summaryrefslogtreecommitdiffstats
path: root/packages/meshbay-node/tests/test_hwaccel.py
diff options
context:
space:
mode:
authorChristophe Besson <cbesson@gmail.com>2026-09-19 08:54:52 +0200
committerChristophe Besson <cbesson@gmail.com>2026-09-19 08:54:52 +0200
commit933daccbcfde7705413d3a10db87d910c650ed42 (patch)
treed9b9372868f8f71c4fcb9d6e85600977fcfc0df6 /packages/meshbay-node/tests/test_hwaccel.py
parent765214c22e7956f9add1501c69e00c918f3e5f2b (diff)
downloadmeshbay-933daccbcfde7705413d3a10db87d910c650ed42.tar.gz
fix(node): node.toml's transfer pools reach the transport
The daemon built the defaults dict for `Roster.node_settings` by hand and left out `max_concurrent_downloads` and `max_concurrent_uploads`. Absent from the dict, both resolved to None, were assigned back onto the config, and the transport skipped them — so node.toml was parsed, validated, and then replaced by `transfers.py`'s own 8. Invisible to anyone who left the value at 8, which is the value the template suggests. There were three copies of that dict and they all disagreed: node_status' was missing those two and `max_upload_gb` besides. One builder now, `config.node_settings_defaults`, and the resolver's key list is a class attribute the tests hold it to, along with the writer's. 1429 passed against a baseline of 1423; the two new behavioural tests fail with the builder reverted. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Diffstat (limited to 'packages/meshbay-node/tests/test_hwaccel.py')
0 files changed, 0 insertions, 0 deletions