Issue #11674 has been updated by Tomoyuki Chikanaga.

Backport changed from 2.0.0: UNKNOWN, 2.1: UNKNOWN, 2.2: UNKNOWN to 2.0.0: UNKNOWN, 2.1: UNKNOWN, 2.2: REQUIRED

----------------------------------------
Bug #11674: `local_variables` returns buffer-overflow garbage with methods with > 10 keyword arguments
https://bugs.ruby-lang.org/issues/11674#change-54823

* Author: KJ Tsanaktsidis
* Status: Closed
* Priority: Normal
* Assignee: Koichi Sasada
* ruby -v: ruby 2.2.3p173 (2015-08-18 revision 51636) [x86_64-darwin15]
* Backport: 2.0.0: UNKNOWN, 2.1: UNKNOWN, 2.2: REQUIRED
----------------------------------------
The following program appears to demonstrate a buffer overflow in `rb_f_local_variables`

~~~ruby
def with_kwargs_10(v1:, v2:, v3:, v4:, v5:, v6:, v7:, v8:, v9:, v10:)
    p local_variables
end

def with_kwargs_11(v1:, v2:, v3:, v4:, v5:, v6:, v7:, v8:, v9:, v10:, v11:)
    p local_variables
end

def with_kwargs_12(v1:, v2:, v3:, v4:, v5:, v6:, v7:, v8:, v9:, v10:, v11:, v12:)
    p local_variables
end

def with_args_11(v1,v2,v3,v4,v5,v6,v7,v8,v9,v10,v11)
    p local_variables
end

with_kwargs_10(v1:1,v2:2,v3:3,v4:4,v5:5,v6:6,v7:7,v8:8,v9:9,v10:10)
with_kwargs_11(v1:1,v2:2,v3:3,v4:4,v5:5,v6:6,v7:7,v8:8,v9:9,v10:10,v11:11)
with_kwargs_12(v1:1,v2:2,v3:3,v4:4,v5:5,v6:6,v7:7,v8:8,v9:9,v10:10,v11:11,v12:12)
with_args_11(1,2,3,4,5,6,7,8,9,10,11)
~~~

Output:

~~~
[:v1, :v2, :v3, :v4, :v5, :v6, :v7, :v8, :v9, :v10]
[:v1, :v2, :v3, :v4, :v5, :v6, :v7, :v8, :v9, :v10, :v11, :!]
[:v1, :v2, :v3, :v4, :v5, :v6, :v7, :v8, :v9, :v10, :v11, :v12, :"\""]
[:v1, :v2, :v3, :v4, :v5, :v6, :v7, :v8, :v9, :v10, :v11]
~~~

Expected output:

~~~
[:v1, :v2, :v3, :v4, :v5, :v6, :v7, :v8, :v9, :v10]
[:v1, :v2, :v3, :v4, :v5, :v6, :v7, :v8, :v9, :v10, :v11]
[:v1, :v2, :v3, :v4, :v5, :v6, :v7, :v8, :v9, :v10, :v11, :v12]
[:v1, :v2, :v3, :v4, :v5, :v6, :v7, :v8, :v9, :v10, :v11]
~~~

There appears to be a buffer overflow, because the symbol :"\"" is next in ASCII order to :!

I'm not familiar with the MRI interpreter internals; I spent a few hours trying to debug the problem but to no avail. It appears that in `vm_eval.c:2072`, `cfp->iseq->local_table_size` is 12 (in `with_kwargs_11`) even though there are only 11 kwargs and no other locals. However, that's as far as I got.



-- 
https://bugs.ruby-lang.org/