Panic in tango benchmarks when running a comparison

When running A-B benchmark comparison, tango panics with undefined tango_init symbol:

> cargo export target/benchmarks/main -- bench

> RUST_BACKTRACE=1 target/benchmarks/main/benchmarks compare target/benchmarks/main/benchmarks -o -p -s 20 --warmup true
thread 'main' panicked at /home/kubouch/.cargo/registry/src/index.crates.io-6f17d22bba15001f/tango-bench-0.6.0/src/dylib.rs:62:60:
called `Result::unwrap()` on an `Err` value: UnableToLoadSymbol(DlSym { desc: "/home/kubouch/git/forks/new-nu-parser/target/benchmarks/main/benchmarks.patched: undefined symbol: tango_init" })
stack backtrace:
   0: rust_begin_unwind
   1: core::panicking::panic_fmt
   2: core::result::unwrap_failed
   3: tango_bench::dylib::Spi::for_library
   4: tango_bench::cli::paired_test::run_test
   5: tango_bench::cli::run
   6: benchmarks::main
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.

Reproduce the comparison with the cargo export and benchmark commands in the report. Start at tango-bench-0.6.0/src/dylib.rs:62 and the tango_bench::cli::paired_test::run_test entry point, then trace why the patched benchmark lacks tango_init. Done means the A-B comparison completes without the reported panic.