Problem

The test_external_crash_reporter_http test was flaky when running under kcov code coverage analysis. The issue is that kcov exits with code 0 even when the process it's monitoring crashes with a signal like SIGSEGV:

kcov: Process exited with signal 11 (SIGSEGV) at 0x7ffff7729c5c
# exit code: 0

This caused the assertion assert child.returncode to fail because 0 is falsy in Python, making the test fail intermittently when running in the kcov CI job.

Solution

Skip the test_external_crash_reporter_http test when running under kcov using the @pytest.mark.skipif decorator. The fix uses the existing is_kcov condition variable from tests/conditions.py:

@pytest.mark.skipif(is_kcov, reason="kcov exits with 0 even when the process crashes")
def test_external_crash_reporter_http(cmake, httpserver, build_args):
    ...

This approach is consistent with how other analyzer-specific behaviors are handled in the codebase (e.g., has_breakpad and has_crashpad already skip tests entirely when kcov is running).

Tests Updated

Only test_external_crash_reporter_http is skipped when running with kcov. Other inproc crash tests work correctly with kcov and do not need to be skipped.

Note: Breakpad and Crashpad tests are already completely skipped when kcov is running via the has_breakpad and has_crashpad conditions, so they don't require updates.

Fixes https://github.com/getsentry/sentry-native/actions/runs/18502219462/attempts/1 and https://github.com/getsentry/sentry-native/actions/runs/18594256344/job/53016350698

cc @jpnurmi
Fixes #1418

Fixes #1418

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.