fix: Better functionality for ipfs add --to-files -w

[ProximaNova]

Make ipfs add + --to-files + --wrap-with-directory work by adding the root CID of what was added to MFS. The is different from past official Kubo which did this: added a literal "." folder to MFS (bug). It's also different from current official Kubo: disallow add to be ran with to-files and wrap. Todo/ideas for this fork: make something like --to-files-cid which only adds root added CID to MFS, whether it's a file, folder, ran with wrap, ran without wrap. (Done: add+to-files+wrap always defaults to adding the CID to MFS and not literal "." and not disabled.)

This update makes perfect sense with --to-files; in other words it does the following:

• Without wrap, the final thing is "[CID] [file/folder name]" -> file/folder name added to MFS.
• With wrap, the final thing is "[CID]" -> CID added to MFS

If a user ran add + --to-files + -w then it's expected behavior to add the final CID to MFS. It being disabled would be unexpected. BTW, https://github.com/ipfs/community/blob/master/CONTRIBUTING_GO.md says something about changing branch name to something other than "master"; I didn't do that.

[guillaumemichel]

Related to #10611 and #10612

[ProximaNova]

Related to 10611 and 10612

(yes it is related to those two. I didn't mention that for some reason.)

Some checks were not successful
4 failing and 12 successful checks

I see this message in regards to and on this pull request after logging in to GitHub. This is insignificant syntax pickiness such as "should omit type bool from declaration; it will be inferred from the right-hand side (stylecheck)" at https://github.com/ipfs/kubo/actions/runs/12703758495/job/35598233653?pr=10658 . I compiled everything before making this pull request and my modified Go code all worked. It compiled or was built with no errors or warnings showing up in the CLI. The resulting binary worked as I described. I am saying all this to point out that those minor "fails" should not detract much from this idea and this functional code.

[guillaumemichel]

Now passing all tests except sharness tests that need to be updated to reflect the change of behaviour.

I am not sure of the value of combining -w and --to-files, which results in adding an unnamed directory in MFS. Filesystems usually use human readable names, so I don't understand why it would be desirable to have a directory named after a CID. It would be like calling mkdir without providing a name for the new directory.