Import legacy Byron addresses

[KtorZ]

Issue Number

ADP-292

Overview

• f71d8a8
  📍 add 'putRandomAddress' to the swagger specifications and API servant definition

• db29bab
  📍 write dummy server implementation for 'putRandomAddress' and fill the gap in various APIs

• 46aa7c5
  📍 add integration tests for byron address restoration

• 3f255e7
  📍 implement server and application handlers for importing an address

• f7ba8fb
  📍 add command for 'address import' mapping to the newly introduced endpoint

• 2ae81f6
  📍 add some integration tests scenario for 'address import'

• 73294ae
  📍 fix non-atomic checkpoint modification in {import,create}RandomAddress
  This is subtle but, everytime we are tempted to write:

  cp <- atomically $ readCheckpoint wid
  {- Do some stuff... -}
  atomically $ putCheckpoint wid cp'

  we should in reality really go for:

  atomically $ do
    cp <- readCheckpoint wid
    {- Do some stuff... -}
    putCheckpoint wid cp'

  because there's no guarantee that the checkpoint hasn't already
  changed between the moment we fetched it, and the moment we insert a
  modification. This could have disastreous effects like actually
  updating a checkpoint in the past while a new one has already been
  created with now oudated data; or worse, in the case of EBB, it'll
  simply override the newly written checkpoint and makes the database
  silently roll back of exactly 1 block!

Comments

$ cardano-wallet-byron address import --help
Usage: cardano-wallet-byron address import [--port INT] WALLET_ID ADDRESS
  Import a random address generated elsewhere. Only available for random
  wallets. The address must belong to the target wallet.

Available options:
  -h,--help                Show this help text
  --port INT               port used for serving the wallet API. (default: 8090)

[KtorZ]

bors try

[iohk-bors]

try

Build failed

• buildkite/cardano-wallet

[KtorZ]

bors r+

[iohk-bors]

Build failed

• buildkite/cardano-wallet

[KtorZ]

bors r+

[iohk-bors]

Build failed

• buildkite/cardano-wallet

[rvl]

Looks good.

[rvl]

Don't the REST dogmas say that PUT is for updating and POST is for creating?

[KtorZ]

Not necessarily. PUT is for pushing an already created resource assuming that a client does have the ability to create such resource entirely. This is the case here where the address pre-exists to the call, and is therefore pushed to the server.

POST would be for creating a resource from partial data, or from a "constructor". There are resources where this is not possible to do differently, like the wallet. We can't possibly allow a client to "put" its own balance or sync progress, hence the POST.

[KtorZ]

Then, for choosing between PUT and PATCH, it's more of a matter of perspective. We are not using PATCH much in the API where it would be legitimate IMO (like when updating the passphrase of metadata, i.e. when doing a partial update of a resource). Yet, we've stick to PUT here because people tend to be somewhat confused by the PATCH method. And, from the perspective of a sub-resource, a PUT is still acceptable for partially replacing an entity. But this is why we write:

PUT /wallets/{walletId}/passphrase

and not:

PUT /wallets/{walletId}

{ "passphrase:" ... }

[KtorZ]

bors r+

[KtorZ]

waiting for Hydra again takes ages :(

[iohk-bors]

Build failed

• buildkite/cardano-wallet