Re: [Hibernate] Middlegen and XDoclet marriage

[One O. <1ov...@mb...>]

We havn=27t actually had much discussion of middlegen on
this list (though it has come up a couple of times)=2E This
recent burst of activity is mainly in response to the
thinking Max has been doing recently and I=27m just kinda
following up on his ideas=2E

I=27d particularly like to know what Konstantin thinks
about this stuff=2C since he is the one currently working
on the Hibernate XDoclet module and perhaps he has the
best idea of what its capabilities will be in the near
future=2E

Now something that occurred to me today is that
*whichever* way we go=2C either Middlegen template or =

Codegenerator=2C it would be helpful for the generated
source to contain XDoclet tags=2E (Currently CodeGenerator
does NOT generate them=2E)

However=2C I am still currently in favor of simplifying
things to=3A


       =3E=3E=3E=3E=3E=3E=3EXDoclet=3E=3E=3E=3E=3E=3E    =3E=3E=3ESchemaE=
xport=3E=3E=3E=3E
java + =40tags             hbm=2Exml             SQL DDL
       =3C=3C=3CCodeGenerator=3C=3C=3C=3C    =3C=3C=3C=3C=3CMiddlegen=3C=3C=
=3C=3C=3C


The the only leg of the roundtripping that is not
fully reproduceable without user interaction is the
DDL -=3E hbm=2Exml (ie=2E Middlegen) step=2E

----- Original Message ----- =

From=3A =22Aslak Helles=F8y=22 =3Caslak=2Ehellesoy=40netcom=2Eno=3E
To=3A =3Chibernate-devel=40lists=2Esourceforge=2Enet=3E
Sent=3A Saturday=2C January 04=2C 2003 3=3A28 AM
Subject=3A =5BHibernate=5D Middlegen and XDoclet marriage


=3E Hi!
=3E =

=3E I=27m new on the Hibernate list=2C so I haven=27t followed your discu=
ssions
about
=3E XDoclet and Middlegen integration lately=2E -But I=27ve seen your Wik=
i=2E As
some
=3E of you probably know already=2C Gavin has developed a Hibernate plugi=
n for
=3E Middlegen that should work with the recent XDoclet module for Hiberna=
te
=3E (which is in XDoclet=27s CVS and the recent 1=2E2b2 release)=2E
=3E =

=3E
http=3A//sourceforge=2Enet/tracker/index=2Ephp=3Ffunc=3Ddetail=26aid=3D66=
0628=26group=5Fid=3D360
=3E 44=26atid=3D415992
=3E =

=3E This means we=27ll have=3A
=3E =

=3E RDBMS -- Middlegen --=3E POJOs with =40hibernate tags -- XDoclet --=3E=
 hbm=2Exml
=3E =

=3E (POJO =3D Plain Old Java Object =3D simple javabean style class)
=3E =

=3E So basically you can start with a database=2C do some configurations =
in
=3E Middlegen=27s GUI=2C and have the java code and hbm=2Exml files gener=
ated=2E
=3E =

=3E Gavin and I have been discussing a different approach though=3A
=3E =

=3E RDBMS -- Middlegen --=3E hbm=2Exml -- Hibernate=27s CodeGenerator --=3E=
 POJOs
=3E =

=3E Both approaches are a mix of bottom-up (you start with a database) +
either
=3E xml-=3Ejava or java-=3Exml=2E
=3E =

=3E According to Gavin=2C Hibernate=27s CodeGenerator is more mature than=

XDoclet=27s
=3E Hibernate module=2E However=2C his current Middlegen plugin advocates=
 the
=3E XDoclet approach (and not the CodeGenerator approach)=2E
=3E =

=3E So my question is=3A Where do we go from here=3F Which approach do =5F=
you=5F
prefer=3F
=3E (Considering that this Hibernate/XDoclet/Middlegen marriage is still =
quite
=3E young and untested)=2E
=3E =

=3E Cheers=2C
=3E Aslak (Middlegen and XDoclet committer)
=3E =

=3E =



---------------------------------------------------------------------
Never lose a fax again, receive faxes to your personal email account!
Visit http://www.mbox.com.au/fax