Just discovered a way to leak an aggregated counter.
step 1:
open two shells connect to the same node: shell1 and shell2 and execute the following in shell1 and shell2 in order:
gproc:reg({a,l,test}, undefined).
step 2:
In shell1 eval an expression that crashes: 1 / 0.
Now we have a dangling {a,l,test} in gproc.
> gproc:lookup_value({a,l,test}).
0
> gproc:unreg({a,l,test}).
** exception error: bad argument
in function gproc:unreg/1
called as gproc:unreg({a,l,test})
Just discovered a way to leak an aggregated counter.
step 1:
open two shells connect to the same node:
shell1
andshell2
and execute the following in shell1 and shell2 in order:step 2:
In shell1 eval an expression that crashes: 1 / 0.
Now we have a dangling
{a,l,test}
in gproc.