Hi Paul,
This correspondence touches on something that I was going to raise - how
do we incorporate experimental F202Y features?
I think your suggestion of -std=f202y is good, we can then make
-funsigned depend on that.
Hmm... one thing. I just read
https://wg5-fortran.org/N2201-N2250/N2234.txt
and I didn't find the unsigned proposal there. Damian, you were there.
Was this considered and rejected, or not considered and postponed
to a future date?
(Anyway, even if this does not make it into F202Y, I am still going
ahead with this. Let other compilers pick it up, or not, as the case
may be).
The reason that I ask is that Reinhold Bader proposed extensions to the
processing of assumed rank objects, which became a DIN proposal to WG5 -
see attached. It made so much sense to me that, in June, I wrote an
experimental implementation. The experimental implementation added the
reshape intrinsic as well as allowing remapped pointer assignment and
associate. A testcase is attached.
In the intervening three months the experimental patch developed bit-rot
to the extent that two chunks would no longer apply. For this reason, I
am very reluctant to set up a separate branch that would have to be
synchronised over a good number of years before 202y becomes the new
standard. Additionally, the array descriptor reform branch suffered so
badly from this that I gave up on it. I propose that we introduce
-std=F202Y, which will prevent experimental F202Y features from being
used unless this standard is selected. That way they remain in mainline
and are readily identifiable for pruning come the day or maintenance as
the F202Y proposals develop. If this is not acceptable, I will open a PR
in which experimental patches can be stashed. Given previous experience,
my preference is very much for the former!
I have also drafted much of the standardese, relative to F2003, and I
suggest that, in memory of Reinhold, this continue as a DIN
proposal. @Thomas, are you willing to carry this forward or should I
transfer it to being a UK proposal?
Unfortunately, with Reinhold passing away, I don't even know who the
DIN member is, if there even is one. So, it is probably best
if you continue this as an UK proposal.
Best regards
Thomas
