Refactor/send form redux cleanup

[PeKne]

Batch of send form redux logic refactoring. I recommend to review this PR commit by commit.

Description

• refactor(suite): signed transaction data stored in redux send form state: signed transaction unserialized data stored in redux state so it does not have to be passed between thunks as an argument

• refactor(suite): send form common thunks are rejected on fail: suite-common sendFormThunks utilize redux-toolkit rejectWithValue for better error handling

• refactor(suite): send form precomposed transaction arguments naming unified: naming of precomposedForm arguments unified for better readability

• refactor(suite): unify send form precomposed-transaction types naming: There are two types of a successful precomposedTransaction. One for Cardano and other for rest of the networks. This commit makes the naming of these types clearer.

• refactor(suite): split pushSendFormTransactionThunk into sub-thunks: Bulky pushSendFormTransactionThunk was split into smaller sub-thunks. Some of theme were needed only by desktop, these were moved to packages/suite package.

• refactor(suite): convert send form drafts amount units refactored

• refactor(suite): redundant precomposedForm state removed from sendFormReducer

[vytick]

I am not deep in the send here, but it lgtm :) I have a few questions though

[bosomt]

See screenshots, same transaction, RBF on but after tx is signed modal reports RBF off

Info:

• Suite version: desktop 24.6.0 (a6fe7cd)
• Browser: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) TrezorSuiteDev/24.6.0 Chrome/118.0.5993.159 Electron/27.3.8 Safari/537.36
• OS: MacIntel
• Screen: 1512x982
• Device: Trezor T2T1 2.7.0 regular (revision 45e8a842a31e62a6d43d7f6ccac62a45e1198ef0)
• Transport: BridgeTransport 3.0.0

[PeKne]

@bosomt The problem should be fixed by: f635113

Please test it again 🙏.

[bosomt]

QA OK

yes fixed and overall OK , i think we can merge it to develop

Info:

• Suite version: web 24.6.0 (31b59dc)
• Browser: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:126.0) Gecko/20100101 Firefox/126.0
• OS: MacIntel
• Screen: 1512x982
• Device: Trezor T2T1 2.7.0 regular (revision 45e8a842a31e62a6d43d7f6ccac62a45e1198ef0)
• Transport: BridgeTransport 2.0.33

[marekrjpolak]

Looks good in general, a lot of work has been done here.

The only thing I'm not quite happy about is the first commit: I prefer passing required arguments explicitly between the stages of a single process, rather than using the global state for it. That way:

• data validation is performed by type checking in development time instead of runtime checks
• functions (mostly thunks in this case) have as few side effects as possible while having as few dependencies as possible at the same time
• less code is needed to pass an argument than to store it into the state, read it somewhere else and validate that it is in fact there
• less effort is needed to understand the code if it's directly visible what the function needs in order to work and what it produces without looking into its implementation (that's why pure functions are so beautiful)

In this case it seems that signedTx and serializedTx are both stored at the beginning of signAndPushSendFormTransactionThunk (or its subroutine) and needed somewhere at the end of it, therefore I don't see a reason for using global state other than hiding the complexity of passing the parameters, which makes the overall complexity bigger instead of smaller imo.

I'm not saying it cannot be merged as it is, but I'll try to look at it a bit more if you don't mind.

[marekrjpolak]

After call I agree that as a prerequisite for send functionality in native app this may make sense. 👍