[xsd-users] Multiple schemas with the same namespace, compiled into a library

[Tony McConnell]

Hi Boris,

Thank you again, for this.

Unfortunately, I need to build an XSD version for a 32 bit platform,
preferably based on the released 3.3.0 source code ... would you have a
patch available for this version to add the different plate capability, or
could you point me to a section of the code please?

I'm building my own XSD 3.3.0 deb package for use within our internal
development machine package repository.

Performing an ack-grep on the source, to try and find places where plates
of 0 might be used in polymorphism code results in:

polymorphism/same-type/test.cxx:102:  const
::xsd::cxx::tree::type_factory_plate< 0, char >
polymorphism/same-type/test.cxx:298:  const
::xsd::cxx::tree::std_ostream_plate< 0, char >
polymorphism/ostream/test.cxx:176:  const
::xsd::cxx::tree::type_factory_plate< 0, char >
polymorphism/ostream/test.cxx:632:  const
::xsd::cxx::tree::std_ostream_plate< 0, char >
polymorphism/comparison/test.cxx:184:  const
::xsd::cxx::tree::type_factory_plate< 0, char >
polymorphism/comparison/test.cxx:188:  const
::xsd::cxx::tree::comparison_plate< 0, char >
polymorphism/supermen.cxx:144:  const ::xsd::cxx::tree::type_factory_plate<
0, char >
polymorphism/supermen.cxx:785:  const
::xsd::cxx::tree::type_serializer_plate< 0, char >

Is this a good place to start? I can see lots other references to plate
templates with parameters of 0, and it looks likely that I could break
things very easily!


On Thu, Apr 11, 2013 at 4:06 PM, Boris Kolpackov <boris at codesynthesis.com>wrote:

> Hi Tony,
>
> Ok, I've implemented this feature and built an x86_64 GNU/Linux binary
> for you to try (I assumed that's what you are using, if not, let me know):
>
>
> http://codesynthesis.com/~boris/tmp/xsd/xsd-4.0.0.a12-x86_64-linux-gnu.tar.bz2
>
> The new option is --polymorphic-plate <num>. The idea is that you need
> assign a different plate (1, 2, 3, etc) to each of your schemas so that
> they each get their own set of polymorphic maps.
>
> Let me know how it goes.
>
> Boris
>