The default value for programs.git.signing.format changed in 25.05
from an implicit "openpgp" to null. Keep the existing gated
mkOptionDefault behavior so the signing block only materializes when
other signing settings are in use, but route the versioned value and
static docs text through the shared state-version helper.
Add a focused current-state-version test that covers a non-empty
signing configuration with no explicit format, alongside the existing
legacy implicit-openpgp and explicit-format tests.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Before 23.05, programs.swaylock.enable implicitly followed whether
programs.swaylock.settings was non-empty. That compatibility path was
still active for older state versions, but it emitted no warning.
Route the default through the shared state-version helper so legacy
users get the standard deprecation warning before the implicit enable
behavior is removed. Add a focused current-state-version test and keep
the existing legacy and explicit-enable coverage in place.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Before 19.03, programs.beets.enable implicitly followed whether
programs.beets.settings was non-empty. That compatibility branch was
still active for older state versions, but it was silent.
Route the default through the shared state-version helper so legacy
users get a consistent deprecation warning before the implicit
enablement is cleaned up. Add focused tests for the legacy and current
default branches alongside the existing beets coverage.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
The default value of programs.password-store.settings changed in 25.11
when Home Manager stopped exporting PASSWORD_STORE_DIR implicitly.
Route the default through the shared state-version helper so users on
older state versions keep the compatibility behavior with a consistent
warning and a single place to remove later.
The existing password-store tests already cover the legacy default, the
current empty default, and explicit override behavior, so no new test
fixture is needed for this migration.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
The gtk4 theme option still inherits from gtk.theme for users pinned
before 26.05, but that fallback was previously silent. Move the
default through the shared state-version helper so the compatibility
branch emits the standard deprecation warning and stays consistent
with other future cleanups.
Add a focused test that covers the legacy inheritance path alongside
the existing current-state-version test for the null default.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Switch xdg.userDirs.setSessionVariables to the shared state-version default helper so older configurations get an explicit warning before the legacy default is removed.
Add focused tests for the legacy and current branches to verify that the session variable exports remain enabled before 26.05 and are disabled by default starting at 26.05.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Replace the inline shellWrapperName migration logic with the shared state-version default helper. This keeps the existing warning and version-dependent default, but moves the message and defaultText formatting into one reusable implementation.
Inline the helper directly at the option declaration so the module stays small while matching the new deprecation-default pattern for future adoptions.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Add a helper for options whose defaults change across home.stateVersion boundaries. This centralizes the warning text and documentation shape so modules do not need to hand-roll the same migration pattern at each call site.
The helper takes legacy and current branches with a runtime value plus optional static documentation text. That keeps the actual default version-gated while avoiding option docs that depend on evaluated config. Add a focused test covering the legacy warning path, the new-value path, and an explicit legacy pin that should not warn.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
now that home-manager always generates an init.lua we can take the
wrapper output instead of reimplemnting stuff in HM (notwithstanding rtp
changes that are not needed in HM).
This is necessary to keep the provider configuration available after
this change https://github.com/NixOS/nixpkgs/pull/487390
Also updated test to reflect ruby disabling
The skills option was creating files under ~/.config/opencode/skill/
(singular) but OpenCode documentation only mentions
~/.config/opencode/skills/ (plural). Both work, but using an
undocumented directory can be confusing.
Fixes: https://github.com/nix-community/home-manager/issues/8907
Upstream change is validating all maintainers against lib.maintainers,
need to merge our maintainer list in to pass doc building.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Add an optional session setting so that session names can be different
than the smug project name. This also enables using template variables
when naming sessions.
Factor the repeated claude-code module patterns into a small set of local helpers instead of open-coding them over and over.
This keeps the behavior the same, but removes duplication in the inline-or-path option declarations, the inline-vs-directory assertions, and the repeated .claude file generation code. The claude-code NMT test slice still passes after the refactor.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Claude Code rejects `--mcp-config` once the Home Manager wrapper injects
it around subcommands, which breaks commands like `claude mcp list`.
Claude Code 2.1.76 fixed `--plugin-dir` so it no longer consumes
following subcommands, so use that path for the generated MCP config
instead.
Generate a plugin directory with a manifest and `.mcp.json`, wrap
`claude` with `--plugin-dir` before user arguments, and snapshot that
wrapper directly in the tests. Keep the existing LSP support in the
generated plugin directory as well, and add coverage for the combined
MCP+LSP case plus the MCP integration merge path.
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
Needs to handle different lua versions better. Determine lua version for
which luaPackages set to reference
Signed-off-by: Austin Horstman <khaneliman12@gmail.com>
To make it easier for contributors to customize their entry: it's easier
to remove the comments (or not since they wont be seen anyway) then
searching for those examples.
I no longer use Zed, and most of the code in the module is not mine.
I'd rather have someone who actually uses the module take it up.
The TODO was resolved a long time ago.
This adds an 'onDispatch' option to the 'submaps' attribute set,
enabling the 'submap = name, ondispatch' syntax which allows submaps to
be closed automatically after a dispatch.
