This is an update. The good news, SDL doesn't need a manifest to work
with Python 2.6. It uses Python's. Why no one just tried using ctypes to
loading my VC 2008 test SDL.dll from Python 2.6 is a mystery. But it
worked. The bad news. I have yet to build all of Pygame's dependencies.
The SDL_ttf configure script tries using SDL.dll and fails. The good
news. I know how to get around this, but it is somewhat ugly, well,
mildly distasteful. Building the dependencies for Python 2.6 will not be
a one-step process; it will take two steps. I will adapt
msys_build_deps.py and add to the Python_2_6 branch. Then I will build
the dependencies. Someone else will have to try and build Pygame using
distutils and Visual Studio 2008 once the dependencies are available.
Lenard
Lenard Lindstrom wrote:
I am unsure what sort of bug report this would be. "Library built with
MinGW causes R6034 error when loading msvcr90.dll installed by Python
2.6 installer"? I have to confirm this first. I admit the problems I
had were with a test program that did not used Python. It may be that
a DLL loaded by Python will use Python's manifest. I should check. And
I am still waiting for a reply to a bug regarding MinGW and distutils.
Lenard
René Dudfield wrote:
I think if you submit a bug report it might get a little more attention.
On Tue, May 6, 2008 at 1:52 PM, Lenard Lindstrom <[EMAIL PROTECTED]>
wrote:
I am waiting for a reply on comp.lang.python as to what manifest is
needed
so SDL can use msvcr90.dll. How long until Pygame can be ported to
Python
2.6 on Windows depends on whether or not such a thing exists.
