From da0e192d7cb7a92df35335c41ec46a7b6d667443 Mon Sep 17 00:00:00 2001
From: BREW <81044362+b-re-w@users.noreply.github.com>
Date: Sat, 3 Oct 2026 19:29:26 +0900
Subject: [PATCH] Test: Do not accept an engine error as a streamed reply

test_streams_tokens_then_eos only checked that some text arrived before
<EOS>, so it passed on mps where every engine call failed: the model
layer turns the failure into an 'ERROR:' message frame. It now rejects
that.
---
 api/tests/test_server.py | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/api/tests/test_server.py b/api/tests/test_server.py
index 60202591..08f9ffe1 100644
--- a/api/tests/test_server.py
+++ b/api/tests/test_server.py
@@ -61,8 +61,10 @@ def test_streams_tokens_then_eos(client):
             break
     ws.__exit__(None, None, None)
 
+    reply = "".join(frames[:-1])
     assert frames[-1] == "<EOS>"
-    assert "".join(frames[:-1]).strip()
+    assert reply.strip()
+    assert "ERROR:" not in reply, reply  # an engine failure arrives as an ERROR message, not as a reply
 
 
 def test_disconnect_mid_stream_stops_generation_and_frees_the_model(client, engine, monkeypatch):