macOS virtualisation refactored and sandboxed in Viable updates | Hacker News Reader

macOS virtualisation refactored and sandboxed in Viable updates

eclecticlight.co

89 pointszdw38 comments

> In case you think that the two apps can be used together to run more than two macOS VMs at the same time, I’m afraid that macOS remains one step ahead of you, and enforces Apple’s licence limit across virtualising apps.

Is this true for all VMs or just VMs running macOS? (How) does the OS know what operating systems the guests are running?

It knows because it's going through the MacOS Virtualization framework. There isn't a similar limit for Linux guests.

It's probably actually really easy to 'enforce' because the method for starting a MacOS VM is completely different than Linux: https://developer.apple.com/documentation/virtualization/vzm...

Yet another example of how macOS is slowly being dumbed down to be more ios like by restricting kernel extensions and forcing developers to use their limited OS APIs.

No.

Too bad I can't run multiple Vms at the same time :(

Virtualization definitely supports this.

This is super useful for me

It's such a shame that Apple feels the need to cripple their devices to protect their revenue. Mac Studio would be a super powerful machine, allowing me to run all my CI and testing on my local Mac, but the 2 VM limit prevents that. So for now I'm stuck with the Intel Mac mini, which doesn't have this stupid limit.

I bet the machines they use for Xcode cloud don't have this dumb limit.

How many versions of macOS do you need for simultaneous CI and testing?

If the testing is automated? As many as fit in memory.

For testing I usually have 4 VMs running the latest version of macOS (spread scross 2 Macs), but they are currently broken after upgrading to the latest version of macOS and Xcode.

I also have a few more VMs running legacy versions of macOS for building stuff that needs to be built on older versions of macOS, but I only run them on demand.

And I use a VM for collaboration (multiple people can code together and share a screen remotely on a designated VM that doesn't have any personal stuff on it).

At most I probably had 6 or 7 VMs running on the Mac mini simultaneously, but if I bought a Mac Studio I'd really want to be able to run more of them. I don't want to have to quit VMs that I'm not currently using because suspend/resume always takes ages and I'm never sure if networks drives are still mounted afterwards.

the macOS license allows only two virtual instances, regardless of CPU architecture ;-(

That's not really relevant though if it is technically possible.. They're not going to sue you for it. If they cared they'd have gone after the hackintosh community a long time ago.

unfortunately it's very relevant if you want to use it for CI in any corporate environment. It's a really annoying limitation.

At least in 2014 or so iCloud was Dell/Oracle.

How is this limit enforced? Couldn’t you do it?

The Virtualization framework has a check inside of it.

Can somebody please jailbreak this.

Apple is such an annoying company. The M1 performance to price ratio definitely made it worth it to buy an MacBook Air but the constant crippling of their product line for product segmentation makes them impossible to like.

They are so consumer hostile, I think it’s going to hurt them in the long run. I’m really thankful that the EU is going to be shaking them a bit on their business policies in Europe.

What happens if/when the EU fails to update their laws on, say, USB-C mandates?

I agree the short term legal framework seems to have benefits, but I fear for the long term impact.

Laws are RARELY every taken off the books. Government doesn’t shrink

What's the long-term fear, here? That Apple will suddenly reverse their stance on iPhone transfer speeds, and design a new super-fast transfer spec for iPhone in 2025? They can still do that. They can do it the same way they did with Thunderbolt, where they work with other industry leaders to propose a standard to the USB-IF consortium. Once everyone is happy with it, then they can roll it out to their products. These are the standards that should be designed democratically, not privately.

All this nail-biting around USB standardization is worrying in the wrong direction. It's like saying that all of our cell carriers should get to choose whatever spectrum they want, because if we don't give them freedom then we might be stifling their innovation. Sometimes, there are more important things to worry about.

They could also theoretically do lightning-3 as an alt-channel over usb-c. The regulation is concerned with charging, not data.

I think your spectrum analogy misses the mark, here’s why:

USB-C is a misnomer. There are actually about 10 different variations, which all conform to the same physical port constraints, but have drastically different behaviors in terms of charging, data transport, and more.

By mandating a USB-C port, I fear we’ve created a pointless monopoly on the design that will stifle future innovation. Is USB-C the “best port in the world”? It’s 2022/2023, the answer may be “yes”… but MicroUSB was “the best port in the world” in 2007-2008, and its downright awful now.

Spectrum is different as it’s a physical medium that the government regulates to ensure the flow of information within that physical medium. Thats more akin to water management or other physical resources, and outside the scope of incentive management for innovations sake.

Is it really? Why shouldn't we also regulate connectors as a physical medium to ensure the flow of information and electricity between compatible devices?

It's a ridiculous comparison, but only because Apple's blatant failure to innovate brought us here. We're not even forcing them to use Thunderbolt or USB-3.0 speeds, it just needs to plug-in the same as other phones.

> Why shouldn't we also regulate connectors as a physical medium to ensure the flow of information and electricity between compatible devices?

I think you misunderstand - regulating the physical medium (the actual shape of the USB-C receptacle and outlet) is all we've done. There is no regulation on how you label/market cables and ports for their capabilities, so there is a 'wild west' right now in that your device has a USB-C plug, but certain features might not be supported by each end of the device, or even the cable. For example, unless you got a pretty nice USB-C<->DP monitor included with your display, your run-of-the-mill $10 Amazon USB-C->DP might not work due to strict signal integrity (read: cable build quality) requirements when MST is absent[0].

0: https://sebvance.medium.com/everything-you-need-to-know-abou...

They already did from micro USB to USB-C. There's a whole process for updates built into this regulation already.

It's not a USB-C mandate. It's a "talk to each other and figure out one solution" mandate. USB-C is just the current result.

Those are good points that I was not aware of. In the literature I read it was focused on “USB-C” as the standard.

Do you have any links to where we can read more?

The current standard is USB-C and there is a list of all device types that need to use the standard from now on. Afterwards, the European Commission shall do market research every 5 years and have the power to update the device type list and charging standard on its own.

Source: the text version of the adopted directive (with translations into all 24 official languages available in the top right corner) - https://www.europarl.europa.eu/doceo/document/TA-9-2022-0338...

No different from ARM, AMD and Intel having product segmentation on their CPU cores.

Apple is the only company among the ones listed that forces you to use their software on select products. That's the product segmentation, and it's really frustrating when hardware like the iPhone/iPad is so excellent but entirely hamstrung by it's OS. There's nothing wrong with segmenting your hardware or writing a pared-back OS for smaller devices, but forcing everyone to use it is a pretty obvious and arbitrary limitation.

Windows Home vs Pro vs .... with "Works best on Windows" OEM agreements.

Those are quite dumb too, but Microsoft's deals with the OEM don't stop me from using Intel's hypervisor framework or iGPU on Linux.

Depends on how much the community has managed to reverse engineer the hardware, firmware, boot lock, and then there is Surface.

The same way that Apple hardware has to be reversed engineered.

Why do you think I'm defending any of Microsoft's products here? When Microsoft releases a locked-down product, it's also as bad as when Apple does it. But even still, there's less reverse-engineering work that goes into getting a Surface device boostrapped than an Apple Silicon one. Writing a custom OS for x86 is lightyears easier than bootstrapping one for ARM.

I don't think, rather pointing out Apple isn't alone.

x86 only happened because of IBM being unable to kill the clones, and going forward many of them will eventually also have ARM workloads.

All OEMs have Apple envy in shipping sealed experiences, just like all 16 bit platforms with exception of PCs.

Hence why they rather ship locked down laptops, phones and tablets than classical desktops.

Well, it's time we stopped the trend dead in it's tracks. IMO, it's fine to seal your experience however you want as long as the bootloader is open on the hardware. If the user isn't trapped in the hardware they own, then it's fair game.

Are you seriously planning to use what the CPU manufacturers oligopoly does in an unrelated market to defend Apple? Because if it’s the case I don’t think it’s even worth my time arguing with you.

It was worth time enough to write that comment. You could have moved on.

Reply on news.ycombinator.com