Update MetaDEx network/transaction fields

[dexX7]

This PR removes the action byte on a transaction level.

The new transaction format:

Add:
[version=0] [type=25] [property] [number] [property] [number]

Cancel-All-At-Price:
[version=0] [type=26] [property] [number] [property] [number]

Cancel-All-Of-Currency-Pair:
[version=0] [type=27] [property] [property]

Cancel-Everything:
[version=0] [type=28] [ecosystem]

There is still some use of action bytes here and there, and until those parts are also updated, it can be completely removed, but this seems to be out of scope for now.

It resolves #47.

[zathras-crypto]

Supported, will modify my RPC branch accordingly when this is merged.

Spec will need updating.

Pinging @achamely @msgilligan @CraigSellars @marv-engine - please raise any objections before we merge this, thanks.

[zathras-crypto]

Note while I do support this 100% - this would nullify all previous MetaDEx test transactions on testnet correct?

[dexX7]

Yes.

Do they matter?

[zathras-crypto]

Not at all :)

[dexX7]

Haha well, that's good to hear. My fear was that there might be someone who already started to build something, based on the assumption there is transaction type 21 with action byte.

[zathras-crypto]

Oh there are - for example we may p*ss off a couple of folks (sorry @patrickdugan !!) who may be working on test integrations - but that doesn't mean this isn't the right thing to do.

Get it right before release :) Plus I haven't seen any objections raised on the thread discussing this proposal.

[dexX7]

It would be extremely helpful, if anyone who might be affected here quickly leaves a note, whether this change would be game breaking.

Although I assume changing the RPC interface has/had a greater impact, since the transaction fields are pretty low level.

[zathras-crypto]

Although I assume changing the RPC interface has/had a greater impact, since the transaction fields are pretty low level.

That's a very, very good point - since that's already been merged (the RPC changes) that should have been the "breaker-of-things" rather than underlying payload.

[achamely]

The omnicore code we were working on was all based off the rpc calls so the underlying action byte change won't be a big upset. We already knew we were going to need to rework our tx21 handling all together once the rpc calls where a little more finalized/fleshed out

[dexX7]

Alright.

Please let me know, if consensus is reached.

[patrickdugan]

This may actually simplify my inventory management across many different currencies, good thinking.

[zathras-crypto]

I raised this at the meeting, consensus is to adopt this change. Feel free to merge at your leisure mate - suggest a spec change is needed to supplement this PR.

[marvgmail]

These new transactions look ok to me.

Is tx 21, version 0 still supported? Until some block number?

[dexX7]

@marv-engine: no, tx 21 doesn't exist anymore. Testnet has no block heights to enable features, and it was never live on mainnet.

If there is any application, which requires tx 21 (for whatever reason), we may provide testnet support for a limited amount of time.

Currently, the old RPC command "trade_MP" (with action byte) is still available for compability, but I think it's going to be removed rather sooner than later, too.