Use metaprogramming to define the kwargs of all Krylov methods

[amontoison]

Thanks @mosullivan93 for your help with metaprogramming!
@dpo @tmigot Julia will not recompile the non in-place methods anymore if we use different keyword arguments.

It will be also easier to maintain.

ps: I found some mistakes in the docstrings of CRLS, CGLS, CGNE and CRMR.

[codecov]

Codecov Report

Patch coverage: 99.71% and project coverage change: -0.01 ⚠️

Comparison is base (cec1dd8) 98.28% compared to head (a6b2b04) 98.27%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #735      +/-   ##
==========================================
- Coverage   98.28%   98.27%   -0.01%     
==========================================
  Files          38       38              
  Lines        6861     6898      +37     
==========================================
+ Hits         6743     6779      +36     
- Misses        118      119       +1     

Impacted Files	Coverage Δ
src/krylov_utils.jl	91.32% <75.00%> (-0.35%)	⬇️
src/bicgstab.jl	100.00% <100.00%> (ø)
src/bilq.jl	100.00% <100.00%> (ø)
src/bilqr.jl	100.00% <100.00%> (ø)
src/cg.jl	100.00% <100.00%> (ø)
src/cg_lanczos.jl	100.00% <100.00%> (ø)
src/cg_lanczos_shift.jl	100.00% <100.00%> (ø)
src/cgls.jl	100.00% <100.00%> (ø)
src/cgne.jl	100.00% <100.00%> (ø)
src/cgs.jl	100.00% <100.00%> (ø)
... and 24 more

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

[github-actions]

Package name	latest	stable
CaNNOLeS.jl
DCISolver.jl
FletcherPenaltySolver.jl
JSOSolvers.jl
LLSModels.jl
Percival.jl
RipQP.jl

[tmigot]

Thanks @amontoison !
It looks good to me. I guess this will increase the pre-compilation of the package though, but that's alright for now.

[amontoison]

No, the functions were already precompiled.
The only difference now is that the non-inplace methods know which in-place version they will call, which means that Julia doesn't need to recompile them at runtime to fully infer the code.
Julia was unable to predict the kwargs....

The in-place methods (without warm-start) were already correctly precompiled.