This PR adresses #166
Error handling in wallet was ported to failure in https://github.com/mimblewimble/grin/pull/713
Using the same error model makes wallet code simpler and may simplify migration to Hyper.
* Fix issue where we have no metadata for a block (#938)
when restarting node before initial sync completed
* Avoid double-locking on add eviction. Fixes#936
* Fix 33c5a983
* Add support for DNS Seed (#940)
* Add support for DNS Seed
* Add port
* Add seed.grin-tech.org
* Remove duplicate IPs
* minimal stratum server
* Modifications for review comments. Move stratum test into its own file, move get_block() into its own rust module, use pool and chain only rather than the entire Miner object
* rustfmt
* cleanup
* cleanup
* Introduce extending_readonly to simplify a forcing and cancelling rollbacks (#945)
readonly views of the txhashset
* Add DNS Seed and make DNSSeed default (#942)
* Add dns seed seeding type
* Add grin-seed.owncrypto.de and make DNSSeed default
* Add email address for each DNS Seed
* [WIP] Core PMMR and API updates to support wallet restore (#950)
* update pmmr to get batch of elements by insertion position
* update pmmr to get batch of elements by insertion position
* add api + chain calls to get traversed outputs back out
* add api + chain calls to get traversed outputs back out
* first pass getting wallet restore to work again with updated utxo-walking api
* Update simulation.md
* Fix Bus Error (core dumped) when validating fast sync txhashset (#956)
This PR fixes#953 by introducing a lock for txhashet_write. It's not enough
to synchronize access to in memory data, files also needs to be protected, so
a general txhashset lock was introduced.
* refactor grin crate into separate modules (#955)
* Add total kernel offset to block api (#954)
* minimal stratum server
* Modifications for review comments. Move stratum test into its own file, move get_block() into its own rust module, use pool and chain only rather than the entire Miner object
* rustfmt
* cleanup
* cleanup
* Merge with grin_grin -> servers code reorg
* Merge with grin_grin -> servers code reorg
* add stratum server stats
This PR fixes#953 by introducing a lock for txhashet_write. It's not enough
to synchronize access to in memory data, files also needs to be protected, so
a general txhashset lock was introduced.
* update pmmr to get batch of elements by insertion position
* update pmmr to get batch of elements by insertion position
* add api + chain calls to get traversed outputs back out
* add api + chain calls to get traversed outputs back out
* first pass getting wallet restore to work again with updated utxo-walking api
* documentation and spelling
* unused values in tests; just .unwrap()
* unneeded macro_use (router)
* unneeded macro_use and use-imports
* unused value -> .unwrap
* unused variables
* clarify that " is required (see 107ea3c4dd)
* first steps to dig into #933 test_p2p issues
* less simulnet build warnings
Our chain orphans data structure now does:
* Eviction based on a max total number of orphaned blocks
* Evicts further away first, then too old
* Avoids all cloning (blocks can be big)
This allows sync to be a little more stupid, only reducing the
number of blocks it requests when orphans get nearly full.
Fixes#897
* Added additional message type size limits
* Externalize sync loop state, check if we received all
* Adjusted some message sizes, log error when too short
* 4x limits to give ourselves more space
* No need for sync to wait if we have enough peers.
* Tolerate a few less headers before considering we got them all.
* avoid repeated truncation in difficulty adjust + minor refactoring
* reinstate needed low-diffculty exception on dampening
* leave consensus breaking change for later one line change; obsolete lower testnet1 max target
* Fix and cleanup of fast sync triggering logic
* New txhashset on fast sync has to be applied, not rolled back
* Do not block if peer send buffer is full, fixes#912
We can re-enable it on some nodes to keep an eye on the network,
but for debug builds this is too expensive. It can take over 60
sec for low-end environments, causing them to never catch up.
I think this is still better than the alternative of telling
everyone to run release builds.
* add hint for gcc-5 for building CUDA plugin
* add N_BLOCks comment for >=6GB GPU cards
* minor logging cleanup
* mark to be kept + reformat as "@ height [hash]"
* show CuckooNN + better wording
