allow deleting only older empty dir without recursion

[kmlebedev]

What problem are we solving?

When write operations are highly concurrent, deletion sometimes deletes files supposedly from empty folders
#4015

How are we solving the problem?

Now not only subfolders are checked, but also the folder itself, that it has been created for more than 3 seconds

How is the PR tested?

make dev
s3cmd --no-ssl --host-bucket=127.0.0.1:8333 --host=127.0.0.1:8333 put fileb s3://test/newFolder/subfolder/fileb
s3cmd --no-ssl --host-bucket=127.0.0.1:8333 --host=127.0.0.1:8333 rm s3://test/newFolder/subfolder/fileb
s3cmd --no-ssl --host-bucket=127.0.0.1:8333 --host=127.0.0.1:8333 ls s3://test/

localy

eaweedfs-filer-1   | I0425 13:14:07.934076 filer_grpc_server.go:40 ListEntries directory:"/buckets/test"  limit:10002
seaweedfs-s3-1      | + ensureDirectoryAllEmpty /buckets/test newFolder
seaweedfs-s3-1      | I0425 13:14:07.935827 s3api_objects_list_handlers.go:440 + isEmpty /buckets/test/newFolder
seaweedfs-s3-1      | I0425 13:14:07.935840 filer_client.go:120 read directory: directory:"/buckets/test/newFolder"  limit:9
seaweedfs-filer-1   | I0425 13:14:07.936346 filer_grpc_server.go:40 ListEntries directory:"/buckets/test/newFolder"  limit:9
seaweedfs-s3-1      | I0425 13:14:07.936771 s3api_objects_list_handlers.go:485 deleting empty folder /buckets/test/newFolder
seaweedfs-s3-1      | I0425 13:14:07.936805 filer_util.go:65 delete entry /buckets/test/newFolder: directory:"/buckets/test"  name:"newFolder"  is_delete_data:true  ignore_recursive_error:true
seaweedfs-filer-1   | I0425 13:14:07.937558 filer_grpc_server.go:286 DeleteEntry directory:"/buckets/test"  name:"newFolder"  is_delete_data:true  ignore_recursive_error:true
seaweedfs-filer-1   | I0425 13:14:07.937863 filer_delete_entry.go:110 deleting directory /buckets/test/newFolder delete chunks: true
seaweedfs-filer-1   | I0425 13:14:07.937939 filer_delete_entry.go:123 deleting entry /buckets/test/newFolder, delete chunks: true
seaweedfs-filer-1   | I0425 13:14:07.939267 filer_grpc_server.go:40 ListEntries directory:"/buckets/test"  startFromFileName:"newFolder"  limit:10001
seaweedfs-s3-1      | I0425 13:14:07.938733 s3api_objects_list_handlers.go:491 - isEmpty /buckets/test/newFolder false
seaweedfs-s3-1      | I0425 13:14:07.941002 error_handler.go:96 status 200 application/xml: <?xml version="1.0" encoding="UTF-8"?>
seaweedfs-s3-1      | <ListBucketResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/"><Name>test</Name><Prefix></Prefix><Marker></Marker><MaxKeys>10000</MaxKeys><Delimiter>/</Delimiter><IsTruncated>false</IsTruncated><Contents><Key>fakedir/</Key><ETag>&#34;d41d8cd98f00b204e9800998ecf8427e-0&#34;</ETag><Size>0</Size><Owner><ID>0</ID></Owner><StorageClass>STANDARD</StorageClass><LastModified>2023-04-12T09:02:40Z</LastModified></Contents></ListBucketResult>

Checks

• I have added unit tests if possible.
• I will add related wiki document changes and link to this PR after merging.
• [ ]