[J3] [EXTERNAL] Re: USE statement – wish: permit specifying additionally the access-spec

[Robert Corbett]

OK, subclause 8.6.1, paragraph 2
makes it clear that both
PRIVATE and PUBLIC can be
specified for an entity in a
scoping unit, and that PUBLIC
wins.  I would assume that the
same rule would apply to
Van's example.

Bob Corbett

> On Nov 15, 2019, at 4:39 PM, Robert Corbett via J3 <j3 at mailman.j3-fortran.org> wrote:
> 
> My first thought was that C815
> prohibits this case.  If the attributes
> PRIVATE and PUBLIC are both
> regarded as the accessibility
> attribute, it is prohibited.  If not,
> there seems to be a missing
> restriction in the existing text of
> the standard.
> 
> Bob Corbett
> 
>>> On Nov 15, 2019, at 3:38 PM, Van Snyder via J3 <j3 at mailman.j3-fortran.org> wrote:
>>> 
>>> On Fri, 2019-11-15 at 23:04 +0000, Bill Long via J3 wrote:
>>> While I'm not part of the Data subgroup, I’m pretty sure that the
>>> design for this feature was not an oversight.  Rather, it likely had
>>> to do with the fact that public and private declarations are limited
>>> to being only in a module, whereas USE statements (often) appear
>>> outside modules. Sure, you could add in a constraint that the
>>> access-spec cannot be specified in a USE statement that is outside a
>>> module, but that is just one extra complication that is not needed
>>> with the current design. 
>>> 
>>> That said, we already permit redundant ways to specify public and
>>> private for many other situations. So I would not be opposed to this
>>> idea.  Some comment from /Data would be informative.
>> 
>> Another reason not to put an access-spec on a USE statement is that a
>> scoping unit can have more than one USE statement for the same module.
>> What is the effect of
>> 
>> use, private :: A
>> use, public :: A
>> 
>> 
>>> 
>>> Cheers,
>>> Bill
>>> 
>>> 
>>> 
>>>> On Nov 15, 2019, at 3:02 PM, Vipul Parekh via J3 <j3 at mailman.j3-fortran.org> wrote:
>>>> 
>>>> On Fri, Nov 15, 2019 at 11:32 AM Tobias Burnus via J3
>>>> <j3 at mailman.j3-fortran.org> wrote:
>>>>> ..
>>>>> I still believe that a use-stmt with access-spec is more readable than
>>>>> use-stmt + access-stmt, but I agree that the functionality is already
>>>>> covered. Hence, I think one should still (re)consider it – but with low
>>>>> priority.
>>>>> ..
>>>> 
>>>> Tobias makes a good point.
>>>> 
>>>> Is it possible to know whether there was a reason besides an oversight
>>>> as to why access-spec with the USE statement was not introduced in
>>>> Fortran 2018 itself when the deficiency with the default accessibility
>>>> for entities accessed from a module was addressed?
>>>> 
>>>> If this is an oversight with Fortran 2018, then how can one exhort J3
>>>> but importantly WG5 to include this and other such "Features that
>>>> address deficiencies and discrepancies" on the Fortran 202X work-list?
>>>> 
>>>> Thanks,
>>>> Vipul Parekh
>>> 
>>> Bill Long                                                                       longb at cray.com
>>> Principal Engineer, Fortran Technical Support &   voice:  651-605-9024
>>> Bioinformatics Software Development                      fax:  651-605-9143
>>> Cray, a Hewlett Packard Enterprise company/ 2131 Lindau Lane/  Suite 1000/  Bloomington, MN  55425
>>> 
>>> 
>>> 
>> 
>> 
>