[RecipesBase] Use interpolation instead of qualification to drop dependance on the namespace of RecipesBase users

[LilithHafner]

We can't assume that folks will have RecipesBase in their namespace.

Fixes #4557

[LilithHafner]

CI is already failing after just a few minutes <3

[BeastyBlacksmith]

We can't expect them to have all those internal functions in the namespace either, I'd say.
The proper fix would be to document that users should at least use

using RecipesBase: RecipesBase, @recipe

[t-bltg]

As Simon suggested, improving the docs is probably the easiest.

With a minor nitpick:

import RecipesBase: RecipesBase, @recipe

[LilithHafner]

The original approach (fb77991) should work if JuliaLang/julia#47726 is fixed upstream.

[LilithHafner]

Failing that, a slightly more verbose approach should work following this pattern

julia> module NanoRecipesBase
           function apply_recipe end
           macro recipe(x)
               quote
                   let M = $NanoRecipesBase
                       M.apply_recipe() = $x
                   end
               end
           end
       end
Main.NanoRecipesBase

julia> module User
           import Main.NanoRecipesBase: @recipe
           try
               println(NanoRecipesBase)
           catch UndefVarError
               println("Good")
           end
           @recipe 1729
       end
Good
Main.User

julia> NanoRecipesBase.apply_recipe()
1729

[LilithHafner]

The proper fix would be to document that users should at least use

As Simon suggested, improving the docs is probably the easiest.

We should only document an unexpected limitation/bug if we can't fix it. This PR should fix it :)

[codecov]

Codecov Report

Base: 90.94% // Head: 90.95% // Increases project coverage by +0.00% 🎉

Coverage data is based on head (e9e12d6) compared to base (497cec4).
Patch coverage: 100.00% of modified lines in pull request are covered.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #4559   +/-   ##
=======================================
  Coverage   90.94%   90.95%           
=======================================
  Files          40       40           
  Lines        7800     7802    +2     
=======================================
+ Hits         7094     7096    +2     
  Misses        706      706           

Impacted Files	Coverage Δ
RecipesBase/src/RecipesBase.jl	94.21% <100.00%> (ø)
src/examples.jl	97.59% <0.00%> (ø)
src/animation.jl	99.04% <0.00%> (ø)
src/backends/gr.jl	91.65% <0.00%> (ø)
src/utils.jl	92.61% <0.00%> (+0.02%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[LilithHafner]

The format check gave a lot of diffs. Is there a way to avoid transcribing all those non-functional-changes manually?

[t-bltg]

https://docs.juliaplots.org/latest/contributing/#Write-code,-and-format

$ julia -e 'using JuliaFormatter; format(["src", "test"])'

[t-bltg]

Nice.

[t-bltg]

Seems like using GlobalRef triggers a segfault apocalypse here 😆

[LilithHafner]

Strange. I guess I'll just revert to 5440348 "fix formatting".

[LilithHafner]

I don't understand the GlobalRef segfaults so I asked on Discourse: https://discourse.julialang.org/t/what-is-a-globalref-and-how-i-use-it/90971

[LilithHafner]

Thanks @ffevotte from discourse for the latest and best workaround for julia#47726

[t-bltg]

Looks good, thanks for getting to the bottom of this with a cleaner solution. Still intrigued by the GlobalRef segfault though 🤔

[LilithHafner]

I'm happy with de5a423, but I believe e9e12d6 contains unnecessarily indirect interpolation.

[t-bltg]

I'm happy with de5a423, but I believe e9e12d6 contains unnecessarily indirect interpolation.

Do you imply that those added interpolations have a runtime performance penalty ?

[LilithHafner]

Do you implicit that those added interpolations have a runtime performance penalty ?

No. There should not be a runtime penalty.

[LilithHafner]

Ultimately, it's just a style/formatting choice. I prefer direct interpolation over module interpolation + qualified reference because it is shorter, but either should work, and y'all are more likely to be maintaining this than me so feel free to pick whatever style you prefer.

[t-bltg]

We disagree on this one, and I explained my position in #4559 (comment).

Thanks for the contribution and the detailed analysis here, always useful to learn about the internals !