Re: [PORTS] alpha/64bit & mkoidname problem (fwd)

>From the calls to postgres in initdb, correct? Has anyone come up
with a decent way of wrapping gdb around postgres in initdb?

Here's what I've got so far -- A file called doit.gdb with the contents:

break DefineIndex
run -boot -C -F -D/usr/local/pgsql/data -d template1

and I've been experimenting with things like this:

cat $TEMPLATE \
| sed -e "s/postgres PGUID/$POSTGRES_SUPERUSERNAME $POSTGRES_SUPERUID/" \
-e "s/NAMEDATALEN/$NAMEDATALEN/g" \
-e "s/OIDNAMELEN/$OIDNAMELEN/g" \
-e "s/PGUID/$POSTGRES_SUPERUID/" \
> /tmp/fifo &
echo foo
gdb -batch -tty /tmp/fifo -x /tmp/doit.gdb postgres

the above certainly doesn't work, and I've gotten it more functional
than this, but never actually functional. If I had better knowledge
of named pipes that would certainly help.

It doesn't look like I'll be able to take care of this by release date
(today, tomorrow)..

Does this mean that alpha is not supported?

On Thu, 26 February 1998, at 11:36:36, Vadim B. Mikheev wrote:

> Try to set break point inside defind.c:DefineIndex() and continue with 's'
> after got there... This is also good to check that args of DefineIndex()
> are Ok - look @ bootparse.y how they get values...
>
> You have to find place where backend tries to get OID of index function name
> (mkoidname).
>
> Vadim