Re: Install a package to alternative location
On Thu, Aug 17, 2023 at 10:20:30PM +0200, Jeanette C. wrote: > my system partition is as good as full, so I wonder if there is a safe > method to install packages to alternative locations. The packages in > question are android-studio-system and android-ndk. Do you have space somewhere else ? One way would be - Find a directory that 1. takes a lot of space and 2. can be moved to some alternate location without affecting the system, e.g. videos or audio files. Or anything that is required only after booting. - Copy it to some alternate location, e.g. an external USB disk. - Delete the original, and replace it by a symlink. - Add the new volume to your /etc/fstab (or mount it manually after login). Ciao, -- FA
Re: Something going wrong with Firefox 116.0.2
On Tue, Aug 15, 2023 at 11:12:08PM +0200, Robin Candau wrote: > Let's either drop it or focus on the potential Firefox issue initially > stated please. Agreed. I had no idea that this was something 'controversial', hence my question. Ciao, -- FA
Re: Something going wrong with Firefox 116.0.2
On Tue, Aug 15, 2023 at 05:08:46PM +0200, Ralf Mardorf wrote: > A pro for Firefox: Brendan Eich left Mozilla > A con for Brave: Brendan Eich owns Brave > I wouldn't use Brave, even if it would be the best browser available, I > even wouldn't use it, if it would be the only browser available. What is your problem with Brendan Eich ? -- FA
blas update
On Wed, Jun 14, 2023 at 08:27:28PM -, Arch Linux: Recent news updates: Felix Yan wrote: > The openblas package prior to version 0.3.23-2 doesnt ship optimized > LAPACK routine and CBLAS/LAPACKE interfaces for compatibility. This > decision has been reverted now, and the ability to choose a different > default system BLAS/LAPACK implementation while keeping openblas > installed is now provided to allow future co-installation of BLIS, > ATLAS, etc. This must be the most confusing advice I ever got :-) As suggested, I tried: pacman -Syu blas-openblas which resulted in: :: Starting full system upgrade... resolving dependencies... looking for conflicting packages... :: blas-openblas and cblas are in conflict. Remove cblas? [y/N] n error: unresolvable package conflicts detected error: failed to prepare transaction (conflicting dependencies) :: blas-openblas and cblas are in conflict So what should I do if I want to keep cblas, lapack, numpy, scipy etc.. working ? Which blas is now 'the default one' ? TNX -- FA
Re: moving /boot
On Thu, Apr 06, 2023 at 03:41:15PM +, u...@net9.ga wrote: > As far as I remember, what you wrote, including reinstalling syslinux, does > the trick. It's what I did, and it did the trick. > Aren't there wiki articles about these subjects? There are, but they can be confusing. The information is spread over multiple pages, and exploring these quickly feels like entering a never-ending rabbit hole. You basically never know if you've seen all of it and if the info you have is complete. Which is why I asked. It's of course a complicated subject, just consider MBR or GUID BIOS or UEFI GRUB or Syslinux Which already leads to (in theory) 8 possible combinations, and then we are ingnoring LVM, RAID, encryption, and probably a lot of other variations which interact in not always very clear ways. Ciao, -- FA
moving /boot
Hello all, I'm having a problem with updating a laptop: the /boot partition is too small for the new ramfs images. The system uses GPT partioning and BIOS boot using syslinux. /dev/sda1 1M Type = BIOS Boot /dev/sda2 /boot 95M Attributes: LegacyBIOSBootable /dev/sda3 / 93G (others for /home, /data and SWAP) One solution would be to move the /boot directory to / instead of giving it its own partition. The syslinux wiki page tells me that the absolute sector address of /boot/syslinux/ldlinux.sys plays a role in the boot sequence, so I suspect that just moving the boot directory and removing /dev/sda2 from /etc/fstab won't be enough. Would re-installing syslinux after that do the trick (and also take care of the LegacyBIOSBootable attribute) ? If not, what would be the correct and safe way to do this ? TIA, -- FA
Re: netctl problem
On Tue, Mar 28, 2023 at 06:54:19PM +0200, Óscar García Amor wrote: > > Scanning the available docs, I found out that systemd-networkd > > is enabled by default. Could this be in conflict with netctl ? > > Yes, in fact, I advise you NOT to use netctl and to use systemd- > netword[1] directly. The reason for the netctl failure turned out to be that dhcpcd was not installed. Which was a bit of surprise, because netctl needs it. The reason for using netctl is that we want to maintain strict manual control. Anything that tries to be clever and connect automatically to whatever network is available, or that requires configuration to opt out of such functionality, is absolutely ruled out. I've been using netctl for ages, so that's what I know. Maybe systemd-networkd can do the same, but all my experience with how systemd services work confirms that they are designed to prefer convenience over security. And those two you can't have at the same time. A bit of history. There are actually two systems, both are used only for one specific very technical purpose. I installed them three years ago and sent them to the user. They both contain some specialised software that is considered a trade secret and highly confidential. So even if I advised the user to update those systems regularly, that was never done, after installation they were never again connected to any network. So imagine what happened when that special software needed an update. Nothing worked, trying to update the systems just failed and whatever we tried just got us deeper and deeper into a rabbit-hole. So we decided to re-install from scratch. The first one was done a month ago, now we are doing the second one. Kind regards, -- FA
Re: netctl problem
Hi Ralph, > Unless the remote hands are used to precisely editing configuration > files, this might be something where a mistake has been made. We verified the dhcp file, and I'm pretty sure it's okay. My friend may be a linux noob, but he's an experienced engineer and learling fast :-) It's not the first time he has to edit a config file. Scanning the available docs, I found out that systemd-networkd is enabled by default. Could this be in conflict with netctl ? It is disabled here on my system (also using netctl), but I don't remember doing that. Maybe it's done by installing netctl. Or it wasn't set enabled by default at the time when I first installed (many years ago). On the laptop, netctl was installed as part of the pacstrap, so it could be that the disabling of systemd-networkd was not persistent. Just guessing ATM. It will be a few hours before we continue the 'remotely supported installation' - we have 6 hours time zone difference... Ciao, -- FA
netctl problem
Hello all, Yesterday I assisted (via a text chat) a linux noob to install Archlinux on a laptop. We got so far that the system can boot on its own, and we can login as root. The netctl package is installed. Also the ethernet-dhcp example is copied to /etc/netctl/dhcp and edited for the correct interface (enp0s25). netctl start dhcp runs without reporting any errors, but then ping 8.8.8.8 says 'network unreachable', and ip link shows that enp0s25 is not UP. The network was OK with the ISO of course. I could probably find out what is wrong if I had my hands on the keyboard, but doing it remotely seems more difficult... So what *could* be the problem or missing ? TIA, -- FA
setting time
Hello all, A friend (who is by no means a Linux expert and lives at the other end of the world) is using an Archlinux system that has its date and time completely wrong - the year is 2019. Is using timedatectl set-time a safe way to fix this ? In other words, would a system time jump cause problems ? In the hwclock manpage there is a warning that -s (which also causes the system time to jump) should not be used on a running system, this is why I ask. The alternative would be to set the HW clock in the BIOS. If I understand things correctly, the new system time will then be set when booting. TNX, -- FA
Re: update warnings
On Wed, Dec 14, 2022 at 08:46:39PM +0100, mpan wrote: > Since pacman does not warn about cycles, I assume that are misreported > warnings about one of the packages in a cycle being installed before > another. Those are usually not an issue. But we would need to know, what the > messages are and which packages they refer to. warning: dependency cycle detected: warning: mesa will be installed before its libglvnd dependency warning: dependency cycle detected: warning: smbclient will be installed before its cifs-utils dependency warning: dependency cycle detected: warning: python-ipykernel will be installed before its python-jupyter_client dependency warning: dependency cycle detected: warning: jupyter-server will be installed before its jupyter-nbconvert dependency warning: dependency cycle detected: warning: openimageio will be installed before its opencolorio dependency warning: dependency cycle detected: warning: lib32-mesa will be installed before its lib32-libglvnd dependency -- FA
update warnings
Hello, A full update gives me a number of warnings about dependency cycles. Can this be safely ignored, or is some special handling required ? TNX, -- FA
Re: [arch-general] ThinkPad X1 Carbon Gen 11 mic not working
On Tue, Jul 05, 2022 at 10:25:14AM +0800, Sadeep Madurange via arch-general wrote: > List of CAPTURE Hardware Devices > card 0: sofhdadsp [sof-hda-dsp], device 0: HDA Analog (*) [] > Subdevices: 1/1 > Subdevice #0: subdevice #0 > card 0: sofhdadsp [sof-hda-dsp], device 6: DMIC (*) [] > Subdevices: 1/1 > Subdevice #0: subdevice #0 > card 0: sofhdadsp [sof-hda-dsp], device 7: DMIC16kHz (*) [] > Subdevices: 1/1 > Subdevice #0: subdevice #0 This means that to record from the built-in mic you need arecord -D hw:0,6 ^ The default (probably -D hw:0,0) would record from the HDA device, which may be wired to an extern mic input or not at all. -- FA
Re: [arch-general] ThinkPad X1 Carbon Gen 11 mic not working
On Mon, Jul 04, 2022 at 04:29:13PM +0200, Iyán Méndez Veiga via arch-general wrote: > In any case, I have a X1 Yoga Gen 6, which is (apart from the screen and > chasis) equivalent to the X1 Carbon Gen 9 (with an Intel Gen 11), and I have > no issues with the mic or the speakers. Everything works out of the box, but > I'm using pipewire. > > Would you consider trying pipewire or is it a no go for you? As far as I'm aware, Pipewire sits on top of ALSA. So in your case I'd expect just ALSA to work as well. Maybe you could test... -- FA