Re: [PATCH] c++: Allow constexpr decltype(auto) [PR102229]

[Jason Merrill <jason@redhat.com>]

On 12/15/21 16:28, Marek Polacek wrote:
> On Wed, Dec 15, 2021 at 04:17:37PM -0500, Jason Merrill wrote:
>> On 12/15/21 15:20, Marek Polacek wrote:
>>> On Mon, Dec 13, 2021 at 10:02:24AM -0500, Jason Merrill wrote:
>>>> On 12/10/21 17:29, Marek Polacek wrote:
>>>>> My r11-2202 was trying to enforce [dcl.type.auto.deduct]/4, which says
>>>>> "If the placeholder-type-specifier is of the form type-constraint[opt]
>>>>> decltype(auto), T shall be the placeholder alone."  But this made us
>>>>> reject 'constexpr decltype(auto)', which, after clarification from CWG,
>>>>> should be valid.  [dcl.type.auto.deduct]/4 is supposed to be a syntactic
>>>>> constraint, not semantic, so it's OK that the constexpr marks the object
>>>>> as const.
>>>>>
>>>>> As a consequence, checking TYPE_QUALS in do_auto_deduction is too late,
>>>>> and we have a FIXME there anyway.  So in this patch I'm attempting to
>>>>> detect 'const decltype(auto)' earlier.  If I'm going to use TYPE_QUALS,
>>>>> it needs to happen before we mark the object as const due to constexpr,
>>>>> that is, before grokdeclarator's
>>>>>
>>>>>      /* A `constexpr' specifier used in an object declaration declares
>>>>>         the object as `const'.  */
>>>>>      if (constexpr_p && innermost_code != cdk_function)
>>>>>        ...
>>>>>
>>>>> Constrained decltype(auto) was a little problem, hence the TYPENAME
>>>>> check.  But in a typename context you can't use decltype(auto) anyway,
>>>>> I think.
>>>>
>>>> I wonder about checking even earlier, like in cp_parser_decl_specifier_seq?
>>>
>>> That _almost_ works except it wouldn't detect things like 'decltype(auto)*'
>>> because the '*' isn't parsed in cp_parser_decl_specifier_seq, only in
>>> declarator.  So the
>>
>> Ah, right.
>>
>>>        if (a != type)
>>>          {
>>>            error_at (loc, "%qT as type rather than plain "
>>>              "%<decltype(auto)%>", type);
>>>
>>> check wouldn't work.  Maybe I could just check if the next token is * or &
>>> and give an error then.
>>
>> No, checking in grokdeclarator makes sense.
>>
>>> Constrained decltype(auto) was a little problem, hence the TYPENAME
>>> check.  But in a typename context you can't use decltype(auto) anyway,
>>> I think.
>>
>> Maybe check PLACEHOLDER_TYPE_CONSTRAINTS in check_decltype_auto instead?
> 
> I've tried that, but that is also true for
> 
> const constexpr C decltype(auto) x2 = 0;
> const constexpr C decltype(auto) fn4() { return 0; }
> 
> where we do want to check if the auto has quals.  Therefore the not very
> pretty TYPENAME check :/.

Aha.  The patch is OK.

Jason