Re: [patch #6637] New hash table implementations

[John Darrington]

On Sat, Sep 27, 2008 at 06:07:52PM +0000, Ben Pfaff wrote:
     
     Follow-up Comment #2, patch #6637 (project pspp):
     
     >Check failed in move elements around in memory (random hash) test at
     tests/libpspp/hmap-test.c, line 300
     I am...puzzled.  This is on what platform?


I'm beginning to think this is a compiler bug.

The problem only manifests itself with gcc 4.3.1 (4.2 and 4.1 seem
ok),   and only when optimisation is enabled ( -O2).

When the problem occurs, I found, by inserting some simple
diagnostics, I find that this line at hmap-test.c:683

        values[i] = e[cur][i].data = i;

doesn't do what one would expect. Because 

 assert (e[cur][i].data == i); 

on the following line fails.

Wierd!


-- 
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://pgp.mit.edu or any PGP keyserver for public key.

signature.asc
Description: Digital signature