Fix segfault in LexiRoute lookahead on classical vertices

[cleitonaugusto]

Description

MappingFrontier::advance_next_2qb_slice walks the vertices of the next cut and,
for any vertex that does not have more than one quantum in-edge, reads
in_edges[0]. A ClExprOp is purely classical and has no quantum in-edges, so
that read goes past the end of an empty vector. It only crashes some of the time,
which matches the reports in the issue.

Building the original file with -D_GLIBCXX_ASSERTIONS makes it abort at
std::vector::operator[](__n=0) called from MappingFrontier.cpp:265.

The fix skips vertices that don't have exactly one quantum in-edge. A classical
vertex has no qubit boundary to move, and this function is only used for the
lookahead when LexiRoute scores swaps (the boundary is restored afterwards), so
skipping it doesn't change what gets routed.

Tests:

• New C++ scenario in test_LexiRoute.cpp and a pytket test in mapping_test.py
  with the circuit from the issue. The pytket test segfaults without the fix.
• Full pytket suite passes locally.
• I also routed 1000 random circuits with ClExprOps and conditionals on a line
  architecture and checked that each qubit and bit sees the same operations in
  the same order after undoing the swaps. No differences.

Bumped TKET_VERSION to 2.1.102.

Related issues

Fixes #2180.

Checklist

• I have performed a self-review of my code.
• I have commented hard-to-understand parts of my code.
• I have made corresponding changes to the public API documentation. (no API change)
• I have added tests that prove my fix is effective or that my feature works.
• I have updated the changelog with any user-facing changes.

[cqc-alec]

Thanks!