search

netconf-wg / restconf

9 stars 4 forks source link

Gen-Art requested early review: draft-ietf-netconf-restconf-09 #46

Closed abierman closed 8 years ago

abierman commented 8 years ago

https://www.ietf.org/mail-archive/web/netconf/current/msg10889.html

abierman commented 8 years ago

PUT: clarify that create via PUT is a MAY not MUST

How to extend? Change protocol revision query parameters can be added in another document

3.4: datastore allows POST not just PATCH

fix table in 3.2 what methods allowed

change text so errors is not a resource type because there is no URL for GET

sec 2.2: add text that says http MUST NOT be used; could check on mailing list

sec 2.3: same issue in NETCONF over TLS; Juergen ended up w/ 1 sentence Kent to consider simplifying this section

3.1: show example of versioning; consider changing 'restconf' to 'restconf-1.0'

4.3: GET If-None-Match 27 will get back Not Modified even if the NACM rules changed so that the representation will be different returned to that client. The ETag (27) is for the resource but ignoring NACM filtering. The client using If-None-Match will not get the updated view after NACM rule changes look at 3.4.1.2 -- highlight issues

abierman commented 8 years ago

sec 2.5: session-based auth: Kent will look into it

sec 5: message vs. method: investigate rest of minor comments: Andy will investigate

abierman commented 8 years ago

ISSUE: PUT: clarify that create via PUT is a MAY not MUST CANNOT FIND ANYTHING IN HTTP RFC THAT AGREES WITH THIS STATEMENT NOT CHANGING THE TEXT -- CURRENTLY MANDATORY FOR SERVER TO ALLOW INSTANCE TO BE CREATED WITH PUT

abierman commented 8 years ago

SEC 2.3: NO REWRITE HAS BEEN DONE; LEAVE FOR KENT

abierman commented 8 years ago

SEC 3.1 now says If the XRD contains more than one link relation, then only the relation named "restconf" is relevant to this specification.

abierman commented 8 years ago

sec 3.1 NO EXAMPLE OF VERSIONING ADDED

abierman commented 8 years ago

OPEN ISSUE: ADDED WARNING TEST THAT NACM AND CACHING DO NOT WORK

Note that the way that access control is applied to data resources is completely incompatible with HTTP caching. The Last-Modified and ETag headers maintained for a data resource are not affected by changes to the access control rules for that data resource. It is possible for the representation of a data resource that is visible to a particular client can be changed without detection via the Last-Modified or ETag values.

abierman commented 8 years ago

There are MUST statements in 1.4 coexistence with NETCONF This intro section has not been changed to remove 2119 keywords

abierman commented 8 years ago

ADDED to 4.4.1: If the data resource already exists, then the POST request MUST fail and a "409 Conflict" status-line MUST be returned.

abierman commented 8 years ago

3.5.1: changed example -- looks worse; looks like the syntax requires client to repeat key leaf names; this is 100% wrong of course; so key1 as the value of key1 is a terrible example

abierman commented 8 years ago

Consider adding an example that shows a 500 response corresponding to a "partial-operation" error in NETCONF, demonstrating what information the client gets to distinguish this 500 from an "operation-failed" error.

NO -- NOT DONE

abierman commented 8 years ago

closed in draft-10