bbrf listen doesn't handle large number of changes well

[0xJeti]

Hi @honoki,

Quite often I'm getting following error when bbrf listen is running and there is some bigger change performed (multiple domains involved etc.):

{"seq":"279880-g1AAAACheJzLYWBgYMpgTmEQTM4vTc5ISXIwNDLXMwBCwxyQVCJDUv3___-zMpiTGJgUl-YCxdiTjVOMjRJNsenBY1IeC5BkaABS_-EGKi8GG2iQaGaQapGMTWsWAB8vKgQ","id":"69.167.164.199",
"changes":[{"rev":"3-84cb819703e4875faa84579fbcbbd510"}],"doc":{"_id":"69.167.164.199","_rev":"3-84cb819703e4875faa84579fbcbbd510","domains":
["xx1.test.com","test.test.com","xx1000.test.com","xx1031.test.com","xx1029.test.com","xx1034.test.com","xx1035.test.com",
"xx1082.test.com","xx1063.test.com","xx110.test.com","xx1207.test.com","xx1057.test.com","xx1220.test.com","xx1101.test.com",
"xx1218.test.com","xx1104.test.com","xx1204.test.com","xx1105.test.com","xx1118.test.com","xx1213.test.com","xx1186.test.com",
"xx1215.test.com","xx1020.test.com","xx1114.test.com","xx1183.test.com","xx1077.test.co
[ERROR] Unterminated string starting at: line 1 column 4082 (char 4081)

I guess the reason is how response from /_changes?feed=continuous;include_docs=true;since= is handled with iter_content(). resp.iter_content() has default limit of 1024 characters so JSON response is cut in the middle. data = json.loads(change) then throws an exception.

                  for chunk in resp.iter_content(None):
                        if chunk:  # filter out keep-alive new chunks
                            chunk = chunk.decode("utf-8")
                            changes = chunk.split('\n')  # make sure we handle individual changes
                            for change in changes:
                                print(change)
                                if(change.startswith('{')):
                                    data = json.loads(change)

[0xJeti]

Changing iter_content() into iter_lines() fixed the issue.

[honoki]

Thanks @0xJeti - I'm running some basic tests and was able to reproduce what you're describing. It seems like your suggestion solves the issue without breaking anything, so I will include this fix in the next release!