[juneja-varun]

Fixes #713.

I upload large files in chunks via a generator (while f.tell() < file_size: requests.post(url, data=chunk_file(f, ...))), and under responses this loop never terminates - the file position never advances, because the generator passed as data= is never actually read. A real network send would consume it chunk by chunk while writing to the socket; responses just stores the generator object as-is in request.body and moves on.

This is the same class of gap as #719/#736 (file-like objects with .read() weren't being consumed either, fixed by reading them in _read_filelike_body) - just for the other kind of streaming body requests supports. Extended that same helper to drain a generic iterator/generator into bytes too.

(The original issue's title points at multipart_matcher, but that's not actually the site of the bug - files= uploads get fully read by requests itself during prepare_body(), before responses ever sees them. The real gap is in the general streamed-data= path, which is what the fix targets.)

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want reviews to match your repository better? Bugbot Learning can learn team-specific rules from PR activity. A team admin can enable Learning in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit da32f59. Configure here.}