[patch #6637] New hash table implementations

[Ben Pfaff]

Update of patch #6637 (project pspp):

                  Status:                    None => Done                   
             Open/Closed:                    Open => Closed                 

    _______________________________________________________

Follow-up Comment #11:

>I found that if I hack the test_moved function so as to declare i,
>j and cur (all the ints) as volatile, then it masks the 
>problem.

You're right, perhaps that is the best approach.

I found an even easier and more innocuous-looking workaround: declare
xnmalloc() with __attribute__((malloc)).

With that change, I pushed the hmap code to the repository, and so I'm
closing this bug.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/patch/?6637>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/