Re: [PATCH] c++: non-dependent .* folding [PR112427]

[Jason Merrill <jason@redhat.com>]

On 11/14/23 10:43, Patrick Palka wrote:
> On Fri, 10 Nov 2023, Jason Merrill wrote:
> 
>> On 11/10/23 10:28, Patrick Palka wrote:
>>> On Fri, 10 Nov 2023, Patrick Palka wrote:
>>>
>>>> On Thu, 9 Nov 2023, Jason Merrill wrote:
>>>>
>>>>> On 11/8/23 16:59, Patrick Palka wrote:
>>>>>> Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK
>>>>>> for
>>>>>> trunk?
>>>>>>
>>>>>> -- >8 --
>>>>>>
>>>>>> Here when building up the non-dependent .* expression, we crash from
>>>>>> fold_convert on 'b.a' due to this (templated) COMPONENT_REF having an
>>>>>> IDENTIFIER_NODE instead of FIELD_DECL operand that middle-end routines
>>>>>> expect.  Like in r14-4899-gd80a26cca02587, this patch fixes this by
>>>>>> replacing the problematic piecemeal folding with a single call to
>>>>>> cp_fully_fold.
>>>>>>
>>>>>> 	PR c++/112427
>>>>>>
>>>>>> gcc/cp/ChangeLog:
>>>>>>
>>>>>> 	* typeck2.cc (build_m_component_ref): Use cp_convert, build2
>>>>>> and
>>>>>> 	cp_fully_fold instead of fold_build_pointer_plus and
>>>>>> fold_convert.
>>>>>
>>>>>> gcc/testsuite/ChangeLog:
>>>>>>
>>>>>> 	* g++.dg/template/non-dependent29.C: New test.
>>>>>> ---
>>>>>>     gcc/cp/typeck2.cc                               |  5 ++++-
>>>>>>     gcc/testsuite/g++.dg/template/non-dependent29.C | 13 +++++++++++++
>>>>>>     2 files changed, 17 insertions(+), 1 deletion(-)
>>>>>>     create mode 100644 gcc/testsuite/g++.dg/template/non-dependent29.C
>>>>>>
>>>>>> diff --git a/gcc/cp/typeck2.cc b/gcc/cp/typeck2.cc
>>>>>> index 309903afed8..208004221da 100644
>>>>>> --- a/gcc/cp/typeck2.cc
>>>>>> +++ b/gcc/cp/typeck2.cc
>>>>>> @@ -2378,7 +2378,10 @@ build_m_component_ref (tree datum, tree
>>>>>> component,
>>>>>> tsubst_flags_t complain)
>>>>>>           /* Build an expression for "object + offset" where offset is
>>>>>> the
>>>>>>     	 value stored in the pointer-to-data-member.  */
>>>>>>           ptype = build_pointer_type (type);
>>>>>> -      datum = fold_build_pointer_plus (fold_convert (ptype, datum),
>>>>>> component);
>>>>>> +      datum = cp_convert (ptype, datum, complain);
>>>>>> +      datum = build2 (POINTER_PLUS_EXPR, ptype,
>>>>>> +		      datum, convert_to_ptrofftype (component));
>>>>>
>>>>> We shouldn't need to build the POINTER_PLUS_EXPR at all in template
>>>>> context.
>>>>> OK with that change.
>>>>
>>>> Hmm, that seems harmless at first glance, but I noticed
>>>> build_min_non_dep (called from build_x_binary_op in this case) is
>>>> careful to propagate TREE_SIDE_EFFECTS of the given tree, and so eliding
>>>> POINTER_PLUS_EXPR here could potentially mean that the tree we
>>>> ultimately return from build_x_binary_op when in a template context has
>>>> TREE_SIDE_EFFECTS not set when it used to.  Shall we still elide the
>>>> POINTER_PLUS_EXPR in a template context despite this?
>>
>> True, we would need build_min_non_dep to also get TREE_SIDE_EFFECTS from the
>> operands.  That might be useful in general for similar situations?
>>
>> I also note that convert_to_ptrofftype uses fold_convert, so the new code
>> could have the same problem if the pointer to member operand is also a
>> COMPONENT_REF.
> 
> Ah, makes sense...  How does the following look then?  I'm not sure if
> we still want to replace fold_build_pointer_plus with build2 if we're
> not going to use that code path in a template context?  Bootstrapped and
> regtested on x86_64-pc-linux-gnu.

We probably do, since we now aren't folding the thing before it.  OK.

> -- >8 --
> -
> Subject: [PATCH] c++: non-dependent .* operand folding [PR112427]
> 
> Here when building up the non-dependent .* expression, we crash from
> fold_convert on 'b.a' due to this (templated) COMPONENT_REF having an
> IDENTIFIER_NODE instead of FIELD_DECL operand that middle-end routines
> expect.  Like in r14-4899-gd80a26cca02587, this patch fixes this by
> replacing the problematic piecemeal folding with a single call to
> cp_fully_fold.  Additionally, don't bother building POINTER_PLUS_EXPR
> in a template context.  This means the returned non-dependent tree might
> not have TREE_SIDE_EFFECTS set when it used to, so we need to make
> build_min_non_dep compensate for this by propagating TREE_SIDE_EFFECTS
> from the original arguments like buildN and build_min do, which seems
> like a generally desirable thing to do anyway.
> 
> 	PR c++/112427
> 
> gcc/cp/ChangeLog:
> 
> 	* tree.cc (build_min_non_dep): Propagate TREE_SIDE_EFFECTS from
> 	the original arguments.
> 	(build_min_non_dep_call_vec): Likewise.
> 	* typeck2.cc (build_m_component_ref): Use cp_convert, build2 and
> 	cp_fully_fold instead of fold_build_pointer_plus and fold_convert.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* g++.dg/template/non-dependent29.C: New test.
> ---
>   gcc/cp/tree.cc                                  | 11 ++++++++++-
>   gcc/cp/typeck2.cc                               |  6 +++++-
>   gcc/testsuite/g++.dg/template/non-dependent29.C | 13 +++++++++++++
>   3 files changed, 28 insertions(+), 2 deletions(-)
>   create mode 100644 gcc/testsuite/g++.dg/template/non-dependent29.C
> 
> diff --git a/gcc/cp/tree.cc b/gcc/cp/tree.cc
> index 417c92ba76f..dc4126f935c 100644
> --- a/gcc/cp/tree.cc
> +++ b/gcc/cp/tree.cc
> @@ -3601,7 +3601,12 @@ build_min_non_dep (enum tree_code code, tree non_dep, ...)
>     TREE_SIDE_EFFECTS (t) = TREE_SIDE_EFFECTS (non_dep);
>   
>     for (i = 0; i < length; i++)
> -    TREE_OPERAND (t, i) = va_arg (p, tree);
> +    {
> +      tree x = va_arg (p, tree);
> +      TREE_OPERAND (t, i) = x;
> +      if (x && !TYPE_P (x))
> +	TREE_SIDE_EFFECTS (t) |= TREE_SIDE_EFFECTS (x);
> +    }
>   
>     va_end (p);
>     return convert_from_reference (t);
> @@ -3636,6 +3641,10 @@ build_min_non_dep_call_vec (tree non_dep, tree fn, vec<tree, va_gc> *argvec)
>       non_dep = TREE_OPERAND (non_dep, 0);
>     TREE_TYPE (t) = TREE_TYPE (non_dep);
>     TREE_SIDE_EFFECTS (t) = TREE_SIDE_EFFECTS (non_dep);
> +  if (argvec)
> +    for (tree x : *argvec)
> +      if (x && !TYPE_P (x))
> +	TREE_SIDE_EFFECTS (t) |= TREE_SIDE_EFFECTS (x);
>     return convert_from_reference (t);
>   }
>   
> diff --git a/gcc/cp/typeck2.cc b/gcc/cp/typeck2.cc
> index 309903afed8..a75f4f8c2d2 100644
> --- a/gcc/cp/typeck2.cc
> +++ b/gcc/cp/typeck2.cc
> @@ -2378,7 +2378,11 @@ build_m_component_ref (tree datum, tree component, tsubst_flags_t complain)
>         /* Build an expression for "object + offset" where offset is the
>   	 value stored in the pointer-to-data-member.  */
>         ptype = build_pointer_type (type);
> -      datum = fold_build_pointer_plus (fold_convert (ptype, datum), component);
> +      datum = cp_convert (ptype, datum, complain);
> +      if (!processing_template_decl)
> +	datum = build2 (POINTER_PLUS_EXPR, ptype,
> +			datum, convert_to_ptrofftype (component));
> +      datum = cp_fully_fold (datum);
>         datum = cp_build_fold_indirect_ref (datum);
>         if (datum == error_mark_node)
>   	return error_mark_node;
> diff --git a/gcc/testsuite/g++.dg/template/non-dependent29.C b/gcc/testsuite/g++.dg/template/non-dependent29.C
> new file mode 100644
> index 00000000000..41bd11ae6b4
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/template/non-dependent29.C
> @@ -0,0 +1,13 @@
> +// PR c++/112427
> +
> +struct A { int m; void f(); };
> +struct B { A a; };
> +
> +template<class T>
> +void f(B b) {
> +  int A::*pd = &A::m;
> +  b.a.*pd;
> +
> +  void (A::*pf)() = &A::f;
> +  (b.a.*pf)();
> +}