Henry
Hmm.. ok but it seems like there is a potential fix here, tho. That's definitely climbing too fast. I have the new room server on the plate for another couple...
## Margaret does too much Another problem is that [margaret](github.com/cryptoscope/margaret) (our flume-like thingy) also does object decoding and exposes it through empty interfaces, which makes it hard to re-use buffers...
Yes! Should probably `internal`ize all these things to make this clearer.
I'm experimenting with making copies of `luigi.Broadcast` with concrete types on [the new rooms implementation](https://github.com/ssb-ngi-pointer/gossb-rooms). Much more readable code..!
Since messages are appended to the root offset log first and indexes updated after that, it's hard to diagnose with a good reproduction story. In the normal cases the KV...
There is no _filesystem check (fsck)_ code on `package sbot` and the `go-sbot` tool using `-fsck` and `-repair. The root cause is still diagnosed even though I tried multiple times......
I tried very hard to get something semi-stable as the NGI Project was wrapping up but there was a lot of rush and things might have slipped through the cracks......
> patchwork has ebt now, so time pretty safe to disable legacy replication I think! Personally I'd like to -1 on that. Feels to early. Our Go port only barely...
IIRC this is related to ssb-ebbt replication. Looks like it doesn’t close down properly..
I think _the parent stream_ is the one returning `Error: stream ended with:13 but wanted:34` but I have no idea why it is closing too soon...