New tokenIds

[zajck]

Replaces #582

Implements necessary changes to support tokenId = <<uint128(offerId)>>.<<uint128(exchangeId)>>
Although initially meant only for preminted vouchers, it turned out it's easier to make the change for all new vouchers.

Main reason to do this is to remove binary search in getPremintStatus, which has now constant time and space complexity.

Additinal implications:

• max preminted range length was reduced to 2^64 which is still more than anyone realistically needs
• no need to store array of preminted offer ids in vouchers, so _rangeOfferIds was removed. Since binary search was removed, there is no need to require that preminted offer ranges are ordered.
• transferPremintedFrom is not needed anymore since regular transfer functions are efficient enough

[anajuliabit]

Besides what I put in the code, could you please add unit tests to validate that tokens created before and after EXCHANGE_ID_2_2_0 are burned correctly?

[anajuliabit]

forgot to choose "request changes"

[zajck]

Tests for token burning before/after the upgrade are in #498 .

[mischat]

I guess, a highlevel comment. The method we added : 'extractExchangeId' states that it is needed when a tokenID is passed in instead of an exchangeID...

when i look for the usage it only seems to be used in the exchangeHandlderFacet, is this meant to be the case ?

And given that the method, in the comments, states that it takes in a tokenID and outputs an exchangeID, i feel like the naming is misleading :

protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:        _exchangeId = extractExchangeId(_exchangeId);
protocol/facets/ExchangeHandlerFacet.sol:    function extractExchangeId(uint256 _exchangeId) internal pure returns (uint256 exchangeId) {

Looking at the code, we don't seem to use these tokenIDs outside of Voucher and the Exchange handler, every other instance of tokenID seems to either be todo with token gating, token based auth, or twinning, which should be fine ...

Anyways

[zajck]

@mischat Re: extractExchangeId

That was meant more for convenience, so users could pass in either tokenId or exchangeId.
But yeah, if we really wanted to support that, I would need to do it everywhere (there are methods in dispute and orchestration handler), and also change all natspec to say "exchangeId can either be exchangeId or tokenId`.

So after your comment I changed my mind and removed extractExchangeId, since all protocol methods are clear that they expect exchangeId. Anyways, whoever is interacting only with the protocol already operates only with the correct exchangeId.
But even if they have only tokenId, they can always easiliy convert it to exchangeId.

I'm running tests locally now and will push the changes soon.

[mischat]

Thanks, this looks good to me

[anajuliabit]

LGTM