Fix `substitute_node`; substitute without re-strashing

[lee30sonia]

• Adds a failing testcase where re-strashing in the propagated substitutions caused problem (segfault in release mode or assertion fail in debug mode) (related to the fix in #560)
• Fixes the above problem in AIG, XAG, MIG, XMG, TIG, fanout_view
• Implements a simpler function to substitute without re-strashing substitute_node_no_restrash (only for AIG, as it is still experimental). This function may be used to directly replace substitute_node in some algorithms (they have the same interface). It could make substitution much faster when there is supposed to be lots of re-strashing and constant propagation. However, it could also cause overhead for algorithms to process redundant nodes. cleanup_dangling should be called in the end to strash nodes again.
• Implements revive_node to revive dead nodes (i.e. nodes marked as dead because their fanout size reduces to 0; they are removed from the hash but kept in the data structure). This function could solve #565.

IDEA: Besides the strategy of substitution without re-strashing, we could also implement another simpler substitute_node which does re-strashing normally but assumes that the network is in topological order (because in that case we don't need the fix in #560) --> future work

[lee30sonia]

Tagging some people with whom I have discussed this issue: @Nozidoali @aletempiac @hriener

[codecov-commenter]

Codecov Report

Merging #586 (9638125) into master (510c9dd) will increase coverage by 0.03%. The diff coverage is 92.59%.

@@            Coverage Diff             @@
##           master     #586      +/-   ##
==========================================
+ Coverage   83.23%   83.27%   +0.03%     
==========================================
  Files         165      165              
  Lines       21508    21558      +50     
==========================================
+ Hits        17903    17952      +49     
- Misses       3605     3606       +1     

Impacted Files	Coverage Δ
include/mockturtle/networks/aig.hpp	93.24% <92.59%> (-0.15%)	:arrow_down:
...nclude/mockturtle/algorithms/dsd_decomposition.hpp	91.54% <0.00%> (+4.22%)	:arrow_up:

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.