-
Notifications
You must be signed in to change notification settings - Fork 110
Conversation
Geth team meetup is next week. I think we better wait for that and split the codebase, rather than pollute go-ethereum/.travis.yml with repo-specific config. I am not sure this will be approved upstream, and is not up to us. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nicely done. I'd add the race detector to p2p
and cmd/swarm
too
If ethereum/go-ethereum#19143 gets merged, then the Note: about keepalive Currently, looking at the test output it seems we might not need the keepalive script if we keep I guess that test and the ones coming after do not produce output because a lot of tests runs in parallel and output can be published only after the tests pass or fail. What do you think? Should we keep the keepalive script or change tests?
|
@nonsense as the code split is undecided and I want to close this issue down, I want to progress and submit upstream. I'm fine with dropping the |
a823a85
to
26c46fe
Compare
I did a rebase/cleanup just to keep the upstream PR cleaner, as we have to interact with others on this. |
`go test -race ./swarm...` sometimes fails because there is no output produced for 10 minutes and Travis just aborts the build. As `t.Log()` holds the buffer until the test does not pass or fail.
26c46fe
to
2a41455
Compare
We had more than a dozen of Data Races in our code base. Go has a nice builtin race detector tool. So we should have a CI job that at least notifies us timely if we introduce a new race. 1st let's just have a daily CRON job triggered on ethersphere/go-ethereum. Later as the job gets stable, we can start using it for ethersphere PRs. resolves: #741
As we only deal with artifacts in the upstream repo.
As we only deal with artifacts there and secretes are not even available on our forks.
As the completion time is around ~15m currently, so we can run it against every push/PR. Still, just on ethersphere.
upstream: ethereum/go-ethereum#19148 Moved there as we need reviews from the geth team, too. |
We had more than a dozen of Data Races in our code base. Go has a nice builtin race detector tool. So we should have a CI job that at least notifies us timely if we introduce a new race.
1st let's just have a daily CRON job triggered on ethersphere/go-ethereum. Later as the job gets stable, we can start using it for ethersphere PRs.resolves: #741
Known Issues
Node.Up
- p2p/simulations: fix intermittent data races related to Node.Up field #1146 also solves