Trash spec 0.2, technical questions

[C. Gatzemeier]

Am Wednesday 01 September 2004 09:03 schrieb Alexander Larsson:

> > There is a lot of things which I think are more complex than are
> > warranted by making subdirectory's vs Nautilus' .Trash-${user} design.
>
> This is why i'm leaning towards the "Use .Trash if set up by sysadmin,
> otherwise try to create $topdir/.Trash-${uid} approach.


Ack. I see the need of a fallback solution in the spec that allows on-same-fs 
trashes.

Was anything said about the go up in the tree as long as you have the right 
permissions to create a valid /Trash approach?

If the permissions on the parent directory are already different the right 
place would be the current directory. It would be a "mv file1 ./Trash/file1
[timestamp]" sort of thing with world-writable ./Trash .

This wouldn't change or lose any permissions so the different ./Trash's could 
still be consolidated into a valid larger tree more upwards with right 
permissions later on.

One drawback of spead trashes would be you'd need to find and list all the 
trash dirs, but then again the spec could define personal indexes for 
implementations to keep up-to date.

Cheers,
Christian