Ensembia/nextcloud-ng_backup
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 17 Wiki Activity

All folders get their original modification time back on a restore

Browse Source 
The backup records the time of the folders with something in them, packed in the meta record,
so an older version still reads the snapshot. A restore through Nextcloud leaves a folder that
was already there alone. The round trip checks what Nextcloud shows, not only the disk.
This commit is contained in:
Dillard Blom
2026-11-07 01:07:09 +01:00
parent 36f2427e9a
commit a8b5e85c9b
8 changed files with 244 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lib/Job/BackupRun.php
+28 -3
View File
@@ -52,14 +52,30 @@ final class BackupRun {
$stats = $state['stats'];
$walker = new TreeWalker($state['roots'], $state['exclude'], function (string $logical) use (&$state, &$stats): void {
self::recordUnreadable($state, $stats, $logical);
});
}, true);
$parent = new ParentCursor($repo, $state['parent'], $state['parentPos']);
$lines = [];
$emptyDirs = [];
$folders = [];
$finishedWalk = true;
foreach ($walker->walk($state['after']) as $logical => $path) {
if (!mb_check_encoding($logical, 'UTF-8') || !Repository::isSafePath($logical)) {
$safe = mb_check_encoding($logical, 'UTF-8') && Repository::isSafePath($logical);
if ($walker->atFolderWithContent()) {
// Only its time: what is in it is reported by itself, and a bad name in the path of
// each of those is what gets reported. The cursor stays for such a name, see below.
$mtime = $safe ? @filemtime($path) : false;
if ($safe && $mtime !== false) {
$folders[] = json_encode([$logical, $mtime], JSON_UNESCAPED_SLASHES | JSON_THROW_ON_ERROR);
$state['after'] = $logical;
if (microtime(true) >= $deadline) {
$finishedWalk = false;
break;
}
}
continue;
}
if (!$safe) {
// A name a snapshot cannot hold (a backslash, a control character, not UTF-8). One
// such path must not make a backup fail, or the whole snapshot unreadable later.
// The cursor stays where it is: the state is JSON, which cannot hold such a name. A
@@ -164,6 +180,11 @@ final class BackupRun {
$writer->write(implode("\n", $emptyDirs) . "\n");
$state['dirSegments'][] = $writer->finish();
}
if ($folders !== []) {
$writer = $repo->blobWriter($stats);
$writer->write(implode("\n", $folders) . "\n");
$state['folderSegments'][] = $writer->finish();
}
$repo->flushPacks();
$state['parentPos'] = $parent->position();
$state['stats'] = $stats;
@@ -186,10 +207,14 @@ final class BackupRun {
$emptyDirs[] = json_decode($line, true, 8, JSON_THROW_ON_ERROR);
}
}
// The times of the other folders too, packed: every look at a snapshot's meta record reads
// this, and a string of a few hundred KiB costs that far less than tens of thousands of arrays.
$folders = isset($state['folderSegments']) ? implode("\n", iterator_to_array($repo->readLines(array_values(array_merge(...$state['folderSegments']))), false)) : '';
$repo->putSnapshot($state['snapshot'],
['time' => $state['started'], 'parent' => $state['parent'], 'label' => $state['label'],
'roots' => array_keys($state['roots']), 'stats' => $state['stats'],
'warnings' => self::warnings($state)] + ($emptyDirs === [] ? [] : ['emptyDirs' => $emptyDirs]) + $state['meta'],
'warnings' => self::warnings($state)] + ($emptyDirs === [] ? [] : ['emptyDirs' => $emptyDirs])
+ ($folders === '' ? [] : ['folderTimes' => base64_encode((string)gzdeflate($folders))]) + $state['meta'],
$repo->readLines(array_values(array_merge(...$state['segments']))));
$state['phase'] = 'commit';
}
lib/Job/TreeWalker.php
+17 -1
View File
@@ -17,6 +17,8 @@ namespace OCA\NgBackup\Job;
* and be merge-joined with the previous snapshot without holding any list in memory.
*/
final class TreeWalker {
private bool $atFolder = false;
/**
* @param array<string, string> $roots name => absolute directory, walked in this order
* @param list<string> $exclude path prefixes ("<root>/<relative>") that are skipped entirely;
@@ -27,9 +29,16 @@ final class TreeWalker {
private array $exclude = [],
/** Called with the logical path of a directory that cannot be listed. */
private ?\Closure $onUnreadable = null,
/** Also yield directories with something in them, right before the first thing in them. */
private bool $withFolders = false,
) {
}
/** Whether what was just yielded is a directory with something in it (see $withFolders). */
public function atFolderWithContent(): bool {
return $this->atFolder;
}
/** @return \Generator<string, string> logical path => absolute path, strictly after $after */
public function walk(?string $after = null): \Generator {
foreach ($this->roots as $name => $dir) {
@@ -75,7 +84,14 @@ final class TreeWalker {
if (is_link($childAbs)) {
continue; // never follow symlinks out of the data directory
}
if (is_dir($childAbs)) {
$childIsDir = is_dir($childAbs);
if (($childIsDir || is_file($childAbs)) && $empty && !$this->isExcluded($childLogical, $childIsDir)
&& $this->withFolders && $logical !== '' && ($after === null || self::compare($logical, $after) > 0)) {
$this->atFolder = true;
yield $logical => $abs;
$this->atFolder = false;
}
if ($childIsDir) {
$empty = $empty && $this->isExcluded($childLogical, true);
// Skip a directory entirely if everything in it sorts before the cursor.
if ($after !== null && self::compare($childLogical, $after) < 0 && !self::isPrefix($childLogical, $after)) {