Support for mortal transactions

[michalkucharczyk]

Sending and monitoring mortal transactions shall be supported.

Following scenario shall be supported: https://hackmd.io/@XuVYQ1rUQjGv8uRzzsdzuw/By_NagidR

[iulianbarbu]

Apparently testing fatp with mortal txs is not that straight forward for me. Even with all the configuration done to setup txs with mortality, like in #43 , for some reason something doesn't go through as expected and down the line txs appear as immortal in polkadot js for a setup where I am starting only rococo nodes (a relaychain only network), or asset-hub-rococo + rococo nodes (parachain network + relaychain). Trying to figure out why since last week, with not much progress.

The good part is that I know where to look to confirm a tx is mortal. Code wise, I ended up reading some code paths in the runtimes (e.g. asset-hub-rococo-local, rococo), which are configured with the CheckEra/CheckMortality tx extension: e.g. here or here, so I would assume things should work correctly in the runtimes. I also tried creating a tx in polkadot JS and ran it against rococo, and it was included in a finalized block, and the tx information showed it is mortal (so polkadot JS creates mortal txs by default I think, since I haven't seen a config for it when creating/submitted the tx).

All in all before doing any runtime debugging to understand if the check mortality extension is executed against txs for asset-hub-rococo-local or rococo-local (for which I'd have to gather some more context), I'd blame first how we build txs based on subxt, or subxt having an obscure bug in terms of SCALE encoding mortal txs. Will keep digging and test my hypothesis.

Otherwise, I am very open to suggestions for how to approach the problem.