- Jan 09, 2021
- Aug 23, 2020
- May 11, 2020
-
-
ale authored
Useful for large blob responses.
-
- Apr 04, 2020
- Feb 11, 2020
- Feb 06, 2020
- Jan 06, 2020
-
-
ale authored
Add the 'connect_timeout' and 'request_max_timeout' configuration fields, to control respectively the initial connection timeout, and the maximum time for each individual request. This allows fine-tuning of the expected performance of specific backends, for example to let optional backends fail fast.
-
- Aug 17, 2019
-
-
ale authored
We're using language-specific Markdown templates to allow simultaneous generation of text and HTML for fancy multipart/alternative emails. The package also supports PGP/MIME for signatures (no encryption yet).
-
- Aug 15, 2019
-
-
ale authored
-
- Jun 30, 2019
-
-
ale authored
The Context in the http.Request is bound to the one we give it when calling balancedBackend.do(), so we were accessing it out of scope when calling the final json.Decode on the response body. This was not a problem for most *small* requests, as the response Body already contains all the data due to having read it along with the headers. However, larger response bodies would cause json.Decode to call a Read with what at that point is a canceled Context, so the Call function would return a mysterious "context canceled" error. (Note that this change introduces a minor, probably less annoying, issue, where we can't reset "resp" between successive calls so we may be invoking json.Decode on a non-pristine object if the first call fails in some circumstances).
-
- Apr 03, 2019
-
-
ale authored
-