fix(widget): dedup rapid completeTask taps on the same task

CompleteWorker.enqueue used a plain WorkManager.enqueue(), which allows
unlimited concurrent OneTimeWorkRequests. A rapid double-tap on the same
row (plausible since the checkbox doesn't visually update until the full
async round-trip -- complete() -> fetchAndPersist() -> updateAll() --
finishes) could spawn two independent, unordered CompleteWorker runs for
the same task, each doing its own fetchAndPersist(); a second worker's
fetch started before the first worker's complete() call had actually
landed server-side could persist a stale snapshot after the first
worker's correct one.

Now uses enqueueUniqueWork("complete_$id", KEEP, ...) so a tap on a task
that already has a completion in flight is dropped rather than racing a
second worker. Different task ids remain independent.

1 files changed, 20 insertions, 1 deletions

diff --git a/android/app/src/main/java/org/terst/doot/widget/work/CompleteWorker.kt b/android/app/src/main/java/org/terst/doot/widget/work/CompleteWorker.kt
index d3017ff..51ad727 100644
--- a/android/app/src/main/java/org/terst/doot/widget/work/CompleteWorker.kt
+++ b/android/app/src/main/java/org/terst/doot/widget/work/CompleteWorker.kt
@@ -35,12 +35,31 @@ class CompleteWorker(context: Context, params: WorkerParameters) :
         const val KEY_ID = "item_id"
         const val KEY_SOURCE = "item_source"
 
+        // enqueueUniqueWork(id, KEEP, ...) instead of a plain enqueue(): a
+        // rapid double-tap on the same row (plausible since the checkbox
+        // doesn't visually update until this worker's full async round-trip
+        // -- complete() -> fetchAndPersist() -> updateAll() -- finishes)
+        // previously spawned two independent, unordered CompleteWorker runs
+        // for the same task. Each does its own fetchAndPersist(), so a
+        // second worker's fetch (started before the first worker's
+        // complete() call had actually landed server-side) could persist a
+        // stale snapshot after the first worker's correct one, leaving the
+        // widget showing outdated data. KEEP means a tap on a task that
+        // already has a completion in flight is simply dropped -- the first
+        // request's result (including its fetchAndPersist/updateAll) is what
+        // takes effect, with no race between two workers touching the same
+        // task. Different task ids still run independently/concurrently,
+        // which is fine since they don't share a row.
         fun enqueue(context: Context, id: String, source: String) {
             val data = workDataOf(KEY_ID to id, KEY_SOURCE to source)
             val request = OneTimeWorkRequestBuilder<CompleteWorker>()
                 .setInputData(data)
                 .build()
-            WorkManager.getInstance(context).enqueue(request)
+            WorkManager.getInstance(context).enqueueUniqueWork(
+                "complete_$id",
+                ExistingWorkPolicy.KEEP,
+                request
+            )
         }
     }
 }