Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦
C
tests: the failed-STOP test waits for the STOP, not for 0.15 s
JustVuggf6665f334b2d
diff --git a/c/tests/test_openai_server.py b/c/tests/test_openai_server.pyindex 89978702..67d03449 100644--- a/c/tests/test_openai_server.py+++ b/c/tests/test_openai_server.py@@ -6750,6 +6750,8 @@ class DispatcherLogprobTailTest(unittest.TestCase):# is answered with a named error rather than having its connection dropped, which# is what an unhandled write error does. Written against this tree's own STOP write# so the pin survives the composition with the checked writer.+ stop_tried = threading.Event()+class DeadOnStop:"""The engine's stdin: accepts the SUBMIT, refuses the STOP."""def __init__(self, process):@@ -6761,6 +6763,7 @@ class DispatcherLogprobTailTest(unittest.TestCase):# the same normalisation `FakeProcess.write` above does.data = bytes(data)if data.startswith(b"STOP"):+ stop_tried.set()raise BrokenPipeError(32, "Broken pipe")return self.process.real_write(data)@@ -6772,7 +6775,12 @@ class DispatcherLogprobTailTest(unittest.TestCase):process.stdout.feed(b"DATA " + request_id + b" 2 -0.5 1 3\nok\n")def terminal():- time.sleep(0.15)+ # DONE only once the STOP was tried, as `_faulted_turn` does. A fixed+ # 0.15 s let a slow runner read DATA and DONE back to back before the+ # idle poll wrote the STOP (macOS CI on dev 1f9a8523): the turn then+ # ended on the recorded fault and no STOP was ever written. The timeout+ # still ends the turn on a build that never writes it, which fails below.+ stop_tried.wait(1.0)process.stdout.feed(b"DONE " + request_id + b" STAT 1 2.5 0 1.0 5 0\n")