Hector Sanjuan

Results 424 comments of Hector Sanjuan

Some questions upon reading the notes above: Is the connection rate limiting also on localhost? i.e. running >16 peers in a single machine will mean they cannot connect to each...

Triage notes: * Not super urgent * There is a address factory in p2p-forge/client that should probably receive "NoAnnounce" and match produced addresses against it. * Community contributions welcome

* Triage notes: * Team member will expand on security considerations from this plus other concerns. * We don't like to provide ways to set arbitrary digest lengths even if...

Hmm yeah, I think we have an example in Rainbow to build using buildx: https://github.com/ipfs/rainbow/blob/main/Dockerfile https://github.com/ipfs/rainbow/blob/main/.github/workflows/docker.yml This implies disabling dockerhub builds and pushing from Github actions instead.

We did this a while ago, forgot to close this one. https://github.com/ipfs-cluster/ipfs-cluster/blob/master/.github/workflows/docker-image.yml

No, but this it worth exploring as an alternative to the --allocations flag. I.e. `--allocations tag:region=eu` or something that limits the set of peers considered to those with certain tags.

I cannot reproduce this. I don't think this is a crash, I think this is a non-0 exit. It is not a panic, yet it is printing the stack trace....

I have no idea... I researched that a bit and did not readily find that you could make go print stacks on non-clean exits with some env-magic.