protocol

Creating a new issue based on lbryio/lbry-sdk#877

• Update ensure_directory_exists() to check that the directory is writable by the current process.
• If it's not, return a useful error
• Add tests to cover the case when
  • download directory does not exist
  • download dir exists but is not writable
  • dir exists an

In the output of list channels the private/public status of the channel is returned, but in the output of pending channels the status is not revealed. To allow for recognizing a pending open channel as a private requires waiting for the channel to confirm.

Ideally the PendingChannelsResponse would include the channel flags or the private/public status of the channel to know which type of chan

The current downloads listed in the install section of the readme are version 0.17 while the current version is 0.18.1.

I think we should not link specific versions here, but link to the release list on Github.

Describe the feature you'd like supported

We should change the CID after enough idle time so that any possible NAT rebinding would be harder to be tracked by an outside observer. The simplest way to implement this would be in the flush send code path. When first starting a flush, check how much time has elapsed since the last send. If it's greater than X (20? 30?) seconds force a CID update

Tokio's automatic cooperative task yielding feature can cause abruptly prevent a task from making further progress. A very busy connection driver might as a result end up failing to fairly serve transmits or timeouts, resulting in bad behavior (e.g. spurious timeouts or congestion window reduction) under heavy load.

Ideally tokio will expose some too

There are multiple functions that return Result<T, ()> in the code, this is equivalent to Option<T>, however instead of changing those to optional types, these should return useful error messages/types so it's easier to debug.

Example: https://github.com/graphprotocol/graph-node/blob/79390c39a13f166999ceea8defc55eedd05265ab/graph/src/data/subgraph/mod.rs#L169

Some of them could just retu

Seems that Firefox (around version 96-97) changed somethign and is now way more aggresive when a gateway does not start returning response immediately.

For example, first load (empty ipfs repo) of http://dist.ipfs.io.ipns.localhost:8080/go-ipfs/v0.11.0 in Firefox 97 produces:

![2022-02-11_19-09](https://user-images.githubusercontent.com/157609/153646056-9bc942ec-bec7-4f0a-b61b-53db7bf6896

I'm mocking a protocol with a function that looks like this:

func foo(
    completion: @escaping (
        _ bar: Int,
        _ baz: Int,
        _ qux: Int
    ) -> Void
)

The generated mocks do not properly remove the line breaks before attempting to put it into a single-line string literal, resulting in the errors Unterminated string literal, `Consecutive statement