| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit e868bfeca5501c7c02f1fc1013ea6cc7e79af043.
Unfortunately, this causes some sites to be outright unreadable since
they seem to rely on a certain set of default colors.
Given that the FOUC is already pretty minimal in Firefox, I think it's
really not worth breaking a bunch of sites just to get rid of the
flashing, at least until we found a better way to prevent it entirely.
Signed-off-by: aszlig <aszlig@nix.build>
|
| |
|
|
|
|
|
|
|
|
| |
The latest Git master version contains a bunch of fixes for yt-dlp
integration and since most of my feeds are actually YouTube channels, it
just makes sense to upgrade.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
| |
This is already defined in the base profile, so we really don't need to
define it twice, espacially because since ee758fdcdbbcae14d7a50d1809c1b2
both definitions are conflicting.
I actually had this removed locally but forgot to stage this for the
aforementioned commit.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The right XKB configuration should be using "us" as its layout and
"dvorak" as the variant, because there are other such variants such as
"it", "fr" or "de" and just using "dvorak" as the layout has very long
been an alias that now has been removed[1] in the latest version of
xkeyboard-config.
[1]: https://gitlab.freedesktop.org/xkeyboard-config/xkeyboard-config/-/commit/470ad2cd8fea84d7210377161d86b31999bb5ea6
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
| |
This is a recent upstream change.
|
|
|
|
| |
The default of this setting changed in foot 1.15.0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I roughly went through the preferences set by Betterfox[1] and checked
whether they'd make sense for my Firefox version.
While most of these options are pretty straightforward, the captive and
connectivity services are features for convenience, eg. when travelling.
However, since the only thing they do is giving UI hints that otherwise
can be found out via a quick detour to the shell or visiting a random
"http://" URL I don't think the convenience outweights repeatedly
pinging detectportal.firefox.com.
[1]: https://github.com/yokoffing/Betterfox
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With newer bsdtar, extraction fails because of this error:
bsdtar: Pathname cannot be converted from UTF-8 to current locale.
bsdtar: Archive entry has empty or unreadable filename ... skipping.
bsdtar: Error exit delayed from previous errors.
The reason is that the current locale isn't set during Nix builds, so in
order for bsdtar to work properly, let's set it to C.UTF-8 so that it
hopefully doesn't convert anything at all (which is what we want).
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The evaluation error is:
error: function 'anonymous lambda' called without required argument 'pname'
This is because the "baseName" attributes of the dotnet tooling have
changed to the more common "pname" attributes that we use throughout
nixpkgs.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
| |
This is handy if we want to pass further arguments to the makeWrapper
call without either replacing installPhase or wrapping twice.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
| |
These package attribute names have been renamed a long time ago and no
longer evaluate in current nixpkgs.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
| |
I occasionally run wireguard setups, sometimes broken beyond repair, and
I really don't care whether those are online or not as long as *any*
interface is alive.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
| |
Right now, nixos-option fails to build with latest nixUnstable and since
I currently don't have a lot of time to properly fix it and also don't
care a whole lot about nixos-option, I decided to override it using the
normal Nix stable package.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Flash of unstyled content[1] is a common issue in browsers and happens
whenever transitioning to a page without the full styles being loaded.
For the most part, Firefox does a good job preventing FOUC, but in some
occasions the default colors still show up. Since I do prefer a dark
color scheme anyway, let's actually make sure that the default colors
represent a dark color scheme.
The reason why I picked the dark colors from startpage.com is purely
because it was the next best to pick some colors from, rather than
personal preference.
[1]: https://en.wikipedia.org/wiki/Flash_of_unstyled_content
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
| |
Don't need it as much as I thought.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The nvme-suspend-quirk patch doesn't apply anymore with kernel 6.3
because kernel 6.3 added a new quirk
(NVME_QUIRK_NO_SECONDARY_TEMP_THRESH) to the nvme_quirks enum.
To fix this, I just edited the patch to refer to the newly added quirk
and increase the bitflag by one.
With this, the patch applies, but I haven't yet built the kernel nor did
I actually test this on the target machine.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
| |
Due to the switch to vuizvui's callPackage in
b9401bd57cfff9955649d5ccd3b346271d787a2e, pkgs.sternenseemann.lib would
suddenly get a self reference as the `lib` argument which is of course
no good.
|
|
|
|
|
|
|
|
| |
Regression introduced by f72625079cc9cea8abe06d1a29b10530da538c15.
I accidentally forgot about the vuizvui namespace in pkgs.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
| |
I had yt-dlp in my local Nix profile since quite a while and didn't use
youtube-dl at all since months. It's been so long that I actually forget
about that, thinking I already did this very change already.
Turns out, that I didn't, hence this commit.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There is no particular reason for this other than me "feeling" that my
version became somewhat old. From looking at the upstream diffs the
changes are mostly fixes.
However, I also updated the QDarkStyle plugin, which does have quite a
few changes and the overall theming now looks lighter than before but
still not in a way that feels like an annoyance. We'll see and revert if
necessary.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Upstream changes from 1.5.26 to 1.5.33:
* Add year selector in popup search
* Autocomplete fixes
* CSSLint: 1.5x faster; supports nested &-selectors and nested @-rules
* CSSLint: new color syntax, removed IE6-8 rules
* CSSLint: new syntax for nested selectors
* Ctrl-/ to toggle line comment will use the standard /* block syntax
*/ to avoid upstream bugs in less and stylus preprocessors
* Don't open installer when browsing a non-raw .user.css URL on github
* Editor: allow to change numbers via keyboard/wheel
* Editor: autocomplete for less preprocessor will show @ variables
* Editor: improvements and fixes for autocomplete
* Editor: new option Arrow keys ↑↓ traverse sections
* Fix CSP patching, color picker hue overflow, section naming, style
size in manager after toggling, dark mode eye dropper
* Fixed userstyles.org install button, again
* Fixes for Stylus pages in Firefox + dark mode
* Fixes for sync
* Hide lint errors for @ vars with less preprocessor
* New beautify options
* New option: toggle .user.css URL installer
* Properly update style preview in tab when its config changes
* Show installed styles in the list of available styles in the popup
finder
* Style manager: added multi-column mode option
* Style manager: added style size
* Style manager: changed Number of applies-to items option minimum
from 1 to 0
* fix: installing github styles in subdirectory inside the repository
* fix: styling of iframes in Firefox
* fix: styling of same-origin iframes in Chrome
* fix: styling of sandboxed iframes in Chrome before 86
* fixed installation on the redesigned userstyles.org
* fixed saving of windowed editor position on a non-primary monitor
* skip & report invalid styles on import
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
| |
I already had the latest Git version laying around in my Nix user
profile, so it's battle-tested enough. The main reason why I did this is
because it has much better support for yt-dlp, which I prefer to
youtube-dl.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
| |
`grub.version` was removed in [1], together with `grub1` in [2].
[1]: https://github.com/NixOS/nixpkgs/commit/30bea8d82d9e9ffa3093106af4527cbfc0663150
[2]: https://github.com/NixOS/nixpkgs/commit/6db74508276ae2c3c48cfea0a7022d01ca273264
|
|
|
|
| |
cc @Profpatsch @aszlig
|
| |
|
|
|
|
| |
These are archived elsewhere and don't need to fill up the backup drive.
|
|
|
|
|
| |
This allows us to get the fix point for scripts from callPackage instead
of building it ad hoc.
|
| |
|
|
|
|
|
|
| |
There's no need to have vlc permanently. I've not encountered issues
with mpv for a very long time, so might as well move vlc to the
backbench.
|
|
|
|
|
| |
I'm not a fan of GTK apps, and now that `termusic` has really good
podcast support I don't need gPodder anymore.
|
|
|
|
| |
This option has been renamed upstream.
|
|
|
|
| |
Magic.
|
|
|
|
|
| |
This has been working just fine on gunnr, so let's see if that's also
true for all my other machines.
|
|
|
|
|
| |
This was removed from nixpkgs due to lack of downstream maintainer.
Also, see https://github.com/NixOS/nixpkgs/pull/225285
|
|
|
|
|
| |
I guess? Not sure what the benefit if this configuration is, but I've
seen it around, so it must be correct, right? Right?? XD
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Recent versions of s6-portable-utils no longer include s6-test
which was deprecated in favor of eltest which supposedly has the
same interface.
I've not tested this commit very thoroughly, but my system now builds
again with recent nixpkgs commits (I think I dependended on s6-test
via the rust writer via nman).
cc @Profpatsch
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit c88fd9eaa12c8a3c06502b09c2056d3c91421952.
The hardware.video.hidpi.enable option was removed a while ago[1]
because it's not clear what a single boolean option should mean, so it
doesn't make sense anymore to make any of our options depend on it.
Forthermore, I'm experimenting with different Wayland compositors at the
moment and most of the stuff that I did here is for Xorg. I expect most
of the stuff to be gone when I'm settled with a Wayland setup that works
for me.
[1]: https://github.com/NixOS/nixpkgs/pull/222689
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
| |
This is just a test for now. I'm sure the config is ridiculous, but
it'll do for the moment.
|
|
|
|
|
|
|
|
|
|
|
|
| |
The syntax file in question queries features from the current terminal
and will bring editing Markdown files to a crawl, freezing Vim for more
than ten seconds.
I haven't encountered a dircolors(1) file within a Markdown fenced code
block yet and even if this should be the case someday it's not worth the
high runtime cost.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
| |
With bash in sh mode the direnv hook still works.
|
| |
|
|
|
|
|
|
|
|
|
| |
wrmsr can be used to disable an erroneous CPU throttling that is
happening with my new battery. Since that needs to be reset after every
power state switch, having the tool in PATH is a great help.
Eventually I'll fix this in a more permanent fashion via a systemd
oneshot service.
|
|
|
|
|
|
|
|
|
| |
Firefox version 111.0 has refactored AutoplayPolicy.cpp a but and there
is now a handy "IsWindowAllowedToPlayByUserGesture" function that is
more convenient in that we just need to patch it to *always* return
false.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
| |
This was dropped in https://github.com/NixOS/nixpkgs/pull/221469
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Essentially gets rid of these warnings:
* The option `services.openssh.permitRootLogin' defined in `...' has
been renamed to `services.openssh.settings.PermitRootLogin'.
* The option `services.openssh.passwordAuthentication' defined in `...'
has been renamed to
`services.openssh.settings.PasswordAuthentication'.
* The option `services.openssh.kbdInteractiveAuthentication' defined in
`...' has been renamed to
`services.openssh.settings.KbdInteractiveAuthentication'.
* The option `nix.readOnlyStore' defined in `...' has been renamed to
`boot.readOnlyNixStore'.
Evaluation leads to the same derivation path as before, so apart from
shutting up warnings this should not change any functionality.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since the default reply regexp depends on the current locale and I
refuse to use German localisation for mutt or even my whole system
(except for a few corner cases such as time format), replying to emails
with "AW: foobar" in the subject gets messy at some point when you get
multiple "Re: AW: Re: AW: foobar".
So allowing both "Re:" and "AW:" prefixes and stripping them should
hopefully make those subjects less messy.
Signed-off-by: aszlig <aszlig@nix.build>
|
| |
|