HNHacker News
TopNewBestAskShowJobs

arghwhat

9,034 karma · joined January 12, 2017

no.
submissionscomments
arghwhat··on RedoxFS is the default filesystem of Redox OS, inspired by ZFS
I also imagine you wouldn't benefit from ZFS there either, even if the ARC wasn't there. You have a single application and can presumably accept occssional data loss (just fetch content upstream). Just need to handle bitrot detection, but there's ways to get around that application-side.

Better to just have the filesystem get out of the way and just focus on being good at raw I/O scheduling.

I wonder if FreeBSD is going to get something io_uring-esque. That's one of the more interesting developments in kernel space...

arghwhat··on RedoxFS is the default filesystem of Redox OS, inspired by ZFS
ZFS managing its own cache that sidesteps the existing pagecache infrastructure and requires dedicating more memory for it to function well than you otherwise would with any other filesystem.

ZFS sidestepping conventional device and mount handling with the way it "imports"/"exports" ZFS pools, usually auto-mounting all datasets straight to your system root by default - if you have a dataset named "tank", it automounts to "/tank".

ZFS operation itself being an inexact science of thousands of per-dataset flags and tunables (again not set through the common path with mount flags), and unless you run something like TrueNAS that sets them for you it's probably best to pretend it's not there.

Common configuration with decent performance commonly involving complexities like L2ARC.

It's far too invasive and resource consuming for a general purpose machine, and does not provide notable benefit there. A dedicated file server won't care about the invasiveness and will benefit from the extra controls.

btrfs fell short by still not having its striped RAID game under control, only being "production grade" on single disks and mirrors for the longest time - probably still?

arghwhat··on RedoxFS is the default filesystem of Redox OS, inspired by ZFS
If it relied on OpenZFS, then I wouldn't be too surprised.

The whole ARC thing for example, sidestepping the general block cache, feels like a major hack resulting from how it was brutally extracted from Solaris at the time...

The way zfs just doesn't "fit" was why I had hope for btrfs... ZFS is still great for a file server, but wouldn't use it on a general purpose machine.

arghwhat··on SonyShell – An effort to “SSH into my Sony DSLR”
The digital camera was invented in 1968, the first auto-focusing SLR is from 1978.

Auto-exposure came a few years before auto focus, but not by much as it still required not only light metering but also multiple mechanical couplings to the lens.

Such cameras (like the famous Canon A-1) were never refered to as "digital cameras", and SLR was used to distinguish from rangefinders primarily.

arghwhat··on SonyShell – An effort to “SSH into my Sony DSLR”
No - it's "single-lens reflex", a single attribute without comma. That is, a mirror redirecting light from a single lens to different targets (viewfinder and film/sensor).

If "four out of five words" was good enough, then an analogue film camera is also a DSLR - it's only missing the "digital" part after all. Seahorse and racehorse is also just one word apart, so basically the same, right?

arghwhat··on SonyShell – An effort to “SSH into my Sony DSLR”
The SDK does not support any DSLR's, and Sony announced its last DSLR camera 10 years ago.
arghwhat··on Find SF parking cops
Care to elaborate, as that sounds rather extreme?

I don't know of a country that requires all bicycle parking in any non-private location to be paid, nor a country that requires payment for roadside parking on country roads outside cities. Heck, even within cities, only the very dense ones seem to require paid parking on smaller roads.

Public space does not imply free of any use, but rather that it is freely used by all. The purpose of paid roadside parking is to reduce demand on what quickly becomes a limited resource in dense cities.

arghwhat··on Find SF parking cops
Here (Denmark), blocking a drive-way would not be a parking violation handled by a parking inspector, but a traffic violation handled by the police. They would contact you directly and have you sprint to your car.

Parking tickets are also considered fees, paid to those managing the parking area (municipality for public roads), as opposed to fines issued by the police or a judge and subject to very specific rules.

arghwhat··on Find SF parking cops
Where I'm from there's no concept of towing on any parking violations. No one has the right to touch your vehicle, and if the police is involved they'd just contact the owner. The police would only move a car if it's a danger to the public.

A tow truck is only something you'd call for assistance, not something you fear seeing.

(Parking fines suck, but the municipal ones are usually more reasonable here, even if they don't always get the rules right. It's the parking companies managing large private parking lots, often for free to the lot owner, that are absurd.)

arghwhat··on Want to piss off your IT department? Are the links not malicious looking enough?
Imagine the legal consequences too, when the services you host make sent personal data to an (otherwise valid) data processor, but surprise the network-wide policy sent traffic went through a random third party that is not part of the Data Processing Agreement and privacy policy given to the end-user/data subject...
arghwhat··on Want to piss off your IT department? Are the links not malicious looking enough?
I do not believe this is a trade-off, I believe this behavior from corporate IT is a primary cause of the problem. I do agree that dealing with users is awful, but that doesn't justify solutions that only make things worse.

The flashing cmd.exe windows are not drivers from Windows Update - this could have been the case as drivers shipped with Windows Update is a total security nightmare running arbitrary code with administrative privileges upon hotplug - but in this context of managed devices, commonly from HP or Lenovo's corporate portfolio, it's usually additional products and changes pushed by group policy or random management software.

The often changing user prompts, looking like they were from some early 2000's hello world example, come from obscure and overlapping management software they remotely deploy which they change at will. You don't need a proprietary solution to remotely upgrade Google Chrome, just specify an enterprise policy with auto-update. You don't need a proprietary solution to prompt users and manage Windows Update, because you have... Windows Update.

The URL checker has no valid benefit, and makes it so people can never learn how to do it themselves. The browser performs the exact same checks with the exact same capabilities through its safe browsing stuff, and corporate IT often has network-level solutions too.

Corporate IT uses emails services that spoof domains and look suspicious, reversing all the phishing training they paid for.

Not to mention that Corporate IT might deploy network-wide solutions like Cisco Umbrella, which is a TLS Man-in-the-Middle attack where you install their root CA on all machines and let them control DNS to randomly redirect all traffic to their servers, effectively undermining the basis of all modern web security for the entire organization.

In general there's a fetish for buying products that has significant negative impact to security, user experience and possibility of training users, usually for the purpose of feigning progress and meeting some targets. Say, they had a ransomware incident, and now they're buying every ransomware product for a few years. Stuff with such buggy kernel code that it deadlocks and makes it impossible to create new processes until you hard reboot. I'm sure that's not a security problem!

arghwhat··on Want to piss off your IT department? Are the links not malicious looking enough?
Modern auth systems do not work by exposing multiple services on a single domain with shared cookies.

Instead, they authenticate using a common auth service (say, auth.google), which by virtue of being a single domain can persist shared cookies for all its consumers. This would yield a valid token (possibly a JWT) that the authenticating application can then use however it would like, including as a cookie on the application's own domain.

Whenever you go to a service that temporarily sends you to a different login domain (often just immediately redirection you back), this is why.

arghwhat··on Want to piss off your IT department? Are the links not malicious looking enough?
Yay for third-party email services that From: be a no-reply address from an entirely different company (and therefore only authenticity validation for that company), and a Reply-To: to some obscure mailbox from the supposed sender. I'm sure that makes perfect sense to most people.

> So the rule of "don't click links in emails" is the only foolproof rule.

The only truly foolproof rule is "don't open emails". Also helps a lot on mental health and associated expenditures!

arghwhat··on Want to piss off your IT department? Are the links not malicious looking enough?
Ah no need, corporate IT already make all URLs malicious looking through some microsoft "secure link" service, and constantly shows everyone shady looking prompts that constantly change and have cmd.exe windows flash in at random.

A phone call from Microsoft about my Norton anti-virus subscription putting me into debt that can only be settled with Nintendo gift cards bought in cash across 16 specific gas stations seem much more legitimate in comparison.

arghwhat··on I just want an 80×25 console, but that's no longer possible
It is very cute. It doesn't take much to use a serial console though, or even enabling SPICE for a fully featured (and detachable).
arghwhat··on I just want an 80×25 console, but that's no longer possible
If the kmscon project wasn't dead, they'd be able to get whatever number of cells and rows they'd like as it doesn't use now-obscure built-in display device modes by setting the font settings.

On the other hand, kmscon is just a graphical terminal emulator and display server built into one. Running `cage foot` (i.e., the very minimal "foot" terminal emulator running under the single-app "cage" display server) and setting foot's settings to whatever you'd like would get you a better (and importantly, fully maintained) experience.

arghwhat··on AMD Open Source Driver for Vulkan project is discontinued
They already work on radv, which is already the better vulkan driver.

This is a matter of AMD no longer wasting time on a pointless duplicate project no-one is really interested in. They can allocate more resources for amdgpu and radv and ultimately do less overall by getting rid of the redundant project.

Win-win.

arghwhat··on Why our website looks like an operating system
I read it in full before posting. Writing about something doesn't make it a better idea afterwards.
arghwhat··on Why our website looks like an operating system
Oh god. It has a pleasant color scheme, but this is an awful idea. By trying to recreate windows and bookmarks in the web app you're at best just implementing redundant features and getting in the way of the native browser features by trying to showcase yours, at worst breaking regular web usage entirely.

Take their right click menu for items to select whether you want an in-app tab or real browser tab. Congrats, you've broken UX by making the native browser right-click menu unavailable on link items, and because you've only implemented this on some things most of your content is not deep linkable as navigation is a cursed in-app feature.

This is as usual a fun tech demo, but it should not be used for anything in the real world.

arghwhat··on Why our website looks like an operating system
> Many websites are free because they survive from ads. Ads make more money if you collect data.

My business is to get money out of other people's wallets and bank accounts. I could get make much money if you just logged into your bank account and approved transactions whenever I told you to, or screamed less whenever I took the wallet out of your pocket on my own.

That there's a way to earn more money does not justify it as legitimate thing to do, and if you can't figure out how to run a service in legitimate ways does not mean that illegitimate ways that attempt to violate its users in secret suddenly become okay.

arghwhat··on Why our website looks like an operating system
No, it is not that. It highlighted an issue, and it makes it painfully obvious when a particular page is being extra ignorant about your privacy and trying to sell it to thousand vendors instead of a handful.

What I don't like about cookie popups isn't the popup (which isn't something the EU law dictated btw), it's that someone thought it was okay to have hundreds of advertisement vendors and data brokers on a single news article, and it's better to know so I can just close the tab and never interact with that webpage again if they're being excessive asshats.

They have failed at enforcing this properly though, in particular with the recent proliferation of "legitimate interest" abuse (it is only legitimate interest if it an implied component to a service I am directly requesting), and the general issue of popups illegally making rejection different from acceptance, intentionally making rejection slow, or even requiring payment to continue without cookies. And yes, the occasionally completely defective prompt.

I do agree that it would be neater if the browser handled this though. Would also be neater if the internet wasn't entirely sponsored by privacy violations. :/

arghwhat··on Danish supermarket chain is setting up "Emergency Stores"
MobilePay starts by being direct bank transactions, and can still use this to some extend. Now it also has credit cards associated though, which depends on Nets to clear.
arghwhat··on Danish supermarket chain is setting up "Emergency Stores"
Nets sold its account-to-account payment business to Mastercard. Nets, which is itself a private company operating in many countries, has not been sold to Mastercard, nor has Dankort, MitID, or others been sold to Mastercard.

The services sold are, for example, Betalingsservice in Denmark and AvtaleGiro/eFaktura in Norway.

To be honest, my concern is that Nets is that Nets is entirely unable to competently provide any services competently and reliably, so while they shouldn't be moved to a US company, I sure do hope that we get everything they do replaced by something else.

Only Nets could somehow manage to make it much more expensive in fees to have a scheduled direct bank transaction (Betalingsservice - for reference, bank transfers are free and instant here) than using a VISA card with fees going to the US. They have also completely ruined Dankort adoption and reliability, and made MitID into a great big mess of a reliability and usability nightmare.

arghwhat··on Hypervisor in 1k Lines
That it's fine to do things just to have fun an mess around, but thats all there is to it. It's not even an OS at the current state.

Other small OS's are actually OS's and can be impressive efforts. Sometimes they bring new concepts or paradigms, but that itself is rarely more than a curiosity without impact. Still fun though, at the very least for the authors.

arghwhat··on Hypervisor in 1k Lines
This is a cool intro to how virtualization instructions work, but people need to understand that it is not revolutionary simplicity or anything like that. It's just a a cool tech demo lacking all the important bits of a modern hypervisor required to make it practical, like paravirtualized drivers for example.

Same when someone claims to have written an OS in 1000 lines and all it does is get you to real mode with VGA graphics and an interactive (but useless) prompt.

(Note that you can benefit from virtualization technology in specialized scenarios outside common hypervisors, but that's not really what's being demoed here.)

arghwhat··on Liquid Glass? That's what your M4 CPU is for
> Do you work on these kind of optimisations for a modern OS

I have worked on modern display servers and application interfaces, and as such also dealt with (but not written) a fair amount of client application render code, generally optimizing for power consumption and latency.

> I honestly couldn't imagine this kind of compositing not happening completely on the GPU or requiring any back and forth between the CPU and GPU.

Well the CPU is always involved, and even responsible for rendering certain assets that are not viable to render on the GPU - fonts in particular are usually rendered by the CPU, with occasional attempts at GPU rendering being made such as servo's pathfinder - but for simplicity let's talk about only widgets rendered by the GPU.

In most cases[^1], a window is rendering to a single "render target" (texture/buffer), handing off a flat and opaque buffer to the display server which the server will either try to display directly (direct scanout) or composite with other windows. In this context, the display server's main purpose is to have exclusive control over display hardware, and apart from compositing multiple windows together it is not involved in render processes.

The application itself when rendering will normally walk its widget tree, accumulate damage and cull widgets to create some form of a render list. Depending on the toolkit and graphics API in question, you'll ultimately end up submitting an amount of GPU work (e.g., a render pass) to render your window buffer (an IOSurface or DMA-BUF), and you then send the buffer to the display server one way or another. The window buffer will become ready later once the asynchronous render tasks complete, and the display server will wait on the relevant fences (on the CPU side, which is responsible for most GPU scheduling tasks) before starting any render task that would texture that buffer or before attempting scanout from the buffer.

The problem with blur is that you have a render task that depends on the full completion of all the prior render tasks, as its shader must read the output buffer[^2] as it completed rendering to an intermediate state. Additionally, other render steps depend on that render task as it has to be overlaid on top of the blur widget, and only after that completes is the buffer ready for the display server. That's a pipeline stall, and because it's on top of the primary content, it's holding up every frame from that app, and due to the blur operation itself an update that before only affected one tile now affects several.

Reading your own output is something you avoid like the plague, and blur is that. If you're used to web/network development, think of it like blocking network roundtrips.

... well this turned out to be a wall of text ...

---

^1: The more advanced case of hardware compositing is where you send a small number of buffers to the display server, e.g. a video buffer, an underlay with some background and an overlay with some controls, and have the display server configure hardware planes using those buffers such that the window is stitched together as the signal is about to be sent to the display. This is not the general case as planes are limited in count and capability, they cannot perform any effects other than basic transforms and blending, and scanout hardware is very picky about what it can use as input.)

^2: One could implement this instead by creating a different render list for just the area the blur needs to sample instead, in the hopes that this will render much faster and avoid waiting on completion of the primary buffer, but that would be an app specific optimization with a lot of limitations that may end up being much slower in many scenarios.

arghwhat··on iPhone Air
The sum of all edge nodes exceed the power in the datacenter, but the peak power provided to you from the datacenter significantly exceed your edge node capabilities.

A single datacenter machine with state of the art GPUs serving LLM inference can be drawing in the tens of kilowatts, and you borrow a sizable portion for a moment when you run a prompt on the heavier models.

A phone that has to count individual watts, or a laptop that peaks on dual digit sustained draw, isn't remotely comparable, and the gap isn't one or two hardware features.

arghwhat··on Jeena's Hyprland Demo
> The only thing I feel Sway’s behind in now is its lack of per-window screen sharing.

That's available in sway/wlroots master with xdg-desktop-portal-wlr master. Will be in next release.

arghwhat··on Liquid Glass in the Browser: Refraction with CSS and SVG
Neat! Doesn't look particularly pleasant as the edges have high contrast single-pixel width artifacts, but maybe it's possible to smooth that out. Could also just be a quirk in chrome's svg/backdrop filter support that they'll fix.
arghwhat··on Liquid Glass? That's what your M4 CPU is for
It's not an uncommon terminology in the WSI portion of graphics APIs, I was just pointing out that it doesn't actually mean anything to the hardware/lower stack. There are only buffers.

(There can be restrictions on which buffer formats and layouts can be used for certain things, scanout being particularly picky, but a regular window can be textured from pretty much any buffer.)

← PreviousPage 7 of 34Next →