Bug introduced in 10.0.0 and fixed in 10.0.1
The bug is also present in 12.2.0 and resolved in 12.3.0 on Mac OS X x86 (64-bit).
If I evaluate, with a fresh kernel
Names["Global`*"]
I get
{"rhs", "z"}
I think this is not caused by any packages in $UserBaseDirectory
or $BaseDirectory
, because if I evaluate the same expression under V9.0.1, which has the same strings for these variables (directories), I get
Names["Global`*"]
{}
As expected.
I was wondering if others are getting the same result. rhs
and z
do not have values, but still I am a bit concerned about the safety of
Remove["Global`*"]
when certain functions may depend on the existence of these symbols in Global`. For example if I set
sys`f[Global`xxxx_] := Global`xxxx
and evaluate
Remove[Global`xxxx]
then the following looks a bit scary
sys`f // Definition
sys`f[Removed[xxxx]:_]:=Removed[xxxx]
sys`f
still works, but I still think it may lead to trouble (e.g. with Save
).
Kirma came up with the nice idea of testing this on the cloud. I get
Apply[And,
MemberQ[CloudEvaluate[Names["Global`*"]], #] & /@ {"z", "rhs"}]
True
But there are many other symbols in the Global`
context there. So, is this a bug?
{FOO, list, msgs, rhs, z}
$\endgroup${"rhs", "z"}
. $\endgroup${list, msgs, rhs, z}
and{"rhs", "z"}
. Win-7 64. $\endgroup$