test: keep the developer's robot out of the suite

[KrishnaH96]

Problem

GlobalConfig reads .env itself, so a ROBOT_IP in the checkout makes every zenoh session dial the robot and stall 1 s; 5 async-module tests fail and the suite runs 3x slower.

Solution

conftest blanks ROBOT_IP/ROBOT_IPS/ZENOH_CONNECT before dimos imports; the dotenv precedence test drops them too. Popping is not enough because pydantic-settings reads the file directly.

How to test

echo ROBOT_IP=10.0.0.1 >> .env
uv run pytest dimos/core/test_async_module_rpc.py dimos/core/coordination/blueprint_config/test_sources.py

19 passed across the six affected files with ROBOT_IP set; 5 failed before.

AI assistance

Used Fable 5.1 extensively for root cause analysis, fix and testing.

Checklist

• I have read and approved the CLA.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

@@            Coverage Diff             @@
##             main    #4103      +/-   ##
==========================================
- Coverage   78.59%   78.59%   -0.01%     
==========================================
  Files        1460     1460              
  Lines      138230   138234       +4     
  Branches    11881    11882       +1     
==========================================
- Hits       108642   108640       -2     
- Misses      26244    26247       +3     
- Partials     3344     3347       +3     

Components	Coverage Δ
Tests	94.59% <100.00%> (-0.01%)	⬇️

Flag	Coverage Δ
OS-ubuntu-24.04-arm	74.77% <100.00%> (-0.01%)	⬇️
OS-ubuntu-latest	75.48% <100.00%> (+<0.01%)	⬆️
Py-3.10	75.25% <100.00%> (+<0.01%)	⬆️
Py-3.11	75.46% <100.00%> (+<0.01%)	⬆️
Py-3.12	75.46% <100.00%> (-0.01%)	⬇️
SelfHosted-Large	30.77% <50.00%> (-0.01%)	⬇️
SelfHosted-Linux	35.68% <50.00%> (-0.01%)	⬇️
SelfHosted-macOS	35.02% <50.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
dimos/conftest.py	81.81% <100.00%> (+0.20%)	⬆️
...core/coordination/blueprint_config/test_sources.py	100.00% <100.00%> (ø)

... and 3 files with indirect coverage changes

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[paul-nechifor]

I think we should ignore everything in .env. I'm checking how it can be done in pytest.

[paul-nechifor]

Added alternative here: #4111

[greptile-apps]

Confidence Score: 5/5

Safe to merge: the changed test bootstrap and dotenv precedence behavior work with inherited local configuration values present.

What we checked:

• Ran a focused pytest probe to validate that inherited environment variables are not used by the connection configuration and that dotenv precedence remains authoritative, as described in the contract for the dotenv-related tests.
• Validated that the generated probe imports the global_config and confirms emptiness for the three connection variables in both process environment and captured configuration, with the existing test passing despite inherited values and the change protection verified in the specified files.

Summary

• Test startup clears inherited robot and Zenoh connection settings before DIMOS configuration is imported.
• The dotenv-precedence test clears inherited values so its temporary .env file remains authoritative.
• Focused before-and-after checks confirmed the changed paths are isolated from local configuration.

_{Reviews (1) · Last reviewed commit: "test: keep the developer's robot out of ..."}

[paul-nechifor]

Added an alternative here: #4111