mirrors/arti.git/tests/shadow, branch arti-v1.7.0

shadow: Update shadow syntax for prop330

2025-09-16T11:55:00Z

shadow CI: decrease onion service xfer sizes

2025-09-03T16:10:59Z

This partially reverts b223d504f6a033600c61e81a0b0a0eea3e43fd44 from arti!3047. Increasing the transfer size exacerbates the failure rate of the onion services too much. We should increase it again once arti#2109 is resolved.

shadow CI: less-verbose logging

2025-08-27T21:29:19Z

With the increased transfer sizes and simulation run-time, the previous log levels resulted in excessively large log files.

CI shadow test: increase transfer sizes to 5 MB

2025-08-27T21:29:19Z

This is the transfer size used in the tor CI chutney tests, to more-fully exercise SENDMEs / congestion control. We also increase the simulation run time to allow time for the larger transfers.

shadow.yaml: move arti build dir to a yaml anchor

2025-08-26T22:33:53Z

Integration testing: use quicktest build

2025-08-26T22:33:53Z

shadow CI: have arti processes log to file instead of console

2025-06-17T16:19:02Z

This is in preparation to switch arti's console logging to use stderr instead of stdout. The test script currently fails the test if there are non-empty stderr files, since that's where rust panics end up getting recorded. It will also make it easier to see if there is any other non-logging output. Also disable trace-level logging for the tor_proto module. It's unclear exactly when/why this was added, but it's probably overkill other than when debugging.

Merge branch 'debug-ci' into 'main'

2025-05-06T14:14:12Z

shadow CI tweaks See merge request tpo/core/arti!2989

shadow CI: enable model-unblocked-syscall-latency

2025-05-05T22:15:51Z

Not expected to solve any immediate issue, but helpful for consistency with the chutney-shadow CI, and I think in the past we've hit busy-loop deadlock without it if/when we try increasing the simulation end time. Similarly, we apply the same tuning of "max_unapplied_cpu_latency" as we use in the chutney-shadow simulation, which gets back some of the simulation stability that enabling model-unblocked-syscall-latency otherwise sometimes sacrifices.

shadow CI test: only require 1 successful HS transfer

2025-05-05T22:08:22Z

As per previous discussion there is some known flakiness around hidden services with TestingTorNetwork enabled: Previously, only requiring 5 of 10 transfers to succeed has been enough to keep this stable, but recently we've gotten unlucky and dipped under this threshold: More investigation is warranted, especially wrt why the problem seems to mostly be affecting articlient-onion-auth, but for now it's better to have a low success threshold than to have folks overriding the CI failure.