@memo
def something():
return [i for i in range(100000)]
and the server returned a 400 Bad Request. It does this when it unexpectedly reaches the end of the payload while still believing that it's reading through the JSON metadata at the top of the request (i.e. before the BLOB payload).
The bug is either:
In the client: somehow incorrectly encoding the metadata header. I don't know how this part works at the moment, but if we have a very large result (as in the example) are we creating a shorter __repr__ style object for result_json purposes? I think maybe the problem is that the sentinel value at the start of the body is incorrectly set.
In the server: my implementation of the decoder might just be wrong somewhere.
I ran this code:
and the server returned a 400 Bad Request. It does this when it unexpectedly reaches the end of the payload while still believing that it's reading through the JSON metadata at the top of the request (i.e. before the BLOB payload).
The bug is either:
__repr__style object forresult_jsonpurposes? I think maybe the problem is that the sentinel value at the start of the body is incorrectly set.