Russell Yanofsky wrote:
>Branko ÄOibej wrote:
>
>
>>...
>>I'm really becoming rather tired of seeing this kind of logic in the
>>code -- _twice_ -- when it properly belongs in config files. The
>>generators are beginning to look like a perpetual hack.
>>
>>
>
>I'm well aware of this problem and I'm planning to do some refactoring to
>alleviate it. But I wanted to get the java and perl bindings building first
>to ensure I don't do it in a python specific way.
>
>
All right. I think this should be a major push in the generator at this
point.
>>>+#ifdef SWIGJAVA
>>>+/* Ignore these function pointer members because swig's string
>>>+ representations of their types approach the maximum path
>>>+ length on windows, causing swig to crash when it outputs
>>>+ java wrapper classes for them. */
>>>
>>>
>>So the comment about getting the bindings to build on Windows was
>>literal -- build, but not work. I wonder if it's all worth the
>>trouble, then?
>>
>>
>
>The struct members in question are function pointers that don't have have
>any thunks associated with them yet. So there is no loss of functionality.
>
>
So the SWIG Java bindings never really worked, on any platform?
-- Brane
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Jul 3 20:02:19 2004