Make the "withdrawMultiple" function work with multiple recipients

[PaulRBerg]

The withdrawMultiple function is severely limited at the moment: it only works if all the streams in the input array have the same recipient. This issue has been previously flagged by @razgraf here, but his criticism at the time was in the context of the proxy target.

Since we don't have a proxy target anymore, the only solution now is to refactor the withdrawMultiple function to implement this behavior. There are two possible approaches:

1. Remove the to parameter and have the smart contract query the recipient for each stream in the input array.
   • This is less gas-efficient but also less labor-intensive on the client side
2. Allow users to specify an array of to addresses.
   • This is more gas-efficient but also more labor-intensive on the client side
   • Technically, this approach unlocks a functionality that otherwise cannot exist, i.e, a recipient with multiple streams could call this function to withdraw to a bunch of different addresses. But this doesn't sound like a super useful feature.

There are pros and cons to each approach. Ultimately, we might want to implement both. WDYT, @razgraf, @andreivladbrg?

At a minimum, I think we should definitely implement at least the first approach, and so I have created this issue: #774

[PaulRBerg]

N.B. I've realized we are missing this feature after being asked by a potential user (Perpetual Protocol):

[razgraf]

Resolved in #785.