fix(providers): review round 1 — streaming parity gaps the collapse exposed

Correctness (4 confirmed + 1 plausible fixed, 2 accepted+documented):

- Anthropic _iter_anthropic_stream handles citations_delta: text-block
  citations now ride the raw block into provider_blocks, as replay
  requires (the retired non-streaming lane preserved them via
  model_dump; the streaming lane dropped them — a pre-existing main-loop
  gap the collapse would have extended to single-shot lanes).
- Anthropic text blocks separate with "\n" at each subsequent block
  start, restoring the retired lane's "\n".join rendering on drained
  lanes AND un-fusing streamed web-search responses in the chat loop.
- response.failed raises typed ResponsesStreamFailedError, listed in the
  provider's retryable_error_names — retry loops treat an in-band
  failure like the wire errors it stands in for instead of
  hard-stopping on a bare RuntimeError (judges keep their heuristic
  fallback after retries).
- drain_stream folds a finish-less stream's terminal citations footer
  (suffix rule: pre-finish info invalidated by any later payload), so
  lax compat servers that never send finish_reason keep their Sources.
- usage max-merge extracted as merge_usage() in _protocol.py — the one
  definition drain uses now and the session's inline consumer adopts on
  #832.

Accepted + release-noted instead of coded around: strict pre-2024
compat servers that 400 on stream_options (such a server already cannot
serve the chat loop; CHANGELOG caveat extended), and repeated-index
parallel tool-call merging on legacy compat servers (identical to the
main loop's accumulator semantics; a shared guard belongs in the #832
unification).

Cleanup: run_with_deadline grows on_abandon (best-effort, cannot mask
the deadline error) and both judges drop the copy-pasted abort
choreography; StreamAbortRef documents the _CancelRef adoption plan;
test_model_turn's fake replays through the shared as_stream adapter;
docs/architecture.md drops the retired Protocol row.

Tests: refusal handler pinned (was advertised, untested); typed-failed
retryability; citations capture; text-block separator (plus the mixed
text+search expectation updated for the separator chunk); finish-less
citation fold; on_abandon firing matrix; StreamAbortRef arrival race.

docs/architecture.md

@@ -609,8 +609,7 @@ LLMProvider (protocol)
 
 | Method | Purpose |
 |--------|---------|
-| `create_streaming()` | Streaming request, yields normalized `StreamChunk` objects |
-| `create_completion()` | Non-streaming request, returns `CompletionResult` |
+| `create_streaming()` | The one transport: streaming request, yields normalized `StreamChunk` objects (single-shot callers accumulate via `drain_stream()` into a `CompletionResult`) |
 | `get_capabilities()` | Per-model flags (`ModelCapabilities`) |
 | `convert_tools()` | Translate OpenAI tool schemas to provider format |
 | `retryable_error_names` | Exception class names that trigger retry |