On 09/10/2013 11:06 AM, Holger Hans Peter Freyther wrote:
On Tue, Sep 10, 2013 at 09:07:32AM +0200, Jacob Erlbeck wrote:
/* return parent node */ -/* MUST eventually converge on CONFIG_NODE */ +/*
- MUST eventually converge either on CONFIG_NODE for every config node or
- on CONFIG_ENABLE for every other user defined node.
- */
The comment is wrong. There is no CONFIG_ENABLE node but the
You're right, it's ENABLE_NODE of course.
comment sounds dangerous too. The ENABLE_NODE might be password protected and I would like to avoid a situation where we come from a 'child' of the VIEW_NODE and end in 'ENABLE_NODE'.
The implementation cares about that (at least for the base nodes). It is not checked whether the go_parent callback does this and there isn't a way yet, to distinguish a view node from an enable node (since there is only is_config_child()/is_config_node() ).
The comment isn't quite explicit about that, but 'user defined node' was meant to refer to node (id's) above CONFIG_NODE, and there aren't any of these nodes that are neither config nor enable (yet).
OTOH, since that doesn't seem to be clear enough I'm going to reword it.
Jacob