fix(llminternal): pair every function call with a result in request contents

[QuentinBisson]

Problem

A function call and its result are two separate session events. When a turn ends between them (a restart, an OOM kill, a client disconnect, a cancellation), the session keeps a function call that no function response answers.

That history is replayed on every later turn, so the session does not recover on its own, and a provider that requires strict pairing rejects the whole conversation. The OpenAI Responses path in this repo sends the dangling call as a function_call item with no function_call_output, which the API rejects. Anthropic, used through a custom model, answers tool_use ids were found without tool_result blocks immediately after. Either way the session stays unusable until it is deleted.

No issue exists for the Go side. The same failure is reported against adk-python in google/adk-python#5856, #3971 and #6582.

Change

Pair every function call with a result while the request contents are assembled, in buildContentsDefault. A call the immediately following content does not answer gets a placeholder result there.

• The stored events are untouched, so recorded history stays intact and a session that is already broken heals on its next turn with no migration.
• It sits above the session service, so one implementation covers every store and every provider.
• Ids are read after the conversion, so a call whose client id was stripped is answered by a response with the same stripped id.
• Pairing is checked positionally, against the immediately following content, because that is the invariant the provider enforces.
• On a turn that is already partly answered, the placeholder joins the leading run of responses. Anthropic requires the results to precede any other block in the message, so appending after a trailing text part would be rejected for the same reason the missing result was.

A call the framework holds open (a long-running tool, an approval, a request for user input) is described as awaiting a response, not as having returned nothing. The difference changes what the model does next: told a tool returned nothing, it reissues the call or proceeds without it; told the call is still awaiting a response, it can wait. Those ids come from LongRunningToolIDs, which lives on the event and does not survive the conversion to contents.

A conversation whose calls are all answered is returned unchanged.

The mirror change for adk-python is google/adk-python#6914.

Testing plan

• go test ./... -count=1 — 158 packages, all pass.
• go build ./..., go vet ./internal/llminternal/, gofmt -l.

internal/llminternal/tool_pairing_test.go drives the request processor over eight histories: the interrupted turn at the end of the history and in the middle, the partially answered parallel turn, the same turn with a trailing text part, the call held open in both directions, calls without an id, and a healthy history that must come back unchanged. Each case also asserts the provider invariant, that the content after a call answers every call it carries; a guard test asserts the same check rejects unrepaired input.

Two TestContentsRequestProcessor_IncludeContents/agentTransfer fixtures end on an unanswered call. They now expect the placeholder result instead of the synthetic continuation turn.