💾 Archived View for tilde.team › ~kiedtl › k1ss › faq.gmi captured on 2022-07-16 at 14:21:34. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2020-11-07)
-=-=-=-=-=-=-
May you find what you seek.
That's up to you to decide.
I'm not going to try and sell you something I'm giving away for free. Have a read of the website and its wealth of information, try KISS in a virtual machine (a chroot is also an option) and come to your own conclusions.
KISS is an acronym for "Keep it simple stupid" (notice *no* comma).
Stupid does not refer to the user. To quote Wikipedia:
The principle is best exemplified by the story of Johnson handing a team of design engineers a handful of tools, with the challenge that the jet aircraft they were designing must be repairable by an average mechanic in the field under combat conditions with only these tools.
Hence, the "stupid" refers to the relationship between the way things break and the sophistication available to repair them.
Packages are updated as quickly as possible while at the same time tested to ensure that no issues arise in the new version. A new version of a package will be held back if it causes issues.
When a package update brings in a bug during the build process or at runtime, it will be patched in KISS until it is fixed upstream. This will typically happen hours after the version is available and doesn't cause a delay.
Nothing prevents you as a user from choosing the update schedule you would prefer. Total control is in your hands. KISS merely provides you with an always up-to-date repository pool.
Every user of the distribution owns their system in its entirety. The management of the distribution which extends to the management of updates to the user's system is entirely optional.
All KISS systems contain the full sources for the distribution and each user has the means of managing and maintaining their machine on their lonesome (or by forming a collective and secondary "upstream").
This includes:
The repository updates are simply a 'git pull' which results in each user having a full copy of the git repositories on their machine at /var/db/kiss/.
As these are all simple shell scripts, the installed "binaries" _are_ the unchanged source code. All of these are self-contained and separate programs.
The 'kiss-outdated' script allows one to check their system for outdated packages against the repology.org database independent of whether or not repology has support for KISS itself.
The 'kiss-export' script allows one to convert an installed package back into a redistributable and installable binary tarball. Simply run 'kiss-export pkg_name' and a tarball will be created in the current directory.
As the website sources are merely plain .txt files, the docs are available in /usr/share/doc/kiss and are readable in less, vim or the program of your choosing.
In addition to the reasoning given for the package manager above, KISS has no lock-in regarding how the machine boots. Were the init scripts to disappear, there would be no real loss.
To continue further, the init scripts need not change. There is no need to update them. Any extensions to them can be made via the /etc/rc.d directory or the /etc/rc.conf file.
What all of this means is that were the GitHub or website to go down, there would be no loss in code, documentation or the distribution itself.
It also extends to each user having everything they will need to continue the distribution for themselves or for other users. A change of git remote is all that is needed in the latter case.
KISS was designed to be maintainable by a single person. The repositories are kept small and focused. There is no requirement or _need_ for infrastructure of any kind.
A user today could choose to go their own way with their system. Everything they need is already in the existing system. The power is in your hands. You are free.
No one has packaged it yet.
Do it yourself.
KISS uses busybox init by default. There is however, no lock-in and the user may use whatever they like. The distribution's boot up scripts are written in an init-agnostic way and act as a base for every init to use.
KISS uses busybox coreutils by default. There is however, no lock-in and the user may use whatever they like. All shell code is written in portable POSIX shell and coreutils invocations.
Barring the above exceptions, users are running KISS without the use of busybox. Popular alternatives are Suckless' sbase/ubase and GNU's set of core utilities.
DRM (specifically the commonly used Widevine) does not work when musl is the C library. The DRM is a binary (typically downloaded by the browser) and it requires glibc.
Nothing can be done to fix this as the DRM is not open source.
Workarounds:
Short answer: With one command? You can't.
The package manager does not do recursive dependency removal on removal of a package. This error-prone automation will not be added to the package manager.
Instead, the workflow is to remove the single package and then look at the output of the 'kiss-orphans' command to see what can now be removed. This command will list all packages which have no relationship with other packages, otherwise known as orphans.
This list may include Firefox and other "end" software so a brain is required when parsing the list. You'll come to learn the relationships between packages and their dependencies and this will eventually become effortless.
The kernel not booting can be a variety of issues. This is almost always related to a configuration issue in the kernel, /etc/fstab or the bootloader.
KISS doesn't use an initramfs so the configuration of the kernel may have different requirements to other distributions.
KISS technically supports booting via an initramfs, it just doesn't require or provide one. As a user you have the means to set this up yourself for your system.
Full disk encryption is also possible without the use of an initramfs in modern kernels (see dm-mod.create).
The initramfs concept is an ugly, complicated and largely optional mess. Thank god it isn't a requirement.
The kernel must be compiled by the user for a variety of reasons.
The user maintains full control over all aspects of their kernel and further, their entire system. There is no lock-in into a specific set of kernel sources, version, use of proprietary firmware, patches, config, etc.
The user decides:
You as a user might actually learn something too. You may come to understand your hardware, what drivers it needs, how the kernel works from a configuration perspective, etc.
You should have a better understanding of this part of the system afterwards and you'll be able fix any issues at this level with relative ease.
Remember, it's only hard the first time. Once a working config is created, no further work should need to be done each time you update your kernel.
If the new kernel version has an issue with your hardware, simply boot another from your backups. If new hardware was added, simply run 'make menuconfig' and add what is needed.
I'll say what I always say. Nothing prevents anyone from providing kernel binaries and an initramfs generation tool for KISS. Just don't wait on the BDFL to do it for you.
Eases distribution maintenance for the BDFL. Shipping a generic kernel demands a humongous, module heavy kernel with support for everything under the sun. An initramfs is then a requirement to boot this damn monstrosity.
There is then a need for a _portable_ initramfs generation tool which to my knowledge doesn't exist. There can't be a dependence on bash or anything outside of core. All initramfs tools are either distribution-specific or non-portable to KISS.
Then there'd be endless support requests for tweaks, additions and removals to the distribution config and the burden of updating the kernel on every release.
Does KISS ship the latest kernel? The long term support kernel? Both? Some users require firmware so we'd need two separate binaries, one for linux-libre and another for regular linux.
What if the latest kernel has issues on some hardware? New builds and binaries must be released with backported patches.
It's a large maintenance burden for something which can simply be solved by the user doing this themselves. The user maintains full control over every aspect of their kernel and is solely responsible for it.
Why should it? See above.
The kernel headers in KISS are pinned to an LTS kernel version and are only updated when there are changes of interest in the kernel or headers themselves (usually by users requesting new features available in the newer headers).
The headers are backwards compatible and are fully usable with a matching or NEWER kernel version. There are two cases where you'd be required to update the headers yourself.
Here's an excerpt from the kernel's documentation which may explain the situation better than I.
Kernel headers are backwards compatible, but not forwards compatible. This means that a program built against a C library using older kernel headers should run on a newer kernel (although it may not have access to new features), but a program built against newer kernel headers may not work on an older kernel.
https://www.kernel.org/doc/Documentation/kbuild/headers_install.txt
This is almost always an error in your CFLAGS/CXXFLAGS. Ensure that you have used -ONUM (CAPITAL O) and not (lowercase o) or (zero 0).
If the above doesn't fix the issue, try building the package with:
$ CFLAGS= CXXFLAGS= LDFLAGS= kiss b pkg
You'll then be able to discern whether or not this was the issue.
This is likely due to you running 'modprobe amdgpu' or equivalent post (or late) boot which causes udev to miss setting up the GPU device in /dev/dri.
There are three separate solutions ranging from easy to hard. Either of the three will solve the issue.
Enabling the service also turns on device hotplug and other features. This isn't the default as there's no need for the daemon to run on a static system.
$ ln -s /etc/sv/udevd/ /var/service
Running 'modprobe' in your inittab or in rc.d will result in it executing *after* udevd has finished setting up devices.
By moving 'modprobe' to /etc/rc.conf, it will load the driver before udevd executes which will allow it to detect and setup the device.
This is trickier but the best solution overall. Instead of using modules, compile the firmware into the kernel.
You'll then not have to worry about udevd/modprobe as the kernel will automatically load the driver as early as possible.
This may include compiling proprietary firmware into the kernel as well which is a little bit finicky but doable.
Most users will typically use 'ffmpeg' as their screenshot tool as they'll already have ffmpeg installed for MPV or Firefox. Other options include 'scrot', 'imlib2' or 'imagemagick' (import).
If you'd like to use another tool, package it yourself. Remember, if no one has packaged something it means that no one has needed said package. It is up to you.