summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAge
* modules/impermanence: Remove pointless `user`/`group` option for userDirsSoispha Peetz5 days
|
* pkgs/ollama: Remove old `acceleration` option (and try to use vulkan)Soispha Peetz5 days
|
* pkgs/rollup: Allow multiple owners in rollup config drop-insSoispha Peetz5 days
|
* modules/dconf: Make possible to disableSoispha Peetz5 days
|
* common/soispha: Enable ollamaSoispha Peetz5 days
|
* modules/openssh: Remove pointless `extra_lib` importSoispha Peetz5 days
|
* hosts/tiamat: Also import soispha's common module profileSoispha Peetz5 days
|
* {modules/mandb,common/*}: Move to vhack namespaceSoispha Peetz5 days
|
* modules/impermanence: Improve usabilitySoispha Peetz5 days
|
* modules/*: Remove all `mkDefaultEnabled` callsSoispha Peetz5 days
| | | | | There is no nice way to make this work with the merged config, as we would now need to differentiate between the host being evaluated for.
* pkgs/rollup: Add config, specifying clear ownership based on targetsSoispha Peetz5 days
| | | | | Thus applies are possible to ignore certain directories for certain targets.
* pkgs/rollup: Support a `--dry-run` flagSoispha Peetz5 days
| | | | That makes testing easier.
* pkgs/rollup: Really sort the patches by numberSoispha Peetz9 days
| | | | I hope this works now.
* modules/gpg: Use a better pinentry programSoispha Peetz9 days
| | | | | | | | The curses one shows up on the latest started terminal, and has serious issues with already started stuff on there (e.g. when trying to run over a running nvim instance). Additionally it's just insecure, as we are using in-band messaging, so an malicious application could draw a similar looking prompt and get me to enter my passphrase in there.
* modules/git: Stop signing commits by defaultBenedikt Peetz9 days
| | | | | My gpg key has expired (around a year ago maybe), and thus I can't use it to sign much stuff anymore.
* modules/git: Correct username and emailBenedikt Peetz9 days
|
* pkgs/rollup: Sort the patch files read from patches dirBenedikt Peetz10 days
| | | | Otherwise we would possibly be applying patch 3 before patch 2 and patch 1.
* pkgs/rollup: Print stderr, if a git command failsBenedikt Peetz10 days
|
* pkgs/rollup: Check if the `--to` directory is actually writableBenedikt Peetz10 days
| | | | | | Otherwise we would create the patches, but fail to apply them. The issue is that we already moved the tag forward, so the user now needs to manually create the patches again (after exiting the sandbox).
* pkgs/rollup: Support creating and applying patches at the same timeBenedikt Peetz10 days
|
* modules/fish: Reorganize the keybindingsBenedikt Peetz10 days
| | | | Insert and Default modes are now more or less how I want them.
* modules/nix/nix_wrapper.sh: Check `/nix/store` instead of `/nix`Benedikt Peetz12 days
| | | | In the direnv sandbox the wrapper fails, because it can't access /nix.
* modules/fish/keymappings: Also escape `#`Benedikt Peetz13 days
|
* pkgs/rollup: Support generating the patchesBenedikt Peetz13 days
|
* modules/nix/nix-wrapper.sh: Don't notifyBenedikt Peetz13 days
| | | | | The notifications don't work in the direnv sandbox and are thus more or less useless.
* modules/fish: Really delete all keymappings before applying the new onesBenedikt Peetz13 days
|
* modules/fish: Define the keybindings as functionBenedikt Peetz13 days
| | | | That's how fish actually wants you to define them.
* modules/nix/nix-wrapper.sh: Ignore `notify-send` failureBenedikt Peetz13 days
|
* modules/nix: Make the nix wrapper actually workBenedikt Peetz13 days
|
* modules/rofi: Change configured terminal to footBenedikt Peetz13 days
|
* modules/backup: Remove `Requires` from timer configBenedikt Peetz13 days
| | | | `systemd` does not understand a Requires= in the [timer] config.
* hosts/apzu: Remove required kernel module of rtw89Benedikt Peetz13 days
| | | | This module does not actually exist (anymore).
* modules/fish: Make the defined keybindings actually workBenedikt Peetz13 days
| | | | | From local testing, these now seem to be (more or less) the same as we had before.
* scripts/build.sh: Don't fail if the result symlinks don't existBenedikt Peetz14 days
|
* pkgs/rollup: Parse git diff directlyBenedikt Peetz14 days
| | | | | It's both easier, and the filterdiff command actually removed renames with similarity 100%. This patch also greatly reduces the parsing code.
* modules/nix: Disable `auto-optimise-store`Benedikt Peetz14 days
| | | | | This setting is apparently bad for performance, and could maybe explain the performance issues I have experience while running a nix build.
* modules/direnv/landrun-fish.sh: Keep the SHELL variableBenedikt Peetz14 days
| | | | Otherwise, `nix shell` opens in bash (the non-interactive one).
* {inputs.nix,unflake.nix}: UpdateBenedikt Peetz14 days
|
* inputs.nix: Add depuplication rules for nearly all possible ways to fetch ↵Benedikt Peetz14 days
| | | | nixpkgs
* modules/{nix,nixos-shell,version}: Use the `sources` and `nixpkgs-unstable`Benedikt Peetz14 days
|
* modules/{taskwarrior,yt}: Move the `lib.mkIf cfg.enable` to the configBenedikt Peetz14 days
| | | | Having it only in the inner attributes is error prone.
* modules/common/soispha: Remove `enable = false` optionsBenedikt Peetz14 days
|
* modules/lutris: Remove gtk2 allowlisting againBenedikt Peetz14 days
| | | | `lutris` no longer depends on it.
* hosts/apzu: Disable the iso boot entryBenedikt Peetz14 days
| | | | | It was already soft disabled, by removing the iso from the tails-iso derivation, so it's better to really disable it.
* scripts/build.sh: Really only delete symlinksBenedikt Peetz14 days
| | | | | I have no idea what I was thinking with the previous `if` conditions. They might be a Chesterton’s Fence, but it works in my computer.
* flake.nix: Add `nvd` to devshellBenedikt Peetz14 days
| | | | That is used for the diffing in the build.sh script.
* pkgs/rollup: Always check the status of git commandsBenedikt Peetz14 days
| | | | | Otherwise, we could continue running even though an underlying git command failed.
* pkgs/rollup: Wrap with the required programsBenedikt Peetz14 days
|
* pkgs/rollup: Actually also parse renames in patchesBenedikt Peetz14 days
|
* pkgs/rollup: Finish first mvpBenedikt Peetz14 days
|