Shorten filenames

src/history/data/metadata.js

@@ -6,10 +6,23 @@ export const getRawResultFilename = function (rawResult) {
 
 // Retrieve filename from a metadatum
 export const serializeFilename = function ({ id, timestamp }) {
+  const idStr = shortenId(id)
   const timestampStr = serializeTimestamp(timestamp)
-  return `${timestampStr}_${id}.json`
+  return `${timestampStr}_${idStr}.json`
 }
 
+// We only keep the last characters of the `result.id` in the filename.
+// This is to keep filenames short since some systems impose a limit.
+// We keep it high enough to prevent collisions though.
+//  - 12 hexadecimal characters is 48 bits of entropy, which has a 50%
+//    probability of collision after 2e7 results with the same `id`, which is
+//    very unlikely
+const shortenId = function (id) {
+  return id.length <= ID_LENGTH ? id : id.slice(0, -ID_LENGTH)
+}
+
+const ID_LENGTH = 12
+
 const serializeTimestamp = function (timestamp) {
   const date = new Date(timestamp)
   const year = date.getUTCFullYear()
@@ -53,4 +66,4 @@ const parseTimestamp = function ({
 }
 
 const RESULT_FILENAME_REGEXP =
-  /^(?<year>\d{4})_(?<month>\d{2})_(?<day>\d{2})_(?<hours>\d{2})_(?<minutes>\d{2})_(?<seconds>\d{2})_(?<id>[\da-f]{8}-[\da-f]{4}-[\da-f]{4}-[\da-f]{4}-[\da-f]{12})\.json$/u
+  /^(?<year>\d{4})_(?<month>\d{2})_(?<day>\d{2})_(?<hours>\d{2})_(?<minutes>\d{2})_(?<seconds>\d{2})_(?<id>[\da-f]{12})\.json$/u

src/history/delta/formats/id.js

@@ -9,9 +9,10 @@ const parseId = function (delta) {
   return delta
 }
 
+// Some stores shorten `id` and only keep the last 12 characters.
 const findById = function (metadataGroups, id) {
-  return metadataGroups.findIndex(
-    ([firstMetadatum]) => firstMetadatum.id === id,
+  return metadataGroups.findIndex(([firstMetadatum]) =>
+    id.endsWith(firstMetadatum.id),
   )
 }