Re: [Modeling-users] bug #839231: Models and import order

[Sebastien B. <sbi...@us...>]

Sorry, I posted too quickly: the 1st patch solved the circular-import
problem, but broke the generated code.

The 2nd patch works as expected --at least it now passes all tests for
python code generation (tests/test_generate_python_code.sh). You'll find
it at:
https://sf.net/tracker/index.php?func=3Ddetail&aid=3D839231&group_id=3D5893=
5&atid=3D489335

If you already applied the first patch, revert it then apply the 2nd
one.

-- S=E9bastien.

Sebastien Bigaret <sbi...@us...> wrote:
> Mario Ruggier <ma...@ru...> wrote:
> > Hello Seb,
> > > Can you confirm that you used the '-B'/--base generation scheme? The
> > > -C/compact scheme does not suffer from this as far as I can see.
> >=20
> > Yes, i use the --base scheme. Sorry I forgot to mention it before, as f=
or
> > me this is hidden inside a little custom script that i use to generate
> > the package.
>=20
> Ok.
>=20
> > OK. I have filed this as bug 839231.
>=20
>   Thanks. I've just attached a patch to this bug report; you'll need to
> patch your local copy and reinstall the framework (templates need to be
> recompiled w/ cheetah-compile). Hopefully this solves the problem --at
> least that's what I see. I still need to redesign some tests so that
> this can be detected in the future, then I'll probably add it to the cvs
> trunk if you confirm that this solves the problem on your side as well,
> whatever order you use in the import-sequence.
>=20
>   Thanks for reporting. FYI this was due to circular imports induced by
>   useless imports in generated MDL/__init__.py. These imports are now
>   done, when necessary, in the dedicated .py modules.
>=20
> -- S=E9bastien.