---- Reported by nengard@gmail.com 2008-07-30 07:50:34 ----

# management Set library management parameters  ** deprecated

on: cgi-bin/koha/members/member-flags.pl

[09:38]  <nengard> what is this permission for?  "management Set library management parameters"  when I set this permission it's the same as just having catalogue permissions
[09:39]  <nengard> do we have a list of what each of these do
[09:40]  <rch> nengard: this should control access to the bulk of the Administration fcns.
[09:42]  * ccatalfo has quit (Read error: 110 (Connection timed out))
[09:42]  <gmcharlt> nengard: but it's basically deprecated in favor of the parameters permission for dealing with the Adminstration form - doesn't control much directly 
[09:43]  <nengard> gmcharlt, it doesn't show the administration link ... should it? or should we have it removed as a permission?
[09:44]  <gmcharlt> nengard: ultimately, it should probably just be removed - parameters permission now covers things



---- Additional Comments From jmf@liblime.com 2008-08-08 05:31:33 ----

Good for 3.2, or whenever we get around to expanding permissions



---- Additional Comments From nengard@gmail.com 2009-04-18 14:04:50 ----

Just wanted to bump this up since we are working on permissions for 3.2 now.



---- Additional Comments From joe.atzberger@liblime.com 2009-05-28 17:44:29 ----

The only template using CAN_user_management is header.inc, as follows:

<!-- TMPL_IF NAME="IndependantBranches" -->
    <!-- TMPL_IF NAME="CAN_user_management" --> (<a href="/cgi-bin/koha/circ/selectbranchprinter.pl">Set</a>)<!-- /TMPL_IF -->
<!-- TMPL_ELSE --> 
    (<a href="/cgi-bin/koha/circ/selectbranchprinter.pl">Set</a>)
<!-- /TMPL_IF --> 

So basically it is just an exception case to IndependentBranches to put the link for choosing a new branch.  But that is bogus, because selectbranchprinter doesn't respect the management permission anyway.  Instead it looks for "circulate_remaining_permissions".

The ability to set a different branch for the session should actually be a subpermission of the general login permissions "catalogue" (Librarian interface), since it should be logically divorced from any circulate function.  



---- Additional Comments From joe.atzberger@liblime.com 2009-05-29 14:54:46 ----

So I now consider the main problem to be that IndepdendantBranches is essentially broken with respect to the way it "secures" the Set link.  

As a granular permission, catalogue=>setbranch would make sense, but that wouldn't fix it for non-granular Indy branches, since they would no longer be able to control the appearance of that link separately (everyone w/ "catalogue" would see it).  

In reality that is all they are controlling, the appearance of the link.  The user can still go to selectbranches.pl and set a new branch (they just don't see the link).  So that represents a security failure.  I'm upgrading the severity accordingly.   

I'm open to suggestions about the best way to fix it, but using a different top level permission (i.e. "management") cannot be it.  That would split the security model for the page, and therefore for the links to the page, as currently seen.

I think selectbranchprinter.pl and circulation.pl need to be refactored.  The branch-setting has to happen at selectbranchprinter and NOT be a post back to circulation.pl.  After that is successful, it can redirect to circulation (or HTTP_REFERER).



---- Additional Comments From gmcharlt@gmail.com 2009-05-29 19:07:58 ----

[13:13]  <atz> gmcharlt: would appreciate feedback on http://bugs.koha.org/cgi-bin/bugzilla3/show_bug.cgi?id=2426
[13:15]  <atz> basically the IndyBranches implementation for controlling "Set branch" is fractured 
[13:16]  <atz> and whatever fix I implement will reveal that... i.e. "yuo broke it!"
[13:29]  <gmcharlt> atz: some thoughts about it
[13:29]  <gmcharlt> 1. this is a good candidate for AJAX - no need to redirect at all
[13:30]  <atz> good thought
[13:30]  <gmcharlt> 2. catalogue->setbranch is OK as a granular permission
[13:30]  <gmcharlt> 3. indybranches is orthogonal to user permissions in this case, but that's OK for now
[13:31]  <gmcharlt> although potentially it could implement the restriction by overriding the value of a particular users setbranch permission
[13:31]  <atz> yeah, system groups will inevitably come into play here
[13:34]  <atz> for (3) are you saying it is OK to leave the manage permission in play for this weird indy branches exception case?
[13:35]  <atz> that would make the bug an INVALID
[13:35]  <gmcharlt> no, the permission should go away
[13:35]  <atz> so basically we'll be showing the link to indy branches users, even if they can't use it
[13:35]  <gmcharlt> and be replaced a test on the indybranch syspref and whether or not the user is a superlibrarian or has cataloge->changebranch
[13:36]  <atz> superlibrarian gets all user permissions already
[13:37]  <gmcharlt> true
[13:40]  <atz> gmcharlt: what do you think about a psuedo permission for setbranch?  i.e., granular permissions would populate it from catalogue:setbranch but it would still exist for non-granular
[13:40]  <atz> ?
[13:40]  <gmcharlt> sounds reasonable
[13:42]  <atz> i'll give it a try



--- Bug imported by chris@bigballofwax.co.nz 2010-05-21 00:51 UTC  ---

This bug was previously known as _bug_ 2426 at http://bugs.koha.org/cgi-bin/bugzilla3/show_bug.cgi?id=2426

Actual time not defined. Setting to 0.0
CC member bchurch@ptfs.com does not have an account here
CC member joe.atzberger@liblime.com does not have an account here