Re: [html-formfu] Performance

2007-06-28 Thread Tobias Kremer
Zitat von Carl Franks <[EMAIL PROTECTED]>: > On 25/06/07, Tobias Kremer <[EMAIL PROTECTED]> wrote: > > (widget) timethis 1000: 7 wallclock secs ( 7.26 usr + 0.35 sys = > > 7.61 CPU) @ 131.41/s (n=1000) > > (formfu) timethis 1000: 63 wallclock secs (59.00 usr + 2.99 sys = > > 61.99 CPU) @ 16.13/s

Re: [html-formfu] Performance

2007-06-28 Thread Carl Franks
On 25/06/07, Tobias Kremer <[EMAIL PROTECTED]> wrote: (widget) timethis 1000: 7 wallclock secs ( 7.26 usr + 0.35 sys = 7.61 CPU) @ 131.41/s (n=1000) (formfu) timethis 1000: 63 wallclock secs (59.00 usr + 2.99 sys = 61.99 CPU) @ 16.13/s (n=1000) Okay, I've done some testing, and here's the r

Re: [html-formfu] Performance

2007-06-26 Thread Carl Franks
Tobias, On 25/06/07, Tobias Kremer <[EMAIL PROTECTED]> wrote: Am 25.06.2007 um 21:46 schrieb Tobias Kremer: > my $widget = HTML::Widget->new(); > $widget->element('Textfield')->name('text')->label('Text')->size > (30) for( 1..20 ); > timethis( 1000, sub { $widget->process } ); Hmmm ... I suppos

Re: [html-formfu] Performance

2007-06-25 Thread Tobias Kremer
Am 25.06.2007 um 21:46 schrieb Tobias Kremer: my $widget = HTML::Widget->new(); $widget->element('Textfield')->name('text')->label('Text')->size (30) for( 1..20 ); timethis( 1000, sub { $widget->process } ); Hmmm ... I suppose that last line should be: timethis( 1000, sub { $widget->process-

Re: [html-formfu] Performance

2007-06-25 Thread Tobias Kremer
Am 25.06.2007 um 21:41 schrieb Tobias Kremer: I'm not sure if what my script is measuring is comparable, but if it is, the results are devastating IMHO. Here's the source: Err, that was a little bit too quick :) Let's try that again ... use HTML::FormFu; use HTML::Widget; use Benchmark qw/:a

Re: [html-formfu] Performance

2007-06-22 Thread Tobias Kremer
> Hmm, are you using the compiled C, or pure-perl TT stash? > Is TT caching on? XS Stash with TT caching on. > Is that 50-60% of just the action subroutine that's being hit, or of > the entire catalyst process? What else is the action doing? Not much else. Here's the entire action: $c->stash-

Re: [html-formfu] Performance

2007-06-22 Thread Carl Franks
On 22/06/07, Tobias Kremer <[EMAIL PROTECTED]> wrote: with C::C::HTML::FormFu:~11 request per second without C::C::HTML::FormFu: ~11 requests per second :( I believe that the many TT INCLUDEs to build the form are causing the bad performance. I measured the stringification time of $form in

Re: [html-formfu] Performance

2007-06-22 Thread Tobias Kremer
> Zitat von Carl Franks <[EMAIL PROTECTED]>: > > > A good start would be to simply remove the FormFu controller and > > create the form manually with load_config_file() - and compare hit > > rates. > > I'm gonna report back with results shortly. If this really is the > cause I'll try to re-implemen

Re: [html-formfu] Performance

2007-06-22 Thread Tobias Kremer
Zitat von Carl Franks <[EMAIL PROTECTED]>: > A good start would be to simply remove the FormFu controller and > create the form manually with load_config_file() - and compare hit > rates. I'm gonna report back with results shortly. If this really is the cause I'll try to re-implement the Catalyst

Re: [html-formfu] Performance

2007-06-22 Thread Carl Franks
On 22/06/07, Tobias Kremer <[EMAIL PROTECTED]> wrote: I'm noticing a tremendous slow-down for all Catalyst controllers which make use of the Catalyst::Controller::HTML::FormFu plugin to utilize FormFu-based forms. My suspicion immediately falls on the Action implementation. I copied how the Fo