Re:Codepad.org, a pastebin that executes code
by davidw on Wednesday March 05, 2008 @11:38AM
(#129749)
Afaik, the only (reasonably common) languages with good sandboxes are Tcl and Java. Like they said above, they're the ones building the sandbox, not the language.
[ Parent
| on HN ]
Re:Codepad.org, a pastebin that executes code
by LogicHoleFlaw on Wednesday March 05, 2008 @03:55PM
(#129992)
Lua has excellent support for building sandboxes for untrusted user code.
Re:Codepad.org, a pastebin that executes code
by davidw on Wednesday March 05, 2008 @04:54PM
(#130039)
Interesting - but what do you mean 'building' sandboxes?
Re:Codepad.org, a pastebin that executes code
by tokipin on Thursday March 06, 2008 @07:34AM
(#130497)
i'm not exactly sure what he's talking about, but there's 2 things Lua has here. one is that the language is functional, so you can replace, alter, or remove the dangerous functions. the other is that you can specify the variable 'environment' of a function with a table (Lua's hash.) so you can construct its environment in an 'opt-in' basis and know for sure everything you're allowing the function to do is safe. and of course these things would be meaningless if built-in functions didn't behave like user functions
Re:Codepad.org, a pastebin that executes code
by jamongkad on Thursday March 06, 2008 @10:57AM
(#130594)
Upmodded and done for Lua.
Re:Codepad.org, a pastebin that executes code
by immad on Wednesday March 05, 2008 @03:12PM
(#129947)
You can map down a few languages to JVM executable bytecode, can't you?
Re:Codepad.org, a pastebin that executes code
by aaroniba on Wednesday March 05, 2008 @08:08PM
(#130173)
In what way does Java provide sandboxing of untrusted code?
Re:Codepad.org, a pastebin that executes code
by wallflower on Wednesday March 05, 2008 @08:46PM
(#130202)
Untrusted Java web start applications and Java applets can't do anything that could harm the real operating system. We had to buy and keep on renewing an expensive code-signing cert so our app could save a local file.
Every once in a while, a security researcher claims to have gotten untrusted bytecode to run outside the boundaries of the sandbox.