Skip to content

Refactor monitor config - #11

Merged
igorcamilo merged 10 commits into
mainfrom
small-adjustments
Jul 28, 2026
Merged

igorcamilo merged 10 commits into
mainfrom
small-adjustments

Conversation

@igorcamilo

Copy link
Copy Markdown
Owner

No description provided.

@igorcamilo igorcamilo self-assigned this Jul 27, 2026
Changed Qt style to adwaita-dark to ensure consistency with GTK themes.
Added pointerCursor settings to unify cursor appearance across
different applications. Updated VS Code settings to manage window
controls and password storage securely under Hyprland.
- Added Lua formatting and linting checks to the repository scripts.
- Updated Lefthook config to include Stylua and Selene for Lua files.
- Introduced selene.toml and stylua.toml for Lua style configurations.
- Modified check-repository.sh to include selene and stylua tools.
- Adjusted check.sh to run Lua formatting and linting checks.
- Added backupFileExtension in flake.nix to handle file collisions.

These changes improve code quality and maintainability for Lua scripts.
Refactor Qt theming settings to use gtk2 for better
visual consistency with GTK themes. Adjust VS Code
argvSettings to ensure secure password storage and
maintain necessary fields for proper functionality
after installation. This change enhances user experience
under Hyprland by addressing keyring issues and
ensuring compatibility with existing themes.
Enhance AGENTS.md and architecture.md to include
speculative project ideas alongside existing procedures.
Refine configuration.nix to improve Qt theme handling
by switching to gtk2 style for better consistency
across applications.

Add ideas.md to outline future project aspirations
and feasibility notes, providing a roadmap for
potential developments in the repository.
Implemented the Ollama service in configuration.nix to enable
local LLM serving for VS Code Copilot. This includes support for
the RDNA4 GPU and necessary model loading configurations.
Updated documentation in ideas.md to reflect the integration
process and manual steps required for VS Code.
Added the Ollama extension from the Marketplace to the VS Code
configuration, enabling local LLM serving in Copilot Chat. The
extension version and hash are pinned manually due to it not being
part of nixpkgs' curated set. This change enhances the local-LLM
setup by allowing users to select the pulled model directly in
Copilot Chat's model picker.
@igorcamilo
igorcamilo marked this pull request as ready for review July 28, 2026 06:04
@igorcamilo
igorcamilo merged commit c736e1b into main Jul 28, 2026
3 checks passed
@igorcamilo
igorcamilo deleted the small-adjustments branch July 28, 2026 06:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant