mirror of
https://github.com/EDCD/EDDN.git
synced 2025-06-04 09:31:21 +03:00
Fixes the regression identified in #165 However the gzip code path can still erroneously think a decompressed request body is form encoded when it is not. This happens when any text in the decompressed body matches the regex: .+=[^\&;]* i.e. some text, followed by `=`, and then some more text, possibly empty, followed by an ampersand `&`, or semi-colon `;`, or the end of the string. `&` and `;` are used to separate key=value pairs in form encoding, the `=` separates a key from its value.