Backwards compability with 1.2.1
A bit of redundancy. There is no need to actually collify, since we'r…
…e flattening anyway. Thanks to Alan for pointing this out.
No reason to use a let now.
Should be private.
Makes enough sense.
Fix a race condition.
Fixed a bug where flatten didn't flatten maps.
Block var too.
Don't bother calling future-call.
This is more of a 5.0 kind of thing
Clean up the sandbox* function. Switch around the argument order of t…
…he function returned by sandbox*. It was stupid before.
Get rid of old whitelist stuff that doesn't work right anymore and ma…
…ke more complex sandboxing possible.
Get rid of ns-init in favor of a :refer-clojure option.
Actually return the permissions.
Pass your own codesource if you want.
Better function for creating permissions.
I must be stoned.
This test is more impressive if it actually does what it is supposed to.
Unmap new-defs if they exceed max-defs as well.
Standard i/o fields blocked.
Add missing test file.
Block Java field access.
Handle security managers appropriately. Unset it after execution.
Fix issue where symbols corresponding to namespaces threw a resolutio…
Added in-ns to secure-tester.