This moves a lot of the test support code into its own package which is only imported for tests,
so it's not shipped as a part of core Corda. The node currently depends on this support code to
compile, although future work could try to separate this out. This change highlights that parts
of production code is dependent on test elements (i.e. dummy keys), and makes it harder for
such accidental crosses to occur later.
An integration test category is also added as part of this work, to contribute towards COR-345.
Add new protocol which manages the entire process of taking a signed transaction ready
for notarisation, through notarisation and onto recording it both locally and informing
remote nodes.
This protocol also optionally can include the ClientToServiceCommand which triggered a transaction
being created, to give the remote nodes context on why a change occurred (i.e. "You are being sent
£100")
Fixup after rebase and fix issue with checking previous deployment of bridges
Correct comments on ArtemisMessagingClient constructor
Fixup rates fix demo
Get rid of when statements
Make NetworkMapCache send modify as well as add//remove events. Make inboxes for nodes persistent.
Suppress warnings
Fix message acknowledgement so that it actually consumes messages properly.
Change queueName to SimpleString to stop lots of wasted conversions
Get rid of spurious import
Tidy up and add comments
Update to include comments on PR
Remove unnecessary import
Try providing a helper interface to encourage enforcing LinearState rules
Fixup after rebase
Change to using Clauses for verifying LinearState standard properties
Fix whitespace change
Tidy up ClauseVerifier after PR comments
Change from SecureHash to a TradeIdentifier class
Change TradeIdentifier to UniqueIdentifier
Add observable for transactions being stored, so the UI can show transactions as they're received, rather than being
limited to the summarised version available from the wallet service.
- Fix thread safety issues in ArtemisMessagingClient. The Artemis API isn't thread safe, but that isn't well documented and it will happily invoke callbacks in parallel.
- Add discussion of how we tackle threading currently in the codebase and make a few other improvements.
- Add a shutdown hook so we stop properly when the user presses ctrl-c
First working Exposed-assisted persistent wallet
Cleaned up Exposed-based persistent wallet
Cleaned up warnings
Fixed up some generic types
Improved comments
Fix up TODO comment
Hikari and config integration
Fix existing tests
Clean up after looking at PR
Clean up commented out lines
Fix initialisation of IRS demo leaving database open
Fix up after rebase
Review feedback. Main change is lazy wallet iteration.
Rebased and incorporated config changes.
Use standardised config loading. Make wallet cash test use persistent wallet.
Added test to ensure wallet retains state in database across instance creation.
Tidy up whitespace and fix bug in test.
Changes include:
- LedgerTransaction is now much more central: it represents a fully resolved and looked-up tx, with the inputs available.
- TransactionGroup and TransactionForVerification are gone. There is a temporary TransactionForContract class for backwards
compatibility but it will also be gone soon.
- ResolveTransactionsProtocol is simplified, and now commits a tx to the database as soon as it's determined to be valid.
- ServiceHub is now passed in more consistently to verification code, so we can use more services in future more easily e.g. a sandboxing service.
- A variety of APIs have been tweaked or documented better.
This change reduces the testing confusion that can occur when cash is issued by one of the parties in a transaction rather than e.g. a neutral third party like a central bank.
Break down what is referred to as "topic" of a message into its component parts. This splits the
general topic from the session ID, so it's clear where a session ID is provided, and whether any
given topic string includes a session ID or not.
