Re: [PATCHES] Users/Groups -> Roles

[Fabien COELHO]

Dear Stephen,


 Attached please find files and patches associated with moving from the
 User/Group system currently in place to Roles, as discussed
 previously.  The files are:

 pg_authid.h
   New system table, contains role information
   To be placed in src/include/catalog, replacing pg_shadow.h

 pg_auth_members.h
   New system table, contains role/membership information
   To be placed in src/include/catalog, replacing pg_group.h


I've looked very quickly at the patch. ISTM that the proposed patch is a 
reworking of the user/group stuff, which are both unified for a new "role" 
concept where a user is a kind of role and a role can be a member of 
another role. Well, why not.


Some added files seems not to be provided in the patch :

sh> bzgrep pg_authid.h ./role_2005062701.ctx.patch.bz2
? src/include/catalog/pg_authid.h
!   pg_namespace.h pg_conversion.h pg_database.h pg_authid.h 
pg_auth_members.h \
! #include "catalog/pg_authid.h"
! #include "catalog/pg_authid.h"
! #include "catalog/pg_authid.h"
! #include "catalog/pg_authid.h"
! #include "catalog/pg_authid.h"
! #include "catalog/pg_authid.h"
+ #include "catalog/pg_authid.h"
+ #include "catalog/pg_authid.h"
! #include "catalog/pg_authid.h"
! #include "catalog/pg_authid.h"
! #include "catalog/pg_authid.h"
! #include "catalog/pg_authid.h"

Or maybe I missed something, but I could not find the pg_authid file?
the '?' line in the diff seems to suggest an unexpected file.

Anyway, from what I can see in the patch it seems that the roles are per 
cluster, and not per catalog. So this is not so conceptually different 
from user/group as already provided in pg.


What would have been much more interesting for me would be a per catalog 
role, so that rights could be administrated locally in each database. I'm 
not sure how to provide such a feature, AFAICS the current version does 
not give me new abilities wrt right management.


Have a nice day,

--
Fabien.

---(end of broadcast)---
TIP 7: don't forget to increase your free space map settings

Re: [PATCHES] Users/Groups -> Roles

[Tom Lane]

Stephen Frost <[EMAIL PROTECTED]> writes:
>   Attached please find files and patches associated with moving from the
>   User/Group system currently in place to Roles, as discussed
>   previously.

I have cleaned this up a bit and committed it.  I normally wouldn't
commit an incomplete patch, but this change is blocking Alvaro's work
on dependencies for shared objects, so I felt it was best to get the
catalog changes in now.  That will let Alvaro work on dependencies
while I sort out the unfinished bits of roles, which I intend to do
over the next day or so.

Many thanks for your work on this!

regards, tom lane

---(end of broadcast)---
TIP 6: Have you searched our list archives?

   http://archives.postgresql.org

Re: [PATCHES] Users/Groups -> Roles

[Stephen Frost]

Greetings,
  
  (Sent this earlier, but afraid it may have gotten caught by the
   too-big bug, so I'm reposting without the files attached, they can 
   all be found at: http://kenobi.snowman.net/~sfrost/pg_role/ ; there 
   are also gzip and uncompressed versions of the unified / context 
   diffs there for those who don't care for bzip2)

  Attached please find files and patches associated with moving from the
  User/Group system currently in place to Roles, as discussed
  previously.  The files are:

  pg_authid.h
New system table, contains role information
To be placed in src/include/catalog, replacing pg_shadow.h

  pg_auth_members.h
New system table, contains role/membership information
To be placed in src/include/catalog, replacing pg_group.h

  role_2005062701.ctx.patch.bz2
Main patch, generated via cvs -z3 diff -c | bzip2
(gzip didn't quite get it under the 100K mark for the list)

  role_milestones
List of milestones associated with my work on Roles support
'*' indicates that the milestone has been met/completed
'?' indicates uncertainty about if something should be done
No marker indicates an item which needs to be done
Note: Documentation needs to be updated

  Again, my apologies for not getting this in sooner, it's been a little
  hectic around here of late.  I'm anxious to get feedback, comments,
  corrections, etc.

Thanks,

Stephen

---(end of broadcast)---
TIP 9: In versions below 8.0, the planner will ignore your desire to
   choose an index scan if your joining column's datatypes do not
   match