No subject

because cstdint will become standard c++ (it is in TR1).

Tom


On Fri, Dec 4, 2009 at 18:03, Bradley Lowekamp <blowekamp at mail.nih.gov> wro=
te:
> Hello all again,
> There are a couple of bug on this topic:
> http://www.itk.org/Bug/view.php?id=3D9426
> http://www.itk.org/Bug/view.php?id=3D1538
> There are significant problem with the definitions in following file:
> http://public.kitware.com/cgi-bin/viewcvs.cgi/Code/Common/itkIntTypes.h?r=
evision=3D1.7&root=3DInsight&view=3Dmarkup
> ITK needs portable fixed with integer types. The complication is that
> several systems do not have a usable 64-bit integer, so addition
> preprocessor conditionals are required. This is not a new problem nor is =
it
> unsolved, but it's a challenge to get a solution working correct everywhe=
re
> ( especially with apple's universal builds ).
> There are currently 3 implementation of this which ITK could use:
> 1) Utilities/vxl/core/vxl_confix.h.in
> which defines types like: vxl_int_8, vxl_sint_32 etc.
> 2) Utilities/kwsys/FundamentalType.h.in
> which defines type integer types
> like:=A0kwsysFundamentalType_UInt32,=A0kwsysFundamentalType_Int32
> 3) C99 stdint.h with a modified version of pstdint.h when needed
> =A0http://en.wikipedia.org/wiki/Stdint.h
> In the downloads notes, it lists the compilers pstdint.h was tested with =
(
> we have more though ).
> VXL already has a system to test for well behave header files and replace
> them on a system by system basis if needed, the use of pstdint is very
> similar.
>
> So the issue remains:
> What do we do with erroneous itkIntTypes.h?
> The types are not defined in a very itk like way; they are in all caps, i=
n
> the global namespace, and the odd use of extern "C" for typedefs. The onl=
y
> place these are used are in the=A0itkMersenneTwisterRandomVariateGenerato=
r.h.
> Also itkMathDetail uses vnl_int64.
> My though is to deprecate the types in itkIntType.h, so that it prints a
> warning if it's include indicating the types were error prone, and there =
are
> new type which should be used instead.
> Then two define new itk types based on 1 or 2.
> Alternatives =A0for new types to use include:
> Just use the C99 types, but I'd estimate that would be more work to get t=
hat
> to work on every compiler.
> Just use vxl types.
> Thoughts and comments?
> Brad
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D
>
> Bradley Lowekamp
>
> Lockheed Martin=A0Contractor for
>
> Office of High Performance Computing and Communications
>
> National Library of Medicine
>
> blowekamp at mail.nih.gov
>
>
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://kitware.com/products/protraining.html
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-developers
>
>