Hello! I'm looking to understand the expected behavior when creating AWS SigV4 signed request, specifically how the literal character of + is treated in query params.

My understanding is that the literal + character is valid in query params, and that the client should encode it as %2B, e.g. something like C++ would be percent encoded as C%2B%2B

It looks like there is a change last year where the key / values of the query params are decoded first, before being re-encoded.

I think the decode + encode essentially assumes that the literal + will always be interpreted and encoded as a space character (%20), which means that the request to be signed will different than the original request, if the original request's query param contains %2B.

Is this the intended behavior?