You are not logged in.
1. Update repo list of packages via apt-get update
2. If there is still error try to install a package with aptitude. Aptitude can help to solve broken dependencies easily.
Yes, you're right. XFCE has a dependency on libpulse0 but that only recommends pulseaudio itself. The GNOME desktop has a hard PA dependency though, perhaps unsurprisingly. No need to remove it anyway: just stop it from starting if you don't want it, as noted by dxrobertson earlier in this thread.
My point was that all the standard desktop environments include PA even if it's as a recommended component rather than a dependency per se.
GNOME also has a hard SystemD dependency via logind and its developers consider X11 as deprecated graphical subsystem. There is no doubt: GNOME is extremely terrible DE by design, UI and UX.
Really good DE such as TDE and LXDE don't depend on PA.
Most of "standard" DE include PA by default as recommended component not for a good reason. The reason is simple: PulseAudio is SystemD of world of sound.
The only people who don't use Pulseaudio these days are hair-shirt minimalists, all the desktop environments have PA as a dependency because it provides a convenient high-level interface for controlling how multiple sources are connected to sinks.
That's not true.
For example, almost all new laptops have HDMI outputs and if pure ALSA is used then it is necessary to configure it to make the inbuilt speakers work, most non-technical users don't know how to do that and so need PA to do it for them.
On every laptop i seen ALSA provided built-in speakers working by default. There is no need to configure ALSA. PA is useless garbage devouring CPU and screwing sound.
You can't run vanilla Devuan because you need custom set of drivers. You can make filesystem via debootstrap and run Devuan with custom kernel for Droid 4. GUI lags may be caused by slow videodrivers.
While Raptor motherboard is probably free of proprietary blobs, the POWER9 CPU is most likely not?
And all of them are not affordable for me.
If you want affordable computer with blobless firmware - go buy Thinkpad T60 and flash libreboot.
I agree with you that X86 standard is versatile and open for vendors to produce X86 hardware, but is the most unfriendly for people who would like to write their own open source boot loader, that is what I primary meant under open source.
One word: coreboot
ToxicExMachina wrote:BIOS is opensource: https://www.seabios.org/
UEFI is also opensource: https://www.tianocore.org/SeaBIOS originally looks like a BIOS for virtual machine guests.
Can SeaBIOS be used directly on any physical motherboard without Coreboot hack which I already mentioned?
Is not SeaBIOS just one of many other possible payloads like GRUB or KEXEC, etc. for Coreboot/Libreboot?Where would be SeaBIOS without Coreboot/Libreboot projects which actually are apposite of what was intended for "openness" (actually lack of openness) of X86 boot loader?
It doesn't matter. Every vendor had own BIOS implementation because it totally depends on hardware by design. BIOS is a part of MS-DOS designed to provide abstraction layer for compatibility purposes. SeaBIOS is another implementation. In this case Coreboot+SeaBIOS == BIOS.
ToxicExMachina wrote:different MIPS cores
Can routers running LibreCMC be treated more secure in terms of my control over their boot loader?
Bootloader is not the thing you should worry about right now.
ToxicExMachina wrote:OpenRISC, RISC-V, old ARM versions implementation, different MIPS cores, etc. You can check some of them at opencores.org
RISC-V is promising project because large organizations decided to support it.I would be glad to try RISC-V, but where to get an affordable board?
What do you mean under old ARM version? Is Cortex A7 old enough to be secure enough?
I was looking for a board with open source boot loader when it would be difficult to inject an invisible and undetectable virtualization trojan on the factory or by a third party blobbed software which could reflash firmwares silently.
1. The most affordable RISC-V board is any cheap devkit with FPGA and i/o ports. You can flash RISC-V based SoC.
2. Older ARM versions - older than modern ARM.
3. Buy board or laptop compatible with upstream coreboot.
ToxicExMachina wrote:If you want REALLY opensource board - find one with RISC-V CPU based on opensource core, and vendor must provide full docs under libre license. You can also try to flash FPGA.
\o/ RISC-V \o/
And there is the J-Core Open Processor.
—▶ https://www.youtube.com/watch?v=o0milqmt4aoI think we need to keep an eye on this one too...
OpenRISC, RISC-V, old ARM versions implementation, different MIPS cores, etc. You can check some of them at opencores.org
RISC-V is promising project because large organizations decided to support it.
This core is also much more interesting than J-Core: https://github.com/riscv-boom/riscv-boom
The browser appears to be webkit based and they don't update the AppImage often enough to cover the steady stream of vulnerabilites associated with that engine so it's probably best not to use it at all.
It's based on qtwebengine. It can use both webkit and blink backends. If they don't update appimage it's very bad because this is the only easy way to install and run such application on any distro.
ToxicExMachina wrote:I would recommend to search for NXP i.MX based boards. However, in general ARM is just a pain. x86 based boards are much more "open source" than any ARM based board.
Please explain how X86 is more open source than ARM?
First of all all of boards you mentioned here are not open source boards.
ARM is almost a pure hardware RISC processor executing supplied instructions directly on its hardware at once.
But X86 is a virtual machine like a Bochs X86 VM generally run on a very high performance hardware RISC processor for which we do not have any documentation at all, look at VIA C3 as an example,
Intel and AMD are the same idea but significantly more optimized for high performance.
It's not criteria of open source. Microcode is conception from the middle of XX century.
Even modern Intel and AMD processors are "RISC" under the hood - they decompose CISC x86 instructions into RISC micro-ops, this isn't a VIA phenomenon
Everything is TOP secret about X86 CPUs actual hardware internals, its hardware implementation, its native RISC like assembly instructions, implementation of X86 virtual machine and backdoors injected into this VM yet on the factory and even BIOS initialization procedure at the very beginning of the boot time has so little documentation that only a few boards have unofficial Libreboot firmware for them from hackers who somehow reversed BIOS fragments and/or schematics and initialization sequences.
It's still doesn't mean open source.
A few more boards have been made compatible with Coreboot and the most of them are too old and most likely some docs just leaked from laboratories to help coreboot developers to reverse engineer the boards boot process (initialization of RAM modules, etc.).
And ARM initialization process is almost completely open source for some boards, though a few code is still unknown too which is executed before uboot just after getting a reset signal. Can it include some trojan running in a TrustZone isolated and undetectable from OS, may be, who knows.
It's also not open source.
Uboot needs to be customized for each board, but a few boards have it open source even officially.
For a few boards like AllWinner A20 we have all software parts open source: Uboot unlike proprietary BIOS or UEFI for X86 and even mainline Linux kernel like for X86, even open source 3D driver is being under development and already works.
In your terms,
BIOS is opensource: https://www.seabios.org/
UEFI is also opensource: https://www.tianocore.org/
And, of course, those pseudo-open-source boards aren't better than any x86 motherboard as an opensource board.
X86 looks very versatile, has outstanding performance, compatibility with huge amount of add on parts, with fully open source OSes like Linux and BSD supported forever, tens of years after hardware release, it still can run the most recent modern up to date OSes like BSD. For example the most up to date BSD 2019 at least in text mode still runs fine on 20 years old Pentium 1/2 1997 and even X works well enough if the board has enough RAM like 200-300 Mbytes.
X86 completely hides its initialization code and most likely virtualization spying invisible trojans in negative and/or zero rings of CPU booted before any OS code and fully undetectable from OS level.
The situation with ARM is much worse. However, if you consider proprietary solutions as opensource, your words has some sense.
If you want REALLY opensource board - find one with RISC-V CPU based on opensource core, and vendor must provide full docs under libre license. You can also try to flash FPGA.
P.S.
I recommend to learn terms first because your delusions leading you destructive way.
ToxicExMachina wrote:I would recommend to search for NXP i.MX based boards. However, in general ARM is just a pain. x86 based boards are much more "open source" than any ARM based board.
In regards to x86 vs ARM you are right, but the issue is there is not enough ARM developer pc's as Linus Torvald mentioned not long ago. x86 just corners the market now until some clever so and so's put together powerful ARM computers not raspberry pi hobby boards.
x86 has own firmware standard. Initially it was BIOS. Now it's UEFI instead of BIOS. There are mostly same hotkeys, 99% same boot mechanism for motherboards from totally different vendors. At the same time every ARM board (even development boards) has own uboot fork, own bugs, own BSP kit incompatible with others, etc. There are enough powerful ARM workstations even today - see ThunderX based computers. The result is still same: proprietary platform. The only advantage is a lot of devices with ultra low power consumption. That's all ARM has. Some terrible things came to x86 from ARM. For example, AMD PSP is just an ARM TrustZone implementation.
I would recommend to search for NXP i.MX based boards. However, in general ARM is just a pain. x86 based boards are much more "open source" than any ARM based board.
You can also run Otter as AppImage.
It's necessary to know what is your videocard and which videodriver you use.
Konqueror from TDE (Trinity Desktop Environment) can do it. It's better than both KDE4/5 and XFCE4 because today TDE is really lightweight DE and XFCE4 is wasting performance like GNOME.
Try to enable multiarch support: https://wiki.debian.org/Multiarch/Imple … _multiarch
The reason of pulseaudio presence is Debian.
I am wondering if there is also an easy upgrade program for switching kernel for Debian/Devuan as on Ubuntu
It's called GRUB.
ToxicExMachina wrote:Besides, if VPN service require custom proprietary VPN client it's a spyware.
Do you know anything about riseup.net? I suggest you learn more about them. From their homepage:
Riseup provides online communication tools for people and groups working on liberatory social change. We are a project to create democratic alternatives and practice self-determination by controlling our own secure means of communications.
They could provide settings for openvpn/wireguard/another vpn client. If they don't want to do that they could pack their client as appimage, tarball or even makeself installer. Snap is anti-community metaproprietary bs (just like flatpak).
It depends on Debian.
I wouldn't recommend to use systemd-based software. For now elogind doing everything good for compatibility but there is no warranty that Poettering won't break logind behavior to make life of an ordinary end user less easy.
Snap format is basically crumpled deb package. You can unpack it yourself. Besides, if VPN service require custom proprietary VPN client it's a spyware.
Pshshshshsaudio and SystemD were made by the same person. Purge pulseaudio first (apt-get purge), then reinstall necessary packages and avoid pulseaudio in dependencies. Some of packages require pulseaudio libraries without pulseaudio itself. You can ignore that packages.