Bruce Momjian <pgman@candle.pha.pa.us>
>
>Oh, this is bad news. I see what you are saying. In 6.5.*, we had
>pg_id, which was used to do this. We still have pg_id, but I assume the
>attempt was to remove reliance ont that in the new initdb.sh. Right
>Peter?
>
>If so, can you suggest a solution under Solaris for getting the user id
>value?
I wonder if pg_id was the result of a similar portability
discussion/problem some years ago ;-)
I'd vote for keeping pg_id, if then we are free'd from all the
portability issues. (getpwent() is portable?)
Keith