Small optimimization of constraint condensation

[fredrikekre]

Use a dict of size n constraints instead of a vector of size ndofs.

[lijas]

I guess it is more memory efficient, but is it faster?

[fredrikekre]

No reason it should be slower 🤷 (compared to everything else in this function I don't think it is noticeable). Just seems nicer to allocate only what you need, in particular since the number of constrained dofs usually is much smaller then the number of total dofs.

[codecov-commenter]

Codecov Report

Merging #417 (d6acbf2) into master (27192eb) will decrease coverage by 0.00%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #417      +/-   ##
==========================================
- Coverage   91.30%   91.29%   -0.01%     
==========================================
  Files          22       22              
  Lines        3046     3043       -3     
==========================================
- Hits         2781     2778       -3     
  Misses        265      265              

Impacted Files	Coverage Δ
src/Dofs/ConstraintHandler.jl	94.65% <100.00%> (-0.04%)	⬇️

---

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 27192eb...d6acbf2. Read the comment docs.

[lijas]

Ok makes sense. I think I use the same trick in create_constraint_matrix, so you can update that as well maybe.
Otherwise this looks good to me.