[J3] [EXTERNAL] an amusing mistake

[Robert Corbett]

The optional comma was present in FORTRAN 77,which is as far back as the CHARACTER data typegoes.  The constraint was added in Fortran 90,when the form and function of declarationstatements were greatly expanded. In Fortran 90and Fortran 95, the constraint was correct,because the double-colon separator could notappear in a expression.  The syntax for arrayconstructors was extended in Fortran 2003 toinclude an optional double-colon separator.The constraint should have been modified inFortran 2003, or a different syntax should havebeen chosen for array constructors.
I am surprised I did not notice the mistakebefore.  I studied the Fortran 2003 standardvery closely when I was still working.

Deleting the optional comma would make lots ofexisting codes nonstandard.  I would rather notdo that.  I am slightly less opposed to doingit now that I am retired and do not have tofield complaints from irate users.
Bob Corbett


    On Tuesday, August 4, 2020, 9:21:36 AM PDT, Clune, Thomas L. (GSFC-6101) via J3 <j3 at mailman.j3-fortran.org> wrote:  
 
 Ditto.


On Aug 4, 2020, at 11:13 AM, Dan Nagle via J3 <j3 at mailman.j3-fortran.org> wrote:
Hi,


On Aug 4, 2020, at 09:10 , Bill Long via J3 <j3 at mailman.j3-fortran.org> wrote:

Perhaps a more reasonable solution would be to just move the optional comma to the Deleted List. (Hence deleting the constraint).



I can agree.


--

Cheers!
Dan Nagle




  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.j3-fortran.org/pipermail/j3/attachments/20200804/23fd1c8c/attachment.htm>