Re: Mason 1.3 autoflush behaviour change

On Wed, Oct 13, 2004 at 09:32:30AM -0700, Jonathan Swartz wrote:
> I suppose we could also change the default to enable_autoflush=>1. I was 
> just hoping to make the higher performance option the default, but maybe 
> that's silly. static_source is the biggest win and it isn't (can't be) 
> the default anyway.

Keeping the defaults the same as Mason 1.2x means that folks are more
likely to upgrade to the newer codebase faster and with fewer hassles.
I'm _thrilled_ that the optimization work has been going on and looking
forward to eventually being able to make RT take advantage of it.

> But Jesse, can I ask how you're using autoflush and why several 
> well-placed manual $m->flush_buffer calls wouldn't work as well?

Because RT is a database-driven application with a fairly large amount
of data behind it, individual page views can sometimes take a fair bit
of time. Turning on autoflush (and turning it off in a few important
places was a really quick way to greatly improve RT's _percieved_
performance. ;) Having to dig through 500+ mason components inserting manual
flushing is really a really unappealing prospect.   Long-term, it may be
a better solution than autoflush, but being forced into it by a newer
Mason appearing on CPAN would be unfortunate.


Jesse


-------------------------------------------------------
This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
Use IT products in your business? Tell us what you think of them. Give us
Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
http://productguide.itmanagersjournal.com/guidepromo.tmpl