Bart Lateur wrote:
> 
> On 17 Jan 2002 09:04:20 -0800, Jeff Zucker wrote:
> 
> >This is the first CPAN release of the pure perl version of the module.
> 
> A first release, and already at 1.0?  (!)

There were several 0.x non-CPAN releases and I wanted to distinguish it
from Jochen's releases.  Besides aren't 1.0 releases known to be buggy? 
Wait for 1.001 if you're cautious :-).  I did get reports from a variety
of testers that it works on linux, sunOS, win32, macOSX so I am somewhat
optimistic.

> I don't know if it even matters... but how does it compare, speedwise,
> to the XS version?

The joins are much faster :-) (since you can't do them in Jochen's
version).

Seriously though, it is undoubtedly slower for some things.  The parsing
slowness can be minimized by judicious use of
prepare-once-execute-many.  The eval_where (and indeed everything) needs
optimization.  I concentrated on getting things to work right (nested
parens and all the variations on joins were a bear) and then I'll go
back and speed things up as I can.  As always, for large apps where
speed is critical, users are better off with a real rdbms.

-- 
Jeff

Reply via email to