[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [leshan-dev] Plan for a 0.9 or 1.0 version of Leshan
|
Hi all,
As planed, we made a lot of changes those last 3 month. We think
it's time to release a 1.0.0M1. No objection :) ?
Except big surprise, most of the API changes was made. The main
missing one is the new QueueMode API which should be available later in
1.0.0M2 or 1.0.0RC1. Now the last effort to release a 1.0.0 will be
mainly focused on californium/scandium.
We strongly recommend to test those changes. This way if there is
any issue (missing API, bug...) we will be able to handle it.
Simon
Le 08/11/2016 à 16:11, Simon Bernard a écrit :
Hi all,
Having a release for Leshan could be great. We're still waiting for
a 1.0 version of the spec, but we identify several points we would
like to see for a first release of Leshan and we should focus on.
You could find the identified issues here :
https://github.com/eclipse/leshan/labels/1.0
This is not specially ordered by priority. (all seems really
essential except #198)
To release Leshan we will also need a Californium release. We
currently depends on the "persistent observationStore" which is in the
2.0 branch.
So if we want to release Leshan we also need to focus on a stable
Californium release. Currently there is a lot changes in Californium,
and I think we should identify a limited number of new feature we want
for this release.
Here is what it seems important to me:
1)This issue seems essentials :
https://github.com/eclipse/californium/issues/104.
2)As I explain here[1], I think that correlate request/response
by session/epoch is really too strong I would also be able to just
made the correlation by Identity in Leshan. Having a way to do that in
Californium by offering a way to implement your own correlator or by
adding a new correlation mode could be a good point.
3)(optional) ObservationStore currently identify an observation
just by token. We should probably use an additional information like
Identity or CorrelationContext to identify it.
This 3 issues focus on security and observe behind NAT.
Simon
[1]https://mailarchive.ietf.org/arch/search/?email_list=core&gbt=1&index=0vwuEnq6IRHJc70l3BTggX6p9vs
