Ensure returned versionId is a string when listing objects

Returned versionId is a string when calling statObject(), as it is read
without transformation (casting) from the headers.

However, the xml parser casts automatically what looks like a number in
the responses. Consequentially, when the S3 storage provider returns a
VersionId with only numbers, listObjects() returns an object with a
versionId as a number, which is inconsistent with what returns other
methods like statObjects() and may result in erratic behaviors for the
program using these methods.

src/internal/helper.ts

@@ -462,6 +462,10 @@ export function sanitizeObjectKey(objectName: string): string {
   return decodeURIComponent(asStrName)
 }
 
+export function sanitizeSize(size?: string): number | undefined {
+  return size ? Number.parseInt(size) : undefined
+}
+
 export const PART_CONSTRAINTS = {
   // absMinPartSize - absolute minimum part size (5 MiB)
   ABS_MIN_PART_SIZE: 1024 * 1024 * 5,

src/xml-parsers.js

@@ -15,10 +15,19 @@
  */
 
 import crc32 from 'buffer-crc32'
+import { XMLParser } from 'fast-xml-parser'
 
 import * as errors from './errors.ts'
 import { SelectResults } from './helpers.ts'
-import { isObject, parseXml, readableStream, sanitizeETag, sanitizeObjectKey, toArray } from './internal/helper.ts'
+import {
+  isObject,
+  parseXml,
+  readableStream,
+  sanitizeETag,
+  sanitizeObjectKey,
+  sanitizeSize,
+  toArray,
+} from './internal/helper.ts'
 import { RETENTION_VALIDITY_UNITS } from './internal/type.ts'
 
 // parse XML response for copy object
@@ -216,12 +225,13 @@ const formatObjInfo = (content, opts = {}) => {
   const name = sanitizeObjectKey(toArray(Key)[0])
   const lastModified = new Date(toArray(LastModified)[0])
   const etag = sanitizeETag(toArray(ETag)[0])
+  const size = sanitizeSize(Size)
 
   return {
     name,
     lastModified,
     etag,
-    size: Size,
+    size,
     versionId: VersionId,
     isLatest: IsLatest,
     isDeleteMarker: opts.IsDeleteMarker ? opts.IsDeleteMarker : false,
@@ -236,7 +246,12 @@ export function parseListObjects(xml) {
   }
   let isTruncated = false
   let nextMarker, nextVersionKeyMarker
-  const xmlobj = parseXml(xml)
+  const fxp = new XMLParser({
+    numberParseOptions: {
+      skipLike: /./,
+    },
+  })
+  const xmlobj = fxp.parse(xml)
 
   const parseCommonPrefixesEntity = (responseEntity) => {
     if (responseEntity) {
@@ -258,7 +273,7 @@ export function parseListObjects(xml) {
         const name = sanitizeObjectKey(toArray(content.Key)[0])
         const lastModified = new Date(toArray(content.LastModified)[0])
         const etag = sanitizeETag(toArray(content.ETag)[0])
-        const size = content.Size
+        const size = sanitizeSize(content.Size)
         result.objects.push({ name, lastModified, etag, size })
       })
     }

tests/unit/test.js

@@ -29,6 +29,7 @@ import {
   partsRequired,
 } from '../../src/internal/helper.ts'
 import * as Minio from '../../src/minio.js'
+import { parseListObjects } from '../../src/xml-parsers.ts'
 
 const Package = { version: 'development' }
 
@@ -2078,3 +2079,56 @@ describe('IP Address Validations', () => {
     })
   })
 })
+
+describe('xml-parser', () => {
+  describe('#listObjects()', () => {
+    describe('value type casting', () => {
+      const xml = `
+          <?xml version="1.0" encoding="UTF-8"?>
+          <ListVersionsResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/">
+            <Name>some-bucket</Name>
+            <Prefix>42</Prefix>
+            <Delimiter>/</Delimiter>
+            <IsTruncated>false</IsTruncated>
+            <EncodingType>url</EncodingType>
+            <KeyMarker/>
+            <VersionIdMarker/>
+            <Version>
+              <IsLatest>true</IsLatest>
+              <VersionId>1234</VersionId>
+              <ETag>"767dedcb515a0e2d995ed95191b75484-29"</ETag>
+              <Key>1337</Key>
+              <LastModified>2023-07-12T14:41:46.000Z</LastModified>
+              <Size>151306240</Size>
+            </Version>
+            <DeleteMarker>
+              <IsLatest>false</IsLatest>
+              <Key>1337</Key>
+              <LastModified>2023-07-12T14:39:22.000Z</LastModified>
+              <VersionId>5678</VersionId>
+            </DeleteMarker>
+            <CommonPrefixes>
+              <Prefix>42</Prefix>
+            </CommonPrefixes>
+          </ListVersionsResult>
+        `
+
+      it('should parse VersionId as string even if number is provided', () => {
+        const { objects } = parseListObjects(xml)
+
+        assert.equal(objects[0].versionId, '1234')
+        assert.equal(objects[1].versionId, '5678')
+        assert.equal(objects[0].name, '1337')
+        assert.equal(objects[1].name, '1337')
+        assert.deepEqual(objects[2], { prefix: '42', size: 0 })
+      })
+
+      it('should parse Size as number', () => {
+        const { objects } = parseListObjects(xml)
+
+        assert.equal(objects[0].size, 151306240)
+        assert.equal(objects[1].size, undefined)
+      })
+    })
+  })
+})