sublist limit

[flycoolman]

I agree with this MVP approach but I do not understand the sublist-limit.
Our server supports the 'depth' parameter but not a sublist limit.

grouping sublist-limit-param-grouping {
description
"This grouping may be used by protocol-specific YANG modules
to define a protocol-specific query parameter.";
leaf sublist-limit {
type union {
type uint32 {
range "1..max";
}
type enumeration {
enum "unbounded" {
description
"Indicates that the number of entries that may be
returned is unbounded.";
}
}
}
default "unbounded";
description
"The 'sublist-limit' parameter limits the number of entries
for descendent lists and leaf-lists.

     Any result-set that is limited includes, somewhere in
     its encoding, the metadata value 'remaining' to indicate
     the number entries not included in the result set.";
}

}

1. The NC and RC drafts do not provide any additional information for this
   parameter or the target list or leaf-list
   Maybe they do not need to say anything, but it is very unclear how the NC
   XPath or subtree filter, and RC target resource
   produce a single list or leaf-list for pagination. It is unclear what
   interoperable filters a client can use on all servers.

2. This sub-list limit applies to all descendant nodes. The server is
   supposed to keep a running count for all descendent lists and leaf-lists
   It seems like this filter would be hard to implement. IMO, this parameter
   should be optional.

3. leaf-lists do not have child nodes. The first description sentence means
   if the target list
   has descendant leaf-lists, they count towards the subentry limit. Where
   are these large leaf-lists that
   need so much special processing? IMO descendant leaf-lists should not be
   counted in the limit.

[billwuqin]

From Andy Bierman:
https://mailarchive.ietf.org/arch/msg/netconf/3xUYErnG9FHRnMVzpunZTKxDtpM/