From ee6f469aa51e2b1045536bdd2fae068c6d2a4ce2 Mon Sep 17 00:00:00 2001 From: Georgios Roros Date: Sun, 1 Mar 2026 19:28:26 -0500 Subject: [PATCH] Fix flaky timing test on Windows CI MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Increase performance test threshold from 100ms to 1.0s. Windows CI runners are slower than Linux, and the 100ms threshold was too tight. The test's intent is to catch obvious regressions (e.g., accidentally re-introducing code generation), not to precisely benchmark. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 --- tests/server/providers/openapi/test_comprehensive.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/tests/server/providers/openapi/test_comprehensive.py b/tests/server/providers/openapi/test_comprehensive.py index a21a764db..5396767c2 100644 --- a/tests/server/providers/openapi/test_comprehensive.py +++ b/tests/server/providers/openapi/test_comprehensive.py @@ -731,9 +731,11 @@ class TestOpenAPIComprehensive: end_time = time.time() - # Should be very fast (no code generation) + # Should be fast (no code generation). Use generous threshold for slow + # CI runners (Windows); the intent is to catch obvious regressions + # (e.g., accidentally re-introducing code generation), not to benchmark. initialization_time = end_time - start_time - assert initialization_time < 0.1 # Should be under 100ms + assert initialization_time < 1.0 # Verify provider was created correctly assert provider is not None