Count a file as deleted only if the previous snapshot of its name had it (closes #304)

A backup counted as deleted every file in the local index under the
source path that the walk did not find. That took in files the
snapshot excludes but another snapshot backs up, and files removed long
ago, which every later backup, and the first backup of another name
over the same path, reported as deleted again.

The backup now looks up the most recent completed snapshot of the same
name and counts only the files it lists that the walk did not find. A
file that is excluded, or inside an excluded directory, is not counted.

The snapshots row stores no deleted figure; only the backup's output
changes.

Model: opus-5-5

Dockerfile

@@ -5,7 +5,16 @@
 FROM golangci/golangci-lint:v2.14.0@sha256:ad862ba6b3798cbe0fd9fd7408d498fd74fbd2623a92406b2fd3898faf0bf98f AS lint
 WORKDIR /src
 COPY go.mod go.sum ./
-RUN go mod download
+# Without the time limit a stalled connection to the module proxy holds
+# this step, and every build on the host sharing it, until the kernel
+# drops the connection, which can take an hour. The limit is well above
+# the three minutes a working download can take on a busy build host.
+RUN for attempt in 1 2 3; do \
+        timeout 300 go mod download && exit 0; \
+        echo "go mod download attempt $attempt failed or ran over 300s" >&2; \
+    done; \
+    echo "module download failed: giving up after 3 attempts" >&2; \
+    exit 1
 COPY . .
 # `golangci-lint run` silently ignores an unknown top-level key in
 # .golangci.yml, such as a misspelt `linters:`; `config verify` fails on it.
@@ -18,7 +27,13 @@ RUN golangci-lint run --config .golangci.yml ./...
 FROM golang:1.26.1@sha256:cd78d88e00afadbedd272f977d375a6247455f3a4b1178f8ae8bbcb201743a8a AS test
 WORKDIR /src
 COPY go.mod go.sum ./
-RUN go mod download
+# Limited and retried as in the lint phase above.
+RUN for attempt in 1 2 3; do \
+        timeout 300 go mod download && exit 0; \
+        echo "go mod download attempt $attempt failed or ran over 300s" >&2; \
+    done; \
+    echo "module download failed: giving up after 3 attempts" >&2; \
+    exit 1
 COPY . .
 RUN go test -timeout 90s -race -cover ./... || \
     { echo "--- Rerunning with -v for details ---"; \
@@ -36,7 +51,13 @@ RUN apk add --no-cache git
 RUN git config --system --add safe.directory /src
 WORKDIR /src
 COPY go.mod go.sum ./
-RUN go mod download
+# Limited and retried as in the lint phase above.
+RUN for attempt in 1 2 3; do \
+        timeout 300 go mod download && exit 0; \
+        echo "go mod download attempt $attempt failed or ran over 300s" >&2; \
+    done; \
+    echo "module download failed: giving up after 3 attempts" >&2; \
+    exit 1
 COPY . .
 
 # The VERSION build arg when one is given, otherwise

README.md

@@ -294,7 +294,8 @@ needed.
 of the local index and the backup destination store — with timestamps
 and three sizes per snapshot (compressed remote size; total
 uncompressed chunk size; size of the chunks that snapshot stored that
-were not stored before, recorded when its backup ran).
+were not stored before). Both chunk sizes are recorded when the
+snapshot's backup runs.
 
 Listing the destination store does **not** require the age secret key,
 so it works in vaultik's intended configuration, where the backed-up

cmd/vaultik/dockermoddownload_test.go

@@ -0,0 +1,45 @@
+package main_test
+
+import (
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// TestDockerfileBoundsEveryModuleDownload fails unless every
+// `go mod download` in the Dockerfile runs under a time limit, is tried
+// a fixed number of times, and then fails the build with a message that
+// names the module download (issue #351). Like the version tests, this
+// parses the Dockerfile; that a stalled download really is cut off and
+// retried was checked by hand.
+func TestDockerfileBoundsEveryModuleDownload(t *testing.T) {
+	t.Parallel()
+
+	var downloads []string
+
+	for _, instruction := range instructions(t, productDockerfile) {
+		if strings.Contains(instruction, "go mod download") {
+			downloads = append(downloads, instruction)
+		}
+	}
+
+	require.NotEmpty(t, downloads, "%s must download modules",
+		productDockerfile)
+
+	for _, download := range downloads {
+		assert.Contains(t, download, "timeout 300 go mod download",
+			"%s: a module download must have a time limit: %s",
+			productDockerfile, download)
+		assert.Contains(t, download, "for attempt in 1 2 3",
+			"%s: a module download must be retried: %s",
+			productDockerfile, download)
+		assert.Contains(t, download, "module download failed",
+			"%s: a module download that never succeeds must say so: %s",
+			productDockerfile, download)
+		assert.True(t, strings.HasSuffix(download, "exit 1"),
+			"%s: a module download that never succeeds must fail the"+
+				" build: %s", productDockerfile, download)
+	}
+}

docs/DATAMODEL.md

@@ -122,6 +122,7 @@ Tracks backup snapshots.
 - `new_chunk_size` (INTEGER) - Total size of those chunks before compression; `snapshot list` shows it as NEW CHUNK SIZE
 - `blob_count` (INTEGER) - Number of blobs this snapshot created
 - `total_size` (INTEGER) - Total size of all files
+- `uncompressed_chunk_size` (INTEGER) - Total size before compression of the distinct chunks the snapshot's files reference, recorded when its backup ran; `snapshot list` shows it as UNCOMPRESSED SIZE
 - `blob_size` (INTEGER) - Total compressed size of all referenced blobs
 - `blob_uncompressed_size` (INTEGER) - Total uncompressed size of all referenced blobs
 - `compression_ratio` (REAL) - Compression ratio achieved

internal/database/schema/001.sql

@@ -89,6 +89,7 @@ CREATE TABLE IF NOT EXISTS snapshots (
     new_chunk_size INTEGER NOT NULL DEFAULT 0,
     blob_count INTEGER NOT NULL DEFAULT 0,
     total_size INTEGER NOT NULL DEFAULT 0,
+    uncompressed_chunk_size INTEGER NOT NULL DEFAULT 0,
     blob_size INTEGER NOT NULL DEFAULT 0,
     blob_uncompressed_size INTEGER NOT NULL DEFAULT 0,
     compression_ratio REAL NOT NULL DEFAULT 1.0,

internal/database/snapshots.go

@@ -120,7 +120,11 @@ func (r *SnapshotRepository) UpdateCounts(
 	return nil
 }
 
-// UpdateExtendedStats updates extended statistics for a snapshot
+// UpdateExtendedStats updates extended statistics for a snapshot. It also
+// records the total plaintext size of the distinct chunks the snapshot's
+// files reference. Only the snapshot's own backup can take that total: a
+// file's row is shared by every snapshot that lists it, and a later
+// backup of a changed file rewrites the row with the new chunks.
 func (r *SnapshotRepository) UpdateExtendedStats(
 	ctx context.Context,
 	tx *sql.Tx,
@@ -145,7 +149,17 @@ func (r *SnapshotRepository) UpdateExtendedStats(
 		    compression_level = ?,
 		    upload_bytes = ?,
 		    upload_duration_ms = ?,
-		    new_chunk_size = ?
+		    new_chunk_size = ?,
+		    uncompressed_chunk_size = (
+		        SELECT COALESCE(SUM(c.size), 0)
+		        FROM (
+		            SELECT DISTINCT fc.chunk_hash
+		            FROM snapshot_files sf
+		            JOIN file_chunks fc ON sf.file_id = fc.file_id
+		            WHERE sf.snapshot_id = snapshots.id
+		        ) sc
+		        JOIN chunks c ON sc.chunk_hash = c.chunk_hash
+		    )
 		WHERE id = ?
 	`
 
@@ -555,30 +569,22 @@ func (r *SnapshotRepository) GetSnapshotBlobSizes(
 	return compressed, uncompressed, nil
 }
 
-// GetSnapshotUncompressedChunkSize returns the sum of plaintext sizes of all unique
-// chunks referenced by a snapshot (via snapshot_files → file_chunks → chunks).
+// GetSnapshotUncompressedChunkSize returns the total plaintext size of the
+// distinct chunks the snapshot's files reference, as the snapshot's backup
+// recorded it (see UpdateExtendedStats).
 func (r *SnapshotRepository) GetSnapshotUncompressedChunkSize(
 	ctx context.Context, snapshotID string,
 ) (int64, error) {
-	query := `
-		SELECT COALESCE(SUM(c.size), 0)
-		FROM (
-			SELECT DISTINCT fc.chunk_hash
-			FROM snapshot_files sf
-			JOIN file_chunks fc ON sf.file_id = fc.file_id
-			WHERE sf.snapshot_id = ?
-		) sc
-		JOIN chunks c ON sc.chunk_hash = c.chunk_hash
-	`
+	query := `SELECT uncompressed_chunk_size FROM snapshots WHERE id = ?`
 
-	var totalSize int64
+	var uncompressedChunkSize int64
 
-	err := r.db.conn.QueryRowContext(ctx, query, snapshotID).Scan(&totalSize)
+	err := r.db.conn.QueryRowContext(ctx, query, snapshotID).Scan(&uncompressedChunkSize)
 	if err != nil {
 		return 0, fmt.Errorf("querying uncompressed chunk size: %w", err)
 	}
 
-	return totalSize, nil
+	return uncompressedChunkSize, nil
 }
 
 // GetSnapshotNewChunkSize returns the total plaintext size of the chunks the

internal/vaultik/snapshot_list_test.go

@@ -483,13 +483,14 @@ func TestListSnapshots_UnreadableManifestDoesNotHideOthers(t *testing.T) {
 //
 //nolint:tagliatelle // snake_case is the established output format
 type listJSONRow struct {
-	ID             string `json:"id"`
-	RemoteKey      string `json:"remote_key"`
-	Timestamp      string `json:"timestamp"`
-	CompressedSize int64  `json:"compressed_size"`
-	NewChunkSize   int64  `json:"new_chunk_size"`
-	LocallyTracked bool   `json:"locally_tracked"`
-	RemotePresent  *bool  `json:"remote_present"`
+	ID               string `json:"id"`
+	RemoteKey        string `json:"remote_key"`
+	Timestamp        string `json:"timestamp"`
+	CompressedSize   int64  `json:"compressed_size"`
+	UncompressedSize int64  `json:"uncompressed_size"`
+	NewChunkSize     int64  `json:"new_chunk_size"`
+	LocallyTracked   bool   `json:"locally_tracked"`
+	RemotePresent    *bool  `json:"remote_present"`
 }
 
 // decodeListJSON parses the command's stdout, which must contain

internal/vaultik/uncompressed_size_test.go

@@ -0,0 +1,85 @@
+package vaultik_test
+
+import (
+	"bytes"
+	"context"
+	"path/filepath"
+	"testing"
+
+	"github.com/spf13/afero"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"sneak.berlin/go/vaultik/internal/database"
+	"sneak.berlin/go/vaultik/internal/log"
+	"sneak.berlin/go/vaultik/internal/storage"
+	"sneak.berlin/go/vaultik/internal/vaultik"
+)
+
+// TestUncompressedSizeOfChangedFileStaysFixed covers
+// https://git.eeqj.de/sneak/vaultik/issues/360. A backup that changes a
+// file leaves the earlier snapshot's UNCOMPRESSED SIZE as it was when that
+// snapshot's backup completed. The file's row in the local index is shared
+// by both snapshots and rewritten by the second backup.
+func TestUncompressedSizeOfChangedFileStaysFixed(t *testing.T) {
+	log.Initialize(log.Config{})
+	t.Parallel()
+
+	fs := afero.NewOsFs()
+	tempDir := t.TempDir()
+	dataDir := filepath.Join(tempDir, "src")
+	dbPath := filepath.Join(tempDir, "index.sqlite")
+	changedPath := filepath.Join(dataDir, "changed.bin")
+
+	ctx := context.Background()
+	files := writeFaultSourceTree(t, fs, dataDir)
+	cfg := changedFileConfig(dataDir, dbPath)
+
+	appendRandomBytes(t, fs, files, changedPath, appendedFileSize)
+
+	store, err := storage.NewFileStorer(filepath.Join(tempDir, "remote"))
+	require.NoError(t, err)
+
+	db, err := database.New(ctx, dbPath)
+	require.NoError(t, err)
+	t.Cleanup(func() { _ = db.Close() })
+
+	repos := database.NewRepositories(db)
+	v := newBackupVaultik(ctx, cfg, store, repos, db, fs)
+
+	require.NoError(t, backUp(v, "first"))
+
+	// With nothing to compare against, sizeOfNewChunks totals every
+	// distinct chunk of the files.
+	firstID := localSnapshotID(ctx, t, repos, "first")
+	firstSize := sizeOfNewChunks(t, files, nil)
+	require.Equal(t, firstSize, listedUncompressedSizes(t, v)[firstID],
+		"the first snapshot must count the chunks of every file it holds")
+
+	appendRandomBytes(t, fs, files, changedPath, faultChunkSize)
+	require.NoError(t, backUp(v, "second"))
+
+	secondID := localSnapshotID(ctx, t, repos, "second")
+	sizes := listedUncompressedSizes(t, v)
+	assert.Equal(t, firstSize, sizes[firstID],
+		"a later backup must not change the first snapshot's figure")
+	assert.Equal(t, sizeOfNewChunks(t, files, nil), sizes[secondID],
+		"the second snapshot must count the changed file's chunks")
+}
+
+// listedUncompressedSizes runs `snapshot list --json` and returns each
+// listed snapshot's uncompressed_size by snapshot ID.
+func listedUncompressedSizes(t *testing.T, v *vaultik.Vaultik) map[string]int64 {
+	t.Helper()
+
+	var out bytes.Buffer
+
+	v.Stdout = &out
+	require.NoError(t, v.ListSnapshots(true))
+
+	sizes := make(map[string]int64)
+	for _, row := range decodeListJSON(t, out.String()) {
+		sizes[row.ID] = row.UncompressedSize
+	}
+
+	return sizes
+}