Leo Balduf

v0.13.0 seems to ignore `Swarm.ConnMgr.Type == "none"`

3

### Checklist - [X] This is a bug report, not a question. Ask questions on [discuss.ipfs.io](https://discuss.ipfs.io). - [X] I have searched on the [issue tracker](https://github.com/ipfs/go-ipfs/issues?q=is%3Aissue) for my bug. - [X]...

storage/redis: reimplement using scripts

2

This possibly fixes #471, and also possibly increases performance, but who knows...

Implement HTTP fullscrapes

1

Fixes #369 (eventually) WIP, but need travis

Storage benchmarks

1

Fixes #447 This also adds a small variant of the announce and scrape benchmarks. From the data we gathered from public trackers, swarms are usually tiny (probably only having one...

udp: read into a large, shared slice instead of a small, per-request slice

4

Before: - use a `BytePool` for 2048-byte slices - use a new slice for each request - return the slice after failure/after the request is finished Now: - use a...

Internal errors might be reported to clients

2

This was the case for the redis storage implementation, and it's probably also true for anything else on the announce path that can produce errors. I think we should come...

With redis coming in, we need a way to clear the storage in between benchmarks.

Evaluate different logging libraries.

2

As it stands now, all our calls to `pkg/log` allocate, regardless of whether or not logging will be done. In addition to that, because we use logrus, logging is pretty...

Currently we reject an (otherwise valid) UDP announce if we don't understand the BEP41 bytes. We could instead ignore the parsing error, or maybe even attempt to just ignore the...