#13 Segmentation fault with certain state

closed-fixed
5
2002-09-11
2002-09-10
No

I am toying around with A* algorithms, and tried to
speed things up by using psyco for the state graph
search loop. This works great, and speeds things up by
some 33%.

However, when a certain starting state is chosen, or
the puzzle randomizer picks such a state, I get a
segfault. I updated to latest CVS (2002-09-09), and
still see the segfault.

Attached is my NxN-puzzle solver, with psyco enabled
and set to a starting state known to cause a segfault.
This package requires the use of PQueue, downloadable
from
http://www.csse.monash.edu.au/hons/projects/1999/Andrew.Snare/#pqueue.

Disable psyco (comment out line 163), or let the
program create a random puzzle (comment out lines 195
and 196), and the algorithm completes.Running out of
time right now, so I don't have a deeper analysis for
you, sorry!

Discussion

  • Armin Rigo

    Armin Rigo - 2002-09-11
    • assigned_to: nobody --> arigo
    • status: open --> closed-fixed
     
  • Armin Rigo

    Armin Rigo - 2002-09-11

    Logged In: YES
    user_id=4771

    one more fix for the infamous psyco_unify() function :-)

     

Log in to post a comment.