[Catalyst] Catalyst speed test results

[Perrin Harkins]

On Mon, 2006-05-22 at 19:32 +0200, Szabo Peter wrote:
> -- Is our test scenario typical?

Is this a serious question?  No, it is not a typical scenario.  Your
reference.pl script looks like a CGI script from 10 years ago.  It
doesn't use templates, doesn't handle errors, doesn't handle database
timeouts, in-lines all the HTML and SQL... I would never allow a piece
of code like this to be used on a project I was involved in.  Meanwhile,
you are comparing it to Catalyst using templates (TT), using an
object-relational module (Class::DBI), and using relatively cleaner code
organized into modules.  You even seem to be using the slower TT stash
instead of the XS one.

There is some overhead from using Catalyst, but your test is not
measuring it.  If you want to measure it properly, then use the same
modules in both places.

- Perrin