- Improve extractItem error messages: distinguish non-attrset elements
from multi-key attrsets, and include the faulty definition via showDefs.
- Use isType instead of raw _type access for order detection.
- Disable type merging (typeMerge = t: null) instead of providing a
functor-based merge. Add test confirming duplicate declarations fail.
This adds a type for name-value mappings that preserve ordering.
Motivating use case: command line flags for package modules /
wrappers / modular services.
The option value can be transformed into a command line in the
correct order.
Additionally, a convenience readOnly option could be provided
to give easy introspection access to the values in an ad hoc
manner.
`emptyValue` for `types.submodule` was `{ value = { }; }`, i.e. none
of the type-declared options or their defaults.
Previously submodules without defs would simply fail for lack of a default,
but since PR #500104 repurposed `emptyValue` as the fallback when no
definitions exist, submodule options without an explicit `default = { }`
(new additions) silently lost their sub-option defaults
(e.g. `requiredFeatures.devnet` in the nixos-test-driver, #511413).
Main change: `emptyValue`: `{ }` -> `{ value = base.config; }`
Additionally, skip `emptyValue`-based default rendering in docs for
types with submodules, because their sub-options are already documented
individually, and forcing evaluation here can break on modules with
invalid or incomplete definitions.
Downstream types that want to use `lib.types.defaultTypeMerge` must include
`wrapped` even though it is deprecated or the internal
`wrappedDeprecationMessage`. This is caused by accessing the
`wrapped` attr in `lib.types.defaultTypeMerge`. The logic should first check
if the attr exists and then if it does check if it is null.
Lists the valid choices in `attrTag` type's `description` field.
Given this description is used in error messages, this serves to make
for somewhat more descriptive errors in the event an option involving
this type turns out not to match.
Signed-off-by: cinereal <cinereal@riseup.net>
This change introduces the suggesting of possible valid option names for
a given invalid option, based on the keys of said invalid option's
parent attrset. That is, `foo.bar.baz` will only be suggested keys from
`foo.bar`, while `(config).baz` will only be suggested keys from the
toplevel.
This reverts commit ce8f304bb6.
The problem was simply a typo (nestedTypes.elemType ->
type.nestedTypes.elemType) ! And CI didn't run for lib in the orginal
PR which is why it didn't get caught.
