When using async uploads, S3Mock incorrectly stores unsigned, chunked uploads without removing meta (chunk, checksum) information when the file is >16KB about s3mock HOT 2 CLOSED

The reason for this is simple:

S3Mock expects the line ends for the chunk information to always be a CRLF / \r\n which is always the case when using AWS SDKs with synchronous uploads.
When using the async / CRT client however, it looks like sometimes, native line end is used, which in my case, on MacOS, is only CR / \n.
S3Mock is then unable to read chunk information, cutting off chunks when decoding the incoming stream.

The description in the AWS documentation is misleading, they claim that chunk information always ends with CRLF...

from s3mock.

so, when using Windows, which natively uses CRLF as the end of line characters, everything should work fine as-is.
This issue is for MacOS and Linux only.

from s3mock.