> *Users* are starting up postmaster? *shrug* ISTM you're left with the
> following choices:
Users of the application that can start up the postmaster of course.
Some poeple dont like using the command line. :P
>
> 1. Work at it until you've figured-out just the right
> combination of shell escapes to get the path though pg_ctl's
> "cooking." The prior caveat applies: If pg_ctl is modified
> relative to that switch's argument handling, it'll break again.
>
> 2. Modify pg_ctl so the argument, in all its processing, is
> preserved with embedded spaces. Of course, the next release of
> pgsql would require re-doing it if pg_ctl is changed.
>
> 3. Talk pgsql's developers into fixing the "problem."
> (If you fix it, as in #2, you could submit a patch.)
>
> 4. Tell your users not to do that. Spaces in pathnames are
> pure evil anyway. A problem generally only found with end-users
> using (only) point-n-drool tools.
>
> Personally, looking at pg_ctl's code, I think #1 is exceedingly
> unlikely to succeed. #2 (and, by extension, #3) looks kind of
> iffy, too.
im probably going to try all of them until i find something that
succeeds. I might just pass it all to the postmaster directly.
thanks!