[core] draft-ietf-core-resource-directory is update atomic

[Jim Schaad]

I am going through and figuring out how to implement this specification and
I came up with a questions on how things work.

1. The document says that PATCH is supported. What about iPATCH?

2. Are PATCH and POST update operations atomic? Is a partial update
permitted or is that not allowed?

Jim

[peter van der Stok]

Thanks Jim,

I try to answer below. I hope I understand the meaning correctly.

Post by Jim Schaad
I am going through and figuring out how to implement this specification and
I came up with a questions on how things work.
1. The document says that PATCH is supported. What about iPATCH?

Good remark. The document should probably state that both are allowed
but iPATCH is preferred.

Post by Jim Schaad
2. Are PATCH and POST update operations atomic?

The spec says that registrations need to be idempotent, that is not
repeated for the other operations (but implied?).
The update changes the lt; you might argue that this is not an
idempotent operation as the effects at a given time changes after a
second invocation of the same update. Nevertheless, I would call it
idempotent.
Idempotency => atomic? (as meaning: all or nothing)

Post by Jim Schaad
Is a partial update
permitted or is that not allowed?

Partial updates are permitted for PATCH and also for POST under the
conditions specified in section 5.4.1
Or is your argument that POST should not be used for partial updates.

Post by Jim Schaad
Jim

Hope this helps.

peter

Post by Jim Schaad
_______________________________________________
core mailing list
https://www.ietf.org/mailman/listinfo/core

[Jim Schaad]

What I was looking at with atomic is along the lines of.

Half way through the update I find that an error will result. Do I throw
away all the of the updates that I have done so far?

Jim

-----Original Message-----
Sent: Monday, August 14, 2017 12:56 AM
Subject: Re: [core] draft-ietf-core-resource-directory is update atomic
Thanks Jim,
I try to answer below. I hope I understand the meaning correctly.

Post by Jim Schaad
I am going through and figuring out how to implement this
specification and I came up with a questions on how things work.
1. The document says that PATCH is supported. What about iPATCH?

Good remark. The document should probably state that both are allowed but
iPATCH is preferred.

Post by Jim Schaad
2. Are PATCH and POST update operations atomic?

The spec says that registrations need to be idempotent, that is not

repeated

for the other operations (but implied?).
The update changes the lt; you might argue that this is not an idempotent
operation as the effects at a given time changes after a second invocation

of

the same update. Nevertheless, I would call it idempotent.
Idempotency => atomic? (as meaning: all or nothing)

Post by Jim Schaad
Is a partial update
permitted or is that not allowed?

Partial updates are permitted for PATCH and also for POST under the
conditions specified in section 5.4.1 Or is your argument that POST should

not

be used for partial updates.

Post by Jim Schaad
Jim

Hope this helps.
peter

Post by Jim Schaad
_______________________________________________
core mailing list
https://www.ietf.org/mailman/listinfo/core