cargo tarpaulin takes forever in the "Mapping coverage data to source" steps

[zacchiro]

$ time cargo tarpaulin --engine llvm &> ~/webgraph-cargo-tarpaulin.log
cargo tarpaulin --engine llvm &> ~/webgraph-cargo-tarpaulin.log  6972,49s user 64,53s system 106% cpu 1:50:12,05 total

Almost 2 hours, which is insane.
Looking at the log, it stands out how the "Mapping coverage data to source" of each test takes 6 minutes (on my laptop):

$ grep -A 1 Mapping webgraph-cargo-tarpaulin.log
Sep 17 17:09:09.499  INFO cargo_tarpaulin::statemachine::instrumented: Mapping coverage data to source
Sep 17 17:15:07.938  INFO cargo_tarpaulin::process_handling: running /srv/src/rust/webgraph-rs/webgraph-rs/target/debug/deps/test_read_webgraph-1d8b146a355a8791
--
Sep 17 17:15:08.153  INFO cargo_tarpaulin::statemachine::instrumented: Mapping coverage data to source
Sep 17 17:21:06.756  INFO cargo_tarpaulin::process_handling: running /srv/src/rust/webgraph-rs/webgraph-rs/target/debug/deps/test_bvcomp-fc6ba07461d5faa0
--
Sep 17 17:21:06.804  INFO cargo_tarpaulin::statemachine::instrumented: Mapping coverage data to source
Sep 17 17:27:06.529  INFO cargo_tarpaulin::process_handling: running /srv/src/rust/webgraph-rs/webgraph-rs/target/debug/deps/ascii_convert-8f34e854de046580

This looks like an upstream issue in tarpaulin, I'm going to report it there.

PS given it took two hours to obtain, the very least I can do is report here for posterity the current code coverage :-)

46.80% coverage, 1287/2750 lines covered