remove shallowCopy for ARC/ORC

[ringabout]

todo:

• changelog

[ringabout]

Two important packages are affected. It seems to be easy to fix them.

[ringabout]

Prs prepared for important packages:

c-blake/cligen#217
mratsim/Arraymancer#562

[ringabout]

My two prs were merged, just needing to wait for their new release.

[Vindaar]

My two prs were merged, just needing to wait for their new release.

I just did an arraymancer PR to also run the CI under ORC. I decided to run it in release mode, so it doesn't take up too much CI time. Locally I only ran the code in debug mode.

Turns out, what is essentially this issue (I think):

#19374

is still alive and kicking and breaks the ORC (and ARC, checked now locally) release tests of arraymancer:

https://github.com/mratsim/Arraymancer/runs/7499159222?check_suite_focus=true#step:18:4781

[ringabout]

Yeah, we don't have ORC CI for arraymancer. The actual packages affected by this PR is numericalnim.

[ringabout]

I just did an arraymancer PR to also run the CI under ORC. I decided to run it in release mode, so it doesn't take up too much CI time. Locally I only ran the code in debug mode.

Turns out, what is essentially this issue (I think):

Yeah, the issue blocks defaults to Orc

[Vindaar]

I see. Is it worth it to push a new arraymancer tag before that issue is fixed then? Personally I'd rather wait until it's fixed, make sure the CI passes & then push a new tag.

[ringabout]

Is it worth it to push a new arraymancer tag before that issue is fixed then? Personally I'd rather wait until it's fixed, make sure the CI passes & then push a new tag.

This is a seperate critical issue for ARC/ORC. I hope this PR can be merged as soon as possible(this week is better).

[Vindaar]

Fair enough, just tagged a new version.

[ringabout]

It seems to be ready since Packages CI / macos-10.15 (batch: 2_3) (pull_request) pass the CI although Packages CI / macos-10.15 (batch: 2_3) (push) doesn't.

[github-actions]

Thanks for your hard work on this PR!
The lines below are statistics of the Nim compiler built from 4c46358

Hint: mm: orc; threads: on; opt: speed; options: -d:release
163556 lines; 15.033s; 841.086MiB peakmem

[Menduist]

Wouldn't it have been better to just make it pop a warning in A/ORC? (Or ideally make it work under ARC/ORC, like an alias to move, but that may not be possible)

[ringabout]

Wouldn't it have been better to just make it pop a warning in A/ORC? (Or ideally make it work under ARC/ORC, like an alias to move, but that may not be possible)

shallowCopy actually breaks code in ARC/ORC, here is a testcase => #20015 And it cannot be an alias for move because of different semantics.

make it pop a warning in A/ORC

See also discussions from #20025