199,9 +10199,6 @@ build_over_call (struct z_candidate *cand, int flags, tsubst_flags_t complain)
if (flags & LOOKUP_NO_CONVERSION)
conv->user_conv_p = true;
- tsubst_flags_t const arg_complain
- = conversion_warning ? complain : complain & ~tf_warning;
-
if (arg_complain & tf_warning)
maybe_warn_pessimizing_move (arg, type, /*return_p=*/false);
@@ -10214,13 +10211,12 @@ build_over_call (struct z_candidate *cand, int flags, tsubst_flags_t complain)
if (DECL_XOBJ_MEMBER_FUNCTION_P (fn))
{
gcc_assert (cand->num_convs > 0);
- static constexpr bool conversion_warning = true;
tree object_arg = consume_object_arg ();
val = handle_arg (TREE_VALUE (parm),
object_arg,
param_index++,
convs[conv_index++],
- conversion_warning);
+ complain);
if (val == error_mark_node)
return error_mark_node;
@@ -10260,11 +10256,14 @@ build_over_call (struct z_candidate *cand, int flags, tsubst_flags_t complain)
&& cand->template_decl
&& !cand->explicit_targs);
+ tsubst_flags_t const arg_complain
+ = conversion_warning ? complain : complain & ~tf_warning;
+
val = handle_arg (TREE_VALUE (parm),
current_arg,
param_index,
convs[conv_index],
- conversion_warning);
+ arg_complain);
if (val == error_mark_node)
return error_mark_node;
@@ -10273,7 +10272,8 @@ build_over_call (struct z_candidate *cand, int flags, tsubst_flags_t complain)
}
/* Default arguments */
- for (; parm && parm != void_list_node; parm = TREE_CHAIN (parm), param_index++)
+ for (; parm && parm != void_list_node;
+ parm = TREE_CHAIN (parm), param_index++)
{
if (TREE_VALUE (parm) == error_mark_node)
return error_mark_node;
diff --git a/gcc/cp/class.cc b/gcc/cp/class.cc
index f3cfa9f0f23..e5e609badf3 100644
--- a/gcc/cp/class.cc
+++ b/gcc/cp/class.cc
@@ -1020,9 +1020,12 @@ modify_vtable_entry (tree t,
\f
/* Check if the object parameters of an xobj and iobj member function
- correspond. This function assumes that the iobj parameter has been correctly
- adjusted when the function is introduced by a using declaration per
- [over.match.funcs.general.4]. */
+ correspond. This function assumes that the iobj parameter has been
+ correctly adjusted when the function is introduced by a using declaration
+ per [over.match.funcs.general.4].
+
+ ??? But it isn't, that's only considered at overload resolution time.
+ cand_parms_match will probably need to check cand->conversion_path. */
bool
xobj_iobj_parameters_correspond (tree fn1, tree fn2)
@@ -1112,29 +1115,10 @@ xobj_iobj_parameters_correspond (tree fn1, tree fn2)
handles xobj parameters of pointer type, we don't have to explicitly
check for that case. */
- /* FIXME:
-
- template<typename>
- struct S;
-
- template<typename>
- struct B {
- int f(this S<void>&) requires true { return 5; }
- };
-
- template<typename>
- struct S : B<void> {
- using B<void>::f;
- int f() { return 10; }
- };
-
- This case is broken, the incomplete type seems to screw with things.
- I'm not sure how to fix that so I'm just noting the issue here, I have a
- feeling it's trivial to do if you know how. */
-
- if (TYPE_MAIN_VARIANT (iobj_param_type)
- != TYPE_MAIN_VARIANT (non_reference (xobj_param)))
+ if (!same_type_ignoring_top_level_qualifiers_p
+ (iobj_param_type, non_reference (xobj_param)))
return false;
+
/* We don't get to bail yet even if we have a by-value xobj parameter,
a by-value xobj parameter can correspond to an iobj parameter provided the
iobj member function is not declared with a reference qualifier.
@@ -8976,9 +8960,9 @@ resolve_address_of_overloaded_function (tree target_type,
documentation for -fms-extensions states it's purpose is to support
the use of microsoft headers. Until otherwise demonstrated, we should
assume xobj member functions are not used in this manner in microsoft
- headers and indiscriminately forbid the incorrect syntax instead of
- supporting it for non-legacy uses. This should hopefully encourage
- conformance going forward.
+ headers and forbid the incorrect syntax instead of supporting it for
+ non-legacy uses. This should hopefully encourage conformance going
+ forward.
This comment is referred to in typeck.cc:cp_build_addr_expr_1. */
if (DECL_IOBJ_MEMBER_FUNCTION_P (fn) && flag_ms_extensions)
/* Early escape. */;
diff --git a/gcc/cp/method.cc b/gcc/cp/method.cc
index da6a08a0304..6a9f03eb8f3 100644
--- a/gcc/cp/method.cc
+++ b/gcc/cp/method.cc
@@ -3400,8 +3400,7 @@ defaulted_late_check (tree fn)
|| TYPE_REF_IS_RVALUE (fn_obj_ref_type))
return false;
/* If implicit_fn's object parameter is not a pointer, something is not
- right. (Or we have finally changed the type of the iobj parameter
- in iobj member functions.) */
+ right. */
gcc_assert (TYPE_PTR_P (TREE_VALUE (implicit_fn_parms)));
/* Strip the reference/pointer off each object parameter before
comparing them. */
@@ -3422,10 +3421,6 @@ defaulted_late_check (tree fn)
{
error ("defaulted declaration %q+D does not match the "
"expected signature", fn);
- /* FIXME: If the user is defaulting an xobj member function we should
- emit an xobj member function for a signature. When we do this, maybe
- we can just synthesize implicit_fn as an xobj member function and
- avoid the dance in compare_fn_parms. */
inform (DECL_SOURCE_LOCATION (fn),
"expected signature: %qD", implicit_fn);
}
--
2.39.3
[-- Attachment #3: 0002-c-adjust-accessor-fixits-for-explicit-object-parm.patch --]
[-- Type: text/x-patch, Size: 8281 bytes --]
From ae3003b20d3e3ab6e50a6d4f2173e10ad9025135 Mon Sep 17 00:00:00 2001
From: Jason Merrill <jason@redhat.com>
Date: Mon, 8 Jan 2024 17:09:54 -0500
Subject: [PATCH 2/2] c++: adjust accessor fixits for explicit object parm
To: gcc-patches@gcc.gnu.org
In a couple of places in the xobj patch I noticed that is_this_parameter
probably wanted to change to is_object_parameter; this implements that and
does the additional adjustments needed to make the accessor fixits handle
xobj parms.
gcc/cp/ChangeLog:
* semantics.cc (is_object_parameter): New.
* cp-tree.h (is_object_parameter): Declare.
* call.cc (maybe_warn_class_memaccess): Use it.
* search.cc (field_access_p): Use it.
(class_of_object_parm): New.
(field_accessor_p): Adjust for explicit object parms.
gcc/testsuite/ChangeLog:
* g++.dg/torture/accessor-fixits-9-xobj.C: New test.
---
gcc/cp/cp-tree.h | 1 +
gcc/cp/call.cc | 3 +-
gcc/cp/search.cc | 19 ++-
gcc/cp/semantics.cc | 14 +++
.../g++.dg/torture/accessor-fixits-9-xobj.C | 119 ++++++++++++++++++
5 files changed, 149 insertions(+), 7 deletions(-)
create mode 100644 gcc/testsuite/g++.dg/torture/accessor-fixits-9-xobj.C
diff --git a/gcc/cp/cp-tree.h b/gcc/cp/cp-tree.h
index dbc71776be3..f3f265a3db8 100644
--- a/gcc/cp/cp-tree.h
+++ b/gcc/cp/cp-tree.h
@@ -7784,6 +7784,7 @@ extern void finish_handler_parms (tree, tree);
extern void finish_handler (tree);
extern void finish_cleanup (tree, tree);
extern bool is_this_parameter (tree);
+extern bool is_object_parameter (tree);
enum {
BCS_NORMAL = 0,
diff --git a/gcc/cp/call.cc b/gcc/cp/call.cc
index 7d3d67600c8..191664ee227 100644
--- a/gcc/cp/call.cc
+++ b/gcc/cp/call.cc
@@ -10815,8 +10815,7 @@ maybe_warn_class_memaccess (location_t loc, tree fndecl,
be more permissive. */
if (current_function_decl
&& DECL_OBJECT_MEMBER_FUNCTION_P (current_function_decl)
- /* ??? is_object_parameter? */
- && is_this_parameter (tree_strip_nop_conversions (dest)))
+ && is_object_parameter (tree_strip_nop_conversions (dest)))
{
tree ctx = DECL_CONTEXT (current_function_decl);
bool special = same_type_ignoring_top_level_qualifiers_p (ctx, desttype);
diff --git a/gcc/cp/search.cc b/gcc/cp/search.cc
index dc76714351f..2b4ed5d024e 100644
--- a/gcc/cp/search.cc
+++ b/gcc/cp/search.cc
@@ -1737,8 +1737,7 @@ field_access_p (tree component_ref, tree field_decl, tree field_type)
return false;
tree ptr = STRIP_NOPS (TREE_OPERAND (indirect_ref, 0));
- /* ??? is_object_parameter? */
- if (!is_this_parameter (ptr))
+ if (!is_object_parameter (ptr))
return false;
/* Must access the correct field. */
@@ -1818,6 +1817,17 @@ reference_accessor_p (tree init_expr, tree field_decl, tree field_type,
return true;
}
+/* Return the class of the `this' or explicit object parameter of FN. */
+
+static tree
+class_of_object_parm (const_tree fn)
+{
+ tree fntype = TREE_TYPE (fn);
+ if (DECL_XOBJ_MEMBER_FUNCTION_P (fn))
+ return non_reference (TREE_VALUE (TYPE_ARG_TYPES (fntype)));
+ return class_of_this_parm (fntype);
+}
+
/* Return true if FN is an accessor method for FIELD_DECL.
i.e. a method of the form { return FIELD; }, with no
conversions.
@@ -1835,15 +1845,14 @@ field_accessor_p (tree fn, tree field_decl, bool const_p)
if (TREE_CODE (field_decl) != FIELD_DECL)
return false;
- tree fntype = TREE_TYPE (fn);
- if (TREE_CODE (fntype) != METHOD_TYPE)
+ if (!DECL_OBJECT_MEMBER_FUNCTION_P (fn))
return false;
/* If the field is accessed via a const "this" argument, verify
that the "this" parameter is const. */
if (const_p)
{
- tree this_class = class_of_this_parm (fntype);
+ tree this_class = class_of_object_parm (fn);
if (!TYPE_READONLY (this_class))
return false;
}
diff --git a/gcc/cp/semantics.cc b/gcc/cp/semantics.cc
index 86e1adf1e71..3299e270446 100644
--- a/gcc/cp/semantics.cc
+++ b/gcc/cp/semantics.cc
@@ -12832,6 +12832,20 @@ is_this_parameter (tree t)
return true;
}
+/* As above, or a C++23 explicit object parameter. */
+
+bool
+is_object_parameter (tree t)
+{
+ if (is_this_parameter (t))
+ return true;
+ if (TREE_CODE (t) != PARM_DECL)
+ return false;
+ tree ctx = DECL_CONTEXT (t);
+ return (ctx && DECL_XOBJ_MEMBER_FUNCTION_P (ctx)
+ && t == DECL_ARGUMENTS (ctx));
+}
+
/* Insert the deduced return type for an auto function. */
void
diff --git a/gcc/testsuite/g++.dg/torture/accessor-fixits-9-xobj.C b/gcc/testsuite/g++.dg/torture/accessor-fixits-9-xobj.C
new file mode 100644
index 00000000000..89be978ae8e
--- /dev/null
+++ b/gcc/testsuite/g++.dg/torture/accessor-fixits-9-xobj.C
@@ -0,0 +1,119 @@
+// PR c++/84993
+// { dg-options "-fdiagnostics-show-caret -std=c++23" }
+
+/* Misspelling (by omitting a leading "m_") of a private member for which
+ there's a public accessor.
+
+ We expect a fix-it hint suggesting the accessor. */
+
+class t1
+{
+public:
+ int get_ratio (this const t1& x) { return x.m_ratio; }
+
+private:
+ int m_ratio;
+};
+
+int test (t1 *ptr_1)
+{
+ return ptr_1->ratio; // { dg-error "'class t1' has no member named 'ratio'; did you mean 'int t1::m_ratio'\\? \\(accessible via 'int t1::get_ratio\\(this const t1&\\)'\\)" }
+ /* { dg-begin-multiline-output "" }
+ return ptr_1->ratio;
+ ^~~~~
+ get_ratio()
+ { dg-end-multiline-output "" } */
+}
+
+
+/* Misspelling of a private member for which there's a public accessor.
+
+ We expect a fix-it hint suggesting the accessor. */
+
+class t2
+{
+public:
+ int get_color (this const t2& x) { return x.m_color; }
+
+private:
+ int m_color;
+};
+
+int test (t2 *ptr_2)
+{
+ return ptr_2->m_colour; // { dg-error "'class t2' has no member named 'm_colour'; did you mean 'int t2::m_color'\\? \\(accessible via 'int t2::get_color\\(this const t2&\\)'\\)" }
+ /* { dg-begin-multiline-output "" }
+ return ptr_2->m_colour;
+ ^~~~~~~~
+ get_color()
+ { dg-end-multiline-output "" } */
+}
+
+
+/* Misspelling of a private member via a subclass pointer, for which there's
+ a public accessor in the base class.
+
+ We expect a fix-it hint suggesting the accessor. */
+
+class t3 : public t2 {};
+
+int test (t3 *ptr_3)
+{
+ return ptr_3->m_colour; // { dg-error "'class t3' has no member named 'm_colour'; did you mean 'int t2::m_color'\\? \\(accessible via 'int t2::get_color\\(this const t2&\\)'\\)" }
+ /* { dg-begin-multiline-output "" }
+ return ptr_3->m_colour;
+ ^~~~~~~~
+ get_color()
+ { dg-end-multiline-output "" } */
+}
+
+
+/* Misspelling of a protected member, for which there's isn't a public
+ accessor.
+
+ We expect no fix-it hint; instead a message identifying where the
+ data member was declared. */
+
+class t4
+{
+protected:
+ int m_color; // { dg-message "declared protected here" }
+};
+
+int test (t4 *ptr_4)
+{
+ return ptr_4->m_colour; // { dg-error "'class t4' has no member named 'm_colour'; did you mean 'int t4::m_color'\\? \\(not accessible from this context\\)" }
+ /* { dg-begin-multiline-output "" }
+ return ptr_4->m_colour;
+ ^~~~~~~~
+ { dg-end-multiline-output "" } */
+ /* { dg-begin-multiline-output "" }
+ int m_color;
+ ^~~~~~~
+ { dg-end-multiline-output "" } */
+}
+
+
+/* Misspelling of a private member, for which the accessor is also private.
+
+ We expect no fix-it hint; instead a message identifying where the
+ data member was declared. */
+
+class t5
+{
+ int get_color (this const t5& x) { return x.m_color; }
+ int m_color; // { dg-message "declared private here" }
+};
+
+int test (t5 *ptr_5)
+{
+ return ptr_5->m_colour; // { dg-error "'class t5' has no member named 'm_colour'; did you mean 'int t5::m_color'\\? \\(not accessible from this context\\)" }
+ /* { dg-begin-multiline-output "" }
+ return ptr_5->m_colour;
+ ^~~~~~~~
+ { dg-end-multiline-output "" } */
+ /* { dg-begin-multiline-output "" }
+ int m_color;
+ ^~~~~~~
+ { dg-end-multiline-output "" } */
+}
--
2.39.3
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609]
2024-01-09 21:56 ` Jason Merrill
@ 2024-01-09 22:34 ` waffl3x
2024-01-09 22:52 ` Jason Merrill
0 siblings, 1 reply; 8+ messages in thread
From: waffl3x @ 2024-01-09 22:34 UTC (permalink / raw)
To: Jason Merrill; +Cc: gcc-patches
On Tuesday, January 9th, 2024 at 2:56 PM, Jason Merrill <jason@redhat.com> wrote:
>
>
> On 1/9/24 15:58, Jason Merrill wrote:
>
> > On 1/6/24 19:00, waffl3x wrote:
> >
> > > Bootstrapped and tested on x86_64-linux with no regressions.
> > >
> > > I'm considering this finished, I have CWG2586 working but I have not
> > > included it in this version of the patch. I was not happy with the
> > > amount of work I had done on it. I will try to get it finished before
> > > we get cut off, and I'm pretty sure I can. I just don't want to risk
> > > missing the boat for the whole patch just for that.
> > >
> > > There aren't too many changes from v7, it's mostly just cleaned up.
> > > There are a few though, so do take a look, if there's anything severe I
> > > can rush to fix it if necessary.
> > >
> > > That's all, hopefully all is good, fingers crossed.
> >
> > Great. Given where we are in the release cycle, I'm thinking to put
> > these with only minimal changes and then do any further adjustments
> > afterward.
> >
> > For this first one, I needed to fix the commit message to wrap at 75
> > columns so that it fits in 80 columns with the initial padding added by
> > 'git log'. I also needed to adjust the ChangeLog entry to please git
> > gcc-verify. And I changed the credit note to a Co-authored-by.
> >
> > The other commit messages only needed wrapping.
>
>
> I just pushed your patches along with these two follow-ons:
Is the type of an implicit object parameter specified elsewhere? I have
looked for it more than once and I could only find that passage, but I
guess [over.match.funcs-1] is what I missed here, so
[over.match.funcs-4] doesn't apply elsewhere.
I have no objection to changing the comment in
iobj_xobj_parameters_correspond if you have a better idea on how to fix
the bug that inspired that comment. Especially since the standard is on
your side for it.
In this case, perhaps the comment should instead say something along
the lines of "this function does not account for using declarations."
That does get a little hairy though if both an iobj and xobj member
function are introduced by a using declaration. Perhaps we should just
optionally take a type that we can compare against? That sounds really
icky, maybe the solution you have in mind will work just fine.
BTW, I definitely assumed TYPE_MAIN_VARIANT was what I wanted in other
places in the patch, so there might be latent errors waiting to happen.
It's probably fine because those areas where I used them are pretty
well tested, but I thought it would be worth mentioning just in case.
I'm very excited to see this get through, thanks again for all the help
and patience.
Alex
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609]
2024-01-09 22:34 ` waffl3x
@ 2024-01-09 22:52 ` Jason Merrill
2024-01-09 23:59 ` waffl3x
0 siblings, 1 reply; 8+ messages in thread
From: Jason Merrill @ 2024-01-09 22:52 UTC (permalink / raw)
To: waffl3x; +Cc: gcc-patches
On 1/9/24 17:34, waffl3x wrote:
> On Tuesday, January 9th, 2024 at 2:56 PM, Jason Merrill <jason@redhat.com> wrote:
>> On 1/9/24 15:58, Jason Merrill wrote:
>>> On 1/6/24 19:00, waffl3x wrote:
>>>
>>>> Bootstrapped and tested on x86_64-linux with no regressions.
>>>>
>>>> I'm considering this finished, I have CWG2586 working but I have not
>>>> included it in this version of the patch. I was not happy with the
>>>> amount of work I had done on it. I will try to get it finished before
>>>> we get cut off, and I'm pretty sure I can. I just don't want to risk
>>>> missing the boat for the whole patch just for that.
>>>>
>>>> There aren't too many changes from v7, it's mostly just cleaned up.
>>>> There are a few though, so do take a look, if there's anything severe I
>>>> can rush to fix it if necessary.
>>>>
>>>> That's all, hopefully all is good, fingers crossed.
>>>
>>> Great. Given where we are in the release cycle, I'm thinking to put
>>> these with only minimal changes and then do any further adjustments
>>> afterward.
>>>
>>> For this first one, I needed to fix the commit message to wrap at 75
>>> columns so that it fits in 80 columns with the initial padding added by
>>> 'git log'. I also needed to adjust the ChangeLog entry to please git
>>> gcc-verify. And I changed the credit note to a Co-authored-by.
>>>
>>> The other commit messages only needed wrapping.
>>
>> I just pushed your patches along with these two follow-ons:
>
> Is the type of an implicit object parameter specified elsewhere? I have
> looked for it more than once and I could only find that passage, but I
> guess [over.match.funcs-1] is what I missed here, so
> [over.match.funcs-4] doesn't apply elsewhere.
That's the only definition I know of. And indeed my statement was
wrong, it also affects the considerations in add_method. But I still
think it would be more complicated in more places to deal with proxy
FUNCTION_DECLs (like we do for inheriting constructors).
> I have no objection to changing the comment in
> iobj_xobj_parameters_correspond if you have a better idea on how to fix
> the bug that inspired that comment. Especially since the standard is on
> your side for it.
>
> In this case, perhaps the comment should instead say something along
> the lines of "this function does not account for using declarations."
> That does get a little hairy though if both an iobj and xobj member
> function are introduced by a using declaration. Perhaps we should just
> optionally take a type that we can compare against? That sounds really
> icky, maybe the solution you have in mind will work just fine.
>
> BTW, I definitely assumed TYPE_MAIN_VARIANT was what I wanted in other
> places in the patch, so there might be latent errors waiting to happen.
> It's probably fine because those areas where I used them are pretty
> well tested, but I thought it would be worth mentioning just in case.
The uses in copy_fn_p seem a little dubious, but there were already
similar ones there, so I think it should be OK.
> I'm very excited to see this get through, thanks again for all the help
> and patience.
Thank you for all your work!
Jason
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609]
2024-01-09 22:52 ` Jason Merrill
@ 2024-01-09 23:59 ` waffl3x
0 siblings, 0 replies; 8+ messages in thread
From: waffl3x @ 2024-01-09 23:59 UTC (permalink / raw)
To: Jason Merrill; +Cc: gcc-patches
On Tuesday, January 9th, 2024 at 3:52 PM, Jason Merrill <jason@redhat.com> wrote:
>
>
> On 1/9/24 17:34, waffl3x wrote:
>
> > On Tuesday, January 9th, 2024 at 2:56 PM, Jason Merrill jason@redhat.com wrote:
> >
> >
> > Is the type of an implicit object parameter specified elsewhere? I have
> > looked for it more than once and I could only find that passage, but I
> > guess [over.match.funcs-1] is what I missed here, so
> > [over.match.funcs-4] doesn't apply elsewhere.
>
>
> That's the only definition I know of. And indeed my statement was
> wrong, it also affects the considerations in add_method. But I still
> think it would be more complicated in more places to deal with proxy
> FUNCTION_DECLs (like we do for inheriting constructors).
You would know the code base better than I, so you're probably right.
The code in add_method probably already handles it just fine, the only
caveat being that it needs corresponding object parameters when
deciding to discard an overload introduced by a using declaration...
uhoh yeah that might not be okay. This probably results in ambiguous
overload resolution, I'll throw together a test case to see if I can
confirm this...
struct B {
void f() {}
};
struct S : B {
using B::f;
void f(this S&) {}
};
int main() {
S s{};
s.f();
}
And yep, you're right, add_method is going to need to care about this
too. In the above test case, the call to f is incorrectly ambiguous.
The overload of f introduced by the using declaration should have been
discarded as the object parameters should correspond.
struct S;
struct B {
void g(this S&) {}
};
struct S : B {
using B::g;
void g() {}
};
int main()
{
S s{};
s.g();
}
While this case works correctly, the overload of g introduced by the
using declaration is correctly discarded.
Hopefully it's not too annoying to solve, I'm not exactly sure how to
go about it (other than the method you've already turned down) so I'll
leave it to you.
Alex
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609]
2024-01-07 0:00 [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609] waffl3x
2024-01-09 20:58 ` Jason Merrill
@ 2024-01-10 17:15 ` Patrick Palka
1 sibling, 0 replies; 8+ messages in thread
From: Patrick Palka @ 2024-01-10 17:15 UTC (permalink / raw)
To: waffl3x; +Cc: Jason Merrill, gcc-patches
Congratulations on landing this impressive work in GCC 14!
On Sun, 7 Jan 2024, waffl3x wrote:
> Bootstrapped and tested on x86_64-linux with no regressions.
>
> I'm considering this finished, I have CWG2586 working but I have not
> included it in this version of the patch. I was not happy with the
> amount of work I had done on it. I will try to get it finished before
> we get cut off, and I'm pretty sure I can. I just don't want to risk
> missing the boat for the whole patch just for that.
>
> There aren't too many changes from v7, it's mostly just cleaned up.
> There are a few though, so do take a look, if there's anything severe I
> can rush to fix it if necessary.
>
> That's all, hopefully all is good, fingers crossed.
>
> Alex
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2024-01-10 17:15 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-01-07 0:00 [PATCH v8 1/4] c++: P0847R7 (deducing this) - prerequisite changes. [PR102609] waffl3x
2024-01-09 20:58 ` Jason Merrill
2024-01-09 21:26 ` waffl3x
2024-01-09 21:56 ` Jason Merrill
2024-01-09 22:34 ` waffl3x
2024-01-09 22:52 ` Jason Merrill
2024-01-09 23:59 ` waffl3x
2024-01-10 17:15 ` Patrick Palka
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).