Activity
axic wants to merge ethereum/evmc
/// This enables implicit conversions from evmc::hex_errc to std::error_code.
zoey-t forked ethereum/solidity
axic wants to merge ethereum/evmc
Maybe add std::isspace(c) == evmc::internal_hex::isspace(c)
too? Probably not needed.
kclowes pull request ethereum/eth-abi
Add make notes command, exclude venv in MANIFEST and gitignore
What was wrong?
Issue #
How was it fixed?
Summary of approach.
To-Do
- Clean up commit history
- Add entry to the release notes
Cute Animal Picture
fredriksvantes push ethereum/ethereum-org-website
commit sha: a6dca216a2ab54b228cea5ac3543c3690bab8ba9
push time in 1 hour agofredriksvantes in ethereum/ethereum-org-website delete branch asanso-patch-1
fredriksvantes pull request ethereum/ethereum-org-website
fredriksvantes merge to ethereum/ethereum-org-website
LGTM!
sawal-timsina forked ethereum/solidity
sawal-timsina forked ethereum/go-ethereum
cameel merge to ethereum/solidity
Remove enforce-via-yul CI run and instead always force via yul in regular test runs.
I noticed when https://circleci.com/gh/ethereum/solidity/1060053 ran out of memory. It's rather wasteful to run the smt tests in this run again - they're already covered and not affected by the option.
We might want to look for other cases like this... Also we could think about restricting this further... the flag really only affects the semantics tests, doesn't it? But most of it is cheap anyways, so the SMT tests are the most wasteful ones I guess.
Looks good. I haven't really reviewed all the test changes, only scrolled through about half of them until github refused to load more diffs. The C++ code looks correct though and tests pass so I'm assuming it must all be fine.
CarlBeek wants to merge ethereum/staking-launchpad
Add Ropsten beacon chain
Address #484
TODOs:
- Check envVar configuration
- some external links are not ready yet (e.g., beacon explorer)
- Port some "post-merge" context from
bellatrix
branch to Ropsten launchpad - The client tutorials are for Goerli. But Ropsten may have different instructions. We may want to hide the testnet tutorial or fix it.
export REACT_APP_ETH_REQUIREMENT=3200384
Are there not 100,012 validators at launch? https://github.com/eth-clients/merge-testnets/blob/main/ropsten-beacon-chain/README.md?plain=1#L28
CarlBeek merge to ethereum/staking-launchpad
Add Ropsten beacon chain
Address #484
TODOs:
- Check envVar configuration
- some external links are not ready yet (e.g., beacon explorer)
- Port some "post-merge" context from
bellatrix
branch to Ropsten launchpad - The client tutorials are for Goerli. But Ropsten may have different instructions. We may want to hide the testnet tutorial or fix it.
Given that we are only going to merge Ropsten once, I think it is fair to just hide the client tutorials for notGoerli (tm).
CarlBeek merge to ethereum/staking-launchpad
Add Ropsten beacon chain
Address #484
TODOs:
- Check envVar configuration
- some external links are not ready yet (e.g., beacon explorer)
- Port some "post-merge" context from
bellatrix
branch to Ropsten launchpad - The client tutorials are for Goerli. But Ropsten may have different instructions. We may want to hide the testnet tutorial or fix it.
Given that we are only going to merge Ropsten once, I think it is fair to just hide the client tutorials for notGoerli (tm).
pipermerriam merge to ethereum/portal-network-specs
Move sections on neighborhood gossip and POKE + adjustments
Moved sections on neighborhood gossip and POKE to the Portal wire specification. Adjusted and removed parts in the Portal wire specification that were already explaining similar concepts.
The neighborhood gossip explanation is specifically adjusted to be now a mix of a local neighbors lookup and a network node lookup.
This is what is currently being done in Fluffy (practically, it is a mix of what was written in state network spec and portal wire, originally history network, spec). It kind of takes the best of both worlds: no additional node lookup if it isn't needed, but still able to (theoretically) propagate a piece of content even if that content is gossiped from only 1 node initially (assuming this node is well connected in the network).
One thing that it still might lack, is propagation to nodes with a very large radius, at least when it uses the nodes from the lookup. For that reason we need to perhaps mix the nodes that we offer the content to. But maybe this will not really be needed and those nodes will be reached anyhow or through other means (e.g. POKE).
Excellent :+1:
cameel wants to merge ethereum/solidity
Remove enforce-via-yul CI run and instead always force via yul in regular test runs.
I noticed when https://circleci.com/gh/ethereum/solidity/1060053 ran out of memory. It's rather wasteful to run the smt tests in this run again - they're already covered and not affected by the option.
We might want to look for other cases like this... Also we could think about restricting this further... the flag really only affects the semantics tests, doesn't it? But most of it is cheap anyways, so the SMT tests are the most wasteful ones I guess.
Just to clarify this - this is not needed after all because the PR initially disallowed omitting compileViaYul
and now it made it default to also
instead. So we don't really need enforcement - the Yul run can only be be skipped if you explicitly request it.
cameel merge to ethereum/solidity
Remove enforce-via-yul CI run and instead always force via yul in regular test runs.
I noticed when https://circleci.com/gh/ethereum/solidity/1060053 ran out of memory. It's rather wasteful to run the smt tests in this run again - they're already covered and not affected by the option.
We might want to look for other cases like this... Also we could think about restricting this further... the flag really only affects the semantics tests, doesn't it? But most of it is cheap anyways, so the SMT tests are the most wasteful ones I guess.
pettinarip push ethereum/ethereum-org-website
commit sha: 035e14364be3cfda341d73b73d4693e9ce5c6716
push time in 2 hours agoasaj forked ethereum/go-ethereum
coveralls issue comment ethereum/solc-js
Fix CLI `--version` test to use version from the binary not `package.json`
Currently tests in cli.ts
are hard-coded to require a binary with version matching package.json
. This not true in general and causes problems in #598.
This PR changes the test to only require the output from --version
to match the output of version()
function in the binary. I also kept a modified version of the previous check as a sanity check - to match that the output actually looks like our version string.
Display human readable type name in conversion error message
Closes #12896.
As discussed, I'm adding in
string Type::humanReadableName()
, which defaults totoString()
and is overridden for error type.