I think the Tor Browser nightly includes Tor master, i.e. this was a pre-release Tor 0.4.3.0-alpha-dev. We should figure out which git commit it was from.
Or if somebody can reproduce it on a recent nightly and tell us which one, then we can rederive the version too.
Managed to reproduce it once, on Linux, with just the tor binary, under valgrind. I started tor, launched an apt update which went to the debian onion service, and ^C'ed at a lucky time.
^CJan 11 08:15:32.000 [notice] Interrupt: exiting cleanly.Jan 11 08:15:32.000 [err] tor_assertion_failed_(): Bug: src/feature/rend/rendcache.c:529: rend_cache_lookup_entry: Assertion rend_cache failed; aborting. (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: Tor 0.4.3.0-alpha-dev (git-5888db496746e2fe): Assertion rend_cache failed in rend_cache_lookup_entry at src/feature/rend/rendcache.c:529: . Stack trace: (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(log_backtrace_impl+0x56) [0x31f386] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(tor_assertion_failed_+0x147) [0x31a3e7] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(rend_cache_lookup_entry+0x112) [0x25acb2] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(rend_client_report_intro_point_failure+0x45) [0x25db05] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(hs_circ_cleanup_on_free+0x94) [0x211f84] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(+0x8a672) [0x192672] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(circuit_free_all+0xca) [0x19307a] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(tor_free_all+0x51) [0x169ab1] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(tor_run_main+0x1125) [0x168725] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(tor_main+0x3a) [0x165fba] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(main+0x19) [0x165b79] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xeb) [0x521b09b] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)Jan 11 08:15:32.000 [err] Bug: src/app/tor(_start+0x2a) [0x165bca] (on Tor 0.4.3.0-alpha-dev 5888db496746e2fe)
==4027== Process terminating with default action of signal 6 (SIGABRT)==4027== at 0x522E7BB: raise (raise.c:51)==4027== by 0x5219534: abort (abort.c:79)==4027== by 0x31FA5D: tor_raw_abort_ (torerr.c:227)==4027== by 0x31A68D: tor_abort_ (util_bug.c:176)==4027== by 0x25ACB6: rend_cache_lookup_entry (rendcache.c:529)==4027== by 0x25DB04: rend_client_report_intro_point_failure (rendclient.c:803)==4027== by 0x211F83: cleanup_on_free_client_circ (hs_circuit.c:631)==4027== by 0x211F83: hs_circ_cleanup_on_free (hs_circuit.c:1274)==4027== by 0x192671: circuit_free_.part.19 (circuitlist.c:1139)==4027== by 0x193079: circuit_free_ (circuitlist.c:1127)==4027== by 0x193079: circuit_free_all (circuitlist.c:1312)==4027== by 0x169AB0: tor_free_all (shutdown.c:123)==4027== by 0x168724: tor_run_main (main.c:1318)==4027== by 0x165FB9: tor_main (tor_api.c:164)
There is definitely some sort of race, where you have to ^C it at just the right time.
Not only macOS-related as visible on Linux, too (tried with 5888db49).
Trac: Summary: Tor 0.4.3.0-alpha-dev asserts when quitting Tor Browser nightly on macOS to Tor 0.4.3.0-alpha-dev asserts when quitting Tor Browser nightly
Trac: Summary: Tor 0.4.3.0-alpha-dev asserts when quitting Tor Browser nightly to Tor 0.4.3.0-alpha-dev asserts when quitting while visiting an onion service
and commit 58879477 (a few commits behind the buggy one) doesn't obviously have the bug. It's hard to claim that for sure though because it's a race that's hard to trigger. So, dgoulet, that's where to start looking. :)