Slog Diagnostics

When the effect list alone can't distinguish gate-deny from cause-shape-reject, swap poke_slab for poke_slab_report. It runs the same poke but also captures every slog emitted at target: "slogger" during the call into a PokeReport:

// Example: asserting on slog diagnostics in your test
use vesl_test::{decode_cause_tag, SlogWarning};

let report = harness.poke_slab_report(slab).await?;

// Bare "did the kernel reject the cause shape?" check.
assert!(report.rejected_cause(), "kernel should have rejected the cause");

// Or assert on which tag the kernel saw before rejecting.
for warning in &report.slog_warnings {
    if let SlogWarning::InvalidCause { noun } = warning {
        assert_eq!(
            decode_cause_tag(noun).as_deref(),
            Some("my-typo"),
            "rejected cause tag should match the hull-side typo",
        );
    }
}

PokeReport fields:

• effect_tags: Vec<String> — same as poke_slab's return.
• slog_warnings: Vec<SlogWarning> — slogs captured during the poke. Each is one of:
  • InvalidCause { noun } — kernel's (soft cause) rejected the cause shape.
  • Other(String) — any other slog emitted at target: "slogger".

Helper: decode_cause_tag(noun) -> Option<String> parses the leading tag from Hoon's printed-noun format. It handles both the cord-decoded [%foo ...] shape and the dotted-decimal fallback.

See Troubleshooting — distinguishing denial paths for which slog shape maps to which failure mode.