Digikam and read-only files

[rcuql3k02]

Hi all,

I notice that Digikam is able to delete files even if they're marked as read only on the file-system.  If a file is marked as read-only on disk, I don't want Digikam to modify it at all.  Is there any way to prevent this behavior?

I'm using Digikam 0.9.0.  I'm happy to submit a bug or wish if that would help.  But I see that bugs on two similar issues are marked as RESOLVED already:
http://bugs.kde.org/show_bug.cgi?id=137770
http://bugs.kde.org/show_bug.cgi?id=122374

Does anyone have any tips for me?

thanks!
alec


--------------------------------------
Protect yourself from spam,
use http://sneakemail.com
_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

[Gerhard Kulzer]

Am Friday 16 February 2007 schrieb [hidden email]:

> Hi all,
>
> I notice that Digikam is able to delete files even if they're marked as
> read only on the file-system.  If a file is marked as read-only on disk, I
> don't want Digikam to modify it at all.  Is there any way to prevent this
> behavior?
>
> I'm using Digikam 0.9.0.  I'm happy to submit a bug or wish if that would
> help.  But I see that bugs on two similar issues are marked as RESOLVED
> already: http://bugs.kde.org/show_bug.cgi?id=137770
> http://bugs.kde.org/show_bug.cgi?id=122374
>
> Does anyone have any tips for me?
>

This is impossible by nature of Linux!

Gerhard


--
Hakuna matata
http://www.gerhard.fr

_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

[rcuql3k02]

Hi Gerhard,

Thanks for your response.

>This is impossible by nature of Linux!

Are you able to expand on this for me?  I don't really understand why it's impossible.  (Sorry!)

What I would like to see is that Digikam checks before deleting a file (or moving a file) and if the file is marked "read-only" then it will not complete the delete/move operation.  Is that possible somehow?

thanks again!
alec.



--------------------------------------
Protect yourself from spam,
use http://sneakemail.com
_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

[Fabien-5]

[hidden email] wrote:
> Thanks for your response.
>> This is impossible by nature of Linux!
>
> Are you able to expand on this for me?  I don't really understand why
> it's impossible.  (Sorry!)

And you're right.
Having a file with read-only mode doesn't prevent it for being deleted,
as long as you have write access to the directory.
Bugs you mentionned were related to the modification of a read-only
file. There used to be bugs around that, but now they are solved.

> What I would like to see is that Digikam checks before deleting a
> file (or moving a file) and if the file is marked "read-only" then it
> will not complete the delete/move operation.  Is that possible
> somehow?

I don't think it should refuse the deletion of a file, but I guess it
could send a popup warning window about that...
Note : konqueror itself doesn't ask anything when moving or deleting a
read-only file.

You can add a new wish in bugs.kde.org if you want.

--
Fabien

_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

[Jakob "Østergaard"]

[hidden email] wrote:

> Hi Gerhard,
>
> Thanks for your response.
>
>> This is impossible by nature of Linux!
>
> Are you able to expand on this for me?  I don't really understand why it's impossible.  (Sorry!)
>
> What I would like to see is that Digikam checks before deleting a file (or moving a file) and if the file is marked "read-only" then it will not complete the delete/move operation.  Is that possible somehow?
>

The operating system kernel (Linux), not the application (DigiKam)
manages file system permissions.

Therefore, if permissions do not work as you expect, it is either your
understanding of the operating system kernel, or, the operating system
kernel, that are at fault. It cannot ever be the application.

In this case, it's not a bug in the Linux kernel either.

Please see:
  http://www.dartmouth.edu/~rc/help/faq/permissions.html

Here you fill find the following snippet, among many other good
explanations of how these things work:
"On a directory, write access means you can add or delete files"

So, in other words; if the directory is writable to you, you can change
the contents of the directory. Files are the contents of a directory, so
if you can modify the directory, you can delete the files in it.

If you are not satistifed with the default Linux kernel file system
permission model, there are other security models available. SELinux for
example, is a very flexible system that will allow you to have much
finer-grained security policies enforced in the system.

But it is not a DigiKam problem  :)

--
Best regards,
    Jakob Oestergaard
    [The SysOrb Team]

_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

[rcuql3k02]

Thanks to Fabien and Jakob for their informative responses, and for the reminder on how permissions work.  

Given that information, I'm happy with things as they are.  I won't submit a new wish.

thanks all!
a.


--------------------------------------
Protect yourself from spam,
use http://sneakemail.com
_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users