TypeError: 'ResponseBody' object is not subscriptable

[guyskk]

I use db.save(doc), it cause error sometimes:

File "/usr/lib/python3.5/site-packages/couchdb/client.py", line 440, in save
    id, rev = data['id'], data.get('rev')
TypeError: 'ResponseBody' object is not subscriptable

and I try to find the reason, in http.py:

369             # Buffer small non-JSON response bodies
370             elif int(resp.getheader('content-length', sys.maxsize)) < CHUNK_SIZE:
371                 data = resp.read()
372                 self.connection_pool.release(url, conn)
373     
374             # For large or chunked response bodies, do not buffer the full body,
375             # and instead return a minimal file-like object
376             else:
377 B->             data = ResponseBody(resp, self.connection_pool, url, conn)
378                 streamed = True

when resp's content-length > CHUNK_SIZE, it will cause error.

[djc]

This needs a bit more detail. I don't think library code is passing a ResponseBody straight into db.save(). Can you create a reduced test case from your code that demonstrates the problem?

[guyskk]

It seems caused by unicode _id, not caused by large data:

def test_save_unicode(self):
    doc = {'_id': '关闭'}
    id, rev = self.db.save(doc)
    doc2 = self.db.get(id)
    id2, rev2 = self.db.save(doc2)
    self.assertTrue(doc['_id'] == id)
    self.assertTrue(doc['_id'] == id2)

self.db.save(doc2) will raise exception, and I just only find {'_id': '关闭'} will cause exception.

>>> '关闭'.encode('utf-8')
b'\xe5\x85\xb3\xe9\x97\xad'
>>>

[guyskk]

it's not really possible for us to resolve the problem
https://github.com/kennethreitz/requests/issues/3098

[djc]

Thanks for working that out. I've filed https://issues.apache.org/jira/browse/COUCHDB-3003 and will close this bug for now.