Matthew Bryan

> I had a quick look at the recipes, and they look good, as far as I can tell. One thing I wondered - why is there a `skip: true...

> I'm not looked in detail how you use/create your test package. Please have a look on the discussion there [cedadev/nappy#59](https://github.com/cedadev/nappy/issues/59) > > In a package usually you don't have...

Hi @jan-car thanks for taking a look, and sorry for the delay I was away this weekend. So I've took a look at GH and I've rebased my MR on...

> (just a side node: there is an alternative dependency solver being worked on that is supposedly quite a lot faster: https://www.anaconda.com/blog/a-faster-conda-for-a-growing-community ) Just gave it a try, still slow...

Looks ideal and quite well developed. Have you been able to test it at all yet ?

Wondering if either there has been progress/planning on this or, if not, should we make a plan? I've been examining the architecture and think I have a reasonable idea of...

It's certainly instructive to try to represent the program flow like this, although the layout probably suffers a bit because I'm figuring it out as I go. I do wonder...

> The IOBackend code is a bit hairy It's not so bad! Although I haven't gone too far into how all the syncs, offsets and mmaps are handled as it's...

A quick question - what's the opinion/thoughts on 'missing' data from a streaming source (especially via UDP streams like the K2). I appreciate that data might come out-of-order and that's...

Another piece of info, registering the `executor.__del__` method to be run with the `atexit` module does correctly allow the script to exit and/or drop to a prompt (and later be...