test: fix flaky test_put_contract_three_hop_returns_response

[sanity]

Problem

test_put_contract_three_hop_returns_response fails intermittently in CI with timeout errors:

• Transaction timed out errors
• Connection already exists errors
• Times out after 120 seconds waiting for PUT response

Latest failure: https://github.com/freenet/freenet-core/actions/runs/19486370415/job/55769443837

Root Cause

The test has a known issue documented in the code itself (crates/core/tests/operations.rs:1938-1941):

// Note: We cannot modify node locations after they're created with the macro,
// so this test will use random locations. The original test had specific location
// requirements to ensure proper three-hop routing. For now, we'll proceed with
// the test and it should still validate PUT response routing.

The test requires:

• Specific node topology for guaranteed three-hop routing (gateway → peer-a → intermediate → peer-c)
• Currently uses random locations which sometimes don't create the required topology
• When topology is wrong, messages don't route correctly and test times out

Impact

• Causes false-positive CI failures on unrelated PRs
• Requires manual re-runs, slowing down development
• Wastes CI resources (10+ minute timeouts)
• Creates uncertainty about actual failures vs flakes

Solution

Make the test deterministic:

1. Use fixed locations for nodes to guarantee three-hop path:

   • Gateway at location 0.1
   • Peer A at location 0.3
   • Intermediate peer at location 0.5
   • Peer C at location 0.7

2. Verify topology before running the test (assert connection graph matches expectations)

3. Replace timeout waits with condition-based polling (related to Replace fixed timing waits in tests with condition-based polling #1943)

4. Add diagnostic logging when test fails to help debug future issues

Related Issues

• Replace fixed timing waits in tests with condition-based polling #1943 - Replace fixed timing waits in tests with condition-based polling
• test: fix all ignored integration tests #2021 - Fix all ignored integration tests

Test Location

crates/core/tests/operations.rs:1922

[github-actions]

🤖 Auto-labeled

Applied labels:

• T-bug (97% confidence)
• P-high (90% confidence)
• E-medium (80% confidence)
• A-developer-xp (95% confidence)
• A-networking (85% confidence)

Reasoning: The report describes a flaky integration test that intermittently times out due to non-deterministic node locations causing incorrect topology and routing; this is a defect (T-bug) that causes CI false-positives and wasted resources, so it warrants high priority. The proposed fix is to make the test deterministic (fixed locations, topology checks, condition-based polling, diagnostics) which is moderate in complexity (E-medium). The issue affects developer experience and CI/testing (A-developer-xp) and is directly related to networking/topology behavior (A-networking). Confidence is high because the issue content clearly identifies root cause and remediation steps.

Previous labels: none

If these labels are incorrect, please update them. This helps improve the auto-labeling system.