I still owe mailing list a full update of my thoughts towards Jackson 3.0.
The gist is that changes accumulated will make it inevitable that
changes are breaking existing usage regarding `jackson-core` and
`jackson-databind` (in the sense that it will not be possible to make
majority of code that works against 2.9 to cleanly compile against
3.0), and this in turns
makes it necessary that new Java packages and Maven coordinates are to
be used for 3.0.
However. Before outlining full picture, a thought occurred: instead of
ending 2.x series with 2.9,
which has been my thinking so far, would there perhaps be benefit from
still doing 2.10, which
would focus NOT on feature additions (in fact, should limit feature
additions to minimum), but
rather would try to:
1. Apply such bug fixes that are slightly riskier than things that can
go in late 2.9.x patches
2. Where possible would "future proof" code so that even if packages
change, some new patterns may be used, to make 2.x -> 3.0 upgrade
little bit more incremental
- In particular, a subset of Builder patterns for JsonFactory and
ObjectMapper could probably supported.
I am not 100% confident that (2) would necessarily work out as well as
I hope, but I think there could be value in sort of prototyping new
construction approach: although Builder patterns are fully implemented
in `master` for 3.0.0, very few developers will be testing that
version.
But 2.10, if released, would be used by many, and as such preview of
new API would get actual testing, feedback, and something that could
improve initial 3.0.0 release.
Thoughts?
-+ Tatu +-
--
You received this message because you are subscribed to the Google Groups
"jackson-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to jackson-user+unsubscr...@googlegroups.com.
To post to this group, send email to jackson-user@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
