8000 test: keep the scaling schedule and setup assistant tests green on main by adeelahmad · Pull Request #648 · adeelahmad/lense · GitHub
[go: up one dir, main page]

Skip to content

test: keep the scaling schedule and setup assistant tests green on main - #648

Merged
adeelahmad merged 1 commit into
mainfrom
claude/project-thread-od4r1g
Oct 18, 2026
Merged

adeelahmad merged 1 commit into
mainfrom
claude/project-thread-od4r1g

Conversation

@adeelahmad
Copy link
Copy Markdown
Owner

Description

Before: main's backend CI is red. Two tests fail in all four FastAPI jobs: test_rules_scale_on_waiting_jobs_and_on_a_schedule and test_a_setup_conversation_sets_the_server_up. That blocks the next release.

After: both pass. Only tests change.

Motivation and Context

  • Scaling schedule test. The test's clock is fixed at 2026-10-18 12:00. The group it creates is stamped with the real time, and a schedule rule's next fire is counted from that stamp. Once the real time passed 18:00 UTC on 2026-10-18, the next 0 18 * * * fire moved to tomorrow, and the 18:01 tick stopped firing. The test now stamps the group with its own clock. It started failing on main at 18:00 UTC today, after e6e96ba had passed.
  • Setup conversation test. The test parses server_status's whole JSON from what the model was sent. The new budgets tool (feat(budgets): the assistant and agents answer where budgets stand #645) and circles tool (feat(circles): circles of trust and sharing with them, with API and MCP #626) made the setup tool list long enough that, at the default 16k window, chat.fit cuts that result in the middle, so it no longer parses. The test now gives the model a 128k default window, so every result fits. Cutting results to fit a small window is intended and unchanged.

Both failures reproduce locally on main and pass with this change (pytest tests/api/test_scaling.py tests/api/test_setup_assistant.py: 19 passed).

Types of changes

  • Bug fix (non-breaking change which fixes an issue)

Checklist:

  • The title is a Conventional Commit.
  • My code follows the code style of this project (ruff check and format pass).
  • I added or updated tests.
  • Docs: n/a, test-only.
  • CHANGELOG: n/a, test-only.

Done questions (AGENTS.md)

  • Goal: main green again so the release can be cut; merges on green.
  • n/a for the rest: no product change.

Generated by Claude Code

The scaling schedule test fixed its own clock at 2026-10-18 12:00 but the
group was stamped with the real time, so once real time passed 18:00 on
that day the schedule's next fire moved to tomorrow and the test failed.
It now stamps the group with the test's clock.

The setup conversation test read server_status's whole JSON from what the
model was sent. The budgets and circles tools made the setup tool list long
enough that, at the default 16k window, chat.fit cut that result in the
middle. The test now gives the model room for every result.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01291QUFHs51wLhGQwp2DAh9
@adeelahmad
adeelahmad merged commit 93fb574 into main Oct 18, 2026
10 checks passed
@adeelahmad
adeelahmad deleted the claude/project-thread-od4r1g branch October 18, 2026 19:55
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.

2 participants

0