fix: normalize transport header extraction and status handling

[kvdhanush06]

Summary

Fixes robustness issues in the Python SDK default transport when handling different HTTP response implementations.

This change:

• adds fallback status resolution using .getcode() and .code
• normalizes response headers into canonical lower-case keys
• removes reliance on response context-manager support

The fix improves compatibility with urllib-style responses, mocked transports, and mixed-case HTTP headers while preserving existing SDK behavior.

Existing SDK response interfaces and downstream behavior remain backward compatible.

---

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation only
• CI / tooling / repo hygiene
• Security fix

---

Changes made

Transport status handling

Updated _default_transport() to resolve HTTP status codes using:

1. response.status
2. response.getcode()
3. response.code

This improves compatibility with:

• urllib
• mocked/test response objects
• alternative HTTP response implementations

---

Header normalization

Normalized all transport response headers into a lower-case dictionary:

headers = {k.lower(): v for k, v in raw_headers}

This removes duplicated casing checks such as:

• X-Request-Id
• x-request-id
• Retry-After
• retry-after

and ensures consistent downstream header access.

---

Response lifecycle handling

Removed direct reliance on:

with response:

The transport now:

• reads response bodies directly
• safely closes responses if .close() exists
• works with non-context-managed response implementations

The same compatibility and normalization logic was applied to urllib.error.HTTPError handling paths.

---

Verification

Validated using:

• reproduction with a fake response object exposing:
  • getcode()
  • getheaders()
  • no .status
  • no context manager support
• full Python test suite

Local verification commands:

python -m pytest

All tests passed locally.

---

Checklist

• My code follows the style of this project (TS strict, no any in public surface; Python type hints + docstrings)
• I have added tests for any new behavior
• All tests pass locally (pnpm test for TS, python -m pytest for Python)
• I have updated the relevant CHANGELOG.md files (typescript/ and/or python/)
• I have NOT bumped the SDK version (maintainers handle that at release time)
• I have NOT added any runtime dependencies (both SDKs are zero-dep by policy)
• I have NOT exposed any private RunInfra endpoint or internal hostname

---

Related issues

Closes #10

[Copilot]

Pull request overview

Improves the Python SDK’s default urllib transport robustness by supporting more HTTP response shapes and normalizing headers for consistent downstream access.

Changes:

• Add fallback HTTP status resolution via response.status, response.getcode(), and response.code.
• Normalize response headers into a lower-cased dictionary using .headers / .getheaders() extraction.
• Remove reliance on response context-manager support and close responses when possible.

Comments suppressed due to low confidence (1)

python/runinfra/init.py:465

• The urllib.error.HTTPError handling path reads the body but never closes the underlying file/response object. HTTPError typically supports .close(), so skipping it can leak sockets/file descriptors on repeated errors. Mirror the success-path cleanup by calling exc.close() (if available) in a finally after exc.read() (including the read-error branch).

            response_headers = _extract_headers(exc)
            try:
                body = exc.read()
            except Exception as read_exc:
                raise _TransportBodyReadError(

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.