From: William S Fulton <wsf@fu...> - 2009-05-27 18:11:50
David Fletcher wrote:
>>>>>> "ob" == Olly Betts writes:
> df> Sure, I can send you the code. I'm not sure everyone wants this, and
> df> the swig-devel mailing list scrubs attachments, too, so perhaps I
> df> should send them to you directly? If you'd like to contact me
> df> directly that'd be great.
> s> Better to put it in the tracker on sourceforge so other people can
> s> look at it too, and there's a public record for the future.
> I can put this into the tracker. Some questions:
> 1. There's a generic intrusive_ptr.i file and, for each supported
> language, a file named boost_intrusive_ptr.i. Right now, there's
> java and python support, and I'm adding perl and PHP. Would you
> like all of these added as a unit? Or would you prefer that I
> break these into language-specific chunks?
One chunk should be fine.
> 2. The best way to test this is to get a copy of libkml, which is a
> Google project. Would you like a snapshot of libkml added to the
> tracker, too? Or, would you rather just have a link to the libkml
In order to test and keep this support working, we need tests that we
can check into our repository. We aren't interested in having a
dependency on libkml, so these tests need to fit into our standard
testing framework. The good news is we already have an extensive
intrusive_ptr test - Examples/test-suite/li_boost_intrusive_ptr.i as
well as a pretty demanding runtime test for the only currently supported
language (Java) - li_boost_intrusive_ptr_runme.java. Porting the runtime
test to the other languages should be fairly easy. It is also a
necessity for me to consider committing the intrusive_ptr support into svn.
Your contributions here will be welcome as I think intrusive_ptr support
is quite important given that it is likely to be in the C++0x standard.