Looking at http://www.bnlearn.com/documentation/man/arcops.html, the way to drop arcs is to use the `drop.arcs`

function.

It doesn't seem to be dropping it though

```
> data(learning.test)
> res = gs(learning.test)
>
> arcs(res)
from to
[1,] "A" "B"
[2,] "A" "D"
[3,] "B" "A"
[4,] "B" "E"
[5,] "C" "D"
[6,] "F" "E"
> drop.arc(res, "A", "B")
> arcs(res)
from to
[1,] "A" "B"
[2,] "A" "D"
[3,] "B" "A"
[4,] "B" "E"
[5,] "C" "D"
[6,] "F" "E"
```

But in the debug logs it seems like it worked.

```
> drop.arc(res, "A","B",debug=T)
* dropping any arc between A and B .
> dropping any arc between A and B .
* (re)building cached information about node A.
* node A.
> found child D.
> found node C in markov blanket.
> node A has 0 parent(s), 1 child(ren), 1 neighbour(s) and 2 nodes in the markov blanket.
...
```

Is this a problem with `R3.4.3`

? How to use arc operations while checking for cycles? I could always replace the `res$arcs`

table, but I want to be able to check for cycles.

```
> version
_
platform x86_64-apple-darwin15.6.0
arch x86_64
os darwin15.6.0
system x86_64, darwin15.6.0
status
major 3
minor 4.3
year 2017
month 11
day 30
svn rev 73796
language R
version.string R version 3.4.3 (2017-11-30)
nickname Kite-Eating Tree
> packageVersion('bnlearn')
[1] '4.3'
```