Python: fix: preserve function-call name when merging streaming deltas

[Osamaali313]

Problem

Content._add_function_call_content (the + operator for function_call content, used by ChatResponse.from_updates / AgentResponse.from_updates to merge streaming deltas) builds the merged function name like this:

name=getattr(self, "name", getattr(other, "name", None)),

Content.__init__ always assigns self.name (defaulting to None), so the attribute is never missing. getattr(self, "name", <fallback>) therefore always returns self.name — even when it is None — and the <fallback> (other.name) is never consulted.

As a result, when two function_call contents are merged and only the second carries the name, the name is silently dropped. This happens for any streaming path where the function name arrives in a later delta than the first chunk:

a = Content.from_function_call(call_id="1", name=None, arguments='{"a":')
b = Content.from_function_call(call_id="1", name="get_weather", arguments="1}")

(a + b).name   # -> None          ❌  (name dropped)
(b + a).name   # -> "get_weather" ✅  (works only when self has the name)

The asymmetry confirms it's an oversight: the sibling fields in the same return use the correct "either side" pattern — exception=getattr(self, "exception", None) or getattr(other, "exception", None) on the very next line, and _add_text_reasoning_content uses self.id or other.id.

Fix

Mirror the adjacent exception field:

name=getattr(self, "name", None) or getattr(other, "name", None),

After the fix, the name is preserved regardless of which side carries it, both-None stays None, and the normal both-set case is unchanged (self wins).

Tests

Extended test_function_call_content_add_merging_and_errors to cover the late-name (a + b and b + a) and both-None cases. Verified RED→GREEN: the new assertions fail on main and pass with the fix; the full tests/core/test_types.py suite stays green (267 passed).

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[github-actions]

Python Test Coverage Report •

File	Stmts	Miss	Cover	Missing
packages/core/agent_framework
_types.py	1226	97	92%	61, 70–71, 125, 130, 149, 151, 155, 159, 161, 163, 165, 183, 187, 213, 235, 240, 245, 249, 279, 705–706, 887–888, 1346, 1421, 1456, 1476, 1486, 1538, 1672–1674, 1959–1964, 1989, 2044, 2049, 2059, 2067, 2074–2078, 2096, 2169, 2182, 2187, 2300, 2323, 2584, 2608, 2707, 2888–2889, 2991, 3220, 3273, 3292, 3331, 3342, 3344–3348, 3350, 3353–3361, 3371, 3460, 3597, 3602, 3607, 3612, 3616, 3702–3704, 3733, 3821–3825
TOTAL	44153	5265	88%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
8856	33 💤	0 ❌	0 🔥	2m 19s ⏱️

[Copilot]

Pull request overview

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

[Osamaali313]

Good catch — fixed in d82c14f. I went with the second option you offered (construct the nameless delta via Content("function_call", ...) rather than the factory), so the public from_function_call(name: str) signature stays strict and the pyright reportArgumentType gate clears. The factory just wraps that same constructor, so the merged objects and assertions are unchanged (test still passes).

I chose this over widening from_function_call to name: str | None because a nameless call is really a streaming-delta intermediate rather than something a caller should mint through the public factory — and production already treats name as optionally-absent on that path (e.g. the content.name guards in _compaction.py / _evaluation.py). But if you'd prefer to make nameless deltas first-class and widen the factory signature instead, I'm happy to switch — just say the word.