[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] About Standas idea



Hi!

On 14 Apr 2003, Adam Klobukowski wrote:

> Ie. by deafault kernel will mount all drives u:/somewhere/
> 
> So we could have: u:/somewhere/a, u:/somewhee/b and so on...
> 
> (symlinks to U:/a U:/b should be made by kernel on bootup 
> automagically to - it is important not to confues users)

Yes.

> And here comes the main thing:
> 
> For some special drives/devices (defined in MINT.CNF for example)

Or implied by the usage of some 3rd party .xfs driver?

> we could have other mounting point, ie:
> 
> u:/somewhere_else/c, u:/even_else/d,

Yes, no problem.

> u:/even/combinated/e

This would need a bit more complicated dynamic structure than it has now 
actually (we would need to handle the unifs folders) which I didn't want 
to touch if possible. Not that it would not be nice to have.

> So by defauld kernel would map all drives to u:/gemdos (and make 
> symlinks to u:/), and if user wants (defines in kernel) it kernel 
> would map it somewhere else...

Hmm, I wanted to have the U:\c -> U:\gemdos\c possibly even in the 
MINT.CNF. If there is no link for the corresponding U:\gemdos\d (yes I 
mean the gemdos\d one - not the U:\d) then it would be automaticaly 
created as U:\d -> U:\gemdos\d (not to confuse users that don't know 
about this new feature). Or we can leave it completely up to the user just 
like the /home -> ext2drive:\home and others... What do you say?

> Final note: changing mount point, mount/unmounti on the fly, 

Yes, the U:\ root would than be able to mount on-fly just by 
creating/renaming a symlink.

The point I had is the dynamic GEMDOS drive -> unifs mount mapping by 
introducing the the fact that access to the C:\ would behave like the 
access to U:\c\ and consequently redirected to the appropriate symlink 
target. Once again C:\ == U:\c\ in my proposal.

> remount as read-only or read-write, remount/mount with special user 
> rights (real needo for FAT -style filesystems) should be implemented 
> too.

This would be quite only a little effort as long as we stay in the U: root 
folder (no tree structure for the .xfs implementation). We can change the 
internal structure and implement the unifs subfolders, but I would propose 
to do it later.

> Why not add flexbile mounting points?

I don't see any difference between my proposal and your additions except 
the permissions and the "u:/even/combinated/e" case. Is there something 
I'm missing here?

regards

STan