On Tuesday 30 May 2006 17:41, Simon Cross wrote: > > I think you misread my example. The test script I attached > > creates a database object and immediately destroys it with > > o.destroySelf() so the net result should be zero because the > > object is removed from both the database and the sqlobject > > cache when it's destroyed. > > My bad. > > However, the reason the objects aren't being freed in your example > code is that you tell the garbage collector not to with > gc.set_debug(gc.DEBUG_LEAK) which also sets DEBUG_SAVEALL.
I know that. But that is just to highlight the problem and make it obvious. Without that it's hidden by the gc. It's just a test script to show the issue, not a part of any real application. > Note that I have personally used SQLObject 0.7.0 to insert millions of > objects into a MySQL database and had the garbage collector > automatically work for me, which is why I'm fairly confident that your > issue goes a bit deeper than a simple set of cyclic references. I have no deep issue really. Just a patch to clean up 2 cyclic references. I know very well that if I disable the leak debugging, the memory usage will stay still, but take my test script for what it really is: a simple script to highlight the issue, not an application with problems. So, are you suggesting that since the garbage collector does the job of finding them, we shouldn't bother to remove some cyclic references we know about and avoid some unnecessary garbage collecting? -- Dan ------------------------------------------------------- All the advantages of Linux Managed Hosting--Without the Cost and Risk! Fully trained technicians. The highest number of Red Hat certifications in the hosting industry. Fanatical Support. Click to learn more http://sel.as-us.falkag.net/sel?cmd=lnk&kid=107521&bid=248729&dat=121642 _______________________________________________ sqlobject-discuss mailing list sqlobject-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sqlobject-discuss
