[ 
https://issues.apache.org/jira/browse/WW-4939?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16494744#comment-16494744
 ] 

Yasser Zamani commented on WW-4939:
-----------------------------------

Do you also have any reference about when and why such practice (random 
constants) is good?

I think to make debugging Struts itself more easier, we can add that random as 
post-fix to current constants which helps debugger when watch the variable 
about it's meaning:

 
{code:java}
    public static final SecureRandom RANDOM = new SecureRandom();

    public static final String ACTION_MAPPING = 
generateUUID("struts.actionMapping");

    public static String generateUUID(String prefix) {
        return prefix + new BigInteger(165, RANDOM).toString(36).toUpperCase();
    }
 {code}

> Use securely generated constants
> --------------------------------
>
>                 Key: WW-4939
>                 URL: https://issues.apache.org/jira/browse/WW-4939
>             Project: Struts 2
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Lukasz Lenart
>            Priority: Critical
>             Fix For: 2.6
>
>
> Right now all the constants are well know and can be used in exploits, ie. 
> {{public static final String ACTION_MAPPING = "struts.actionMapping";}}
> Instead of using string literals we should generate random strings at runtime 
> to avoid using literals directly in exploits. Users can still use the 
> constants in their code but not in dynamic expressions.
> {code:java}
>     public static final String AUTH_TOKEN = generateUUID();
>     public static String generateUUID() {
>         return new BigInteger(165, RANDOM).toString(36).toUpperCase();
>     }
> {code}
> This will probably break backward compatibility but using string literals 
> instead of the constants by the users is a bad practice anyway.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to