Seth Michael Larson
Seth Michael Larson
FYI I'm working on this right now. Will submit a PR once I get a bigger chunk existing of tests passing.
@sigmavirus24 Taking the old branch that was mentioned in #51, bringing it up to date with the module, adding Python 2 and 3 compatibility, fix all tests, add new tests....
@sigmavirus24 I thought I read the issue closely but I will re-read. The branch defines a `MultipartEncoderGenerator` object that is used in the same way that the tool-belt code would...
@pquentin I guess depending on how we accomplish this one it doesn't necessarily have to be a breaking change. We can deprecate `encode_multipart_formdata` in favor of a function that produces...
@shadycuz It sounds like you're on the right path!
@EnricoMi We're not adding new features to 1.x as there won't be a 1.27 release. All new features must go into 2.0.
@EnricoMi We don't have a date in mind, we're trying to finish [these blocking issues](https://github.com/urllib3/urllib3/milestone/6) before a release though.
Okay, first of all thank you for writing up all of this and doing tons of research. I apologize that I've taken a long time on getting an official response...
Certainly could work! Are there many users of decode_content=False? I can only remember seeing pip use that.
`encode_multipart_formdata` function uses twice as much memory as the amount of data being sent and doesn't perform well on streams of data as you'd expect. Could this be the root...