Cache results of kempner_with_data and other related computations

[fingolfin]

This has only a tiny impact for our "standard example" from PR #238 but still seems useful. Tests are done against current Oscar.jl master

Before:

julia> @time scalar_test(g)
 18.572885 seconds (233.76 M allocations: 24.483 GiB, 10.40% gc time, 5.98% compilation time)

julia> @time scalar_test(g)
 17.501142 seconds (230.64 M allocations: 24.328 GiB, 10.84% gc time)

After:

julia> @time scalar_test(g)
 18.635003 seconds (233.66 M allocations: 24.477 GiB, 10.71% gc time, 6.01% compilation time)

julia> @time scalar_test(g)
 17.423853 seconds (230.54 M allocations: 24.322 GiB, 10.94% gc time)

So it just saves ~100000 allocations, but that's not nothing.

---

I think shorter & quicker micro benchmarks are useful for quick turnaround, so here is a new "standard example" we can copy & paste around:

Before:

julia> using GenericCharacterTables, Chairmarks

julia> g=generic_character_table("3D4.0");

julia> @b scalar_product(g[3],g[7])
3.098 ms (75572 allocs: 3.831 MiB)

julia> @b scalar_product(g[20],g[21])
77.750 ms (995574 allocs: 92.981 MiB, 13.43% gc time)

After:

julia> using GenericCharacterTables, Chairmarks

julia> g=generic_character_table("3D4.0");

julia> @b scalar_product(g[3],g[7])
3.035 ms (75572 allocs: 3.831 MiB)

julia> @b scalar_product(g[20],g[21])
77.214 ms (994886 allocs: 92.936 MiB, 13.85% gc time)

So no change in the first scalar product, and 690 fewer in the second (which is 0.07% ... sigh).

So we really need to go back to some real profiling.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.15%. Comparing base (4d562bb) to head (8012233).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #258      +/-   ##
==========================================
+ Coverage   95.10%   95.15%   +0.04%     
==========================================
  Files          14       14              
  Lines         961      970       +9     
==========================================
+ Hits          914      923       +9     
  Misses         47       47              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[fingolfin]

Seems there is a problem with Oscar 1.1. I'll look into it. In the meantime: any thoughts on this PR, @SoongNoonien ?

[fingolfin]

Made it work with Oscar 1.1 (see new code comment)

[SoongNoonien]

In the meantime: any thoughts on this PR

I'm not so sure if it is worth it. I think this makes the code less nice.