WIP
[Linux] Is Neofetch's memory output correct?
"*insert program* shows different memory amounts to Neofetch."
The correct way to calculate memory usage (and what Neofetch currently does) on Linux is the following:
MemUsed = Memtotal + Shmem - MemFree - Buffers - Cached - SReclaimable
Until late 2016 htop, conky, Screenfetch, Neofetch and a lot of other programs were displaying the incorrect memory usage amount. Almost all of the programs I could find have been updated to fix this but there are still a few programs displaying incorrect usage. Those of you using older versions of these programs are probably where this question comes from.
Source: KittyKatt/screenFetch/issues/386
TL;DR: Yes
[Linux] Neofetch reports inaccurate distro name
It's supposed to be [insert a downstream distro here], but it shows [insert an upstream distro here] instead!
When Neofetch detects a Linux distro it first looks for the lsb_release command before searching for the /etc/os-release file. Since some downstream distros mostly utilize their upstream distro's repositories they'll include the upstream distro's version of lsb_release. The prominent case is with Antergos and Arch with lsb_release installed, Antergos will be detected as "Arch" instead.
When the upstream's version of lsb_release is executed the output will be that of the upstream distro. (Arch output on Antergos and etc). This means that unless the downstream distro supplies their own version of lsb_release in their own repository, the output will be the upstream distro's.
We had deleted support for lsb_release but more complaints came up because it proved to be very useful in detecting a wide number of distros. Almost all distros we officially support provide their own lsb_release package which makes lsb_release output the correct information. See issues/493
So, shouldn't you add an exception rule in cases like these?
A good distro will always provide their own lsb-release file, otherwise they will be practically seen as the upstream distro with fancy modifications.
So, is there any workaround?
Yes, actually. You'd have to remove your lsb-release (or any equivalent) package. (That is, if the distro filled their /etc/*-release properly. Contact your distro's maintainers if they don't).
[Linux/FreeBSD] Why doesn't getgpu show my exact video card model?
If your lspci | grep "VGA" (Linux) output looks like this:
01:00.0 VGA compatible controller: NVIDIA Corporation Device 1401 (rev a1)
Instead of this:
01:00.0 VGA compatible controller: NVIDIA Corporation GM206 [GeForce GTX 960] (rev a1)
or if your pciconf -lv | grep -B 4 -F VGA (FreeBSD) output looks like this
vgapci0@pci0:0:2:0: class=0x030000 card=0x00000000 chip=0xbeef80ee rev=0x00 hdr=0x00
vendor = 'InnoTek Systemberatung GmbH'
device = 'beef'
class = display
subclass = VGA
Instead of this:
vgapci0@pci0:0:2:0: class=0x030000 card=0x00000000 chip=0xbeef80ee rev=0x00 hdr=0x00
vendor = 'InnoTek Systemberatung GmbH'
device = 'VirtualBox Graphics Adapter'
class = display
subclass = VGA
Then you're affected by the issue.
There are two possible causes. This can be caused by your /usr/share/misc/pci.ids\* (Linux) or /usr/local/share/pciids/pci.ids\* (FreeBSD) files being outdated and you can fix it by running
sudo update-pciids (Linux only)
If the list is not updated even after running the command above, that means your GPU is brand new, and no one has registered your GPU to the pci-ids repository.
You can submit an issue/contact us through gitter, or you can add it yourself to https://pci-ids.ucw.cz/. Make sure you have the right vendor ID and the device ID by checking lspci -nn (Linux) or chip=0xaaaabbbb with aaaa as your device ID and bbbb as your vendor ID (FreeBSD) and read the guidelines.
- Home
- General
- Info
- Images
- Ascii
- Other