Snapshot synced v3.6.1 working state as baseline

Machine-sync snapshot: the working tree carried the full 3.0.0-3.6.1 arc (licensing/FileManager/DatabaseExplorer removal, all new replacement screens and REST controllers) that had not been committed on this machine. Captured here as a safety baseline before starting the Adminify-parity revamp. Co-Authored-By: Claude Opus 4.8 <[email protected]>
2026-07-12 08:41:18 -05:00
parent 07d77bda57
commit e2db2ad4b5
88 changed files with 8944 additions and 2730 deletions
@@ -0,0 +1,98 @@
+import { ref, onUnmounted } from 'vue';
+
+// Phase weights for a rough overall-progress percentage. Backup and
+// restore jobs use different phase sets (see BackupEngine::effective_phases
+// on the PHP side) — only the phases relevant to whichever job is active
+// contribute, everything else is ignored.
+const PHASE_WEIGHTS = {
+  db_dump: 40,
+  file_zip: 40,
+  upload_remote: 20,
+  download_remote: 20,
+  db_restore: 50,
+  file_restore: 50,
+};
+
+const PHASE_LABELS = {
+  pending: 'Starting…',
+  db_dump: 'Backing up the database…',
+  file_zip: 'Archiving files…',
+  upload_remote: 'Uploading to remote storage…',
+  download_remote: 'Downloading from remote storage…',
+  db_restore: 'Restoring the database…',
+  file_restore: 'Restoring files…',
+  done: 'Done',
+  failed: 'Failed',
+  cancelled: 'Cancelled',
+};
+
+/**
+ * Drives a backup/restore job to completion by repeatedly calling the
+ * REST "step" route — a setTimeout-recursion loop (not setInterval), so a
+ * slow step can never overlap the next poll for the same job. The
+ * server-side wp_cron safety net keeps the job moving even if this stops
+ * polling (e.g. the tab is closed mid-backup) — this composable is purely
+ * the foreground "watch it happen live" experience, not the only thing
+ * keeping a job alive.
+ */
+export function useJobPolling(api) {
+  const job = ref(null);
+  const isPolling = ref(false);
+  const error = ref('');
+  let timer = null;
+  let stopped = false;
+
+  const percent = () => {
+    if (!job.value) return null;
+    const weight = PHASE_WEIGHTS[job.value.phase];
+    if (!weight) return job.value.status === 'done' ? 100 : null;
+    // No fine-grained progress within a phase is cheap to compute
+    // client-side (server tracks byte/row offsets, not percentages) — an
+    // indeterminate bar per phase, weighted by how much of the whole job
+    // that phase represents, is a reasonable approximation without a
+    // second round trip just to compute a number.
+    return null;
+  };
+
+  const label = () => {
+    if (!job.value) return '';
+    if (job.value.status === 'failed') return job.value.error_message || 'Failed';
+    return PHASE_LABELS[job.value.phase] || PHASE_LABELS[job.value.status] || '';
+  };
+
+  const poll = async (jobId) => {
+    if (stopped) return;
+    try {
+      const res = await api(`/backup/jobs/${jobId}/step`, { method: 'POST' });
+      job.value = res.job;
+    } catch (e) {
+      error.value = e.message || 'Request failed';
+      isPolling.value = false;
+      return;
+    }
+
+    if (!job.value || ['done', 'failed', 'cancelled'].includes(job.value.status)) {
+      isPolling.value = false;
+      return;
+    }
+
+    timer = setTimeout(() => poll(jobId), 1500);
+  };
+
+  const start = (jobId) => {
+    stopped = false;
+    isPolling.value = true;
+    error.value = '';
+    poll(jobId);
+  };
+
+  const stop = () => {
+    stopped = true;
+    isPolling.value = false;
+    if (timer) clearTimeout(timer);
+  };
+
+  onUnmounted(stop);
+
+  return { job, isPolling, error, start, stop, percent, label };
+}