Last modified: 2014-11-17 10:34:40 UTC

Wikimedia Bugzilla is closed!

Wikimedia has migrated from Bugzilla to Phabricator. Bug reports should be created and updated in Wikimedia Phabricator instead. Please create an account in Phabricator and add your Bugzilla email address to it.
Wikimedia Bugzilla is read-only. If you try to edit or create any bug report in Bugzilla you will be shown an intentional error message.
In order to access the Phabricator task corresponding to a Bugzilla report, just remove "static-" from its URL.
You could still run searches in Bugzilla or access your list of votes but bug reports will obviously not be up-to-date in Bugzilla.
Bug 14636 - Blocking/Permissions Rewrite
Blocking/Permissions Rewrite
Status: NEW
Product: MediaWiki
Classification: Unclassified
User blocking (Other open bugs)
All All
: Low enhancement with 5 votes (vote)
: ---
Assigned To: Nobody - You can work on this!
Depends on: 16447 16644 18807
Blocks: 674 4995 15821 25400 28530 22316
  Show dependency treegraph
Reported: 2008-06-24 23:51 UTC by Chad H.
Modified: 2014-11-17 10:34 UTC (History)
11 users (show)

See Also:
Web browser: ---
Mobile Platform: ---
Assignee Huggle Beta Tester: ---

Basic work (9.88 KB, patch)
2012-05-01 18:25 UTC, Khaled El Mansoury

Description Chad H. 2008-06-24 23:51:38 UTC
Currently, our blocking/permission system is rather ad-hoc. In looking at some ways it could use improvement, it is becoming increasingly clear that a re-do is in order. To do this, I have composed several ideas:

1) Blocking should be permission-based. If a user can upload, read, write, they should be able to be blocked from uploading but not writing. This allows for finer-grained blocks which allow users to still contribute yet be prohibited from doing something they're having problems with. The major hurdle is keeping current blocks in place (I would suggest making all current blocks a general "block from everything").

2) Checking permission should be a 1-stop-shop. With the aforementioned grained blocks, we'll see an easier integration with the general permissions check. User::isAllowed() will be able to check if a user is _really_ allowed to do something. This eliminates the usual 2-step of "check if blocked," then, "check if has permission." 

I'm keeping in mind that we'll need to preserve current abilities (rangeblocks, autoblocks, etc). Additionally, the Special:Blockip interface needs to remain as simple as possible. I'm planning to start a BlockingRewrite branch, but I'd like some input first. Thanks.
Comment 1 Andrew Garrett 2008-06-25 10:30:37 UTC
I was going to do this at some stage, but I'm unlikely to get around to it until the end of the year due to real-life commitments. If you do go ahead before me, here are a few things to consider:
1. All permissions checks can already done, in Title::getUserPermissionsErrors, which also returns the reason that the action can't be done.
2. Performance - we can't scan the block table per-pageview, so any checks need to be done based on an index. My idea was to have a separate table for conditions which cause a block to occur, and use an index on that.
3. It's possibly a good way to write per-page blocking (bug 870).

I'm sure there were others.
Comment 2 Andrew Garrett 2008-08-14 05:40:25 UTC
That's bug 674, by the way.
Comment 3 Platonides 2010-01-29 22:15:09 UTC
Also for the rewrite: determine a syntax to give the permissions per-namespace.
Comment 4 John Mark Vandenberg 2011-04-28 12:20:24 UTC
Bringing the current systems closer together would be great.  There are not many changes needed to implement the current block options as rights functionality.

One step is to
* add a 'can edit own user talk' permission
* remove/deprecate $wgBlockAllowsUTEdit

Then the existing system can be replicated with an autopromote group 'blocked' that uses APCOND_BLOCKED (r52083), and revokes 'edit' and grants 'editowntalk' for wikis with $wgBlockAllowsUTEdit.

That would mean the three rights currently rights under admin control ('createaccount', 'sendemail', 'editowntalk') are proper permissions and the storage details can be abstracted under the permissions API, which may continue to access them from ipblocks table, or uses a new user_rights_override table which would require the block logic is updated as well.
Comment 5 Khaled El Mansoury 2012-04-27 06:28:12 UTC
So if we work backwards, how would the new permissions be set via LocalSettings.php?
I thought of something like this:

$wgGroupPermissions['users']['createtalk'] = true; //As normal: Permission granted
$wgGroupPermissions['somegroup']['createtalk'] = false; //As normal: This group does not have the permission to do so, but the permission maybe granted from another group the user is present in, otherwise of course performing the above action is denied.
$wgGroupPermissions['somegroup2']['createtalk'] = -1; //If a user is present in this group, then regardless of whether they have permission granted for the above action from another group, they would always be denied to perform the above action.

This would of course get rid of $wgRevokePermissions.

Any thoughts on the above?
Comment 6 Khaled El Mansoury 2012-05-01 18:25:46 UTC
Created attachment 10491 [details]
Basic work

OK. As I briefly mentioned on IRC, here is my contribution on an attempt to rewrite the permission system.
I know that this is nowhere near a solution, but I see it as a start and just would like some feedback on the work, and maybe I could build from there.

The system works just like I described it in Comment 5, it builds on top of the usual $wgGroupPermissions, but gets rid of $wgRevokePermissions.

For the following example, if the user is in group 'noupload', no matter whether he already has permission to upload from another group or not, he would always be denied to upload.

$wgGroupPermissions['noupload']['upload'] = -1;

So, any feedback is really appreciated! Thanks.
Comment 7 Khaled El Mansoury 2012-05-05 23:01:08 UTC
Comment on attachment 10491 [details]
Basic work

Patch with several modifications submitted to Gerrit for review: (Change I0886845d)
Marking attached patch as obsolete.
Comment 8 Sumana Harihareswara 2012-05-07 01:44:41 UTC
Thanks for putting it in Git/Gerrit!

Note You need to log in before you can comment on or make changes to this bug.