Tamas K. Papp

yes, I think a PR with dispatching to `deepcopy` would be fine. I am curious, what's the use case?

I think you need `julia-actions/setup-julia@v2` for the LTS to be recognized. If you commit everything, then run ```julia PkgSkeleton.generate("/path/to/PGFPlotsX"; templates = [:github]) ``` it should update all workflow files to...

> So this type of recursive copying done by deepcopy is unexpected for data that is provided by the user. I am not sure what the best solution would be...

> like the return value of mv etc which was unrelated to the function With this I agree, I just prefer `nothing` instead of `return`. But feel free to format...

Still could not figure out a workaround for this. Perhaps something like ```@docs MyMod.MyMod # hide ``` would be nice (but currently does not work).

I would be happy to make a PR, but would appreciate some advice on 1. how to handle this for the interface (an ignore list in `makedocs`? a `# hide`...

Reviving this old issue. I would still find this functionality useful. As for the API, a list of symbols to ignore, to be provided as an argument to `makedocs`, would...

Sorry for the stupid question, but if the `Manifest.toml` is regenerated anyway then does it need to be part of the repository? I am facing the same problem to a...

> Current approach: include a Manifest, but try to update it if something is broken. I am under the impression that it is regenerated unconditionally. I tested with #54 too....

Note that breaking it down works: ```julia function g(a) @cast _[i, j] := a[j][i] end function h(r) z = map(g, r) @cast _[i, j ⊗ k] := z[k][i, j] end...