Followup to bug 326506 -- this comment got lost somehow.

git-svn-id: svn://10.0.0.236/trunk@193402 18797224-902f-48f8-a5cc-f745e15eee43
This commit is contained in:
bzbarsky%mit.edu
2006-04-02 22:00:08 +00:00
parent 9ff3edeea6
commit f8006936fc

View File

@@ -178,18 +178,26 @@ interface nsIPrincipal : nsISerializable
* principal. Principals are equal if they are the same object, they
* have the same origin, or they have the same certificate ID.
*
* Thus a principal subsumes itself if it is equal to itself.
* Thus a principal always subsumes itself.
*
* The system principal subsumes itself and all other principals except
* the non-principal.
* The system principal subsumes itself and all other principals.
*
* The non-principal is not equal to itself or any other principal, and
* therefore does not subsume itself.
* A null principal (corresponding to an unknown, hence assumed minimally
* privileged, security context) is not equal to any other principal
* (including other null principals), and therefore does not subsume
* anything but itself.
*
* Both codebase and certificate principals are subsumed by the system
* principal, but no codebase or certificate principal yet subsumes any
* other codebase or certificate principal. This may change in a future
* release; note that nsIPrincipal is unfrozen, not slated to be frozen.
*
* XXXbz except see bug 147145!
*
* Note for the future: Perhaps we should consider a certificate principal
* for a given URI subsuming a codebase principal for the same URI? Not
* sure what the immediate benefit would be, but I think the setup could
* make some code (e.g. MaybeDowngradeToCodebase) clearer.
*/
[noscript] boolean subsumes(in nsIPrincipal other);