Agreed, sounds like a plan. We should probably output a warning about
deprecated interface usage and it's consequences.
2017-06-08 15:49 GMT+03:00 Vladimir Ozerov :
> Alex,
>
> Sergi suggested to deprecate this interface and it is already deprecated.
> However, internally it is still used and in
Alex,
Sergi suggested to deprecate this interface and it is already deprecated.
However, internally it is still used and in particular we depend on it in
IGFS inetrnals. It will require some time to refactor IGFS.
Let's just throw an exception if both AffinityKeyMapper and affinity fields
mappings
Vladimir,
Your concerns sound reasonable to me. However, I have a feeling that there
was a specific reason why added CacheKeyConfiguration to the
IgniteConfiguration and not to the CacheConfiguration. Looks like indeed,
our original approach is not flexible enough anymore and I am +1 for the
sugge
Folks,
(Yakov, this is partially about your problem with DML)
I found one nasty problem with our "affinity key field" concept. Currently
it works as follows:
1) Affinity keys are stored in binary metadata
2) Metadata is populated when object get's serialized for the first time.
3) There is an exc
