feat(prune): R1 prune-reconcile pure core — (owned ∩ present) − referenced

Add REAPER-free/filesystem-free prune orphan-set core + additive
BankBook::referencedPaths() union query, with full unit coverage.
This commit is contained in:
2026-07-26 18:02:42 -04:00
parent b6464f78c8
commit 3cfa9239d2
6 changed files with 426 additions and 0 deletions
+63
View File
@@ -0,0 +1,63 @@
#pragma once
// prune_reconcile — the pure core of Phase R (Reclaim), Wave 1. The safety-critical
// "which files are orphans" decision, computed with NO filesystem I/O and NO REAPER
// types. The mirror of view_mode_model's reconcile(liveGuids), one level DOWN: it
// reconciles FILES ON DISK against REFERENCED FILES (the union across every bank),
// where reconcile reconciled membership entries against live tracks.
//
// PURE MODULE (CLAUDE.md §load-bearing split): NO REAPER types, NO SWELL, NO
// vendor/ includes, NO filesystem calls. Standard library only. Unit-tested outside
// the DAW — this is the safety-critical part (it decides which bytes get deleted in
// R2/R3), so it is hard-tested here before any I/O exists.
//
// -- The one computation ------------------------------------------------------
//
// orphans = (owned ∩ present) referenced
//
// * present — files enumerated in the resolved current bank folder (R2 shell).
// * referenced — every project-relative path referenced by ANY bank in the book,
// pool included (union across the whole book — see BankBook::
// referencedPaths). A file referenced by any bank — including via a
// COPY into a second bank — is NEVER an orphan (the prune null test).
// * owned — the owned-file manifest: the files the bank system itself created
// (OwnedFileManifest). A present-but-unowned (hand-dropped) file is
// NEVER reclaimed — prune reclaims only the system's own leavings.
//
// The three settled guardrails fall straight out of the set algebra:
// * ∩ present — never proposes deleting a file that is not on disk (an owned-
// but-absent manifest entry yields no orphan, no error).
// * ∩ owned — never a hand-dropped file (ownership attribution, fork R-D).
// * referenced — never a file any bank references (union safety, prune null test).
//
// -- Path representation: EXACT-STRING match (safety-critical) -----------------
//
// Every path in the model is a project-relative string compared VERBATIM: Sample.
// relativePath, OwnedFileManifest::contains (p == relativePath), and BankIndex all
// use raw std::string equality — no separator normalization, no case-folding, no
// trailing-slash trimming. This core MATCHES that convention exactly: it compares
// the raw strings the shell supplies. Feeding a consistent spelling across the three
// inputs is the R2 shell's contract (it enumerates the folder, unions the book, and
// reads the manifest against the SAME resolved current folder). Diverging from exact
// match here (e.g. case-insensitive compare) would be the unsafe direction — it could
// let one spelling of a referenced file be treated as an orphan under another.
#include <string>
#include <vector>
namespace reasampler {
// Computes the prune orphan set: (owned ∩ present) referenced.
//
// Returns the subset of `present` that is BOTH owned AND unreferenced, in the ORDER
// they appear in `present` (deterministic output — mirror of the insertion-order
// determinism the index / manifest keep; the R2 dry-run reports a stable file list).
// Duplicate spellings within `present` are de-duplicated in the result (a folder
// enumeration yields distinct names, but the core does not rely on that).
//
// Pure: no I/O, no REAPER, no hidden state. All three inputs are project-relative
// path strings, compared by exact std::string equality (see header note).
std::vector<std::string> pruneOrphans(const std::vector<std::string>& present,
const std::vector<std::string>& referenced,
const std::vector<std::string>& owned);
} // namespace reasampler