Chris Travers wrote:
> I am going to make another suggestion here.  We should probably
> suggest that no reporting functions should run as security definer.
> This allows the same data to be accessed relationally as would be
> retrieved by the function.

I'd like to see the project go one step further and declare that no 
read-access functions should run as SECURITY DEFINER, only write-access 
functions.

I can't think of any real-world example I've ever run into where 
column-level permission granularity couldn't solve this.  (Well, except 
for really badly-designed databases that needed, effectively, row-level 
security!)

And there shouldn't be any functions that both read and write (I think).

--
-Adam Thompson
  <[email protected]>
  (204) 291-7950

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Ledger-smb-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ledger-smb-devel

Reply via email to