Skip to content

fix: concurrent-run MCP test flake, one to_zone, doctor $TZ warning, README job wording - #40

Merged
ggapp1 merged 4 commits into
mainfrom
fix/carry-items
Oct 29, 2026
Merged

ggapp1 merged 4 commits into
mainfrom
fix/carry-items

Conversation

@ggapp1

@ggapp1 ggapp1 commented Oct 29, 2026

Copy link
Copy Markdown
Owner

Four small items left over from the whole-project review and 6b-2b.

  • test(mcp): the reconnect tests found their fixture server with a machine-wide pgrep -f, so concurrent pytest runs SIGKILLed each other's servers (3 concurrent runs: 25 failures). Scoped to this process's children with pgrep -P; 3 concurrent runs now pass 50/50 each. Test-only.
  • refactor(config): to_zone() is the one place the naive-UTC → zone conversion is spelled; App.in_timezone and the dashboard's _in_zone both delegate. Behaviour unchanged; mutating it to bare .astimezone() fails three tests under TZ=Pacific/Auckland.
  • fix(doctor): when the zone comes from the host and $TZ is set, doctor warns (exit code still 0): the shipped systemd/launchd units do not set $TZ, so the service may resolve a different zone. Pinned-zone behaviour unchanged.
  • docs(readme): scheduled jobs may call the model, so they are no longer listed in the model-independent data plane.

Not addressed: _check_day_boundary cannot see an operator who pinned a zone different from their old per-app value; that needs a stored marker and is design work.

@ggapp1
ggapp1 merged commit a58cfa3 into main Oct 29, 2026
6 checks passed
@ggapp1
ggapp1 deleted the fix/carry-items branch October 29, 2026 18:08
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant