Remove GeometryBasics/StaticArrays

[sjkelly]

I realize GeometryBasics is barely used here. It is a quite costly dependency, and is only used in a few places, where a simple Tuple seems to suffice. Removing it as a dependency completely removes the need for Plots to depend on StaticArrays, shaving a substantial amount of time off using:

Before:

  3.803301 seconds (8.99 M allocations: 601.643 MiB, 10.07% gc time, 20.31% compilation time)
elapsed time (ns): 3803300770
gc time (ns):      382894522
bytes allocated:   630868505
pool allocs:       8986011
non-pool GC allocs:1970
malloc() calls:    201
realloc() calls:   4
GC pauses:         7
full collections:  2

---

julia> @timev using Plots
  4.005920 seconds (8.99 M allocations: 602.993 MiB, 10.34% gc time, 21.30% compilation time)
elapsed time (ns): 4005919780
gc time (ns):      414263105
bytes allocated:   632284409
pool allocs:       8990504
non-pool GC allocs:2083
malloc() calls:    201
realloc() calls:   4
GC pauses:         6
full collections:  1

After:

  2.442962 seconds (5.31 M allocations: 341.848 MiB, 6.64% gc time, 32.28% compilation time)
elapsed time (ns): 2442961858
gc time (ns):      162169900
bytes allocated:   358453201
pool allocs:       5304483
non-pool GC allocs:1820
malloc() calls:    133
realloc() calls:   4
GC pauses:         6
full collections:  1

----

julia> @timev using Plots
  2.842176 seconds (5.31 M allocations: 342.093 MiB, 14.46% gc time, 27.25% compilation time)
elapsed time (ns): 2842176139
gc time (ns):      411022416
bytes allocated:   358710657
pool allocs:       5304476
non-pool GC allocs:1834
malloc() calls:    133
realloc() calls:   4
GC pauses:         6
full collections:  2

[codecov[bot]]

Codecov Report

Base: 81.10% // Head: 80.85% // Decreases project coverage by -0.25% :warning:

Coverage data is based on head (acf2987) compared to base (1a95d8b). Patch coverage: 42.85% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #4287      +/-   ##
==========================================
- Coverage   81.10%   80.85%   -0.26%     
==========================================
  Files          28       28              
  Lines        7324     7317       -7     
==========================================
- Hits         5940     5916      -24     
- Misses       1384     1401      +17     

Impacted Files	Coverage Δ
src/Plots.jl	87.80% <ø> (ø)
src/examples.jl	98.46% <ø> (+1.53%)	:arrow_up:
src/init.jl	100.00% <ø> (ø)
src/recipes.jl	65.79% <ø> (-1.56%)	:arrow_down:
src/utils.jl	75.54% <ø> (-0.08%)	:arrow_down:
src/components.jl	89.14% <42.85%> (-0.26%)	:arrow_down:
src/consts.jl	75.00% <0.00%> (-25.00%)	:arrow_down:
src/backends.jl	65.74% <0.00%> (-1.23%)	:arrow_down:
src/backends/hdf5.jl	83.02% <0.00%> (-0.74%)	:arrow_down:
... and 11 more

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

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

[sjkelly]

Any reason why Benchmarks is failing? Is it caching a Manifest?

[BeastyBlacksmith]

Any reason why Benchmarks is failing? Is it caching a Manifest?

Seems it is looking for GeometryBasics for the run of the master version and can't find it, since it is not in the Project.toml anymore...

[sjkelly]

I have not much time to look into BenchmarksCI more. I am wondering if this were merged it would then run okay?

[sjkelly]

@BeastyBlacksmith thanks for looking at the BenchmarksCI!

[BeastyBlacksmith]

@BeastyBlacksmith thanks for looking at the BenchmarksCI!

Yeah, well.. this is a mess though. It ought to be easier...

[BeastyBlacksmith]

So benchmarks are green now. Not sure how to go about ci, looks kind of unrelated

[sjkelly]

Amazing! It seems CI has been failing a few days now?

On Wed, Aug 24, 2022, 08:24 Simon Christ @.***> wrote:

So benchmarks are green now. Not sure how to go about ci, looks kind of unrelated

— Reply to this email directly, view it on GitHub https://github.com/JuliaPlots/Plots.jl/pull/4287#issuecomment-1225650524, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAP5MDCFETLVT2PQZ2R62CTV2YH7FANCNFSM54XZ5KBQ . You are receiving this because you authored the thread.Message ID: @.***>

[BeastyBlacksmith]

So benchmarks are green now.

Looking at the results though.. I question the fidelity of the benchmark

[BeastyBlacksmith]

Better now

[t-bltg]

CI was fixed in https://github.com/JuliaPlots/Plots.jl/pull/4319.

I will merge this PR when ci is green (after merge of https://github.com/JuliaPlots/Plots.jl/pull/4320), since the ref images changed.

[BeastyBlacksmith]

I'd wait with merging until we checked whether we can keep existing functionality with StaticArraysCore (see https://github.com/JuliaPlots/Plots.jl/pull/4287#discussion_r931254890)

[t-bltg]

Noted, postponed.

[sjkelly]

StaticArraysCore does not cover that dispatch case I commented on. Any GeometryBasics and StaticArrays dep would end up in downstream plotting recipes packages instead. Though the StaticArraysCore dependency is cheap, it isn't going to help here.

[t-bltg]

@BeastyBlacksmith, merge ?

[sjkelly]

I have updated the docs here: https://github.com/JuliaPlots/PlotDocs.jl/pull/300

[j-fu]

:+1: