okay, so let's add the following then to section 4, in the explanation of the "differences" output parameter:
"When a datastore node in the source of the comparison is not present in the target of the comparison, this can be indicated either as a "delete" or as a "remove" in the patch as there is no differentiation between those operations for the purposes of the
comparison. "
I will post this in a -06 shortly. Please let us know if this addresses your concerns or if there is anything else.
Hi Alex,
I think the only “problem” with using both “remove” and “delete” is that it could be confusing (when should one be used and not the other). Adding some text to say they’re the same for the diff operation is good enough
for me.
Regards,
Reshad.
Hi Reshad,
re: question 1: As you indicate, there may be no distinction between indicating a "remove" or a "delete" in the patch. Right now it would be acceptable to return either. If we want to eliminate this freedom, which one would you prefer be used? Shall we
remove the possibility for "delete" and just cover it using "remove"?
Note that the place where this is specified in the model is as part of a condition that specifies when the source value should be included. If we want to rule out that diff can return either "remove" or "delete" (indeed they are synonymous), we would need
to add text to the container description that when a data object is present in the target of the comparison but not the source, that "remove" should be used to indicate that.
The model would be changed follows. Please confirm if this looks good to you & we'll incorporate it.
OLD
container differences {
description
"The list of differences, encoded per RFC8072 with an
augmentation to include source values where
applicable.";
uses ypatch:yang-patch {
augment "yang-patch/edit" {
description
"Provide the value of the source of the patch,
respectively of the comparison, in addition to
the target value, where applicable.";
anydata source-value {
when "../operation = 'delete'"
+ "or ../operation = 'merge'"
+ "or ../operation = 'move'"
+ "or ../operation = 'replace'"
+ "or ../operation = 'remove'";
description
"The anydata 'value' is only used for 'delete',
'move', 'merge', 'replace', and 'remove'
operations.";
}
reference "RFC 8072: YANG Patch Media Type";
}
}
}
container differences {
description
"The list of differences, encoded per RFC8072 with an
augmentation to include source values where
applicable. Where a difference include a data object
in the target that is not present in the source,
this shall be indicated as a 'remove' operation
in the patch, not as a 'delete' operation.";
uses ypatch:yang-patch {
augment "yang-patch/edit" {
description
"Provide the value of the source of the patch,
respectively of the comparison, in addition to
the target value, where applicable.";
anydata source-value {
when "../operation = 'merge'"
+ "or ../operation = 'move'"
+ "or ../operation = 'replace'"
+ "or ../operation = 'remove'";
description
"The anydata 'value' is only used for 'merge',
'move','replace', and 'remove' operations.";
}
reference "RFC 8072: YANG Patch Media Type";
}
}
}
On 9/15/2020 4:04 PM, Reshad Rahman (rrahman) wrote:
Hi Alex,
I will review the latest version.
See below for questions/responses.
On 2020-09-15, 5:19 PM, "yang-doctors on behalf of Alexander L Clemm" <yang-doctors-bounces@xxxxxxxx on behalf of ludwig@xxxxxxxxx> wrote:
Hello Reshad, hello YANG Doctors,
thank you for your review! Please find my replies inline, <ALEX>. We
have also just posted -05 (thanks, Yingzhen, for doublechecking my
updates).
--- Alex on behalf of coauthors
On 9/7/2020 7:06 AM, Reshad Rahman (rrahman) wrote:
> <Here's the same message with hopefully more readable formatting>
>
> Review of rev -04 by Reshad Rahman
>
> The document is clear and well-written. While some issues have been identified, they can be resolved quickly.
>
<snip>
> Questions
> 1. YANG model: does the operation “delete” make sense for a diff operation? If it is kept, it’d be good to have some text explaining that for a diff operation, “delete” and “replace” are the same? If they’re not the same, please also add some text….
<RR> I actually meant "delete" and "remove".
<ALEX> Here we are simply referring to the basic YANG-patch edit
operations per https://tools.ietf.org/html/rfc8072#page-11. Those are
in turn derived from <edit-config> operations per
https://tools.ietf.org/html/rfc6241#page-37. I am not sure we need add
to explain those, as we are directly referring to YANG-patch.
</ALEX>
<RR> The operations are indeed well defined in RFC8072 (copied below), but they are defined from the perspective of YANG-Patch. So for YANG-Patch "delete" and "remove" are different operations, but from a diff comparison I believe they are the same (the resource must exist since it's being returned in a diff)
+-----------+-----------------------------------------------------------------+
| delete | delete a data resource if it already exists; if it |
| | does not exist, return an error |
| | |
| remove | remove a data resource if it already exists |
+-----------+-----------------------------------------------------------------+
> 3. YANG model P9, for the “uses path:yang-patch”, why not have a reference to RFC8072 (is it because the description above mentions RFC8072)?
<ALEX> We are clearly referencing RFC 8072; are you suggesting to put a
reference substatement below the uses statement? It looks a little
strange to me but sure, we will add it.
<RR> Not needed.
> 4. Section 7 mentions rate limiting requests per client. Should there be a “global” rate-limiting too, i.e not client-specific?
<ALEX> I am not sure this is really needed as I think the number of
management clients will in general be fairly limited to begin with, but
we can certainly add it. How about the following text:
OLD:
One possibility for an implementation to mitigate against such a
possibility is to limit the number of requests that is served to a
client in any one time interval, rejecting requests made at a higher
frequency than the implementation can reasonably sustain.
NEW:
One possibility for an implementation to mitigate against such a
possibility is to limit the number of requests that is served to a
client, or to any number of clients, in any one time interval, rejecting
requests made at a higher frequency than the implementation can
reasonably sustain.
<RR> Good with me.
</ALEX>
> 5. Wondering if section 8 should be in an Appendix (or even removed)? Also, the method suggested doesn’t seem to guarantee that the difference persisted for the “dampening” time.
<ALEX> Personally, I do think it makes sense to include a brief
discussion of possible further extensions. I suggest to keep the
section if it's okay with you, or perhaps leave it to the chair whether
they have a preference to remove it.
</ALEX>
<RR>Whatever the WG/chairs decide is fine with me.
Regards,
Reshad.
