Improve generation of typealiases

[kimikage]

This is yet another solution for the issue which is going to be fixed by PR #232.

This also includes the preparation for #228. I don't believe typechar is used in other packages, but I've added depwarn because there are cases where showtype is used.

Edit:
This PR is not intended to replace #232, but to demonstrate a different approach. However, in any case, I believe that some fixes will be required for #228.

[codecov]

Codecov Report

Merging #233 (953001e) into master (fe2135d) will decrease coverage by 0.11%.
The diff coverage is 94.11%.

@@            Coverage Diff             @@
##           master     #233      +/-   ##
==========================================
- Coverage   95.68%   95.57%   -0.12%     
==========================================
  Files           6        6              
  Lines         742      746       +4     
==========================================
+ Hits          710      713       +3     
- Misses         32       33       +1     

Impacted Files	Coverage Δ
src/FixedPointNumbers.jl	95.43% <87.50%> (-0.32%)	⬇️
src/deprecations.jl	50.00% <100.00%> (+50.00%)	⬆️
src/fixed.jl	96.89% <100.00%> (ø)
src/normed.jl	95.47% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fe2135d...953001e. Read the comment docs.

[kimikage]

I am going to try to see if the symbol-generating "macro" is effective.

[kimikage]

I've tried several metaprogramming techniques and generated function seems to be the simplest.

julia> versioninfo()
Julia Version 1.6.0-DEV.1050
Commit fd57021b33 (2020-09-25 18:40 UTC)
Platform Info:
  OS: Windows (x86_64-w64-mingw32)
  CPU: Intel(R) Core(TM) i7-8565U CPU @ 1.80GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-10.0.1 (ORCJIT, skylake)

julia> @time using FixedPointNumbers
[ Info: Precompiling FixedPointNumbers [53c48c17-4a7d-5ca2-90c5-79b7896eea93]
  2.279809 seconds (213.42 k allocations: 13.772 MiB, 0.25% gc time) # master
  1.179185 seconds (209.43 k allocations: 13.267 MiB)                # this PR

printing ARGB32(cf. #232 (comment))

julia> versioninfo()
Julia Version 1.5.2
Commit 539f3ce943 (2020-09-23 23:17 UTC)
Platform Info:
  OS: Windows (x86_64-w64-mingw32)
  CPU: Intel(R) Core(TM) i7-8565U CPU @ 1.80GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-9.0.1 (ORCJIT, skylake)

julia> @benchmark dump(io, $img) # master
BenchmarkTools.Trial:
  memory estimate:  1.70 MiB
  allocs estimate:  9216
  --------------
  minimum time:     740.600 μs (0.00% GC)
  median time:      1.450 ms (0.00% GC)
  mean time:        1.666 ms (16.23% GC)
  maximum time:     181.050 ms (0.46% GC)
  --------------
  samples:          3000
  evals/sample:     1

julia> @benchmark dump(io, $img) # this PR
BenchmarkTools.Trial:
  memory estimate:  1.70 MiB
  allocs estimate:  9216
  --------------
  minimum time:     622.900 μs (0.00% GC)
  median time:      1.344 ms (0.00% GC)
  mean time:        1.562 ms (17.51% GC)
  maximum time:     202.086 ms (8.59% GC)
  --------------
  samples:          3211
  evals/sample:     1

The original reason for submitting this PR is never to speed up printing colors, but I am not in a position to negate the need for the specialization anymore.

I will modify this PR as you asked and will not merge this PR unless there is a consensus.

[kimikage]

The essence of this PR is to reduce the precompile time and to support two-character type prefixes.
I don't think it's selfish to merge this, as this PR has been open for over two months...