* Revert "Fix not downloading old blocks (#8642)"
This reverts commit d1934363e7.
* Make sure only one thread actually imports old blocks.
* Add some trace timers.
* Bring back pending hashes set.
* Separate locks so that queue can happen while we are importing.
* Address grumbles.
* Mark test helpers and test-only specs as cfg(test)
* Use test-probe to conditionally compile test helpers
* Remove test probe and directly use features tag
* Update `add_license` script
* run script
* add `remove duplicate lines script` and run it
* Revert changes `English spaces`
* strip whitespaces
* Revert `GPL` in files with `apache/mit license`
* don't append `gpl license` in files with other lic
* Don't append `gpl header` in files with other lic.
* re-ran script
* include c and cpp files too
* remove duplicate header
* rebase nit
* Remove public node related settings
* Fix tests
* Unwrap accounts provider in all rpc apis
* Unwrap AccountProvider in all cli places
* Fix rpc tests
* Custom Error Messages on ENFILE and EMFILE IO Errors
Add custom mapping of ENFILE and EMFILE IO Errors (Failure because of missing system resource) right when chaining ioError into ::util::Network::Error to improve Error Messages given to user
Note: Adds libc as a dependency to util/network
* Use assert-matches for more readable tests
* Fix Wording and consistency
* network-devp2p: Test for discovery bucket insertion.
All test values are randomly generated and the assertions are checked manually.
Test fails because distance metric is implemented incorrectly.
* network-devp2p: Fix discovery distance function.
The Kademlia distance function (XOR) was implemented incorrectly as a population count.
* network-devp2p: Refactor nearest_node_entries to be on instance.
Optimizations are possible with more access to the discovery state.
* network-devp2p: Fix loss of precision in nearest_node_entries.
* network-devp2p: More efficient nearest node search.
The discovery algorithm to identify the nearest k nodes does not need to scan
all entries in all buckets.
* Lots of println to figure out what eth_getBlockByNumber does/should do
* Remove debugging
* Print warnings when fetching pending blocks
When calling `eth_getBlockByNumber` with `pending`, we now print a deprecation warning and:
* if a pending block is found, use it to respond
* if no pending block is found, respond as if if was a request for `Latest`
Addresses issue #8703 (not sure if it's enough to close it tbh)
Additionally as to the port, the new command line option
now allows the user to specify the network interface the P2P-Parity
listens, too. With support for 'all' and 'local' like in all other
versions of this flag. Default is 'all' (aka ).
* Set the request index to that of the current request
When setting up the chain of (two) requests to look up a block by hash, the second need to refer to the first. This fixes an issue where the back ref was set to the subsequent request, not the current one. When the requests are executed we loop through them in order and ensure the requests that should produce headers all match up. We do this by index so they better be right.
In other words: off by one.
* Allow post-homestead forks to be specified in CommonParams
* Fix all json configs
* Fix test in json crate
* Fix test in ethcore
* Fix all chain configs to use tabs
Given we use tabs in .editorconfig and the majority of chain configs.
This change is done in Emacs using `mark-whole-buffer` and `indent-region`.
* Add license and readme
* Use pure rust implementation
* Bump version to 0.1.1
* Don't use C, prefer the pure Rust implementation
* Add test for `write_keccak`
* Bump version
* Add benchmarks
* Add benchmarks
* Add keccak_256, keccak_512, keccak_256_unchecked and keccak_512_unchecked – mostly for compatibility with ethash
* Remove failed git merge attempt from external git repo
Cargo.lock updates
* whitespace
* Mark unsafe function unsafe
* Unsafe calls in unsafe block
* Document unsafety invariants
* Revert unintended changes to Cargo.lock
