## 🤖 Release

* `pklr`: 0.4.1 → 0.4.2

The `deprecated` field added in
[#58](#58) is scoped to `pub(crate)`
([#61](#61)) so `ObjectSource` no longer
exposes new public surface. cargo-semver-checks still flags the struct
as no-longer-externally-constructible-via-literal, but in practice
nothing downstream constructs `ObjectSource` literally — it's an
internal-ish struct describing evaluator state — so this ships as a
patch release. The release-plz auto-detected major bump is overridden
here manually.

### Changelog

#### Fixed

- *(eval)* make @deprecated lazy — only warn on access
([#58](#58))
- *(value)* scope `ObjectSource.deprecated` field to `pub(crate)`
([#61](#61))

---
This PR was originally opened by
[release-plz](https://github.com/release-plz/release-plz/) as 0.5.0;
manually retargeted to 0.4.2.

<!-- CURSOR_SUMMARY -->
---

> [!NOTE]
> **Low Risk**
> Low risk: this PR only updates release metadata (version numbers and
changelog) with no functional code changes.
> 
> **Overview**
> Publishes release `v0.4.2` by bumping the crate version from `0.4.1`
to `0.4.2` in `Cargo.toml`/`Cargo.lock`.
> 
> Updates `CHANGELOG.md` with the `0.4.2` entry noting the fix to make
`@Deprecated` warnings lazy (warn only on access).
> 
> <sup>Reviewed by [Cursor Bugbot](https://cursor.com/bugbot) for commit
5ebfa23. Bugbot is set up for automated
code reviews on this repo. Configure
[here](https://www.cursor.com/dashboard/bugbot).</sup>
<!-- /CURSOR_SUMMARY -->

Closes hk discussion #878.

The auto-generated pkl/Builtins.pkl had bindings of the form

  check_byte_order_marker = Builtins["builtins/check_byte_order_marker.pkl"].check_byte_order_marker

whose RHS reads a property annotated @deprecated. With pklr's old
eager handling those warnings fired on every load; even after pklr
0.4.2's lazy @deprecated, the field-access on the deprecated stub
still warns at module-evaluation time because the Builtins.pkl
binding itself performs that access.

Two coordinated changes:

- bump pklr to 0.4.2 (lazy @deprecated, warning fires only on field
  access — see jdx/pklr#58, jdx/pklr#61).
- have scripts/gen_builtins.py special-case the deprecated aliases:
  skip them in the auto-generated loop and emit them as @deprecated
  bindings that point at the canonical byte_order_marker step. The
  Builtins.pkl bindings no longer access a deprecated property at
  load, so the warning only fires when a user references
  Builtins.check_byte_order_marker / .fix_byte_order_marker
  directly — preserving the migration nudge.

End-to-end with HK_PKL_BACKEND=pklr:
- import Builtins.pkl, no use of deprecated aliases → silent.
- use Builtins.check_byte_order_marker → warning fires.

…#880)

## Summary

- Closes [discussion #878](#878).
- With `HK_PKL_BACKEND=pklr`, every load of `Builtins.pkl` printed
`[pklr] WARNING: property 'check_byte_order_marker' is deprecated` and
the same for `fix_byte_order_marker`, even when the user's `hk.pkl` did
not reference those aliases.
- Combine the upstream pklr fix with a generator change so the warning
only fires when a user actually references the deprecated aliases.

## Two coordinated changes

1. **Bump `pklr` 0.4.1 → 0.4.2**
([jdx/pklr#58](jdx/pklr#58),
[jdx/pklr#61](jdx/pklr#61)). pklr now treats
`@Deprecated` lazily — the warning fires on field access, not on module
evaluation.
2. **`scripts/gen_builtins.py`**: special-case the deprecated aliases.
Skip them in the auto-loop and emit them as `@Deprecated` bindings that
point at the canonical `byte_order_marker` step. With this,
`Builtins.pkl`'s own bindings no longer access a deprecated property at
load, so even the lazy field-access warning stays silent until the user
explicitly references `Builtins.check_byte_order_marker` /
`.fix_byte_order_marker`.

The deprecated stub files (`pkl/builtins/check_byte_order_marker.pkl`,
`pkl/builtins/fix_byte_order_marker.pkl`) keep their `@Deprecated`
annotations untouched — the migration nudge is preserved for direct
importers.

## Test plan

- [x] `mise run test:cargo` — 145 passed.
- [x] End-to-end with `HK_PKL_BACKEND=pklr`:
- import `Builtins.pkl` without using deprecated aliases → **silent**
(was 2× warnings per load).
- access `Builtins.check_byte_order_marker` → warning fires once per
evaluation.
- [x] `hk` pre-commit hooks pass.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

<!-- CURSOR_SUMMARY -->
---

> [!NOTE]
> **Low Risk**
> Low risk: changes are limited to a dependency patch bump and
generation/mapping of builtin aliases to reduce deprecation warnings,
with no impact on core hook execution logic beyond alias resolution.
> 
> **Overview**
> Stops `Builtins.pkl` from triggering deprecation warnings on load when
using the `pklr` backend.
> 
> This bumps `pklr` from `0.4.1` to `0.4.2` and updates
`scripts/gen_builtins.py` to *skip* generating deprecated builtin
bindings in the main loop, then re-add them as explicit `@Deprecated`
aliases pointing to the canonical `byte_order_marker` builtin.
> 
> The pre-commit migration map is updated so `check-byte-order-marker`
and `fix-byte-order-marker` now resolve directly to
`Builtins.byte_order_marker` instead of the deprecated alias names.
> 
> <sup>Reviewed by [Cursor Bugbot](https://cursor.com/bugbot) for commit
549c070. Bugbot is set up for automated
code reviews on this repo. Configure
[here](https://www.cursor.com/dashboard/bugbot).</sup>
<!-- /CURSOR_SUMMARY -->