core/backup/archive.go

package backup

import (
	"archive/zip"
	"context"
	"crypto/sha256"
	"encoding/base64"
	"encoding/json"
	"fmt"
	"io"
	"os"
	"path/filepath"
	"sort"
	"time"
)

// Section maps a zip path prefix to an on-disk directory to be archived.
type Section struct {
	Prefix string
	Dir    string
}

// BuildInput is the full set of inputs for BuildArchive.
type BuildInput struct {
	Source   Source
	Dump     []byte
	Password string
	Sections []Section
	// Extra holds additional fixed zip entries (zip path -> contents), written
	// after database.dump and before section files.
	Extra map[string][]byte
}

// backupFile pairs a zip path with its source path on disk.
type backupFile struct {
	zipPath  string
	diskPath string
}

// collectSectionFiles walks a section directory and returns zipPath→diskPath
// pairs. A missing directory is skipped silently (count 0, no warning), matching
// the cms collectSectionFiles semantics. Files that cannot be accessed inside an
// existing directory append warnings and are skipped.
func collectSectionFiles(ctx context.Context, section Section, warnings *[]string) ([]backupFile, error) {
	if section.Dir == "" {
		return nil, nil
	}
	if _, err := os.Stat(section.Dir); os.IsNotExist(err) {
		return nil, nil
	}

	var files []backupFile
	err := filepath.Walk(section.Dir, func(path string, info os.FileInfo, err error) error {
		if ctx.Err() != nil {
			return ctx.Err()
		}
		if err != nil {
			*warnings = append(*warnings, fmt.Sprintf("Failed to access %s: %v", path, err))
			return nil
		}
		if info.IsDir() || !info.Mode().IsRegular() {
			return nil
		}

		relPath, err := filepath.Rel(section.Dir, path)
		if err != nil {
			*warnings = append(*warnings, fmt.Sprintf("Failed to get relative path for %s: %v", path, err))
			return nil
		}

		files = append(files, backupFile{
			zipPath:  section.Prefix + filepath.ToSlash(relPath),
			diskPath: path,
		})
		return nil
	})
	if err != nil {
		if ctx.Err() != nil {
			return nil, ctx.Err()
		}
		*warnings = append(*warnings, fmt.Sprintf("Failed to walk %s: %v", section.Dir, err))
	}
	return files, nil
}

// BuildArchive streams a backup zip archive to w: a manifest.json, the
// password-encrypted database dump, any Extra entries, then the files of each
// configured section in order. It returns the manifest and any non-fatal
// warnings encountered while collecting/writing section files.
//
// Zip entry ordering is: manifest.json, database.dump, Extra (sorted by path),
// then section files in section order. Counts for the well-known section
// prefixes (media/, brand/, icons/, styles/, plugins/) are recorded in the
// manifest.
func BuildArchive(ctx context.Context, w io.Writer, in BuildInput) (*Manifest, []string, error) {
	var warnings []string

	if err := ctx.Err(); err != nil {
		return nil, warnings, err
	}

	// Checksum of the unencrypted dump.
	checksum := computeChecksum(in.Dump)

	// Encrypt the dump.
	encryptedDump, encParams, err := EncryptWithPassword(in.Dump, in.Password)
	if err != nil {
		return nil, warnings, fmt.Errorf("encryption failed: %w", err)
	}

	// Collect files for every configured section, preserving section order.
	sectionFiles := make([][]backupFile, len(in.Sections))
	countsByPrefix := make(map[string]int)
	for i, section := range in.Sections {
		files, err := collectSectionFiles(ctx, section, &warnings)
		if err != nil {
			return nil, warnings, err
		}
		sectionFiles[i] = files
		countsByPrefix[section.Prefix] += len(files)
	}

	manifest := &Manifest{
		Version:      FormatVersion,
		Format:       FormatName,
		ExportedAt:   time.Now().UTC(),
		Source:       in.Source,
		DatabaseSize: int64(len(in.Dump)),
		MediaCount:   countsByPrefix["media/"],
		BrandCount:   countsByPrefix["brand/"],
		IconsCount:   countsByPrefix["icons/"],
		StylesCount:  countsByPrefix["styles/"],
		PluginsCount: countsByPrefix["plugins/"],
		Encryption:   encParams,
		Checksum:     checksum,
	}

	manifestJSON, err := json.MarshalIndent(manifest, "", "  ")
	if err != nil {
		return nil, warnings, fmt.Errorf("manifest marshal failed: %w", err)
	}

	if err := ctx.Err(); err != nil {
		return nil, warnings, err
	}

	zipWriter := zip.NewWriter(w)

	manifestEntry, err := zipWriter.Create("manifest.json")
	if err != nil {
		return nil, warnings, fmt.Errorf("failed to create manifest.json in zip: %w", err)
	}
	if _, err := manifestEntry.Write(manifestJSON); err != nil {
		return nil, warnings, fmt.Errorf("failed to write manifest.json: %w", err)
	}

	dumpEntry, err := zipWriter.Create("database.dump")
	if err != nil {
		return nil, warnings, fmt.Errorf("failed to create database.dump in zip: %w", err)
	}
	if _, err := dumpEntry.Write(encryptedDump); err != nil {
		return nil, warnings, fmt.Errorf("failed to write database.dump: %w", err)
	}

	// Extra fixed entries, in a stable (sorted) order for reproducibility.
	// Failure contract (cms parity — the cssManifestsFile block in
	// buildBackupArchive): a failed Create skips the entry silently, a failed
	// Write appends a warning. Extra entries are regenerable extras; a problem
	// with one must never abort the archive.
	if len(in.Extra) > 0 {
		names := make([]string, 0, len(in.Extra))
		for name := range in.Extra {
			names = append(names, name)
		}
		sort.Strings(names)
		for _, name := range names {
			entry, err := zipWriter.Create(name)
			if err != nil {
				continue
			}
			if _, err := entry.Write(in.Extra[name]); err != nil {
				warnings = append(warnings, fmt.Sprintf("Failed to write %s: %v", name, err))
			}
		}
	}

	// Section files, with periodic context checks (every 10 files).
	written := 0
	for _, files := range sectionFiles {
		for _, bf := range files {
			if written%10 == 0 {
				if err := ctx.Err(); err != nil {
					return nil, warnings, err
				}
			}
			written++

			src, err := os.Open(bf.diskPath)
			if err != nil {
				warnings = append(warnings, fmt.Sprintf("Failed to read %s for zip: %v", bf.diskPath, err))
				continue
			}

			entry, err := zipWriter.Create(bf.zipPath)
			if err != nil {
				closeBestEffort(src)
				warnings = append(warnings, fmt.Sprintf("Failed to create %s in zip: %v", bf.zipPath, err))
				continue
			}

			if _, err := io.Copy(entry, src); err != nil {
				warnings = append(warnings, fmt.Sprintf("Failed to write %s to zip: %v", bf.zipPath, err))
			}
			closeBestEffort(src)
		}
	}

	if err := zipWriter.Close(); err != nil {
		return nil, warnings, fmt.Errorf("failed to close zip: %w", err)
	}

	return manifest, warnings, nil
}

func closeBestEffort(c io.Closer) {
	_ = c.Close()
}

// computeChecksum calculates the SHA-256 hash of data, base64-encoded —
// byte-identical to the cms computeChecksum so the existing cms restore path
// (verifyChecksum) accepts archives built by this package.
func computeChecksum(data []byte) string {
	hash := sha256.Sum256(data)
	return base64.StdEncoding.EncodeToString(hash[:])
}