melchebo

Btw, the bug itself was introduced in 30656c5e by Claude Code, and I guess.. improper human review and testing.

https://github.com/RsyncProject/rsync/commit/30656c5e

Someone using AI to bisect recent rsync. https://github.com/themgt/rsync-compare-link-dest-341-343-re...

Someone trying to fix it with more Claude Code: https://github.com/RsyncProject/rsync/pull/930

Related ticket: https://github.com/RsyncProject/rsync/issues/915

I'd recommend putting in more regression testing in the commit before 30656c5e, and rebase it forward while keeping functionality.

First question, bomb shelter??

Next what popped in my head, “Military grade means, made by the lowest bidder.”

Do not connect it to the internet. Problem solved.

Basically anything in a social network needs to learn to defend itself against threats. Make computer a hermit, and it can go without updates for a long time.

(Oh, but you don't like that? Well, Microsoft doesn't like getting in the news for some worldwide botnet of all Windows 10 machines. I bet they'll figure this out sooner or later.)

Maybe you'll like this, if you download this .html and view it, you'll see a random projection from a random spot on earth.

https://gist.github.com/HenkPoley/8fa7a4a8e25f106585584463c1...

Sadly mostly north oriented.

Did some more vibe coding: https://gist.github.com/HenkPoley/0a0eac0e81c53145dec8c19568...

Browser versions tend to burn more power.

LookingGlass' small portrait display is more like (2D) lenticular. Their larger displays use a micro lens system that directs pixels out in beams in different direction. So you also get parallax in the up & down direction.

I wonder what kind of rig is needed for recording that. It has to be at a least a few different viewpoints.

https://techcommunity.microsoft.com/t5/it-transformation/mic...

For the moment (well, february) the improvement with PHP 8.0 JIT is minimal: https://www.phoronix.com/scan.php?page=news_item&px=PHP-8.0-...

But usually a JIT is made in a way that you can then move forward with your optimizations in small tested steps. So it makes sense for JITv1 to have a very similar performance as the non-JITed version. It probably generates highly familiar code. First make it do the same, then optimize.

Slightly concerning that a few of their Netty benchmarks show high error counts. Usually meaning it could not handle the load, and behaved odd.

There a Turkish ham radio operators replying on the blog that they know of nothing of this kind. Maybe this "news" was made up?

Seemingly the only relevant reply in this comment section. That said, the article is kind of vague and has a few big misunderstandings unrelated to Tononi's Phi and Information Integration Theory.

Phi is basically a complexity metric that tries to put a number on the effect of memory in a system. I'm unsure wether this has any relation to 'consciousness', but it's at least interesting to measure effects of having a low or high Phi.

For example functional programming languages favor a low Phi. Everything is defined by the input, memoization is possible (sort of what you call 'mocks' in other programming languages). There are no side-effects (memory of past events elsewhere), or at least they have to be invoked explicitly as 'unsafe' operations.

Also, code that calculates Phi: https://github.com/wmayner/pyphi/blob/develop/pyphi/examples...

What do you think of their recent Swift Playground programming education effort ?

Chris Granger (of Light Table fame) and his bunch at Eve are working on something like that. There are a bunch of presentations on YouTube. They are more looking at it abstractly as a possible programming / data model, than purely 'compile an Excel sheet to a programming library'.

Eve: https://twitter.com/with_eve Chris Granger: https://twitter.com/ibdknox

Or donate some of the first blocks (s)he mined.

Just read the paper, it's explained. Or read the presentation, it uses pictures.

Basically they run R, a single threaded statistics tool which is setup to hog a core, and in some other cgroup a wildly multithreaded tool. If you have a NUMA system (check with `lstopo`) then it's possible that the scheduler thinks the many tasks in one domain of cores is balanced with just R on one core of another domain. Meaning you can have several (ex: 7 out of 8) cores idle. It has to do with the way hierarchical rebalancing is coded, and that their 8x 8-core AMD machine has a deep hierarchy.

Presentation PDF: http://www.i3s.unice.fr/%7Ejplozi/wastedcores/files/extended...

Most recent Intel cores actually appear to be slightly slower at integer performance. http://imgur.com/a/2fiLF

A lot of their bugs are induced when you have a cgroup with multithreaded/-process tasks and a cgroup with just a few using processor time.

Just testing one benchmark will not show it, unless you have something else running too.

It seems like everything with buffer/process queues needs a good polish to optimize for fairness.

E.g. this is just being readied: http://blog.cerowrt.org/post/fq_codel_on_ath10k/

(also note higher throughput as result)

You check wether your system in NUMA with `lstopo`.

That might be what they call the 'overload on wakeup' bug. Maybe try the patch? I've read that the patch additionally needs a small fix, the goto label position went missing.

Probably just before rcu_read_unlock() in that function: http://lxr.free-electrons.com/source/kernel/sched/fair.c#L51...

This research applies to 'NUMA' systems. Commonly servers with multiple physical CPUs that each have a connection to their own memory banks. They can access memory of the other CPU by requesting it, but that take time. So the process scheduler has to take that into account. Usually by keeping processes a slight bit affixed to the place where it was started.

Off-topic, but high-performance long running processes are mainly programmed in C, C++ & Java. Maybe stuff like Rust and Swift in the future. Fortran if you are doing mathematical computation, but then you'd probably already use it if you need it.

For what I estimate that you mean with high traffic on PHP or node systems on multiple servers, probably you want to look at Elixir and it's Phoenix web framework. It's more appropriate for responsiveness (as in low latency). And less boilerplate than Java. |> http://www.phoenixframework.org/docs/overview

From what I understand from this presentation the 'scheduling domain' abstraction is reused through different layers of the hierarchy. So for example the two hyperthreads on one logical core are also modeled as 'scheduling domain'.

https://events.linuxfoundation.org/images/stories/slides/lfc...

The scheduler looking at an idle core decides wether to steal work from an overloaded neighbor. It will only compare over the interconnects in the figure (between domains).

E.g. the the two cores in the dark grey box can steal work from each other. But they will only see load averages of the neighbouring domain. In certain cases the current scheduler calculates the load figures sort of odd, so the idle core decides that a neighboring overloaded 'scheduling domain' is not overloaded.

It should probably mainly be seen as a proof of concept for their scheduler decision visualization tools. Which are not public for the time being. That should make checking and fixing bugs easier in the future.

The conference where this will be presented starts next week. So I think they'll post it after they catched up some sleep after that.

From the looks of it, this mainly applies if you are running a NUMA system. Which you probably are not. Though it may apply in other places where 'scheduling domains' are used to model hardware (e.g. hyperthreading and maybe big.LITTLE).

The last paragraph of that segment says they pick it because it's the first in the list of idle cores. E.g. if a core is no longer idle it removes itself from the list, and it adds itself at the end, so effectively it's ordered in deminishing idletime time.

Though from my understanding of the patch they walk over all CPUs which is not precisely constant time, and it doesn't seem to be just the list of idle CPUs; for_each_online_cpu(_cpu): https://github.com/jplozi/wastedcores/blob/master/patches/ov...

That said from what I grasp of the code it also doesn't bail out when power management is enabled, so maybe the public code or the paper are not at the same stage of development.

eBay really dug themselves into that hole when they introduced the star rating system. Basically they mapped everybody who wasn't actually known to be bad onto 5 stars. Which is why you get messages begging you to rate them 5 star.

Personally 5 stars would be when the box it came in had water soluble ink and it made a mess on your clothes, but the company offered to pay the dry cleaner costs. Even though there was no expectation for them to know that the shipping company would leave package in the rain / a puddle for so long.

Anything above "meh, it didn't crash and burn immediately, I guess it's okay" is 5 stars on eBay.