Skip to content

Inspecting archives

Use crushr info when you want to inspect an archive without extracting it.

Basic forms:

crushr info crushr info --list crushr info --entry crushr info --find crushr info --propagation

info is an introspection surface. It does not extract files, mutate the archive, or write to the filesystem.

crushr info

This shows the archive contract and archive-level truth surface in human-readable form.

Typical sections include:

  • archive path
  • format version
  • global flags
  • preservation profile
  • structural summary
  • verification summary
  • whether strict extraction is supportable

crushr info --list

Lists archive contents without extraction.

  • metadata/index-driven
  • fail-closed
  • deterministic

If crushr cannot prove the listing, it does not guess.

Use:

  • --entry for a single path
  • --find for search

crushr info --entry

Exact logical-path lookup without extraction.

crushr info archive.crs --entry src/main.rs crushr info archive.crs --entry src/main.rs --json

Fields include:

  • logical path
  • trust class
  • payload verified
  • metadata complete
  • extent count
  • size bytes
  • payload BLAKE3
  • logical range (start..end) in both hex and decimal form
  • identity source
  • strict extraction supported
  • reason (if non-canonical)

Trust classes:

  • canonical
  • metadata_degraded
  • recovered_named
  • recovered_anonymous
  • unrecoverable

Notes:

  • payload integrity is independent from metadata completeness
  • logical range is logical, not archive-physical
  • JSON mode is the authoritative machine-readable surface
  • not-found is explicit and deterministic (no fallback to search)

crushr info --find

Deterministic search over known identities.

crushr info archive.crs --find src crushr info archive.crs --find .rs --json

Behavior:

  • substring match (default)
  • deterministic lexical ordering
  • no extraction
  • no identity guessing

Search scope:

  • canonical
  • metadata_degraded
  • recovered_named

Not included:

  • recovered_anonymous (no stable identity)

Output:

  • default → CLI presentation
  • --json → structured output

Reserved flags:

  • --find-mode substring
  • --find-limit

Unsupported modes produce deterministic errors.

crushr info --propagation

Explains why entries are impacted.

crushr info archive.crs --propagation crushr info archive.crs --propagation --json

Answers:

  • what corruption was detected
  • which structures are affected
  • which entries are impacted
  • why they are impacted
  • whether canonical extraction is blocked
  • what trust-class outcomes are supported by current evidence

Modes:

  • default → human-readable explanation
  • --json → full propagation graph

crushr info --propagation

Use this when you need deterministic dependency and impact explanation without extraction.

bash crushr info archive.crs --propagation crushr info archive.crs --propagation --json

Human mode (--propagation) is operator-facing and includes bounded sections for archive context, detected corruption, impact summary, and entry impacts.

For dependency-dense entry impacts in human mode:

  • dependencies are rendered as multi-line rows
  • first 4 dependencies are shown in deterministic report order
  • any remainder is shown as + <n> more
  • this summarization is human-mode only

JSON mode (--propagation --json) remains machine-readable and separates:

  • dependency graph (nodes, edges)
  • currently detected corruption (detected_corruption)
  • activated impact from current corruption (activated_impacts)
  • per-entry dependency + impact explanation (entry_impacts)

Per-entry explanation includes:

  • direct vs propagated dependencies
  • bounded dependency reasons (for example requires_index, requires_block_payload)
  • whether canonical extraction is currently blocked
  • trust classes supported by current evidence only

The report is explanatory only. It does not imply repair behavior and does not make speculative survivability claims.

Rules:

  • deterministic
  • evidence-based
  • no speculation
  • no repair
  • no survivability prediction

info vs verify vs extraction

info:

  • what exists
  • what can be proven
  • what can be inspected
  • whether strict extraction is supportable

verify:

  • whether strict extraction requirements are satisfied
  • whether recovery is required

extract:

  • what was actually restored
  • final trust-class outcomes

Common workflow

crushr info archive.crs crushr info archive.crs --list crushr info archive.crs --entry src/main.rs crushr info archive.crs --find src crushr info archive.crs --propagation

Then:

  • crushr verify
  • crushr extract
  • crushr extract --recover

Summary

Use info to answer:

  • what exists?
  • what can be proven?
  • is strict extraction supportable?
  • what do we know about this entry?
  • why is this entry impacted?

Use verify to answer:

  • will strict extraction succeed?

Use extract to answer:

  • what actually restored?