-
Notifications
You must be signed in to change notification settings - Fork 157
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Feat/evmc6 #57
Feat/evmc6 #57
Conversation
govendor fetch github.com/ethereum/evmc/bindings/go/evmc@=v6.0.2
Submods, gomods, oh my! (Mostly vendor/dep management) I used a submodule with go mod's replace directive because I want to first establish an MVP implementation of the existing PR as-is. Issues found during testing, design and architectural questions are unstable and still open for discussion. Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
As noted in the comment, this may not be the right gitm commit -S -s -m core/vm: Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
Just a little walkaround of what we've got here. This introduces a new submodule at Consequences of this are:
I don't love this. I would prefer build (and test) requirements to be as simple and standard as possible. Since we're eternally-pegged at v6.3.1, a solve for this would be to include the dependency as an in-place fork. Precedent for this pattern exists in the codebase already in
Again, I think this is ugly. IMO |
Signed-off-by: meows <[email protected]>
…/evmc.git And rm -rf evmc/.git .gitignore Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
…lative uses Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
Docker image complaining about 'bash' not being available. Signed-off-by: meows <[email protected]>
'make test' should run the standard go tests. It shouldn't do anything fancy. This moves the evmc-specific testing to one place, ensuring that tests both with and without the example_vm.so are run. Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
Since emvc/ is now a copy-cloned package instead of a submodule, this is no longer necessary. Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
This is congruent to go-ethereum implementation, although it does seem a little redundant/tangential. Signed-off-by: meows <[email protected]>
I'd like to remove Github Actions as the CI for EVMC6 tests, since that's all its currently used for, and I think it'd be good to have our CI tests consolidated at one provider, if possible. This keeps CI visibility simple and predictable. Signed-off-by: meows <[email protected]>
This improves CI readability. Signed-off-by: meows <[email protected]>
Remove the subjectivity a little, more descriptive. Signed-off-by: meows <[email protected]>
…patible Test failed on Travis because unknown operator. https://travis-ci.org/github/etclabscore/core-geth/jobs/710376273 Signed-off-by: meows <[email protected]>
… on Travis We can stick with Github Actions for now to avoid the CI headache. Signed-off-by: meows <[email protected]>
…tateTest This installs a new test for --evmc.evm flag. It uses the ethereum/evmcone C++ standalone EVM .so artifact. Tests for Constantinople, Istanbul, and Phoenix are skipped because they are not supported by this latest-possible evmcv6-compatible EVMOne version. Signed-off-by: meows <[email protected]>
This moves the logic of skipping forks from adhoc in the StateTest runner (where it awkwardly used the testMatcher method in a 'special' way), to still using the testMatcher in a special way, but now more descriptively as a logic facet of the StateTest type itself in the Subtests method. Signed-off-by: meows <[email protected]>
Signed-off-by: meows <[email protected]>
…config This change is not related to the EVMC feature, and does not change test behavior, so striking it. IstanbulBlock existing is a no-op, since difficulty did not change at that fork anyways. Signed-off-by: meows <[email protected]>
Rel #55 Support EVMC6