Hi devs, I'm finishing the generic script macro (based on jsr 223) and the groovy macro based on it. It's working pretty well but I have a last issue: what do we do about security ? I can't commit as it is because, like the "old" groovy macros, you can do almost everything so we have to protected the script macros from "simple" users. Now how do we do that ? 1) The quicker is to use the programming right with some bridge like the old groovy macro but it would be better to get rid of this special right in the new architecture. 2) we could also use the JAVA security policy to run the scripts in a "sandbox" but I think it means that the XWiki platform called by a scripts is also running in the same limited sandbox (?). 3) groovy has it's own support of JAVA security policy (see http://groovy.codehaus.org/Security) which give us the best control on it but it means it's only groovy 2) and 3) means also for us to define the exhaustive list of allowed or forbidden JVM api. I'm -1 for 1), I'm not sure for the details of 2) and I would prefer to be as generic as possible so i don't like 3) very much. I did not looked deeply on 2) and 3) yet, it's just a first mail to know if someone already played with theses concepts in JAVA. So WDYT ? Any other ideas ? -- Thomas Mortagne