Good point! I was relying on the fact that our field names are unique for
now, but there's no guarantee it will stay that way. So yes, set "field" to
the full path, and ban "." from field names.

 But the path should go all the way to the root dictionary.  Eg, for a bad
mode,

   "field": "meta.cost-type.cost-mode"
   "value": "foo"

or for a bad IP address in an endpoint cost request,

   "field": "endpoints.srcs"
   "value": "ipv4:a.b.c"

- Wendy


From:  "Y. Richard Yang" <[email protected]>
Date:  Wed, October 9, 2013 16:56
To:  Wendy Roome <[email protected]>
Cc:  IETF ALTO <[email protected]>
Subject:  Re: [alto] New error format in draft 20

> If the invalid value is a subfield, the server must set "field" to the
> subfield name and "value" to the invalid subfield value, converting it to a
> string if needed. For example,  if the "cost-mode" subfield of the "cost-type"
> field is invalid mode "foo", the server should set "value" to "foo", and
> "field" to "cost-mode", not "cost-type".
> 

An example used in the discussions will say:

"field" : "cost-type.cost-mode"
"value" : "foo"

The reason for the "." subfield structure is to avoid future cases of
ambiguous subfield name, for example, two fields with different names, but
both have a subfield with the same name. We do not have any two subfields
with the same name in the current design, but future extensions might and
hence cause problem for the general structure. For the "." notation to be
valid, we need to avoid "." in field names. What do you think?


_______________________________________________
alto mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/alto

Reply via email to