Mark Schmale

Is there any way the client could detect its a proper replay id instead of just checking for /#\d+/? I guess, if possible, we should prevent creating links in messages...

I really like the idea of making. I'm currently failing to understand how this could work for decoding but encoding looks pretty good. It also enables lots of other use...

I guess you are using the angular framework in the browser. For this you would have to compile `bencode` using something like browserify or webpack since we are using APIs...

Okay, I just guessed (hoped) that webpack would use [feross/buffer](https://github.com/feross/buffer) to implement the Buffer api in browsers. Looks like I was wrong on that point then, sorry! I'd have to...

Yeah we kind of where working on that in ~2014 but never really made any good progress that does not break the interface or destroy performance for all other cases....

At first: sorry for the delay. The holidays... I see your point. Albeit that "i123ei123e" (AFAIK) is not valid bencode (since its actually two values), there are (at least) two...

in case we do (b), the parser would work as it currently does. There would be a bencode.decode.resume() function that checks if the internal cursor is at the end of...

while strict parser should choose 3., you should currently be able to call decode.next() and get the next result. Thats pretty close to what a `resume()` would be. Throwing an...

I think this needs a new merge from master (with conflict solving in encode.js) and we need to remove the float test since this breaks with bigint.

Thinking about it again. 2^53 bits are about 1 Petabyte. I guess we can assume that there just are no chunks with a defined size of that size?