`borg2 mount`: Change top directory names when mounting a full repo

[PhrozenByte]

When mounting a full repo with borg2 mount, the top-level directories are currently named {name}-{id}.

This makes sense because archive names are no longer unique in borg2, but using the archive ID (aid) isn't very user-friendly. It's just a hash, so if one wants to mount or inspect one or more specific archives, one first has to run borg2 repo-list to figure out which aid belongs to which archive. This is a bit cumbersome.

This only affects mounting the whole repo. Mounting a single archive or a specific path isn't affected.

I'd like to propose two possible alternative approaches:

1. Use a better hardcoded directory name format.
   We could discuss which metadata is most useful and include that in the directory name. I think the archive start time is probably the most useful addition, so something like {name}-{id}-{start} would already make things much easier. WDYT?

2. Make the directory name format configurable.
   Add an option (I'd probably prefer a mount option over a CLI option) that lets users define the directory name format themselves, using the same placeholders that are already supported by commands like borg2 repo-list.

   The default could stay {name}-{id}, or it could switch to something like {name}-{id}-{start}.

[ThomasWaldmann]

Doesn't it use the archive creation time already as mtime for the corresponding directory, so you can even sort by that column in a file manager.

Yes, it could be configurable, but it would also be good to be sane-by-default.

When creating the directory names, we have access to ALL archive metadata, but I would avoid putting too much information into the name, especially if it can be represented by other directory metadata also.