[
https://issues.apache.org/jira/browse/MYFACES-4021?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15041435#comment-15041435
]
Moritz Bechler commented on MYFACES-4021:
-----------------------------------------
Sorry for the delay. I was actually refering to the server side state saving
mechanism where only the view key is transferred to the client. The
serialization/deserialization to/from the session is not as bad, usually an
attacker cannot control that (and one may assume that things in the session
will be serialized/deserialized by something else anyway).
That view key is encoded as a string by KeyFactory, but if I didn't totally
miss something, serialized afterwards (HtmlResponseStateManager.getSavedState()
-> DefaultStateTokenProcessor.decode() -> StateUtils.reconstruct() ->
StateUtils.getAsObject()) and that I think might not be totally necessary. The
the deserialization on restore gives a vector for all the nice deserialization
attacks out there (if USE_ENCRYPTION=false).
> blacklist
> org.codehaus.groovy.runtime.,org.apache.commons.collections.functors.,org.apache.xalan
> in MyFacesObjectInputStream
> -----------------------------------------------------------------------------------------------------------------------------
>
> Key: MYFACES-4021
> URL: https://issues.apache.org/jira/browse/MYFACES-4021
> Project: MyFaces Core
> Issue Type: Bug
> Reporter: Romain Manni-Bucau
> Priority: Blocker
>
> https://github.com/apache/incubator-batchee/commit/cfd133c309c21a82fb24cfcc9a7c2365aee4678a#diff-acd0bc06477ce776b0ad8fdda76f8b7eR56
> mecanism can be used
> (due to recent vulnerability discovered in [collections], spring, groovy we
> can't suppose we don't run with these libraries so we need this fix as well)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
