Re: python reference leaks
Brian Kelley <[email protected]>
| Newsgroups | gmane.comp.db.metakit |
|---|---|
| Message-ID | <[email protected]> |
This will be a good bug to quash. I will have some time to look at it next wee as well. I guess I am lucky in that my metakit connections are all persistent :) Could you add a del stor before the gc.collect and see if there is any difference? There shouldn't be, but I have had issues with not explicitly deleting storages. On a related note, I have also been thinking of rewriting the metakit interface using swig. Swig didn't handle C++ very well when Gordon McMillan wrote the original interface. Unfortunatly, Gordon is no longer supporting metakit or his interface code which will progressively make using metakit with newer versions of python difficult. Swig has caught up in leaps and bounds and even allows python classes to subclass C++ classes which is brilliant. This might be a largish project if we want to keep the interface as nicely pythonic as Gordon was able to do. Brian _____________________________________________ Metakit mailing list - [email protected] http://www.equi4.com/mailman/listinfo/metakit