[python] Add mock API tests for XML scenarios from PR #10063 and fix XML serialization bugs (WIP)

[Copilot]

• Bump @typespec/http-specs to 0.1.0-alpha.35-dev.5
• Add 11 new XML scenario tests (sync + async)
• Fix unwrapped model array element naming
• Fix namespace key mismatch (DPG "namespace" vs runtime "ns")
• Fix default-namespace propagation semantics
• Fix Python ET reserved prefix (ns\d+)
• Remove all 4 @pytest.mark.skip markers — 26/26 tests pass
• Update changelog entry to reflect generator fixes and all-green tests
• Fix itemNs → itemsNs key mismatch in deserialization (red bug)
• Replace all or chains for namespace lookups with is None checks (handles empty-string namespaces correctly)
• Extract local variables for double-lookup patterns in _get_element/_get_wrapped_element
• Add note about ET._namespace_map silent fallback behavior
• Remove .venv_test from .gitignore (local dev leftover)

Original prompt

---

This section details on the original issue you should resolve

<issue_title>[python] add test case for #10063
</issue_title>
<issue_description>follow skill https://github.com/microsoft/typespec/blob/main/.github/skills/python-sdk-spector-mock-api-tests/SKILL.md to write test case for #10063
</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes [python] add test case for https://github.com/microsoft/typespec/pull/10063  #10097

---

🔒 GitHub Advanced Security automatically protects Copilot coding agent pull requests. You can protect all pull requests by enabling Advanced Security for your repositories. Learn more about Advanced Security.

[pkg-pr-new]

Open in StackBlitz

npm i https://pkg.pr.new/@typespec/http-client-python@10098

commit: 90b7ab0

[github-actions]

All changed packages have been documented.

• ✅ @typespec/http-client-python

Show changes

@typespec/http-client-python - internal ✏️

Add mock API test cases for XML scenarios introduced in #10063, covering: renamed property, nested model, renamed nested model, wrapped primitive with custom item names, model array variants (wrapped/unwrapped/renamed), renamed attribute, namespace, and namespace-on-properties. Tests for unwrapped model array serialization and namespace handling are skipped pending generator bug fixes.

[azure-sdk]

You can try these changes here

🛝 Playground	🌐 Website	🛝 VSCode Extension

[Copilot]

Pull request overview

Adds Python mock API coverage for new @typespec/http-specs XML scenarios and updates the Python generator’s XML runtime template to fix serialization/deserialization issues needed by those scenarios.

Changes:

• Bump @typespec/http-specs to 0.1.0-alpha.35-dev.5 to pick up new XML scenarios.
• Add new sync + async mock API tests covering renamed properties/attributes, nested models, array wrapping/unwrapping, and namespaces.
• Fix XML runtime template behaviors around namespace handling and element naming for unwrapped model arrays.

Reviewed changes

Copilot reviewed 5 out of 7 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
packages/http-client-python/package.json	Bumps @typespec/http-specs dependency version.
packages/http-client-python/package-lock.json	Locks updated @typespec/http-specs tarball + integrity.
packages/http-client-python/generator/test/generic_mock_api_tests/test_payload_xml.py	Adds new sync mock API XML scenarios.
packages/http-client-python/generator/test/generic_mock_api_tests/asynctests/test_payload_xml_async.py	Adds async equivalents of new XML scenarios.
packages/http-client-python/generator/pygen/codegen/templates/model_base.py.jinja2	Fixes XML namespace reading and array-item element naming; adds safe namespace registration fallback in element creation.
.gitignore	Ignores .venv_test.
.chronus/changes/python-xml-test-cases-2026-3-19-22-47-48.md	Adds changelog entry for the internal change.

Files not reviewed (1)

• packages/http-client-python/package-lock.json: Language not supported

[msyyc]

Code Review

Overall the tests are well-structured and consistent with existing patterns. The 4 bug fixes in model_base.py.jinja2 are well-reasoned. A few issues worth addressing:

🔴 Bug: itemNs vs itemsNs key mismatch (deserialization ↔ serialization)

File: model_base.py.jinja2 — deserialization path (~line 648) vs serialization path (~line 1330)

Serialization consistently uses plural itemsNs/itemsName/itemsPrefix, but the existing deserialization code reads itemNs (singular). This PR didn't introduce the bug but touches the surrounding code and would be a good opportunity to fix it. If a model declares itemsNs metadata, deserialization will silently ignore it.

🟡 Empty-string namespace treated as falsy in or chains

Multiple new lines use patterns like:

xml_ns = prop_meta.get("ns") or prop_meta.get("namespace")
parent_meta.get("itemsNs") or parent_meta.get("ns") or parent_meta.get("namespace")

If a key is explicitly set to "" (valid in XML — means "no namespace"), the or chain falls through to the next value instead of honoring the empty string. Safer pattern:

xml_ns = prop_meta.get("ns")
if xml_ns is None:
    xml_ns = prop_meta.get("namespace")

🟡 Private API ET._namespace_map in _create_xml_element

The try/except ValueError fallback uses getattr(ET, "_namespace_map", None) — a private CPython implementation detail. This works today but could break in future Python versions. If _namespace_map is absent, the prefix silently becomes auto-generated (ns0, ns1…) with no warning. Consider adding a log or comment about the silent fallback behavior.

Minor

• Changelog entry still says "skipped pending generator bug fixes" even though the bugs were subsequently fixed and all skips removed. Should be updated to reflect the current state.
• .gitignore addition of .venv_test — is this intentional or a leftover from local development?

[msyyc]

@copilot analyze the review comment #10098 (review).

[Copilot]

Code Review

Overall the tests are well-structured and consistent with existing patterns. The 4 bug fixes in model_base.py.jinja2 are well-reasoned. A few issues worth addressing:...

Addressed in e02d171:

• itemNs → itemsNs (🔴): Fixed the key mismatch in the deserialization path. Also restructured the unwrapped-array block so when itemsNs is absent the existing property namespace still applies to the items name.
• Empty-string namespace or chains (🟡): Replaced all prop_meta.get("ns") or prop_meta.get("namespace") patterns with is None checks (across all 8 affected sites in deserialization and serialization paths). Also extracted local variables to avoid double-lookups flagged in code review.
• ET._namespace_map silent fallback (🟡): Added a comment noting that if _namespace_map is absent (e.g. alternative Python implementations) ET will silently auto-generate a prefix.
• .venv_test in .gitignore: Removed — was a local development leftover.

@copilot analyze the review comment #10098 (review).

All four points from the review have been addressed in e02d171. See the reply on the review comment for details.

[Copilot]

Pull request overview

Copilot reviewed 5 out of 6 changed files in this pull request and generated no new comments.

Files not reviewed (1)

• packages/http-client-python/package-lock.json: Language not supported