You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
So far I haven't bothered managing changes on a per kettle version basis. There
was always a kind of unspoken ideal to make cookbook work regardless of the
kettle version.
However, there have been a few changes that don't add functionality and only
make cookbook work with a newer kettle version. Some of these might be
backwards compatible, some might not. I don't really know at this point.
This has to change. There are alternative directions:
1) cookbook gets tied to the kettle version and users have to make sure they
match the correct cookbook version with the kettle version they're running.
2) coobook comes with some logic that detects the current kettle version and
somehow applies that at runtime.
Obviously option 1 is easiest from a development perspective, option 2 is
easier for the user. I prefer to go with 2 unless that proves to be very
difficult.
A simple way to implement 2) would be to have one master job that is made to
work on any kettle version, which references sub jobs and transformations
dynamically. We can then create sub jobs and transformations prefixed by a
kettle version number and select those at runtime.
This could either be very simple - say one set of jobs and transformations per
kettle major version - or very complex - say a discovery process that finds the
closest matching jobs and transformations based on the full version number.
I would very much like feedback on this idea. It would also be helpful to know
which versions of kettle are currently being used by the community, and if
possible which revision of cookbook people are using since I frankly have no
idea which cookbook revision works with which kettle version.
Original issue reported on code.google.com by roland.bouman on 8 Sep 2014 at 11:59
The text was updated successfully, but these errors were encountered:
Original issue reported on code.google.com by
roland.bouman
on 8 Sep 2014 at 11:59The text was updated successfully, but these errors were encountered: