Freaky

    -% scan-build10 make
    scan-build: Using '/usr/local/llvm10/bin/clang-10' for static analysis
    ...
    scan-build: No bugs found.

And back to nix commands, libxo is used by a chunk of the FreeBSD base tools to offer output in JSON, amongst other things: https://github.com/Juniper/libxo

    -% ps --libxo=json,pretty
    {
      "process-information": {
        "process": [
          {
            "pid": "52455",
            "terminal-name": "5 ",
            "state": "Is",
            "cpu-time": "0:00.00",
            "command": "-sh (sh)"
          },

    -% uptime --libxo=json,pretty
    {
      "uptime-information": {
        "time-of-day": " 8:34p.m.",
        "uptime": 1730360,
        "days": 20,

> https://hur.st/bloomfilter/

Blush. However will I cope with all this extra traffic?

One common way to improve bloom filter cache performance is to divide them into blocks - elsewhere is mentioned doing it to the cache line, but it would be interesting to see how much performance would be gained with a more naive approach, for instance, splitting the filter into 4KiB pages.

I've done this for disk-backed filters, but never looked to see if it improved performance generally.

> But those are not science fiction.

Stories based on physics and maths are not science fiction? Are you feeling OK?

> If you tell a story that is within the limits of what we know to be possible with the knowledge that we have right now, you will end up with a very boring story.

Um. Most stories in general stay within reasonable limits of what we know to be possible - are most stories very boring? You basically seem to be arguing that all stories that aren't fantasy are boring, which is clearly untrue.

> Interstellar, which makes for a nice story

Interstellar is tediously boring rubbish :P

> For example, try to write a story where the entire premise is that someone manages to construct an Alcubierre drive. You'd get ...Star Trek: First Contact.

What... no? The phase-space of all possible Alcubierre-drive fiction is not "Vulcans come to visit", any more so than the phase-space of all possible stargate fiction is "Ra gets quite angry".

Does mentioning your faeries are made out of atoms stop them being fantasy?

> That's how it goes with sci-fi, innit. You can come up with anything you like as long as it's obvious that it's just advanced technology, not magic.

Sure, I mean, even Star Wars is still considered within the genre despite it all being completely made-up - but the point of hard sci-fi is you're not just making it up and rubbing science-words on your endless stream of arbitrary plot contrivances.

... the primary weapon of the Trisolans in the first book are subatomic ansible faeries.

I thought it was kind of funny, because the whole point behind the Dark Forest theory is that it emerges because aliens are so unknowable, there can never be proper trust between them. Yet a relatively young and slowly-developing race is literally more sure of what humanity is doing than we are of our neighbouring countries.

"Hard sci-fi", pfft :P

https://www.tinc-vpn.org/documentation-1.1/Simple-Peer_002dt...

Maybe we just need the right sort of developmental pressure.

Yeah, it's just k single-bit lookups - ideally you do something to get them into clusters, like dividing the database into sub-filters, so you're doing random lookups into, say, a 32KB chunk instead of a whole 2GB filter.

> Did you look into Cuckoo filters as well?

Cuckoo filters look like an interesting alternative and looking at them more closely is on the to-do. I don't think they'd have any significant space savings, though - they're similarly about 75% the size of the equivalent bloom filter. Maybe they'd be faster for lookups?

I'd also be interested in playing with matrix filters[1], which supposedly get close to the theoretical limits for these sorts of structures. Implementing them seems rather more involved, sadly - particularly given the only reference I can find is a fairly inscrutable CS paper. Show us the code damnit.

[1]: https://arxiv.org/abs/0804.1845

With that sort of FP rate it's not really much use beyond filtering API calls. I'd suggest 2GB[1] as a more sensible minimum. A compressed filter can get this down somewhat.

> Binary-searching the whole dump is surely faster.

Not really. log2(500M) is ~29, k for a suitably sized bloom filter's only 23. Interpolation search can get you a result in more like 10 seeks, but a bucketed bloom filter can get your lookup down to a single read.

Having spent a fair bit of time faffing about with this stuff I ended up settling[2][3] on Golomb compressed sets[4], which can get the full list with a 1-in-10 million FP rate into 1.5GB.

[1]: https://hur.st/bloomfilter/?n=500M&p=1.0E-7 [2]: https://github.com/Freaky/gcstool [3]: https://github.com/Freaky/ruby-gcs [4]: http://giovanni.bajo.it/post/47119962313/golomb-coded-sets-s...

I made a start on a Ruby port too: https://github.com/Freaky/ruby-gcs - I have vague plans to finish it off and write a Rodauth (http://rodauth.jeremyevans.net/) plugin for it.

    res, err := http.Get("wcpgw")
    defer res.Body.Close()

    ...

    panic: runtime error: invalid memory address or nil pointer dereference
    [signal SIGSEGV: segmentation violation code=0x1 addr=0x40 pc=0x5ed9af]

When Python 3 turned up and it was on average slower, my first thought was "well, good luck getting people to adopt that".

https://developers.cloudflare.com/1.1.1.1/dns-over-https/clo...

> Is there any plan to do a source-code release of any of the games you've re-released?

> We have released the source to Strife: Veteran Edition[1] and we are planning on releasing more code in the future.

1: https://github.com/svkaiser/strife-ve

> I have put the team on a hiatus while we reassess our path so that we can return to our vision. We are taking a break, but NOT ending the project. Please accept my personal assurance that we will be back and stronger than ever. System Shock is going to be completed and all of our promises fulfilled.

> "Our intention is to ship exactly the game that was promised, with as much of the features that were promised as we can, in a timeframe that will get it out as fast as we can. Our expectation is probably Q1 of 2020."

> Kuperman gave that date with the kind of expression and shrug that conveyed: no, that date isn't what anyone wants to hear. But it's practical. He was unequivocal when I asked if another two years of development would put Nightdive in a financial bind.

> "No it doesn't. We've also had renewed interest from some publishing partners, for whom the more complex and costly game was something of a challenge for them, that this is what they were hoping we would provide and renewed their interest."

    command="/usr/local/bin/borg serve --append-only --restrict-to-path=/home/backup/bla",restrict ...

Do you get enough control do to that on rsync.net?

https://www.tarsnap.com/man-tarsnap-keymgmt.1.html

e.g. you may give each host their own write-only key so they can automatically create new snapshots, while keeping the keys that permit reading and deleting old snapshots on separate machines with separate passphrases.

They're system-pulls off eBay, and appear to be some specialist OEM part - note they don't appear on ark, they go L5630 straight to L5640.

Not listed :/