Remove StableSet and StableDict

[jakobnissen]

These two data types were originally added to prevent nondeterminism when compiling regex. Instead, prevent nondetminism by relabeling the minimized machines.

Cherry-picked from the v1 PR

See issues #19, #106

[codecov[bot]]

Codecov Report

Merging #107 (ff0f66b) into master (69d16f5) will decrease coverage by 10.27%. The diff coverage is 73.77%.

@@             Coverage Diff             @@
##           master     #107       +/-   ##
===========================================
- Coverage   90.76%   80.49%   -10.28%     
===========================================
  Files          16       16               
  Lines        1808     1866       +58     
===========================================
- Hits         1641     1502      -139     
- Misses        167      364      +197     

Flag	Coverage Δ
unittests	80.49% <73.77%> (-10.28%)	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/Automa.jl	100.00% <ø> (ø)
src/edge.jl	58.97% <52.94%> (-41.03%)	:arrow_down:
src/machine.jl	79.04% <100.00%> (+6.20%)	:arrow_up:
src/traverser.jl	100.00% <100.00%> (ø)
src/sset.jl	0.00% <0.00%> (-100.00%)	:arrow_down:
src/sdict.jl	0.00% <0.00%> (-95.53%)	:arrow_down:

Help us with your feedback. Take ten seconds to tell us how you rate us.

[CiaranOMara]

Can we expect the dot files to be consistent, or do they need some form of mapping/ordering too?

[jakobnissen]

@CiaranOMara they are consistent for Machine's, but not for NFA's or DFA's. Is that a problem? Then I can consider making the latter deterministic, also.

[CiaranOMara]

Not sure if it's a problem. The nfa2dot and dfa2dot functions were using StableDict. So, those functions may be producing inconsistent dotfiles. Could it be as trivial as changing for s in keys(serials) in /src/dot.jl to for s in traverse(nfa.start) or traverse(dfa.start) to get the StableDict behaviour?