id(cli): pkgs/id does not compile on main — PR #3738 inserted parse_result_limit between Command's derive and the enum

Evidence (verified 2026-11-01, main e49c0be)

pkgs/id/src/cli.rs lines ~199–212 on current main:

#[derive(Subcommand, Debug)]
/// Parses a result limit (`--first` / `--last` / peek `--lines`): zero is
/// rejected with a clear error instead of silently producing an empty result
/// or a header-only preview (#3059). ...
fn parse_result_limit(raw: &str) -> Result<usize, String> { ... }

pub enum Command {

The parse_result_limit function (added by PR #3738, "id(cli/repl): reject zero result limits at both boundaries (#3059)", merged 2026-11-01 09:03 CST as 5fa0650) was inserted between #[derive(Subcommand, Debug)] and the pub enum Command it belongs to. The derive now targets a function and the enum has no derive.

cargo test --lib in pkgs/id on main therefore fails with 24 errors, including:

• error[E0774]: derive may only be applied to structs, enums and unions (the misplaced attribute)
• error[E0277]: cli::Command doesn't implement Debug (×15, from test code formatting {other:?})
• error[E0277]: the trait bound cli::Command: Subcommand is not satisfied
• error[E0609]: no field format / format_error on Result<FindArgs, anyhow::Error> (cascade)

Full log from the discovering run: laptop /tmp/test-4789.log (queue task 4789's verification build, 2026-11-01).

Impact

pkgs/id does not compile on main at all — every id change is unverifiable and any CI leg building pkgs/id is red until this is fixed. Discovered while verifying task 4789 / PR #3748 (directory_auth, #1732), which is unrelated and blocked on this for its test run.

Fix

Move parse_result_limit (with its doc comment) so #[derive(Subcommand, Debug)] sits directly above pub enum Command again — e.g. place the function after the enum or above the derive. No behaviour change: the function body and all value_parser = parse_result_limit call sites stay as PR #3738 wrote them.

Acceptance

• cargo check / cargo test --lib in pkgs/id compiles on main.
• The id(find/search/show/peek): a zero count is accepted everywhere — '--first 0'/'--last 0' report 'no matches found' for queries that do match, and 'peek --lines 0' renders a preview with no content lines #3059 zero-limit rejection still works (existing tests from PR id(cli/repl): reject zero result limits at both boundaries (#3059) #3738 pass).
• No other behaviour change.