You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The difference between the expected and actual are the additional results tags. Since the results is already marked as an array, the second set of partial results should not create another array element. This is causing the issue with deserializers like Newtonsoft.Json#1261
The text was updated successfully, but these errors were encountered:
When chunked=true is sent, multiple JSON bodies are sent. I don't see the error. Some JSON parsers have difficulty with that and throw an error, but they really shouldn't.
We can't really change the format. First, sending the JSON chunked but as a single JSON body would defeat the purpose of chunked results. You would not be able to read any of the points until the entire JSON body was consumed. Second, that would break backwards compatibility.
I'm not sure this is a bug with InfluxDB as much as a problem with the JSON parsing library.
Close loop: this is an issue with default buffering by .Net http client. Without HttpCompletionOption.ResponseHeadersRead parameter, the GetAsync waits until entire response is buffered at client side. Which defeats the purpose of chunking. Setting the HttpCompletionOption and reading the response as a stream line by line works like a charm.
Latest Nightly (1.3.0~n201703250800)
Steps to reproduce:
chunked=true
, andchunk_size=10
Expected behavior:
Actual behavior:
The difference between the expected and actual are the additional
results
tags. Since the results is already marked as an array, the second set of partial results should not create another array element. This is causing the issue with deserializers like Newtonsoft.Json#1261The text was updated successfully, but these errors were encountered: