PSPP-BUG: [bug #12931] ONEWAY transposes output rows/columns

[Ben Pfaff]

URL:
  <http://savannah.gnu.org/bugs/?func=detailitem&item_id=12931>

                 Summary: ONEWAY transposes output rows/columns
                 Project: PSPP
            Submitted by: blp
            Submitted on: Mon 05/02/05 at 06:28
                Category: Output Driver
                Severity: 7 - Major
                  Status: None
             Assigned to: jmd
             Open/Closed: Open

    _______________________________________________________

Details:

The ONEWAY procedure assumes that hash tables happen to iterate in an
expected order.  If they don't, then output rows or columns are transposed. 
I changed the hash table implementation and now the tests fail.  I am pretty
sure that the hash table is correct and that the ONEWAY code needs fixing.

This comment in ONEWAY makes me assume it's a ONEWAY bug:

/* FIXME: Potential danger here.
   We're ASSUMING THE array is in the order corresponding to the 
   hash order. */

Perhaps you could use hsh_sort() or hsh_sort_copy() to help with the
solution.







    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?func=detailitem&item_id=12931>

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