* don't force logging syncs
* Add failing test
* wip discovery engine
* re-add chronicles sinks
* wip
* move network related stuff to own folder
* move peer related stuff to own folder
* extract discovery into it's own engine
* update imports
* move pending blocks into engine module
* add top level exports
* update imports
* update import paths
* update imports
* support for inflight request filtering and tests
* use `remove` instead of `del`
* fix sorting in `selectCheapest`
* re-org test file structure
* fix to use discovery engine
* file re-org
* fix compilation
* fixup discovery to use async handlers
* more re-org
* rework with support for discovery engine
* add logging
* use defaults
* wip: reworking with discoveryengine
* wip: more test fixes
* more logging
* use ordered table
* use `bt` for blocktype Block
* fix tests
* make tests work with discovery engine
* expose all node components
* fix to work with discovery engine
* wip
* propagate cancellation in listBlocks
* start/stop disc engine in blockexc engine
* remove disc engine start/stop
* wire up discovery engine
* misc comments and imports
* pass discovery to dagger node
* set sleep timers
* unused imports
* misc
* don't spawn a task, await it
* don't await handlers
* trace logging
* reduce default sleep time
Co-authored-by: Tanguy <tanguy@status.im>
* rework discovery with async queues
* increase max message size for large manifests
* increase sleep time to 100 millis
* pass config
* check for nil on start/stop
* fix tests and split out discovery tests
* don't auto mount network
* add discovery tests
* rework moc discovery
* move discovery moc to disc dir
* don't force logging syncs
* don't force moc discovery on all tests
* rework discovery with methods
* add top level utils file
* don't use asyncCheck
* don't pass entire blocks to list blocks calback
* spelling
* - don't send want reqs to peers reporting the cid
- Don't request blocks directly on presense update, use `requestBlock`
* bug, nodes should not have blocks in local store
* Add failing test
* prefetch blocks so that download isn't serial
* if request already pending, return the handle
* fire discovery if no peers report block as have
* only query discovery if not enough nodes for cid
* wrap async req in template
* use non awaiting version of queue routines
* rework E2E tests as unittest
* re-add chronicles sinks
Co-authored-by: Tanguy <tanguy@status.im>
* rework discovery with async queues
* misc style changes
* increase max message size for large manifests
* use upraises and avoid exceptions on key access
* increase sleep time to 100 millis
* pass config
* make list blocks trigger a callback on each block
* check for nil on start/stop
* fix tests and split out discovery tests
* don't auto mount network
* add list block tests
* add discovery tests
* rework moc discovery
* move discovery moc to disc dir
* don't force logging syncs
* don't force moc discovery on all tests
* rework discovery with methods
* add top level utils file
* don't use asyncCheck
* don't pass entire blocks to list blocks calback
* spelling
* wip: adding request for storage endpoint
* wire in erasure coding
* fix tests for erasure coding
* put type definitions into separate file
* integrate erasure coding
* change run/shutdown to start/stop
* temporary sleep, otherwise the fsstore blocks
* feat: introduce LRU cache
Replace `MemoryStore` with LRU caching mechanism.
`lrucache` library was forked to https://github.com/status-im/lrucache.nim.
Co-authored-by: Eric Mastro <eric.mastro@gmail.com>
# Conflicts:
# dagger/dagger.nim
# dagger/stores.nim
# dagger/stores/manager.nim
# tests/dagger/blockexc/testengine.nim
# tests/dagger/helpers/nodeutils.nim
# tests/dagger/testnode.nim
# tests/dagger/teststores.nim
* feat: introduce --cache-size CLI option
Allow for a value of `0` to disable the cache.
# Conflicts:
# dagger/dagger.nim
* allow dynamic block size in cache
allow block size to be variable/dynamic
update lrucache to use updated lrucache dep
Using removeLru proc, added tests
* Refactor CacheStore init block
Co-authored-by: Michael Bradley, Jr <michaelsbradleyjr@gmail.com>
* use PeerInfo in event handlers
* use CidV1 and raw multicodec as default
* add block stream abstraction
* raises defect
* adding dataset abstraction
* move blockstream into own dir
* reorg files and fix tests
* rename dataset to blockset
* wip
* wip
* adding basic test for treehash algo
* run blockset tests along with with the rest
* remove obsolete contents
* fix chunker tests
* rename bitswap and move to stores
* rename bitwsap to blockexc and move to stores
* moare project structure reorg
Dmitriy Ryajov
markspanbroek
Tanguy
Eric Mastro