Edit: files.devuan.org is  s - l - o - w .  Unless you want to party like it's 1999, choose a faster mirror. They're listed on this page: https://www.devuan.org/get-devuan

#2
Expert install on old (2011) Thinkpad. It asks twice for me to supply the wireless firmware on a usb. I already know that the iso contains the exact firmware I need, so I decline. Twice.

I unplugged the ethernet cable to make sure I could do a wireless install. Installer loads the firmware and detects the nearby wireless signals. I could not authenticate with my router. It turns out that it works fine with no wireless password or if I change the password to one that contains only lower case letters. The original pass had numbers and special characters.

It's not done yet, but it seems to be going along fine.

#1
Full desktop install in qemu yesterday worked. Wicd is gone, and nothing currently replaces it. Wired connection was working with dhcp automatically set up in /etc/network/interfaces.

Xfce did not ask on first login if I wanted default layout or bare layout. Launchers on lower panel exist but are not set to anything. I'll see if I get the same thing on the second install.

You're running into the changes made to su in buster/beowulf. See the release notes for fixes. https://files.devuan.org/devuan_beowulf … _notes.txt  It's in the "What's new" section near the top.

The print server task includes cups. You could set up the printer by going to localhost:631 in a web browser. When you do any admin tasks, it will ask for root login.

I went back and forth between sysvinit and runit a few times before I realized that elogind was not running. I set it to start in runlevels 2-5 with sysv-rc-conf, but it still didn't run. I ended up reinstalling elogind libpam-elogind and libelogind0 and the delays are gone. It seems to be working normally now.

Aha! I was afraid you might use those lines I posted. You know, the ones that I prefaced with "These lines won't work." 

You must add a debian source line. One like this:

deb http://debian.csail.mit.edu/debian experimental main

You can pick one closer to home if you're not near Boston. Check a *debian* website for more mirrors.

Edit: Re-read the part in my first post about "Be careful!"

(No prob, HoaS. We bring different parts of the message.)

HoaS beat me to it: How are you looking?

I'm not sure which package you need, but this one has a version 21.

apt policy mesa-common-dev 
mesa-common-dev:
  Installed: (none)
  Candidate: 18.3.6-2+deb10u1
  Version table:
     21.0.2-1 1
          1 http://debian.csail.mit.edu/debian experimental/main amd64 Packages
     20.3.5-1 50
         50 http://sledjhamr.org/devuan/merged ceres/main amd64 Packages
         10 http://debian.csail.mit.edu/debian sid/main amd64 Packages
     20.3.4-1 50
         50 http://sledjhamr.org/devuan/merged chimaera/main amd64 Packages

I wouldn't expect the minimal-live to boot on any uefi system. It's not uefi-compatible. Please try the amd64 desktop-live iso. It has both 32 and 64-bit uefi bootloaders. It should boot on your system.

If you choose to install it, I think you'll need to install the grub-efi-ia32 package first, then run the installer and let it install the bootloader.

I don't understand why you think cfdisk defaults to gpt. I also don't see where you can change the partition label on a partitioned disk. It just reports what is there. On an unparitioned disk, the first thing cfdisk does is give you a choice of partition table types. I had to zero the first few MB of a disk before running cfdisk to see that.

cfdisk lists all the internal disks if you run it without any argument. gdisk will ask you to type in the disk you want. cfdisk is for msdos partitioned disks, gdisk is for gpt disks. You can use both in the same session if you want.

For setnet.sh: Setup interfaces, select eth0, Conf, DHCP, done. I usually forget about that and run 'dhclient eth0'.

The new version of refractainstaller will be packaged as soon as one of our local git wizards shows up. It's all set to build. It'll be in ceres within a few days and in chimaera next week. It probably won't make it into beowulf, but I'll post the deb packages on my sourceforge site.
https://sourceforge.net/projects/refracta/files/tools/

The command shown above would be used if you want to get the actual source code of the package to examine it or to rebuild the package differently. If you just want a copy of the installable packages, the command would be

apt download name-of-package

You could do this for every installed package to get a full set of .deb packages. dpkg -l will list all installed packages.  Something like this might work for you.

apt download $(dpkg -l | awk '/^ii/ { print $2 }')

The reason you're having trouble finding devuan packages on devuan servers is because most of the packages in devuan get pulled directly from debian and magically merged with the few pure devuan packages to give the impression that we have a repository of 60,000 packages. The software that does the merging is called amprolla. Our own packages are at pkgmaster.devuan.org. Debian's packages are at packages.debian.org.

Another option would be to download the debian dvd set, the devuan server iso and maybe the devuan desktop dvd iso. We don't provide a full dvd set like debian does, so it's possible that some packages might still be missing. But I think most of what you want will be in the debian set. You will have to pull these packages off the media manually to use them or maybe set up a local instance of both debian and devuan repos and amprolla to do the merge and filter out banned packages.

Source code for debian packages can also be found in their git repo, salsa.debian.org
Source code for pure devuan packages can be found at git.devuan.org

apt-cacher might do what you want, too. I don't know much about it.

Edit: This might be helpful. How to set up a devuan package mirror:
https://pkgmaster.devuan.org/devuan_mir … hrough.txt

Those are the same settings I've used in virtualbox.

No, that's not what I was saying, but I'm pretty sure that would work. What I meant was if rEFInd was already installed to the hard drive and you did a netinstall without adding a bootloader, refind would see that installed system and be able to boot it.

rEFInd sees all the bootloaders in the efi partition, whether they work or not, plus it sees linux kernels that could be booted. If there's a live-CD in the optical drive, it sees that and will boot it. I just tested that with rEFInd that I have on a 256mb usb stick that's probably older than a fair number of devuan users. I keep it around in case I need it to boot something.

You got this same error on hardware (macbook pro) and in virtualbox.

grub-install: error: unable to identify a filesystem in hostdisk//dev/sda; safety check can't be performed.

Is your virtualbox emulating a macbook pro? If not, then what are you doing, exactly? I installed from the minimal-live in qemu yesterday and it all worked. Does the virtual disk use msdos or gpt partition table? Are your partitions really formatted with filesystems?

Check the sha256sum on the download and check that the burn was good.

Edit: How did you get the minimal live to boot on the MBP? That iso does not work with uefi.

If you run blkid without any arguments, it will show all drives and partition, and should say that the efi partition is vfat.

If you can't get grub to cooperate, here some information about using rEFInd:
https://www.lifewire.com/dual-boot-linu … ot-manager
https://www.rodsbooks.com/efi-bootloaders/index.html

Update:  I forked live-config to create live-config-runit which sets up console autologin in the live session. The code for that is in /lib/live/config/0190-runit. It works, but I did not replace the live-config init script with runscripts.

Today I tried making runscripts for live-config. It's set up as a one-shot, and it runs, but it does not run early enough. A comment in the init script says it must start before mountkernfs and in my sysvinit system, there is rcS.d/S01live-config which starts before S02mountkernfs. That link is not present in the runit system.

/etc/runit/1 contains:

/lib/runit/run_sysv_scripts '/etc/rcS.d'

So, /lib/runit/run_sysv_scripts will start the init scripts linked in rcS.d.
If the init script is not executable, then live-config does not run until the desktop comes up, so autologin does not happen, and you get a login screen. I've been making the init scripts non-executable, mainly to make it easier to see which ones are managed by runit when I look in /etc/init.d.

Is there another way in runit to make a script run first? I know a runscript can state what must be started before it, but I haven't seen a way to get to say what it must start before. (everything, in this case).

'update-service --remove live-config && chmod +x /etc/init.d/live-config' set it back to the way it was. Now autologin works again in the live-isos I make.