NixOS on Framework Laptop
kvark.github.io
kvark.github.io
My eyes were good enough then that I was running it at native 2880x1800 resolution (no HiDPI business), and I was finally happy with the resolution I could get while mobile (I had been using a 30" 2560x1600 as my workstation monitor for a couple of years at that point, and this was an upgrade over even that!)
These days I use a 55" 4K OLED and a 32" 6K Pro Display XDR as my workstation monitors. For anyone considering this, I'd say wait for the 40" 4K OLEDs to come out (or wall mount the 55" and set your desk half a meter or so away from the wall). I sit just a little more than arm's length away from mine, and that's great for immersive flight simulator experience, but not so great for normal daily desktop use. (I'm keeping mine on my desk so I can keep using it along with powered sit-stand.)
As for why it's not so great; it's mostly a resolution issue. 4K isn't enough these days, and isn't high enough DPI for the amount of my field of view it takes up at this distance. I would also be quite happy with an 8K OLED that's, say, 55" or 65"... but only when I can also get one that drives at 120FPS, so that will be a bit of a wait.
My ideal would probably be something similar to what Linus has set up. He uses long cables from a nearby closet, so even when the workstation fans are blasting he doesn't hear it. For now, I use 2 and 3 meter cables plugged into a mid tower sitting on the far side of my desk on the floor.
For why I'm not doing the workstation-in-another-room approach: The 4K@120fps HDR OLED relies on HDMI 2.0 cables, which are only recently available at 3 meter lengths. Linus is using a not-commercially-available prototype optical HDMI repeater to get longer cabling.
For normal use the system is near-silent and at full blast it's below the noise floor with a headset playing something at a reasonably low volume (and not distractingly loud even if I just have the headset on with nothing playing).
You need to use Brigadier to download the Boot Camp drivers to get the display brightness controls widget. Unfortunately, there isn't an option for Linux.
Some good discussion on this:
https://bill.harding.blog/2020/04/26/linux-touchpad-like-a-m...
https://bill.harding.blog/2019/03/25/linux-touchpad-like-a-m...
My preference is to disable acceleration in favor of very high sensitivity for physical mice, to disable small touchpads (like on old Windows laptops) and retain a substantial but smooth acceleration curve for large touchpads (like on a Mac).
I think the DEs often let you adjust how many lines a single scroll increment scrolls by, but not the size of movement that triggers a scroll increment on the touchpad. I think depending on your touchpad driver (evdev is usually more flexible than, say, synaptics) and maybe your choice of display server (Wayland vs. Xorg), you can adjust the other aspect of touchpad scrolling.
Marketing teams actively distort are public opinion.
Your comment is remarkably dismissive IMO.
You'll survive
https://wiki.archlinux.org/title/TLP https://wiki.archlinux.org/title/Laptop https://wiki.archlinux.org/title/Power_management
Even if you're not running arch, the archwiki is an amazing source of information
The worst bit was how loud the fan used to run.. if it had spun any faster you'd think it would start hovering.
That being said, I think there could be a good use case for "meta" NixOS distributions. I'm working on such a thing for a NixOS based RetroArch type thing. But could see a PopOS like configuration, focused on on an “out the box experience”
Anyone with the hardware is welcome to contribute to nixos-hardware!
Even "no-op empty configs" are good imo as they show "there's no magic sauce needed".
Created an issue for this for anyone interested!
I plan on keeping my M1 for a 2-3 more years, then we will have to see what Framework has come up with then.
I like them a lot in principle (peripherals that can be switched to what I need? Yes please. Upgradable RAM? thanks), but I also won't compromise on battery time, performance or "linux works outside the box". Battery time could have been an easy fix if they had made it so I could switch the battery without taking the laptop apart.
https://guix.gnu.org/manual/devel/en/guix.html#Services
That is, the only difference in practice would be replacing 'herd' with 'systemctl' when stopping and starting services manually. Anything else, such as modifying a service, has to go through the configuration system.
Similarly, for service authors, whether the API you design ends up creating a systemd unit file or shepherd service object is an implementation detail and would be mostly transparent.
import framework.nix {}my code is possible if people want it :) the real value of NixOS is hardware and OS config can be turned into libraries!
And then Linux in general is notorious for having issues with USB WiFi adapters. More info on that can be found here https://github.com/morrownr/USB-WiFi
We allow runaway complexity to obscure the inner workings of an important software package, to the point, where having source access no longer makes a practical difference to a "casual" power-user or even a software engineer. I can source-dive runit, s6, OpenBSD's rc, or OpenRC and within a weekend, have a pretty OK understanding, fix a bug, or add a feature. Having software as complex as systemd taking so many roles, and making itself so ubiquitous and irreplaceable, makes me feel that I no longer have any control or understanding of my system, bugs or misfeatures just become things you live with, or you riddle yourself with arcane workarounds. Portability suffers, monoculture takes root.
I've lost faith in "mainstream" Linux distros. I use macOS and OpenBSD for workstations, and Alpine where I need e.g. Docker. It must either be simple enough to understand, or provide enough utility/convenience that the lack of (practical) source access is compensated. systemd fits neither category.
Boot Loader Interface, Discoverable Partitions Specification, etc.
I’m sure the implementation details are very interesting.
But where is the “Configure and Start a Service” option? Which is arguably the most important.
For a quick intro to new services, see the examples here: https://www.freedesktop.org/software/systemd/man/systemd.ser... (they should cover most basic needs) then read above for the details on the parameters. You stick the file in /etc/systemd/system/your_thing.service and enable/start it.
No, it's not as easy as "man systemd.service".
What's probably most useful in that documentation is the "hello world" example, which is buried at the bottom of the documentation. Also, key details are elsewhere, in "man systemd.exec" and "man systemd.unit".
The systemd docs a great comprehensive reference, but the man pages are poor introductory documentation.
Some useful tips that are hard to find:
* man systemd.directives documents every directory and will tell you which man page as the full documentation for it.
* man systemd.index lists every man page related to systemd.For example, `systemd-analyze` is a good tool for boot performance analysis, so + for systemd, but if I want/need to use perf for boot performance analysis, then systemd stands in my way, while for initd it's just yet another boot mode.
However having used it, NixOS seems like it mitigates the worst aspects of systemd. It gets rid of that that /etc/systemd /lib/systemd symlink "cleverness" for overriding/enabling units. Being functional makes it so that when systemd mysteriously breaks - for example if there is a loop in dependencies - it's easy to track down the change that caused it. And rather than splaying the config out in a bunch of random "unit" files, it's all contained in the nix config. About the worst thing I can say is that defining your own service is a bit more obtuse that it needs to be, with the arbitrary terms of the systemd unit format carrying over into the Nix config.
Just use `systemctl cat`, it prints the whole effective configuration.
I'm no fan of SysV init either, and I appreciate the advancements that systemd does bring to the table. But it is hard to shake the feeling of it having done the system software equivalent of sucking in a bunch of bloated javascript frameworks. Sure they made life easier for the developer, but everyone else has to live with the mess.
For enabling/sequencing, I'd say the preferable way would be to have a top level config file that pulls in the specified units explicitly rather than stitching together units based on their internal contents (akin to persistent structure rather than mutable cells). Juxtaposition is the most syntactically powerful operator and prevents loops intrinsically. Whereas by splaying dependencies in unit files, if you make one errant backreference the system is likely to not boot at all. I'm sure there's another clever command to check for that, but shrug.
No, for that you would use `systemctl edit UNIT...` to create or edit a drop-in file, e.g. to add local overrides for an existing service, or `systemctl edit --full UNIT...` to open a copy of the existing unit file for editing which will replace the packaged version. Add `--force` to create a new unit file for a service which doesn't already exist.
I've been doing fine on Debian dropping files in /etc/systemd, with a possible hook to `systemctl enable`. A rudimentary ad-hoc object system with overrides etc for every program is like the exact opposite of what I want when doing sysadmin. I think NixOS's approach does a good job of coping with systemd, in that it bundles all that complexity up in a way so it doesn't propagate, despite still having to look up what the Wants/Needs/Likes/Prayers/Requires/Binds designations mean every time I touch the config.
Don't forget `systemctl daemon-reload` to update the in-memory state when the unit files change. The `edit` command takes care of that for you. It also provides a template for the override files (when not using `--full`) which shows the current settings.
You are free to just drop files in /etc/systemd, of course. This is a fully-supported workflow. The `edit` command is only there for the convenience of the system administrator.
> A rudimentary ad-hoc object system with overrides etc for every program is like the exact opposite of what I want when doing sysadmin.
Are we talking about systemd, or NixOS? Their approaches to configuration really aren't all that different in this regard but you still seem to prefer one over the other. Or is the problem perhaps that you have two different programs trying to manage the configuration?
Personally, from the POV of distributions other than NixOS, I think the systemd approach is far superior to alternatives without overrides, where the only choice is to clone & modify the vendor-provided scripts. With overrides you can fine-tune the parts that matter for your environment without taking over the maintenance of the entire script (or unit file). For example, it's more likely to keep working after an upgrade.
Personally I'd rather clone and modify a full config file rather than only overriding portions. Once I own it, I want to keep owning it - it's more straightforward to track down a problem due to my own assumptions being invalidated, than due to a distribution's assumptions being invalidated. And it's much nicer to read one file for the settings, than having to jump between a few different ones and know the overriding rules (referring to configuration in general here, not just systemd which provides the tools to work with them).
Good point about the daemon-reload. I actually don't have many custom systemd unit files, and they generally don't change. But to be fully correct and not require manual intervention or a full reboot I do need to include that step (although I'm moving away from Debian towards NixOS so shrug).
The biggest recurring issue is with graphic cards that have poor support in Linux and require firmware blobs and with WiFi cards that have no free drivers / firmware.
It is trivial to use vanilla Linux with these machines as Guix System lets you use any other package for the `kernel` field in the configuration. The "nonguix" channel (hosted on Github) provides packages for vanilla Linux ("linux") in various versions, and also includes firmware packages that you can use in your config file.
If that's the only thing keeping you from using Guix System then I'm happy to say that it's a hurdle that is very easy to overcome.
Intel, AMD, and Nvidia GPUS - all of them need firmware blobs to work as expected. Intel and AMD offer free drivers but not free firmware.
And if you want to WiFi 5 (ac) or WiFi 6 (ax), chances are that you'll need non-free drivers/firmware as well.
> The "nonguix" channel (hosted on Github) provides packages for vanilla Linux ("linux") in various versions, and also includes firmware packages that you can use in your config file.
Ah, that's good to know.
I'll probably give the Guix package manager a shot some day.
If the whole reason you're doing it is to migrate back to upstart's syntax then it seems that would be not so useful, you might as well go in the other direction: write an upstart generator for systemd, and then you can just plug that into ordinary NixOS.
Edit: Actually I think at some point NixOS did support upstart? Not sure whatever happened to that.
There is no chance that my Nix changes would affect existing configurations. They would be completely self contained.
I had some users too but it still wasn't really worth it. I can't suggest it as a business model or a hobby project, you are basically going to be highly constrained by everything else on the system and your setup will just start looking like everything else eventually. Systemd also works fine as a session manager for whatever.
Those other things you mention are confusing to me, gvisor and KVM are mostly orthogonal to container management. And FreeBSD jails don't work on Linux.
Google Cloud uses gvisor for their K8s offering; AWS and Fly.io use firecracker for their container offerings.
Cgroups in v2 can be delegated easily and cleanly. As well as namespaced. Systemd or no systemd. Systemd just makes your life harder if you want to do rootless containers without integrating with them.
IIRC the GP comment was asking about NixOS, not Nix. If you have everything already going through managed K8s or firecracker then I don't understand what you are using NixOS for. You could just install Nix on some other distribution that uses whatever init/container setup you want.
"Cgroups in v2 can be delegated easily and cleanly. As well as namespaced."
Right but none of those other things that were mentioned support cgroups delegation at an OS service level, only systemd does. Unless they have added this recently and I missed it. And if you're just using this to run a hypervisor then you're bypassing all that completely.
I don't see what you mean systemd makes it harder, you have to do basically the same process in any container manager if you want delegation. This is part of the design of cgroupsv2, it's not something systemd came up with. I'm sorry if I'm asking stupid questions but I honestly am really confused what your use case is, and your explanations are just making me more confused, so maybe something got lost here.
Delegating a cgroups doesn't take systemd, by the way. You can do it with mkdir().
Systemd puts user processes in a part of the cgroups tree where they can't work freely, by default. You have to use a systemd unit or the dbus api to break out of that box. That is a design decision on system's part. If I was putting user processes in cgroups, I would give them space to work amongst themselves.
AFAIK you actually should not be using mkdir, because it is racy to have more than one process writing to the toplevel hierarchy. It's only safe to do that from the cgroup manager. That's what I've seen with all the existing implementations anyway.
"Systemd puts user processes in a part of the cgroups tree where they can't work freely, by default. You have to use a systemd unit or the dbus api to break out of that box. That is a design decision on system's part."
It's a design decision that was made because of cgroupsv2. You should really read systemd's documentation on this if you haven't, it describes in detail why this is. https://systemd.io/CGROUP_DELEGATION/
Edit: and also the kernel documentation on delegation. https://www.kernel.org/doc/html/latest/admin-guide/cgroup-v2...
In particular, any other container manager that wants delegation needs to do this same thing using a similar mechanism that you initiate by configuring it somehow. (It doesn't have to be a unit file or dbus call obviously.) Processes cannot just freely break out of their cgroup, that wouldn't be secure. Sorry if you know all this already, maybe a reader doesn't. Or maybe this is helpful to you if you want to add this to startup, I don't know.
"If I was putting user processes in cgroups, I would give them space to work amongst themselves. "
Ok now you lost me again... is this not exactly the purpose of cgroup delegation?
Source? And define top level hierarchy? The top level isn't really special.
This is why locks exist btw.
I'm not convinced you have to manage cgroups from one process. That is not how most implementations do it.
>Processes cannot just freely break out of their cgroup, that wouldn't be secure.
Processes can freely migrate ("break out" if you insist on that terminology) if they have write access to one level above them in the tree. Systemd does not put you in that situation, but it could.
E.g. if you are in /user/<uid> you have nowhere to go, but if your are in /user/<uid>/default you can go to /user/<uid>/<container_runtime>/<unique_container_id>/default.
"Because the resource control interface files in a given directory control the distribution of the parent’s resources, the delegatee shouldn’t be allowed to write to them."
You could add locking but that would be basically doing what systemd/docker/runc does and adding a new API on top of it which is then available through D-Bus or whatever. The top-level hierarchy is the top of any cgroup tree, if you have two processes writing to that without synchronization then they will potentially stomp over each other's values. You technically can go and run "sudo mkdir" inside your docker's toplevel cgroup but that would probably break things.
"if they have write access to one level above them in the tree. Systemd does not put you in that situation, but it could."
Well it does do that if you turn on delegation by giving you your own sub-tree, it doesn't do it by default because most services are not container managers and don't need an additional sub-tree. If you don't do that then write access is not allowed by the cgroups API, see above. It would not really make sense to allow a child process to say "I am going to take 100% of the cpu controller now and you get none sorry", that would defeat the purpose of cgroups.
1 of the reason why I like MBP is the sturdiness and build quality, but i'm ready to go on diy stuff and leave apple for my next laptop. I was watching Dave2D video but when I saw the spring of the hinge when opening the laptop and the flex on the screen, it kind a scares me. How does it go for all of you? How's the build quality? also how's the track pad compared to mbp?
I don't want to own a laptop where opening the screen with 1 hand feels like it's going to break it. Kind a feels like dell inspiron all over again
It's much better than any other laptop I've used aside from a MacBook Pro. The screen does flex a bit but unless you're actively applying torque to it, you won't notice. It's totally fine when opening and closing it.
Looking at Dave's video, the only thing that's holding me back was when he try to flex the screen :(
Honestly the build quality of MBP is so good that it's holding me back from changing into something else. Hopefully V2 of framework will come with the sturdiness, no brainer for sure
Opening it with one hand feels great. The thought of it "feeling like it's going to break" has never crossed my mind.
In cases where you may damage the device, part of the beauty is that all of the parts are sold by Framework and come with replacement guides and any tools you may need (usually just the included screwdriver).
I feel like a total idiot at least a couple times a week while I’m holding my phone in one hand, saying “hang on, let me check” and doing weird contortions to get my laptop open with one hand using my belly to hold the bottom down.
The track pad is basically the same story. No real complaints but it could be better.
I like this aside and I noticed the same when I tried to use Gnome a couple of years ago. I asked a coworker what he recommended, he showed me KDE and I've never looked back. The scaling was great too when I needed it for my MBP's retina display.
Avoid doing this unless you have no choice.
As I said, this setup has two big drawbacks
- I can't use my laptop display without adjusting the font size - Almost all GUI apps look strange because the text is large and icons are small. For example, the uBlock Origin icon on Firefox is tiny and sometimes I make mistake and click elsewhere.
Scaling fonts and not the display is a hack but considering fractional scaling is even more messy, I've resorted to using this method.
This is why I said that I won't buy the FrameWork laptop because it needs fractional scaling.
The most obvious solution that fixes all of these problems is to buy displays that work with 2x scaling. A combination of 15 inch 4K or 13 inch 3000x2000 laptop display and a 27 inch 5K external display is the ideal combination because both of them would work with 2x scaling.
I’ve seen this warning in the Ubuntu docs. Can someone explain exactly why a GPU driver would cause issues with disk encryption and UEFI boot?
[1] https://askubuntu.com/questions/1023036/how-to-install-nvidi...
Of course the low-quality binary blobs Nvidia ships can always contain some annoying bugs, so maybe this was fixed already and the author is just cautious or maybe I will run into those problems eventually.
I don't think it's a problem with the current iterations of Nvidia/Linux, though.
As for what can cause the problem, I have several ideas:
- Secure boot and code verification breaking driver loading
- The Nvidia driver, loaded during early startup, freaking out about config files not existing
- Some kind of handover problem between the fallback VGA driver and the proprietary blob that wouldn't exist if the boot files would be available from the bootloader directly
- Nvidia is known to track boot state in their GPU firmware (one way they used to prevent virtualizing consumer GPUs) so it could be related to a second attempt to initialize the GPU
Unencrypted boot leaves me vulnerable to a malicious actor taking my laptop, modifing grub and/or linux kernel to send and/or store my password to unlock the boot volume and then handing my laptop back and me typing in my un-lock password
Possible, but I view that threat as low
The issue I have seen is that a LOT of BIOS (at least on desktop machines) seem to just let you reset the secure boot enforcement status via the BIOS reset jumper... which defeats the point although it does raise the physical access requirement to a longer duration.
I’d love to put an underpowered qcom or nxp arm64 chip in and get insane battery life, but that doesn’t seem likely for a while since it needs 4x usb4 :/. Even the Xavier nx SoM only has 1x4 lane pcie..
I don't think it's rude. It's about choice, just make sure it's clear to the users what sort of peripherals they should expect to work.
That said I think it would be almost more fun to put the arm computer into the USB C module.
Your numbers are clearly suspicious. Did you ever take measures to circumvent the CPU throttling bug under Linux on your X1E? See [1].
Also, regarding your battery comments, when my X1E G1 was new I got great battery life, around 7-8 hours. I still get 5ish.
> requires LUKS full-disk encryption, and having it setup conflicted with NVidia’s proprietary driver.
This just seems bizarre and hard to believe to me; once mounted your LUKS drive is no different (to software) than any other block device.
[1] https://wiki.archlinux.org/title/Lenovo_ThinkPad_X1_Extreme#...
I'm hoping to start with nix+home-manager at some point, maybe on top of PopOS; anyone have experience with that setup?
It's also worth noting that many of the benefits of Nix are directly driven from being able to write custom Nix expressions. So if you want to reap all of its benefits, it's definitely worth learning. The core language is small and the experience is not so far off from writing JSON, so that part is relatively easy to pick up. The actual work lies in learning how packaging is done in Nixpkgs. This is best done by grepping around the Nixpkgs repository for examples. Nixpkgs' policy of having all packages in a single repository really shines here.
One of the things I really appreciate about Arch is the flexibility it offers. I'm not forced to use GRUB if I don't want to, unlike Fedora and Debian. I'm also not forced to see dm-crypt UUIDs on mounted disks like Dracut does. I can choose what I want. This latter issue was present in NixOS as well the last time I tried it.
Oh, I'm also not a fan of how unbearably slow the nix package manager is compared to pacman.
I do recognize the advantages of declarative configuration of your entire operating system though.
Building a declared environment takes quite a while and switching a NixOS or home-manager generation takes even longer. It was my main complaint about NixOS when I was starting out.
The thing I realised though is that the main scenario where you actually need the speed is when you want to try things out and for that, Nix' ad-hoc environments are much better suited.
When I want to try a new program (btop for example), I simply run `nix-shell -p btop` rather than adding it to my declared environment and sitting through a `nixos-rebuild test`. That's usually faster and, more importantly, cleaner than installing the package to a global environment with a regular package manager.
MacOS has second tier support to Linux. I'm sure PopOS will work well
https://github.com/NelsonJeppesen/nix-lifestyle/tree/main/ni...
The only factor that I could think of that might make a difference across distros is the availability of hardware video acceleration in browsers. But that gap is likely to close soon, if there are any right now.
I should probably chime in on the forum, but per that thread I tried Wayland but it didn't go so well, particularly with an external monitor plugged in. Too many misbehaving and misscaled apps, I remember having issues with vscode (scaling) and either kitty or alacrity (can't move or adjust window).
Great to hear about the 8 hour duration as well, I’m absolutely loving this new toy!
Are third parties able to make replacement parts and dongles with Framework’s blessing? Or are they trying to lock in the parts market?
Looks amazing!
Desktop wise, I’ve been using PopOS for 7-8 months now, so it’s quite great to have that same setup in a portable fashion now.
- changing default shortcuts from CMD to CTRL. Is swapping the positions enough?
- Janky scrolling/no inertia ( i'm not on a framework (hp x360 spectre), is the framework better?
- all the small tuning software i have (Karabiner for remapping, Amethyst for tiling, BitBar etc)
- is there a quick look equivalent on PopOs?
I strongly recommend not to do this.
I did it and it was a nightmare: some of mac's hotkeys are like linux hotkeys but reversing Meta & Ctrl, but not all of them. I adjusted the most egregious hotkeys, but there were always more. This went on for around a month. In the end, it was impossible for me to get used to the differences: my muscle memory would betray me many times over the day, and it wasn't getting better.
Then I deicded to just reset everything back to my distro's defaults. It was painful for a couple days. I made a few mistakes in the next week... and then it all got in muscle memory. I can even switch from working on a mac to working on linux without issues now.
I should note to the OP that I almost exclusively use my HHKB[0] for all of those operating systems, which places the CTRL key where the Caps Lock normally is, abstracting caps lock functionality to a function key.
When I do use a different keyboard (or the laptop’s native keyboard), it takes a few minutes to adjust and every now and then I’ll accidentally tap the wrong key, but overall I would say negative impact is negligible
[0] https://www.hhkeyboard.com/uk/products/ (I have the Pro2, and for some reason the official site is not loading for me. This keyboard has lasted me 4+ years and I use it for gaming, coding, and everything in between. By far my favorite keyboard I have ever used, the toupre keys feel soooo so satisfying. Amazon link here: https://www.amazon.com/dp/B07K9QHF4P/)
For me, I just adjust to using CTRL instead of CMD. It only takes me a day or two of consistent use for the muscle memory to switch over.
The most painful part is if I try and go back-and-forth between the Mac and Linux laptop, that causes pain. But if I just switch pretty completely over to the linux environment, the muscle-memory switches over pretty quickly.
If you don't need to go back-and-forth on an ongoing basis, I'd recommend just pushing through the transition period.
PopOS has a decent tiling manager built in now. (Icon in top right corner by default) though I should note it isn’t as “hardcore” as classic tiling managers and is instead built primarily for approachability over pure power.
Not exactly sure what you mean but quick look. If you mean spotlight, PopOS has a (imo better) equivalent bound by default to the Super key. If you mean the applications overview, they have that as well by default on a button in the top left of screen. Can also be bound to a shortcut.
Quick Look in this case is the quick open of files through finder by tapping 'Space'. Might be some equivalents, i'll look around.
I've used Nvidia binary drivers with LUKS full disk encryption for years, on laptops and on desktops, and it's just fine. I can't even think of a possible way they would interact.
The framework laptop looks interesting, but there are so many misunderstandings or inaccuracies in this writeup that I personally do not trust the benchmarks or other details to be accurate.
Funny, I much prefer a full-size Right-Shift key, which usually requires compact up/down arrow keys. I hate typing on keyboards with tiny, stunted little right-shift keys. I use Right-Shift much more than the arrow keys.
This, in turn, is a trade-off for manufacturing. It will increase the cost of everything slightly. The keyboard unit itself as it will be larger, maye be harder to produce. Machining or producing the chassis now requires accounting for an additional cut.
This is all solvable, yes. Though it is harder for smaller scale productions. It is likely they are working with vendors that, themselves, only have "entirely rectangular" keyboards as this is what the market ends up wanting for cost reasons.
I think the asymmetric, row-staggered layout that keyboards have discourages this.
Though, a fancy technique many keyboard-enthusiasts will use is to have dual-use keys. e.g. Tap-Hold dual use. The key acts the same when tapped, but acts as a modifier key when held.
e.g. changing Capslock key to escape when tapped, Ctrl when held is probably the easiest to think of.
The best use of tap-hold is for Home-Row modifiers. Putting alt/command/ctrl/shift as modifiers of asdf (and symmetrically, jkl;).
This looks super interesting for a long-term laptop, but small arrow keys bring me so much day-to-day anguish I wouldn't want to suffer through them long-term.
Of course, Framework doesn't want to have to manufacture different chassis or whatever for them either, so it's believable and makes sense. (And I'm sure it's true on some level for competitors, even Apple - surely it's basically compatible and swappable, they just don't sell them.)
What about AMD?
Good news, suspend resume has been broken on my AMD ThinkPad, at least for me, since kernel version 5.10 came out although it didn't really work reliably before either.
Whenever, I dare enter systemctl suspend, my screen goes blank, fans start spinning at the highest speed. A hard reboot is needed at that point. The journal is filled with these error messages.
[drm:amdgpu_ib_ring_tests [amdgpu]] ERROR IB test failed on gfx (-22)
EDIT: Looks the author of the blog post isn't able to do suspend resume on his AMD ThinkPad T495 either, which is good news.
PS: waiting for shipment to Brazil (and maybe a 15'' model)
I don't know about the compilation procedure but how hard is it to spin up a beefy CPU intensive server ($300/mo = $0.2 per 30 min use), let it git pull your work, compile, upload the binary to some storage, notify you and shut down automatically?
This way, you can close your laptop and move while compiling.
Much cheaper than not being able to choose the laptop you want the most.
These days, you can be a bit more creative than do it all on your machine.
Related famous last words: https://news.ycombinator.com/item?id=9224
Also, what expense? $0.2 per compile instead of buying a laptop with better CPU?
I thought the same - then I bought in September 2021 a "Lenovo Ideapad 5 Pro 16ACH6" which is when I found out that the "Realtek Semiconductor Co., Ltd. Device 8852"/"Realtek RTL8111E Ethernet LOM"/"Lenovo Device 4852" needs a brand new driver/module, respectively is not compatible with anything else.
I lost a lot of hair until I found "https://github.com/lwfinger/rtw89.git" (thaaanks a lot!!!) which made it work (reliable in my case, even with notebook "suspend-to-RAM").
As of kernel 5.14.5 (on Gentoo) embedded support for that NIC is not available, nor I think in 5.14.12 (not 100% sure - I did not install it, just peeked into it) :(((
Cool to see you getting NixOS running, Arch on the Framework has worked flawlessly. So much nicer than the X1.
From the software freedom point of view it's very much a no go, although they are apparently going to release the firmware sources and looking into coreboot[2].
[1]: https://community.frame.work/t/does-intel-chip-have-intels-m...
[2]: https://community.frame.work/t/free-the-ec-and-coreboot-only...
boot.kernelPackages = pkgs.linuxPackages_5_12;
but slowly 5.12 and 5.13 were both removed as specific kernel versions in nixpkgs, so it's
boot.kernelPackages = pkgs.linuxPackages_5_14;
now which still works.
I eventually plan to just remove the line and use the default kernel, once the default kernel is bumped up to 5.12
There's a ton of interest, just hoping that it could happen in some form:
https://community.frame.work/t/any-chance-of-trackpoint/1026
I first discovered frame.work on hacker news, watched a ton of youtube videos and got all jazed up and bought one, then waited for it to ship (about a month since they do it in batches). I actually never had been this excited about a product as this, and would read up on their site daily, to prepare.
Anyway, it arrived, and I assembled my first laptop! The hardest part was hooking up the wifi, but with a little thinking it worked. Took about 10 mintues, with kids running around the background.
Having a month to ponder all of the flavors of linux, I was initially jazed about NixOS as well. I'm not a daily linux user and reading through the documentation was discouraging. There doesn't seem to be a simple way to install it that I have discovered. Eventually I discovered Fedora Silverblue which has the same concept of an immutable OS. I still want to try NixOS, but I will need help with that. Another nice thing about these devices, you are free to install whatever you want on them, unlike Apple which locks it down. The Fedora Silverblue installation was about 4 minutes, minimal decisions to make. So that is what I ended up with.
At first I tried version 34. This did not work immediately with the WiFi, and that had me concerned (what if I didn't hook it up right?). LOL.
So, I tried the RAWHIDE release (which will be version 35). Mostly I did this to see if the wifi worked and it did! After playing around with it, I started getting used to it. It works with wifi, the fingerprint reader (great), bluetooth, trackpad (see comments below).
Build Quality. The Framework build quality is excellent. It's much lighter than the MBP, and I think the reason for feeling a bit more 'flexible' ie bendable is the lack of glue. I think the glue also helps the accoustics. The speakers are definately a disapointment compared to the MBP. That said, I would trade ability to swap out battery for feeling more solid. Maybe this can be improved, but to me it's not a huge issue.
Trackpad. I am not entirely used to linux, so at first I didn't know what was a hardware or what was a software issue. The two finger scrolling works, as does three finger navigation, but scrolling on the browser sucks. Two issues, acceleration is a little wonky (no rubber band feel), and scrollbars are always visible. The claim is that wayland does not have some universal scrolling system but relies on each application to implement it. Firefox, chrome, Ungoogled Chromium all seem to act the same. Some of this might be Fedora 35 which is still being developed. That being said, I've tried recent windows machines this experience is better than that. I think the MBP does have a smoother experience here, but this is usuable for sure.
I am now starting to get used to the linux experience (gnome). I like the screen, wish it was a bit bigger, but certainly usuable.
I have yet to get a second or third monitor working on here.
As a basic example, I could have a list of basic packages that I want on all of my machines, and then depending on which machine my config is being deployed to, I might extend or shrink that list using actual code, because that list is represented as a an array or a linked list and I can directly manipulate that (though, at least in Nix I think everything's immutable, so you would technically only use that initial list to create a new list from it).
Took about 6 minutes, is that long nowadays?