To debug memory corruption/leaks you have a couple of options. 1. libumem. You can link the library to libumem and take that route. 2. watchmalloc(3MALLOC). You can use LD_PRELOAD to use this allocator and see if anything shows up.
-Sanjay Clay Baenziger wrote: > Hi all, > I've written a Python<->C bridge for libaiscf. It provides a few > ways to interact with SMF. I'm looking for some folks to look the code > over for memory management issues, for any potential SMF issues or any > other issues others might see. This code has not be whacked to be > c-style happy, so don't worry about that unless you really want. > I am having issues with the delete section of > AIservice_set_subscript(), if anyone has any suggestions what might be > happening (dbx stack trace included). This is triggered when one tries > to delete a property via: > del(libaiscf.AIservice(libaiscf.AISCF(),"test")['status']) > > Thank you, > Clay > > Webrev is at: > http://cr.opensolaris.org/~clayb/10740/webrev_prelim1/ > > How the module works: > --------------------- > To load the module, one runs > import libaiscf > > Then to create an SMF instance object, one can run > (for svc:/system/install/server:default) > instance=libaiscf.AISCF() > (for svc:/system/install/server:someThingElse) > instance=libaiscf.AISCF("someThingElse") > > Further, to create an AI service object, one runs: > service=libaiscf.AIservice(instance,"serviceName") > > To see a property key under the service, one can do: > service['status'] (which returns the string value or a KeyError as a > dictionary would if the key doesn't exist) > To set or change a property under the service, one can do: > serivce['status']="off" > > All actions query SMF and no data is cached in case something changes > under the consumer. All actions are handed to SMF when executed. Other > functions implemented can be read in the PyDoc output for the module, > here attached. > ------------------------------------------------------------------------ > > _______________________________________________ > caiman-discuss mailing list > caiman-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
