Re: [Patch, Fortran] gfc_notify_std cleanup

[Tobias Burnus <burnus@net-b.de>]

Mikael Morin wrote:
> On 17/07/2012 12:03, Tobias Burnus wrote:
>> On 07/17/2012 11:42 AM, Janus Weil wrote:
>>> +    case GFC_STD_F2008_TS:
>>> +      strcat (buffer, " TS 29113:");
>>> +      break;
>> That's currently correct. However, there is another post-Fortran 2008
>> Technical Specification in preparation. (Coarray extensions.)* How
>> should be handled? "TS 29113 or TS 12345:"?
>>
> Future extension:?
> To be standardized feature:?
> Bleeding edge stuff:?

Well, those names are nice – until we have Fortran 2015 and Fortran 2020.

> Fortran 2015:?
>
> If they are both in the next standard revision, they have to be under
> the same flag anyway.

Well, they will be in the next standard, however, the goal is to be able 
to compile with a -std= flag, which does not include compiler extensions 
but those features. I have met a couple of users, who just wanted to 
have Fortran 95 + allocatable components/dummy arguments (TR 15581:2001) 
and not any Fortran 2003 feature. However, as gfortran doesn't have a 
suitable -std=, they had to use -std=f2003.

Okay, for Fortran 2015, that is probably not a problem as it seems as if 
only few features will be added. There is not much time according to the 
schedule [1], only until mid-2015 – and a lot of time will be spend on 
the coarray TS (due: mid-2013). Additionally, there seems to be the 
feeling in the committee that one shouldn't move forward too quickly, 
given that finding a Fortran 2003 compiler is difficult and having no 
compiler with full Fortran 2008. [2]

In that sense, one could rename -std=f2008ts into -std=f2015 and have 
the "f2015ts" as alias. But otherwise, Fortran 2008 plus TS and Fortran 
2015 are different beasts.

Tobias

[1] ftp://ftp.nag.co.uk/sc22wg5/N1901-N1950/N1925.txt
[2] Especially the UK's position is along those lines 
ftp://ftp.nag.co.uk/sc22wg5/N1901-N1950/N1923.pdf