Leon Timmermans
Leon Timmermans
Can't argue with this. It's probably not easy to solve though :-/
I can't reproduce this (on a Septemer 5th rakudo)
This sort of thing is very necessary. I wish it was prettier to implement though. I'm currently trying to reduce the amount of contextual variables, as it tends to lead...
> would that be fine for merging immediately? I guess. Possibly my problem is that I don't fully understand your code; tests may be helpful here.
I'm not sure perldelta is really the right place for this.
We don't usually update perldelta like that (except minor edits like typos and broken links). Perhaps perlexperimental is a more logical place.
@leonerd probably some ideas about where the best place would be
That is quite a surprise. It seems PDL has its own [typemap evaluator](https://metacpan.org/dist/PDL/source/Basic/Gen/PP.pm#L1036) that will need to be adapted to this change.
> Are you asking me to change the code? Absolutely. Even if this change is reverted you should change the code (but in a different way); it's broken for output...
Yeah I will do that.