[GRASS5] v.in.ascii

Hamish hamish_nospam at yahoo.com
Thu Jun 23 21:07:00 EDT 2005


Right, valgrind is done for 'v.in.ascii -zt'

I used the first 250,000 points from the sample lidaratm2 dataset.

  http://bambi.otago.ac.nz/hamish/grass/memleak/v.in.ascii/


Three un-freed calls sum to 91% of the still reachable memory at program
termination.  (roughly 300 bytes lost per vector point)

RTreeNewNode (node.c:47)
dig_alloc_line (struct_alloc.c:112)
dig_alloc_node (struct_alloc.c:46)


A test following children processes is running now, but I think the 
above three are the major culprits. Will add these results to the page
later in the day. (but I'm running with -t so no DBF stuff)



Hamish




More information about the grass-dev mailing list