| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
I'm bored of the names my machines have. Luckily NixOS makes changing
names as easy as changing underwear! So, let's use Valkyrie names
instead. Also, change the console font to something nicer, like the
default `Lat2-Terminus16` font!
|
|
|
|
|
|
|
|
| |
I've had these packages in here disabled in case I'd ever have to get
back to them. I've not had to for, what feels like, an eternity.
Additionally to slimming the config down, removing them will also
somewhat slim down closure size since Nix will build them anyway even
though they're not actively used.
|
|
|
|
|
|
| |
The previous commit was wrong in that it was not the correct way to
include the 'light' package. According to the manual, this is the
correct way.
|
|
|
|
|
|
| |
Since the 'modesetting' driver does not work with 'xbacklight' we have
the choice of either 'light' or 'backlightctl'. For now we'll choose
'light' and see how it goes.
|
|
|
|
|
| |
The release notes for NixOS 19.03 state that the 'intel' driver has
been superseded by the 'modesetting' driver.
|
| |
|
|
|
|
|
|
|
| |
This reverts commit 292099429d0bab51f0cb634c6bc5c9745c4c1383.
Well, looks like iwd ist still not ready for primetime with my
university's eduroam implementation.
|
|
|
|
|
|
| |
Now, it's not yet clear if iwd will work with my uni's eduroam service
at all. Either way though, this is in preparation for the switch and
test once I'm near an eduroam access point.
|
|
|
|
| |
These packages are unnecessary currently.
|
|
|
|
|
|
| |
I've become increasingly aware that my configurations drift, and while
this is still not amazing in terms of DRY and whatnot, it's all I can do
for now.
|
|
|
|
|
|
|
| |
I've set up a RPi to act as a print server, but the printers are
configured as RAW, so the burden of having the correct drivers installed
lies with the client, which is fine considering the current network
setup.
|
|
|
|
|
| |
Now that the mutt configuration works like a charm again, we can finally
remove claws-mail.
|
|
|
|
|
| |
Apparently only three primary DNS servers allowed, so we'll reduce it to
whatever DHCP offers plus two of our own definition.
|
|
|
|
|
|
|
|
| |
The original idea was to have globally reachable nameservers as
fallback, but really this isn't what I wanted. I want to have
nameservers additionally to the ones provided by DHCP, but in case the
one provided by DHCP does not deliver the resolve, the other primary
nameservers here will be queried instead.
|
|
|
|
|
|
|
| |
Attribute set entry wasn't closed with a semicolon.
Signed-off-by: aszlig <aszlig@nix.build>
Cc: @devhell
|
|
|
|
|
| |
Let's try to take the DHCP supplied nameservers and add defined ones as
fallback in systemd-resolved.
|
|
|
|
|
| |
Family members won't use jabber clients, but will happily use Signal. Oh
well. Better than nothing.
|
|
|
|
| |
Additionally let's enable Brother HL-1110 drivers.
|
|
|
|
| |
Let's see how cloudflare's DNS behaves.
|
|
|
|
|
| |
This is part of getting a consistent mutt configuration going. Here we
remove all scattered mutts.
|
|
|
|
|
| |
0.3 seems a bit extreme for my taste, so I'm hoping 0.2 will be abit
more subtle but still distinguishable.
|
|
|
|
|
|
| |
I'm spending too much time at Uni and dnscrypt will not work properly in
this network and is causing lookup timeouts, so I'm going to have to
disable it until I find a solution.
|
|
|
|
|
| |
Instead of just running compton when i3 starts, we'll try to run it as
an actual service here.
|
|
|
|
|
| |
DNSCrypt doesn't cope well with Uni's network configuration. We need
Google nameservers so dnscrypt can fetch its information. Meh.
|
|
|
|
|
| |
One of many attempt to follow in making the code more readable and DRY
if possible.
|
|
|
|
|
| |
I find myself turning this on manually more and more, so might as well
turn it on automatically.
|
|
|
|
| |
I'll need some docker love for a bit, so here it is.
|
|
|
|
|
|
|
| |
DNSCrypt is pretty cool so far, but networkd is using the nameservers
adviertised through DHCP on respective interfaces which defeats the
point of DNSCrypt a bit. So, unfortunately we'll have to work around it
this way.
|
|
|
|
|
| |
It seems that dnscrypt is working, so for now I'll remove Google's
nameservers.
|
|
|
|
| |
I want to test this and see how it performs.
|
|
|
|
|
| |
Forgot that I do like to use this at Uni when I can't be bothered with
the IPMI webinterface (which I have yet to update from Java to HTML5).
|
|
|
|
| |
Forgot that my volume controls are bound to pamixer, so reenabling it.
|
|
|
|
|
|
|
| |
In an effort to make my footprint a bit smaller I've taken a bit of time
to disable pkgs I'm sure I won't be needing anytime soon. However, I'm
keeping them commented for historic reasons and in case I do need
something back I don't have to look far.
|
|
|
|
|
| |
Running kernel 4.14 allows btrfs with zstd compression, so let's enable
that and also remove "ssd" enforcement on eris.
|
|
|
|
|
|
| |
And here I've been sitting, wondering what the hell was wrong and why it
would wait ages to access a partition that doesn't (as it turns out)
exist. Duh.
|
|
|
|
| |
This is due to a bug and should be reenabled once solved.
|
| |
|
| |
|
|
|
|
|
| |
Using polybar now I feel there is no need for pasystray and I would also
like to have compton draw me shadows under windows.
|
|
|
|
| |
Uni has some bonkers nameservers. This should hopefully fix it.
|
|
|
|
|
|
|
| |
Since rofi 1.4.0 the old way of theming has been dropped in favor of a
more complex and fine-grained theme engine. These theme files are rather
large and so it makes sense to move to the theme configuration to rofi's
XDG_USER_CONFIG_DIR.
|
|
|
|
| |
Sorry.
|
|
|
|
| |
Let's switch to networkd.
|
|
|
|
|
| |
More often than not connman is not working reliably; on this machine at
least.
|
|
|
|
|
|
|
|
|
| |
Since this is a reinstall (thanks Toshiba BIOS people /s) I've opted to
remove f2fs and replace it with btrfs. After each new kernel fsck.f2fs
would take forever to check the filesystem before the system was usable,
and this doesn't happen with btrfs (or any other FS AFAIK). Also, I
really miss --reflink. We're also removing 'elevator=noop' here since
we're going to use BFQ.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Signed-off-by: devhell <devhell@mailfresser.de>
|