> (2020-11-15) There won’t be any support for native ARM Homebrew installations for months to come. See both #7857 (comment) about CI infrastructure and #7857 (comment) about GCC for details.
Tbh it’s kind of weird to me that Apple didn’t sponsor the work needed for this, and would ensure that it happened sooner.
I mean, Apple work closely with Adobe and other software companies to ensure a good experience for the end users. At least I think they do.
And there are quite a lot of people using and depending on Homebrew.
Take myself for example, an Apple customer. I use an iPhone and a MacBook Air. But many of the tools available through Homebrew are so essential to me that without Homebrew I likely would have stuck with Linux and FreeBSD for my laptop too like I used to instead of using a MacBook Air at all.
And it’s not just one laptop either. In the future when I can afford to I am going to buy more Apple computers.
I don’t understand why people are expecting day one support for applications like Homebrew.
By its very nature Homebrew is going to hit every edge case that can possibly exist during this transition process. It also has a huge number of indirect dependancies, also driven by its nature as a package manager. Meaning a huge amount of work across many projects is needed to get Homebrew “working” for everyone.
We’re at the start of Apples two year transition to ARM CPUs. If important thing X doesn’t work right, and it’s important to your workflow, then buy an Intel Mac.
Otherwise have patience and wait for people work through the issues. Besides in a years time Apple are undoubtedly going to release the M2, M3 etc with a whole slew of “pro” devices better suited for developers.
As to why Adobe gets a helping hand. Simple, Adobe’s customers almost certainly make up a bigger portion of Apple’s revenue than developers. Plus normal people understand what photoshop is, and understand how important it is to work on these new Mac. The general public hasn’t even heard of Homebrew, never mind understands how impressive it would be to get such software working on day 0.
> I don’t understand why people are expecting day one support for applications like Homebrew.
I'm not paying the Homebrew developers anything so I've got no right to expect anything or criticise whatever they choose to do with their time and their project.
But if they did want compatibility it's not true that they're 'at the start of Apples two year transition to ARM CPUs' - you have been able to access development machines for many months. I got one and I don't maintain anything nearly as important as Homebrew.
The reason Apple do developer previews, betas, and early access to hardware is so that you can have day-one support, if you want it.
Homebrew does not control the software that it helps distribute. There are hundreds of pieces that need to work together for this to work, so Homebrew developers can at most prepare their infrastructure, which I don't think is a big issue here. Moreover, the number of developers with access to an M1 chip until last month was very small. For most of these people, the M1 transition process is just starting.
For certain developers... for a $500 fee. I doubt that many of the developers of packages on homebrew had access to the kit nor the inclination to buy it even if they did. They don't have the userbase nor the financial incentives to be ready for apple silicon day 1.
The number of developers with access to an A12Z running macOS was also pretty small, so I don't see how calling out the model number of the chip changes OP's point.
Actually, Homebrew developers have been working on this for months (in their spare time, too, given that they are all volunteers); but the reality is that this is a gargantuan task, 6 months simply wasn't enough.
So no, I don't think "you can have day-one support, if you want it" is accurate at all.
> Adobe’s customers almost certainly make up a bigger portion of Apple’s revenue than developers
Do we have any actual numbers on that? I feel like developers (and power-users generally) have internalized this notion that we're a rounding error compared to the general public (perhaps a holdover from the old days, when regular folks didn't use the internet or computers at all). But in 2020, software development is a massive portion of the economy, and Macs seem to be disproportionately popular with startups and developers.
A little lazy googling puts the graphic design industry at $12.7 billion, and the software development industry at $388.98 billion.
Money aside, far more people care about Photoshop than care about Homebrew. Optics-wise, it makes more sense.
I hope this next year shows Apple helping out pivotal projects like Homebrew. If they expect their own developers to be using their own hardware, then they will certainly have to make it work.
I was only a student when they switched to Intel, so I wasn't aware of any issues facing developers back then. I'm guessing they'll follow the same playbook, though.
There is a reason why designers, photographers, video editors and influencers are the prioritised. Get them on your side and it's free marketing for you.
> I don’t understand why people are expecting day one support for applications like Homebrew.
Because an important segment of Apple's customers--developers--rely on Homebrew for critical functionality.
Of course it's difficult to move an entire ecosystem to a new architecture, and Apple may have made the right move here. But it's not very hard to understand why people who stand to gain so much from the new architecture, but can't use it yet because the ecosystem isn't there yet, are a little annoyed.
> We’re at the start of Apples two year transition to ARM CPUs. If important thing X doesn’t work right, and it’s important to your workflow, then buy an Intel Mac.
> Otherwise have patience and wait for people work through the issues. Besides in a years time Apple are undoubtedly going to release the M2, M3 etc with a whole slew of “pro” devices better suited for developers.
There's also, arguably, a pretty pronounced bleeding-edge trend over the past 12-18 months or so, across a number of device categories: foldables everywhere, new game consoles, multiple open-spec Linux phones, M1 Macs, etc.
And if you're not itching to early adopt, the timing for all of these is actually kind of great. Because parallel to this, there are multiple ecosystem-level changes hitting over the next several months that should translate to big usability jumps for any of the ones that get a next version. (Though probably not the consoles, since those tend to run longer lifecycles by design.)
Of particular note: by the time the M2 or M3 has ironed out the macOS ARM ecosystem, you can also reasonably expect high-end devices as a category to have 6GHz WiFi and a modern spec for Bluetooth audio.
If I were apple I would of shipped a dozen developer transition mac minis to them free of charge at least a few months ago, as part of developer relations, to kickstart the process if they haven't done something similar.
Apple's support of my software company was terrible and shameful.
I got a dev unit (mac mini). It had beta 2 on it. We started the porting effort and had questions. Forums were the only place they said answers could be had. Never got any answers.
When beta 10 was released, it just wouldn't install. We tried for a week then saw a post on the forums that gave a hint about Mac-to-Mac reset of the firmware and general machine state. We had to upgrade another machine to Catalina to use it (inconvenient, but not a killer).
When 11.0.1 came out it wouldn't install and now the machine is a brick. We ordered a production mini with the M1 and it will be here in late December. While we sit with the brick and can't do any porting.
At every step we hit brick walls and had zero help from Apple. It was beyond frustrating.
I am not sure about the OP, but M1 Macs are not available in store anywhere in Canada, or at least not in Alberta and BC. Ordering by mail is the only option here at the moment.
Apple doesn't really have any interest in making it easier for people to use Macs for anything that doesn't benefit their closed ecosystem - they are still ok with you using it for open source or enterprise work so long as they don't have to do anything special. That part of their stance has been very consistent.
People have been consistently rewarding them for this direction by buying more of their stuff and relying on open source projects to fill in the gaps without really acknowledging the obvious problem that Apple has made it increasingly unattractive to use their hardware to do anything other than Apple stuff.
To me it matters most that I be able to do things that I want easily on the platform I choose - in that sense PC hardware with Linux is the best choice I have, with Windows coming in 2nd due to how much Microsoft has done to make their OS attractive for wide variety of use cases that are not necessarily to Microsoft's direct benefit.
Didn't know that, thanks - have they contributed to homebrew in the past? If they have then then it's just sloppiness to release hardware without contributing a port and maybe later on we will see it happen?
Homebrew works today (w/ build all packages instead of prebuilt bottles available), however they have a policy of not having downstream patches, so that source trees upstream have to have the Apple Silicon patches.
For packages such as GCC, that of course will take a lot of time. For Java, there are other supported runtimes like Azul Zulu, but Homebrew just wants OpenJDK and not much else.
Perhaps they could (ab)use the versioning system, vending it as openjdk@azul or something? I'm not an expert on how Homebrew does dependency resolution.
> Apple doesn't really have any interest in making it easier for people to use Macs for anything that doesn't benefit their closed ecosystem - they are still ok with you using it for open source or enterprise work so long as they don't have to do anything special. That part of their stance has been very consistent.
Besides being used internally as mentioned, making the Alpha Geeks (to use Tim O'Reilly's phrase) happy would be prudent for their continued success. If the Alpha Geeks aren't using your product, they will not recommended it to their friends when asked for computer advise, and it will not be installed in relatives' home where the Alpha Geeks tend to do tech support.
The Alpha Geeks was how Apple probably got more popular that PCs in the first place, and they may now have a lot/enough inertial to keep moving forward, but Microsoft and PCs used to be the 800-pound (400 kg) gorilla, and they're less so now.
There's a lot of Unix-y stuff being used in academy, research, and HPC: plenty of folks there would be just as willing to switch to a PC with Ubuntu/Debian/CentOS (which run the world's HPC clusters) if macOS gets in the way of their workflow if Homebrew or Macports don't work.
>The Alpha Geeks was how Apple probably got more popular that PCs in the first place
At which point in time was that?
From the statistics I could google in a few seconds, for the last 20 years, MacOS was always hovering around 10-15% of the worldwide desktop PC market share with Windows at around 75-85% and linux/BSD the rest of the scraps.
> There's a lot of Unix-y stuff being used in academy, research, and HPC: plenty of folks there would be just as willing to switch to a PC with Ubuntu/Debian/CentOS (which run the world's HPC clusters) if macOS gets in the way of their workflow if Homebrew or Macports don't work.
Absolutely. I'd say Linux is the default for science and ML!
All of the Apple products that have achieved broader success have been derided by ‘Alpha Geeks.’ The original candy coloured iMacs, iPods, iPhones, iPads, were all attacked by geeks.
If Microsoft is less successful in the home market now, it is because Windows is seen by many consumers as too techy, and too difficult.
But Apple is very big now. They can now afford to not court the "Alpha Geeks" since they already have captured the mainstream market(at least in the US).
You may say that's not true. But just look at their actions and not their marketing speak.
It's better than nothing but clearly it doesn't seem to be a priority for them and the rest of my point still stands - they don't do nearly enough outside of their priorities.
Contribute patches for all the blockers mentioned towards native homebrew in this post?
This is classic of the all or nothing attitude that anti-Apple and anti-Mac people have. Unless Apple does some Herculean task that shows they're worthy, they're shit.
I consider that developing software for the Apple platform is a free job to benefit Apple. It rarely pays, and you need to be always on the look for ways Apple is breaking the platform on their frequent releases. I personally wouldn't do that unless I had a very clear motivation.
You know Boot Camp exists, right? That is a huge amount of work, done specifically by Apple themselves, in order to support running completely different OSes on their hardware.
You can run homebrew Rosetta though, right? Native is, of course, the preferred direction, but this doesn’t mean you’re hosed on an M1 MacBook by any means.
Yeah... as much as I tend to dislike Apple and their decisions, this is a complaint I don't really understand... especially since the M1 with Rosetta 2 supposedly runs software better than the Intel macs did anyway! I actually chalk this up to the weird fetish people who use Apple products tend to have with things being exactly "modern" as opposed to being stable and functional or even performant (which tends to lead them to pester people to upgrade to new things on the Apple treadmill constantly rather than support new features or use cases... Windows and Android users don't do this and I frankly think are much better off--both professionally and psychologically--because of it). Put elsewise: Apple did do the work of porting homebrew... by writing Rosetta 2, and they probably tested it on Homebrew internally to make sure it was compatible with lots of Homebrew packages.
Yes. And it has worked perfectly for me so far, as if I was still using brew on an Intel mac. People complaining that they can't do development on an M1 mac because of the lack of homebrew support either don't realize you can simply run it under Rosetta, or are just complaining for the sake of it.
I am holding off purchasing multiple ARM Macs until homebrew works properly. We need them for our development and ops work, and now these uber-cool Macs are out, buying old Intel ones has become a no-no too, so we don't buy any desktops/laptops at all!
If you read this and work at Apple, you can show this message to your bosses and I hope that it could entice the company to sponsor Homebrew with hardware + engineer time to help the project go forward. I am pretty sure than homebrew generates 100s of millions of sales for Macs (100k computers sold instead of Linux laptops if homebrew did not exist seems reasonable).
Hundreds of millions of Macs is the right order of magnitude to measure total sales–most Mac users are not developers using Homebrew. And the relevant people are almost certainly already aware of the existence of Homebrew.
If you read this and work at Apple, you can show this message to your bosses and I hope that it could entice the company to sponsor Homebrew with hardware + engineer time to help the project go forward.
As you can read further down the thread, the Homebrew team was granted test units during the summer and had M1 Macs donated to them once they shipped.
How is it acceptable that Apple donates just now? Even more so when Apple themselves use Homebrew. tags:#apple
edit: I don't mind the foreseen responseless downvotes. This is more of a personal note on yet another reason to stay away from APPL after having problems with iPhone 6 CPU throtling and MacBook Pro keyboard fiasco.
I'm not following the logic. Because Homebrew was developed 11 years ago on generally available hardware by people who already had such hardware, Apple... owed them free hardware? For what purpose? Just to say thank you?
Homebrew isn't a cross-platform thing. It's written by MacOS users for MacOS users. (Although I guess it now also supports Linux.) The project didn't need anything from Apple as long as Apple was using generally-available hardware.
Once Apple announced new hardware, they donated hardware to avoid the brew devs needing to buy new hardware. I'm not sure why anyone would expect anything before then.
Development, mantainance, support and servers cost money and time.
The least I would expect from a tax evading megacorp like Apple was to support the development of a tool that's so widespread and important to developers on macs. That would be the smart thing to do.
The fact that homebrew is specific to macs only adds insult to injury because that's even more reason to support such a tool and strengthen their ecosystem.
Throwing a bone so late in the game with no extra support other than a few beta machines worth a single week of one developer's income is laughable.
I shouldn't be surprided by Apple's stinginess. Last time they got hacked by a one-of-a-kind security research team which found dozens of vulnerabilities, the team was paid within 6 digits what could easily sell for 7 ou perhaps 8 digits to state actors. https://news.ycombinator.com/item?id=24718078
Poeple's beef with Apple are rarely unrooted. Just read the links in this thread alone and my experience with iPhone 6 and macbook's keyboard.
Travelling far to a "genius store" to show a stuttering iPhone only to be told that the battery is still not broken enough for warranty despite the phone having to slow down to protect the underprovisioned underplaned battery is like being called a clown and a liar.
And diminishing tax evasion denounced by multiple unbiased entities to "people can't reason with Large numbers" is a disservice to those who could use those billions in health care.
Even the Developer Transition Kit / Universal App Quickstart Program has been handled poorly. Bunch of teething and hardware failure/DOA issues and arbitrary limitations with near zero engagement from Apple.
Nothing like the original Intel transition program, I'm willing to bet they're going to give credit instead of hardware for the exchange this time around too. It's an entirely different company now.
Apple doesn't NEED to subsidize or keep developers happy anymore. They've already won us and the marketshare. SE, vertical integration, leapfrog CPU and battery life.. They don't have any competitor, so why waste money investing in the ecosystem? We're going to build for them anyway.
Sure. They just developed a new chip, with a new 5nm process, at to drop all the devs off a cliff and hang them out to dry. That will sell units. If they don't work the issues out they'll have to partially shelve the design and go grovelling back to Intel for CPUs because things will continue not to work until it affects the profits. I would be incredibly surprised to see "near zero engagement" with major vendors on this, have (e.g. Adobe) devs written about big issues, or just FOSS wranglers who feel their Free project is undervalued?
I agree there's been a big shift in focus. I'm seriously unhappy about the Amazon/Google level of ad-down-throat-ramming I have to put up with on my Apple TV. But I don't think they're throwing the towel in on hardware and software innovation.
>And there are quite a lot of people using and depending on Homebrew.
People are under no obligation to upgrade immediately to big sur. I'm holding out on Catalina for a few months with older hardware. M1 is a big change and I'm not surprised it will initially break a few things.
Also the point of having third party app ecosystem is that the first party doesn't have to maintain them and yet still receive some benefit from them.
It's homebrew's own choice that they have 100% of everything on Apple before they ship. They're simply lazy and don't want to have to deal with support of people complaining why something doesn't work. This is the problem with Homebrew and why people shouldn't use it.
> Tbh it’s kind of weird to me that Apple didn’t sponsor the work needed for this, and would ensure that it happened sooner.
They could donate something, but why should they? It’s community project not run by Apple. We, users, who actively use it to make professional work, should be responsible for supporting community projects like that. :)
AFAIK, Apple does or did sponsor MacPorts and uses or used it internally; but maybe I'm wrong.
Apple should have a vested interest in developers on their platform and thus ports of popular F/OSS; but obviously, developers, graphic artists, music and other content producers aren't as significant a customer group relative to consumers as they used to be. I wonder how the decline of the Mac platform for professional work hits Apple's baseline assuming the historical draw of content creators to the platform also brought consumers and has been used by Apple in advertising, but I guess Apple knows better and simply makes so much more money with iDevices.
Apple sponsored much of the original work on MacPorts, but these days it's run as an independent project of which a handful of maintainers are still from Apple.
The first patches for Apple silicon came from Apple engineers, FWIW.
Apple's a large company, and like many large companies different people use different things ;) Since teams are largely autonomous with regards for tooling, I am sure that the choice of Homebrew or MacPorts for dependency management comes down to the choice of one engineer from the early stages of a project. For teams that use neither themselves developers are free to choose either for their machines. I would assume that new hires mostly pick Homebrew to use when given the choice.
Apple should have a vested interest in developers on their platform and thus ports of popular F/OSS…
Lets not forget that pretty much everything that's part of the Darwin layer (a.k.a. Unix) is already running natively on M1 Macs, including Ruby, Python, Perl, Vim, Emacs, etc. Straight from Apple.
I wonder how the decline of the Mac platform for professional work hits Apple's baseline assuming…
Apple's Mac revenue hit an all-time high just last quarter and that was before M1 Macs shipped. Yes, a product line that's about to be 37 years old in January 2021 hit $9 billion in revenue during a global pandemic [1] and jumped 39% [2]. I doubt there's a decline in professional usage of the Mac; even if there were, entry level M1 Macs that outperform professional Intel-based machines from Apple and others will address that.
but I guess Apple knows better and simply makes so much more money with iDevices.
The iPhone is brings in the most money but the Mac ($9 billion) made more than the iPad ($6.7 billion) and the Apple Watch + HomePod ($7.8 billion).
I kind of disagree. For the same reason that shops provide free parking, it would make sense for Apple to put effort into the de-facto package manager for macOS.
They don't "provide macOS for free" any more than they provided the T key on your mac keyboard for free: their software is only designed to run on their hardware (for which they also attempt legal restrictions to prevent you using it elsewhere in addition to cryptographic DRM techniques) and they really only want you running their software in their hardware (which they already go to lengths to enforce on iOS and which they increasingly are pushing onto macs)... the software may as well be firmware (and in fact that is what Apple calls it on iOS) and is part of the vertically integrated product they sold you; so being annoyed about this is at least as legitimate about being annoyed that they didn't implement well any other feature of this laptop you are paying over a thousand dollars for (and, as stated by the person you are responding to, not just in their response to you but also in the post you responded to, the response is to just take their money elsewhere, which is _great_).
I think people still has this belief that power users is still an important target market for Apple. Basing on their hardware decisions and software support, I don't think this is true anymore.
Really weird indeed. The fact that they don't have an official package management system these days is already odd, but not helping the most popular one is just shooting yourself in the foot.
Correct me if I'm wrong, but I believe Windows didn't have one until a few months ago..? Also, the Mac App Store has been around for 10 years, Homebrew for 11 years and MacPorts for 18 years, so it's not like there was an urgent need to build one.
The Mac App Store is for app delivery, similar, but not exactly a replacement.
A package management system must take care of versions, dependencies, such as, this version of MySQL needs OpenSSL version X, Readline Y, etc. Make sure it doesn't mess with Python's OpenSSL/readline decencies which are both different from the ones used by the OS, and so on.
> I mean, Apple work closely with Adobe and other software companies to ensure a good experience for the end users. At least I think they do.
If they are, they're not doing a very good job; Adobe stuff is reportedly kind of buggy even under Rosetta, and historically Adobe has been very, very slow with the x86_64 transition, the Intel transition, the MacOS 10 transition, and so on.
A bit tongue-in-cheek, but Apple only just launched the public beta phase of MacOS on ARM. There is plenty of time for developers to work on this until the release of the stable product (M2 or whatever it will be named).
Pessimistically: the first public release of anything is basically a beta test.
Semantically this is not the case because they've "released", sure, but I don't think there's anything meaningful you can do until you scale out like this.
Homebrew makes Macs more attractive for more people, with little "loss of control" from Apple. I'm sure Apple (people) themselves use Homebrew internal.
> it’s kind of weird to me that Apple didn’t sponsor the work needed for this
Apple has lost its way.
They should not only be helping with this, but SHIPPING more stuff like this themselves.
And "pro" has lost its meaning. It is now for "youtubers who want nice stuff" and don't mind someone else controlling their experience, as opposed to engineers and scientists and creative people who are actually creating the future.
> But many of the tools available through Homebrew are so essential to me that without Homebrew I likely would have stuck with Linux and FreeBSD for my laptop
Apple is sending you a message on your value to them as a customer by the fact that they didn't bother :-)
Have you tried WSL2 lately? You can get way more powerful systems with a Linux environment and soon to arrive GPU pass through.
My experience with WSL2 is that it’s a bajillion little broken stuffs here and there. If you are not doing the same exact kind of web development that everyone else is, it’s not a smooth experience.
Anyway it’s not about value if customers or anything, they’re just jackasses. Apple could afford to fix 10,000 different concerns simultaneously, and many companies do achieve that level of support, at least proportionate to the resources others have. There isn’t some strategy here.
Tried WSL2 for a while but there was still a lot of things that needs polishing.
Finally decided on dual booting Pop!_OS and couldn't have been any happier. It's as close to MacOS as I can get in a desktop system. Due to Proton, I can now play games now too!
>Submit PRs to fix things. Almost every issue we have had so far has been already known. We know things aren't working. We need help fixing things not telling us what isn't working.
Sorry, I did consider it, but knowing that HN now populates the first /<name> as well as the domain for some known sites where it's meaningful, such as GitHub, I opted for the issue's title alone.
That @highlight doesn't have any effect, unless he's already reading the thread for other reasons and sees it with his eyes (apparently not yet :). Emailing them using the Contact link in the footer to have a title corrected works well (and more rapidly) though.
Sorry if this is a little bit off-topic but I can't find this answer anywhere. Can I run Docker on M1 right now? Even through Rosetta 2? Or does it not work that way? I want to buy an M1 laptop but need to be able to run some simple images on it.
Unfortunately you cannot run Docker on the M1 yet, even with Rosetta 2. There are some hacky workarounds for manually launching single images, but I don’t think that’s what you’re looking for.
There is no packaged way to run docker on m1 yet. Once Docker is supported, it will be running ARM native - so you will likely use QEMU to emulate other architectures like x86 and x86_64 if you don't have ARM-based containers available.
MacPorts itself mostly works. Many of the basic packages that are already cross-platform build fine. Some of the more complex languages and runtimes have issues building, or fail at runtime.
I've been following this daily since the issue was created. It's quite fascinating watching the progress on this. Once everything works, I will get a new laptop
Yeah I wonder if there’s a good way to keep it updated. I found an older blog post that puts their updating in .zshrc but I feel there must be better ways.
Why does Brew have all of those dependencies?
fwiw, it works fine right now if you install two versions, one for rosetta and one for apple silicon and let the latter compile things from source.
Tbh it’s kind of weird to me that Apple didn’t sponsor the work needed for this, and would ensure that it happened sooner.
I mean, Apple work closely with Adobe and other software companies to ensure a good experience for the end users. At least I think they do.
And there are quite a lot of people using and depending on Homebrew.
Take myself for example, an Apple customer. I use an iPhone and a MacBook Air. But many of the tools available through Homebrew are so essential to me that without Homebrew I likely would have stuck with Linux and FreeBSD for my laptop too like I used to instead of using a MacBook Air at all.
And it’s not just one laptop either. In the future when I can afford to I am going to buy more Apple computers.