Right now the docs correspond to what's on master. We need docs for each daemon release so developers can see what's available in the version they are using.

https://github.com/jimporter/mike may be helpful.

Because in CliClient.resume() we don't account for the case, when there is an exception in one of the async processing stages:

void resume() {
	...
	try {
		client.startAsync(state -> {
			printer.ifPresent(p -> p.print(state));
			if (!options.shouldSeedAfterDownloaded() && state.getPiecesRemaining() == 0) {
				runtime.shutdown();
			}
		}, 1000); // <-- CompletableFuture.w

Currently we don't have any mechanism to limit the maximum number of clients that could be handled simultaneously.

This feature should be designed properly. Here is some clue: https://redis.io/topics/clients#maximum-number-of-clients

Steps to reproduce:

1. hc init my-dna
2. cd my-dna
3. hc generate zomes/kebab-case
4. hc package

Output:

> CARGO_TARGET_DIR=${CARGO_TARGET_DIR:-/tmp/my_first_app/target} && echo $CARGO_TARGET_DIR 
"CARGO_TARGET_DIR=${CARGO_TARGET_DIR:-/tmp/my_first_app/target} && echo $CARGO_TARGET_DIR "
/home/paul/target
> CARGO_TARGET_DIR=${CARGO_TARGET_DIR:-/tmp/my_first_app/target} 

众所周知，tracker也是会被封的，所以每天获取有效的tracker能提高效率，至于tracker列表的获取，看这个
https://hao.su/3278/

Currently lokinet uses llarp_time_t (which is milliseconds) everywhere for times. Part of the justification was that it makes serialization easier, but conversely it makes other parts of the code a bit more complicated.

It would be nice to refactor so that std::chrono duration types get used instead (for compiler-defined duration cast math), being mindful of the existing serialization requir