Andreas Schildbach
Andreas Schildbach
This splits out more than half of the class (1148 lines!) which should make further refactorings easier.
Problems with the `parent` fields, and their `getParentTransaction()` accessors: * They pull information about how/where Input/Output are used into Input/Output themselves, and with it quite some business logic that depends...
The idea of this patchset is to check if base has no unwanted dependencies to libraries or bitcoinj modules. It was created this way: - Remove all references of non-base...
`getAddrSeeds()` returns hardcoded IPv4-addresses of peers to try in case any of the other peer discovery methods (mainly DNS) can't be used. Problems: * The data is likely outdated, as...
The idea is to make sure Instants that are transmitted over the P2P network never have milli- or nanoseconds. That way, we cannot loose them on the way. A `checkBitcoinTime()`...
I had been eagerly renaming recently added `get*Instant()` accessors to `*Time()`. Turns out I didn't really understand the convention of omitting the "get" is only used for final fields, or...
Currently the tests in `bitcoinj-core` are limited to JDK 8, just like our [runtime requirements](https://github.com/bitcoinj/bitcoinj/blob/master/README.adoc#technologies). However for building we already require JDK 11, so we technically could allow Java 11...
This is meant for discussion if and how we want to modernize our network stack (likely *after* 0.17). Once upon a time, we used [Netty](https://netty.io/) and had only blocking I/O....
We keep getting this error in the unit tests that do socket IO. An example below. The issue is always intermittent. Possible mitigations: 1. Don't use socket IO in unit...
**Current state** Currently we represent seed/key/chain creation times as a long (seconds since epoch). In cases where a creation time is not known, we use `0`. In cases where a...