Shivaram Lingamneni

It was the [Gluster Native Client](https://docs.gluster.org/en/v3/Administrator%20Guide/Setting%20Up%20Clients/). In principle I think this condition can arise with any networked storage layer: 1. write(2) partially succeeds, returns an error 2. the filesystem is...

I think I was hoping for something compatible with [errors.Is](https://pkg.go.dev/errors#Is)? Is there an obstacle to that?

The Gluster corruption happened again (still no logs AFAIK), so I did a first pass at a recovery tool here: https://github.com/slingamn/buntrepair

Thanks so much! I agree that this is more likely than concurrent file access. We were thinking that if concurrent file access were causing this, the second half of the...

Here's the offending line: https://github.com/spaolacci/murmur3/blob/f09979ecbc725b9e6d41a297405f65e7e8804acc/murmur32.go#L129 My understanding of the issue: the sanitizer expects that an `unsafe.Pointer` should be valid for a full word-sized read. So on a 64-bit arch, a...

Couple initial impressions: 1. Can you eliminate the while-true loop in run.sh? Is it just for resilience to crashes? 2. golden.py should ideally exit when bspwm exits --- I think...

It was probably a `wifi.ap.new` but I'm not entirely sure. The reason I think it's a data race is: 1. It happened inside [json.Marshal](https://golang.org/pkg/encoding/json/#Marshal); the contract for `json.Marshal` makes it...

Yeah, this one really surprises me because whenever I've seen a data race cause panics in the past, it's always been map access. The time window in which a map...

We're based on 3ac520c744ee3f and seeing a similar issue: https://gist.github.com/slingamn/34dd0eb3aedbb3f7763188843682db6e Thanks for your time!

Definitely worth considering if it can help persuade people to abandon their terminal clients or bouncers :-) Joining the channel would be gated on having the `snomasks` operator capability (the...