Check why the bytecode lengths are changing in the "extra_file_input" bug tests

[kuzdogan]

During the PR #1879 we introduced a check on the bytecode lengths during verification ie. if bytecode lengths don't match we assume these contracts don't match.

However after implementing this the test "should detect extra file input bug when optimizer is enabled" started to fail because apparently the bytecode lengths are different for the onchain contract vs. recompiled with extra files.

To my understanding this shouldn't have happened because the bug is only affecting the AST IDs, and the bytecode lengths should've stayed the same. We should investigate what the real issue is or if our understanding was wrong.