Steve Singer <ssinger at ca.afilias.info> writes:
> -Slony now requires that your libpq be built with --enable-thread-saftey 
> on all platforms.  Previously this was only required on some platforms( 
> AIX, Solaris).  However we have determined that it is required on all 
> platforms.  This should not be an issue for postgresql installs from 
> binary packages that are already compiled with --enable-thread-safety 
> but might require users who have built their own postgresql from source 
> (without --enable-thread-safety) to recompile.  (See bug #40)

Just to quibble a little bit on details, I'm not certain that this
necessarily requires that the --enable-thread-safety specifically be
used.

What's necessary is that libpq be thread-safe, which may be validated
via a libpq function, PQisthreadsafe().

On many (perhaps most, perhaps, but not *for certain* all, hence my
quibble) platforms, for PQisthreadsafe() to return true requires that
PostgreSQL (well, libpq, specifically) be configured with
--enable-thread-safety.  But I'm not certain that this is always
necessary.  It appears that some platforms pass PQisthreadsafe() without
"configure --enable-thread-safety" having been explicitly requested.

Quibble done :-).

> A complete list of changes can be found in the RELEASE notes file.
>
> We would ask that people give the release candidate a try and report 
> both successes and failures to us (either through the mailing list or 
> off-list email).

Agreed!

Lots of stuff has gotten cleaned up for a 2.0.5 release; some validation
of success/failure would certainly be appreciated.  It has been a pretty
long road, and I certainly hope 2.0.5 "actual" proceeds mighty soon.
-- 
output = reverse("ofni.sailifa.ac" "@" "enworbbc")
Christopher Browne
"Bother,"  said Pooh,  "Eeyore, ready  two photon  torpedoes  and lock
phasers on the Heffalump, Piglet, meet me in transporter room three"