6 ms·

A fair warning to developers: You're getting into an adventure with M1. From docker images not built for M1 with segmentation faults on qemu (eg. Liquibase for

A fair warning to developers: You're getting into an adventure with M1. From docker images not built for M1 with segmentation faults on qemu (eg. Liquibase for spring developers), to _significant_ troubles trying to make React native apps build with the M1 and XCode. Don't get me wrong, I have a Macbook air M1 and I love it, but it hasn't been a love without pain. Also, the magsafe feels like it comes too late. Almost like a political feature in response to the EU measure of enforcing a single charging cable.

urthor 5y ago
It's very difficult to justify the purchase because ultimately I've switched to thin client for serious development long ago. I guess I'll use the unrivaled performance per watt to drive Jetbrains.
- PedroBatista 5y ago
  > I guess I'll use the unrivaled performance per watt to drive Jetbrains. So you're advocating for nuclear power?
syspec 5y ago
You can still charge using USB-C
- sergioisidoro 5y ago
  Yes, that is true. And I hope that would be the default (at least in Europe). But the timing of this launch (when all new machines were already using USB-C), with the decision of the European commission is quite a coincidence. Which makes the feature feel a bit out of place, or just a good upselling opportunity for Apple.
  - vinay427 5y ago
    Honestly, I’m not sure I see this. They made the charging cable USB C at one end, shipping a standard USB C PD adapter and even mentioning that a USB C to C cable can be used with the same adapter to charge the same laptop. I don’t think they could have done much more without compromising MagSafe, and this allows for the same portability of cables and adapters that people like about USB C PD.
- interestica 5y ago
  Which has me wondering. The MagSafe brick is 140w and current USBC can only do 20vX5a (100w). New USBC version can do 240w apparently - but will the USBC on these laptops be able to pull 140w+ via the USBC? edit: more digging says that the MagSafe will be using USBC PD 3.1 spec - which presumably means 240w potential or at least as much the brick. I only wonder because I have stupid issues like the Microsoft Surface Book2 that tries to pull more power than the provide power brick can handle! https://usb.org/sites/default/files/2021-05/USB%20PG%20USB%20PD%203.1%20DevUpdate%20Announcement_FINAL.pdf https://usb.org/sites/default/files/2021-05/USB%20PG%20USB%2... edit 2: okay, while the Block will be PD3.1, the Thunderbolt 4 ports (USBC) only follow PD2.1, so they will likely be limited to 100w. https://www.theverge.com/2021/10/19/22734233/apple-140w-macbook-charging-brick-gan-usb-c-pd-3-1-third-party-chargers https://www.theverge.com/2021/10/19/22734233/apple-140w-macb...
adam_arthur 5y ago
Is there any reason Apple couldn't implement a magsafe type connector but using USB-C form factor? By that I mean, you could use any USB-C compatible cable in the same port (just without the magnetics) It seems like they could have gotten the best of both worlds that way. Am I taking crazy pills?
- jagger27 5y ago
  USB-C is already a friction fit. Somehow adding magnets to it would just make it attach more strongly, which isn't the point of MagSafe. I feel like making a slightly loose USB-C cable but with magnets would be a bad user experience, and probably against the USB spec. I don't really get what you're suggesting.
  - adam_arthur 5y ago
    Surely an alcove could be created such that the usb c connector creates a firm fit once the magnets latch on. e.g. magnets pull retractable flaps open such that the magsafe USB-C can be retracted easily. When using regular USB-C the flaps are locked in place, so works in that case too. I'm sure it's possible, but perhaps not worth the effort. Not one of humanity's most difficult problems.
    - jagger27 5y ago
      “Retractable flaps” sounds like it’s begging for a -gate suffix. Flapgate just rolls off the tongue. Apple doesn’t like moving parts, for obvious reasons. I think Apple did exactly the right thing here. If I have my MacBook hooked up to my desk monitor, I’ll just use a single USB-C cable for power, video, mouse, and keyboard like I do today. That kind of desk setup isn’t a tripping hazard. If I take my laptop to a coffee shop and need to charge it, I’ll use MagSafe because it’s more likely to be tripped over or yanked in that situation. What’s key is that I only needed power at the coffee shop. What else would be going over the cable in that kind of setting?
- nomel 5y ago
  You can find these on amazon, for example: https://www.amazon.com/Ansbell-Magnetic-Adapter-Transfer-Compatible/dp/B0831KYRD2 https://www.amazon.com/Ansbell-Magnetic-Adapter-Transfer-Com...
buzzert 5y ago
What’s the problem with building React Native apps?
zmmmmm 5y ago
I'm curious about the docker image situation. I had thought that the M1 could natively emulate x86 instructions. So why then can you not run the native x86 docker images? Is it a virtualisation issue?
- bronson 5y ago
  It does, which is slow, then the disk emulation makes it slower. Yes you can do it, but it’s not much fun. Improvements are on the way.
- metabagel 5y ago
  Apparently, it does work in most cases, but aarch64 images have better performance and are more stable. https://docs.docker.com/desktop/mac/apple-silicon/ https://docs.docker.com/desktop/mac/apple-silicon/
- johndoe0815 5y ago
  The M1 does not emulate x86 instructions (in hardware). Emulation is provided by Rosetta 2, a combined ahead-of-time and just-in-time emulation that can run x86-64 macOS binaries on Aarch64. However, there is still some hardware support for Rosetta 2 in the M1, e.g. to support strong memory ordering. More details on Rosetta 2 internals can be found at https://ffri.github.io/ProjectChampollion/ https://ffri.github.io/ProjectChampollion/
- fulafel 5y ago
  There is no hw emulation, it's all software.
harshitaneja 5y ago
React native issue is resolved in the 0.66[1] finally. It was walking on egg shells dealing with a weird combination of rosetta and arm configs. I have not faced docker issues except in the first few months but maybe that's just our workflow where we already had aws graviton instances. [1] https://reactnative.dev/blog/2021/10/01/version-066#better-support-for-apple-silicon-xcode-13-and-ios-15 https://reactnative.dev/blog/2021/10/01/version-066#better-s...
b15h0p 5y ago
Do you have more information about that Liquibase issue?

urthor 5y ago

It's very difficult to justify the purchase because ultimately I've switched to thin client for serious development long ago. I guess I'll use the unrivaled performance per watt to drive Jetbrains.

PedroBatista 5y ago
> I guess I'll use the unrivaled performance per watt to drive Jetbrains. So you're advocating for nuclear power?

syspec 5y ago

You can still charge using USB-C

sergioisidoro 5y ago
Yes, that is true. And I hope that would be the default (at least in Europe). But the timing of this launch (when all new machines were already using USB-C), with the decision of the European commission is quite a coincidence. Which makes the feature feel a bit out of place, or just a good upselling opportunity for Apple.
- vinay427 5y ago
  Honestly, I’m not sure I see this. They made the charging cable USB C at one end, shipping a standard USB C PD adapter and even mentioning that a USB C to C cable can be used with the same adapter to charge the same laptop. I don’t think they could have done much more without compromising MagSafe, and this allows for the same portability of cables and adapters that people like about USB C PD.
interestica 5y ago
Which has me wondering. The MagSafe brick is 140w and current USBC can only do 20vX5a (100w). New USBC version can do 240w apparently - but will the USBC on these laptops be able to pull 140w+ via the USBC? edit: more digging says that the MagSafe will be using USBC PD 3.1 spec - which presumably means 240w potential or at least as much the brick. I only wonder because I have stupid issues like the Microsoft Surface Book2 that tries to pull more power than the provide power brick can handle! https://usb.org/sites/default/files/2021-05/USB%20PG%20USB%20PD%203.1%20DevUpdate%20Announcement_FINAL.pdf https://usb.org/sites/default/files/2021-05/USB%20PG%20USB%2... edit 2: okay, while the Block will be PD3.1, the Thunderbolt 4 ports (USBC) only follow PD2.1, so they will likely be limited to 100w. https://www.theverge.com/2021/10/19/22734233/apple-140w-macbook-charging-brick-gan-usb-c-pd-3-1-third-party-chargers https://www.theverge.com/2021/10/19/22734233/apple-140w-macb...

adam_arthur 5y ago

Is there any reason Apple couldn't implement a magsafe type connector but using USB-C form factor? By that I mean, you could use any USB-C compatible cable in the same port (just without the magnetics) It seems like they could have gotten the best of both worlds that way. Am I taking crazy pills?

jagger27 5y ago
USB-C is already a friction fit. Somehow adding magnets to it would just make it attach more strongly, which isn't the point of MagSafe. I feel like making a slightly loose USB-C cable but with magnets would be a bad user experience, and probably against the USB spec. I don't really get what you're suggesting.
- adam_arthur 5y ago
  Surely an alcove could be created such that the usb c connector creates a firm fit once the magnets latch on. e.g. magnets pull retractable flaps open such that the magsafe USB-C can be retracted easily. When using regular USB-C the flaps are locked in place, so works in that case too. I'm sure it's possible, but perhaps not worth the effort. Not one of humanity's most difficult problems.
  - jagger27 5y ago
    “Retractable flaps” sounds like it’s begging for a -gate suffix. Flapgate just rolls off the tongue. Apple doesn’t like moving parts, for obvious reasons. I think Apple did exactly the right thing here. If I have my MacBook hooked up to my desk monitor, I’ll just use a single USB-C cable for power, video, mouse, and keyboard like I do today. That kind of desk setup isn’t a tripping hazard. If I take my laptop to a coffee shop and need to charge it, I’ll use MagSafe because it’s more likely to be tripped over or yanked in that situation. What’s key is that I only needed power at the coffee shop. What else would be going over the cable in that kind of setting?
nomel 5y ago
You can find these on amazon, for example: https://www.amazon.com/Ansbell-Magnetic-Adapter-Transfer-Compatible/dp/B0831KYRD2 https://www.amazon.com/Ansbell-Magnetic-Adapter-Transfer-Com...

buzzert 5y ago

What’s the problem with building React Native apps?

zmmmmm 5y ago

I'm curious about the docker image situation. I had thought that the M1 could natively emulate x86 instructions. So why then can you not run the native x86 docker images? Is it a virtualisation issue?

bronson 5y ago
It does, which is slow, then the disk emulation makes it slower. Yes you can do it, but it’s not much fun. Improvements are on the way.
metabagel 5y ago
Apparently, it does work in most cases, but aarch64 images have better performance and are more stable. https://docs.docker.com/desktop/mac/apple-silicon/ https://docs.docker.com/desktop/mac/apple-silicon/
johndoe0815 5y ago
The M1 does not emulate x86 instructions (in hardware). Emulation is provided by Rosetta 2, a combined ahead-of-time and just-in-time emulation that can run x86-64 macOS binaries on Aarch64. However, there is still some hardware support for Rosetta 2 in the M1, e.g. to support strong memory ordering. More details on Rosetta 2 internals can be found at https://ffri.github.io/ProjectChampollion/ https://ffri.github.io/ProjectChampollion/
fulafel 5y ago
There is no hw emulation, it's all software.

harshitaneja 5y ago

React native issue is resolved in the 0.66[1] finally. It was walking on egg shells dealing with a weird combination of rosetta and arm configs. I have not faced docker issues except in the first few months but maybe that's just our workflow where we already had aws graviton instances. [1] https://reactnative.dev/blog/2021/10/01/version-066#better-support-for-apple-silicon-xcode-13-and-ios-15 https://reactnative.dev/blog/2021/10/01/version-066#better-s...

b15h0p 5y ago

Do you have more information about that Liquibase issue?