Dicklesworthstone d816f5ba8b Harden source command timeouts ...

Run SSH and source-management child commands in their own process group and collect stdout/stderr through bounded reader threads. On timeout, cass now terminates the whole child group and waits for pipe drain with a second bound instead of leaving grandchildren alive to hold pipes open.

Apply the same bounded process cleanup semantics to remote rsync detection, source install, probe, and sync paths so source operations fail closed with useful timeout errors rather than hanging status or indexing workflows.

Extend the real-index performance harness timeout helper to kill process groups and cover shell-grandchild pipe holders, matching the production failure mode the release needs to guard against.

2026-06-07 13:00:45 -04:00

..

doctor_e2e_runner.rs

feat(doctor): doctor v2 — archive-first, candidate-based recovery model

2026-05-06 17:17:14 -04:00

doctor_fixture.rs

feat(doctor): doctor v2 — archive-first, candidate-based recovery model

2026-05-06 17:17:14 -04:00

e2e_log.rs

chore(repo): reorganize evidence trees + harden tracked-artifact hygiene

2026-05-06 12:33:33 -04:00

mod.rs

test(doctor): add scripted e2e artifact runner

2026-05-05 00:50:27 -04:00

search_asset_simulation.rs

skill-loop: derive simulation failure error

2026-04-25 15:01:40 -04:00

timeout.rs

Harden source command timeouts

2026-06-07 13:00:45 -04:00