On 7/28/2007 11:43 PM, Phil Thompson wrote:
The leak that your test case demonstrated had nothing to do
with /Transfer/ or
sipTransferObj. If you still have a leak then I need another test
case.
You mentioned that you fixed a memory leak in SIP with mapped types
and /Out/ annotations. This is one of the two leaks in my testcase (when
calling the function with the /Out/ annotation). But in the other leak
in my testcase, there is no /Out/ annotation, only a return value.
Will that be fixed as well with tonight snapshot?
Sorry, I misread your email. Both tests now use the same amount of memory.
The other leak is down to you using sipConvertFromInstance() instead of
sipConvertFromNewInstance(). I'd forgotten that a transfer object of NULL
means different things to the two functions.
OK that's fine, I verified that the original problems are fixed with
latest snapshot and that fix.
I offer stl.sip for inclusion in SIP. It covers my own needs right now
(vectors) but shall I need to wrap more STL containers (sets, maps,
ecc.) I would updated it with support for them. I guess that it might be
useful for people wrapping C++ code with SIP.
There is a high degree for redundancy in it, as you can see. You might
want to think about it as a way to improve SIP itself.
--
Giovanni Bajo
_______________________________________________
PyQt mailing list [email protected]
http://www.riverbankcomputing.com/mailman/listinfo/pyqt
