> Thanks so much. Now index problems have gone from my LinuxPPC box!
Whoo hoo! Thanks Vadim! Indices are much happier on my Linux box also :)
> Only remaining problem seems somewhat related to view. Still the
> select_views test and "select * from pg_user" produces core dumps.
Sadly, that is still true also. As I mentioned earlier, the actual
segfault happens when memory is free'd toward the end of a query. But of
course there is some problem earlier when the memory was allocated...
This is probably unrelated, but is it expected that the system views
have rule names prefixed with "_ret" whereas a view I create myself has
a name prefixed with "_RET"? Should initdb surround some of these names
in double quotes while it is doing its brute-force shadowing?
- Tom