Our Transfer-Encoding: chunked implementation was not actually correct, and as a consequence was not robust to servers that improperly add leading zeroes to chunk lengths. This is fixed here by more closely conforming to the RFC spec ourselves, and removing a few parse states that don't match the BNF defined there.
Our Transfer-Encoding: chunked implementation was not actually correct, and as a consequence was not robust to servers that improperly add leading zeroes to chunk lengths. This is fixed here by more closely conforming to the RFC spec ourselves, and removing a few parse states that don't match the BNF defined there.
Fixes #23.