This is the mail archive of the fortran@gcc.gnu.org mailing list for the GNU Fortran project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Legacy patches - should I wait?


On Wed, Jan 23, 2019 at 09:43:26AM +0000, Mark Eggleston wrote:
> Both features are supported by PGI Fortran CE 18.10 without the need for 
> any compiler switches.

I'll reframe from commenting on my experience with PGI.

Can you point to actual in-the-wild code that uses these
nonstandard forms?  Adding something simply because PGI
supports it seems dubious to me. 

> I agree that users should be encouraged to write standard conforming 
> code that is why the patches are only enabled using 
> -fdec-format-defaults and -fdec-blank-format-item and in turn by -fdec. 
> They are intended for the case where there is a large codebase and it 
> would take considerable effort to fix the code.

Do these nonstandard forms occur in DEC Fortran?  If not, these
should probably be hidden behind -std=legacy.  I tried a google
search for both, but could not find any information.

> When would be a good time to submit the patches for consideration, now 
> or after the current development stage (regression and documentation 
> fixes only) is over?

If the changes are hidden behind an option and you have
done regression testing, I suppose you can submit them now.
The last status report for 9.0 suggested a mid-April release.

OTOH, if you want to follow the rules of GCC, then submit a
bug report, attach the patch, and wait for stage 1.

-- 
steve


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]