
On 11-12-2014 06:14, Royce Lv wrote:
For filename like "imgzip.tar.gz", we will get prefix for "imgzip.tar" and suffix ".gz" then name will be "imgzip.tar-clone-1.gz", because some image disk image may be compressed for download.
* the new name would be "imgzip.tar-1.gz" :) The "-clone" tag isn't added in that case. I agree that this is an "ugly" name, even though it solves the issue we're working on: not being able to download files with the same name. But what automatic rule can we use to generate a nice name for every file we download? Let's try it with "http://www.domain.com", "http://www.domain.com/foo", "http://www.domain.com/foo.zip", "http://www.domain.com/foo.tar.gz", "http://www.domain.com/foo.important.zip". It's hard to guess what the real extension is. I was looking at the behaviour of browsers (e.g. Chrome, Firefox), and I noticed they add " (1)" to the file name when it already exists. We can do that as well, even though I don't think it's standard to how we've been naming things in Kimchi: "my-template-vm-1", "my-vm-clone-1". I think "my-vol.tar.gz (1)" would look out of place. How about "my-vol.tar.gz-1" and ignore file extension completely? Do you have any other suggestion?