Pedro Roma wrote: > Hi Markus > The thing is, when I run this module with valgrind, I get memory leacks (not > much but I'm getting it) so it means that I should free it at the end. > To cite from a previous post, Glynn replying to me [1]:
start citation --> Explicitly freeing memory when termination is imminent is a waste of resources (e.g. CPU cycles, and I/O bandwidth in the case where free() causes swapped-out memory to be swapped back in). It's colloquially referred to as "rearranging the deckchairs on the Titanic". ... Valgrind highlights *potential* issues; it isn't omniscient. [Nor is it an angry god which must be appeased with a sacrifice of SSH keys.] <-- end citation As I understand, you still have to investigate potential issues if they are real issues. As I said before, I think in this case it's not an issue because memory is allocated only once and releasing this memory would be "rearranging the deckchairs on the Titanic". Markus M [1] http://lists.osgeo.org/pipermail/grass-dev/2009-April/043504.html _______________________________________________ grass-user mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/grass-user
