On 16 November 2010 10:17, Alex Clark <[email protected]> wrote:
>
> I'm actually now interested in both approaches. While transmogrifier
> gets a bad rap for being "heavy" I think there are times when you do
> want to reach for something more than a "one off" script.

For the record, Transmogrifier is really quite lightweight. I think
this "bad rap" comes from people who have not really looked at what it
is or what it does and make a bunch of assumptions. It *does* require
that you spend an hour or so learning its core concepts. I'm not sure
that's an unreasonable expectation. If you understand dicts and
generators, then I think you're pretty much there. If you want to
write your own blueprints, you need to know how to register a utility.

In my own experience, once I'd understood those core concepts, writing
a migration process using both out-of-the-box Transmogrifier
blueprints from various packages and my own custom blueprints for
client-specific stuff took a lot less time, was a lot easier to test,
and a lot less error prone than custom scripts. I've written my fair
share of ad-hoc migrations.

Transmogrifier is built on the principle that every migration is
different. It is not a one-size-fits-all solution. Rather, it is a
nice abstraction that lets you focus on writing the bits that *are*
specific to your use case, and reuse well-tested, easy-to-understand
components for things like reading from a database, traversing
filesystem directories, constructing Plone content, dealing with
workflow states, etc - but only when they make sense.

Martin
_______________________________________________
Product-Developers mailing list
[email protected]
http://lists.plone.org/mailman/listinfo/product-developers

Reply via email to