At 01:04 16.04.00 +0800, James Henstridge wrote:
>> > Possibly there will be no patch necessary anymore with
>> > the 0.7er series, but only a special makefile, because
>> > the new auto generation fixes most of the issues releated
>> > to the M$ compiler as well ...
>>
>> Sounds good... Anyone interested in getting it up on gcc?
>
>The issues on the MS platform are not compiler dependent, but linker
>dependent. The problem is that on MS platforms, initialisation of static
>structures occurs before dynamic linking. This means that if a static
>structure on an MS platform references a global variable or function from
>a DLL, the linking will fail. You must fill in the members in the
>structures that come from DLLs from program code instead.
>
I'm not quite sure if this is all correct. Although the compiler
messages and my explanation seem to back up this interpretation,
another work around mentioned at the Python FAQ is compiling with
M$VC as c++. At least than theses problems would be cured.
But my patches would possibly become much bigger ...
>The new pygtk should take this into account, so it should compile with
>both MS and gcc compilers (as it is standard ANSI C).
>
A little off-topic: Why are there so few people using the nice
c++ capbilities in the GCC world?
Hans
-------- Hans "at" Breuer "dot" Org -----------
Tell me what you need, and I'll tell you how to
get along without it. -- Dilbert
-
To unsubscribe: echo "unsubscribe" | mail [EMAIL PROTECTED]
