hi, On Wed, 6 Jun 2007, Perrin Harkins wrote:
On 6/6/07, Octavian Rasnita <[EMAIL PROTECTED]> wrote: > I have tested HTML::Template::JIT, but HTML::Template::Compiled was much > faster than it. Sorry, but I suspect there's a mistake in your test. Possibly you counted the time for JIT to do the initial compile, which is slow but only happens once. HTML::Template::Compiled is fast, but it's not as fast as JIT.
well, in my tests (i'm the author of HT::Compiled, so of course i will advertize it =) it was faster in some cases than JIT (e.g. if the template was big enough). and yes, i mean faster *after* the initial compilation. it's also a bit faster than H::T::Pro (but only in the best case, case_sensitive and with memory cache). (i commented out JIT of my benchmark script, though, because it generated errors someday and didn't stop. also i realized that it changed the $_ variable after the run.)
harder to debug template coding mistakes with, and HTML::Template is fast enough.
well, many people say, why optimize code if the database is slow anyway. i don't like that. if you can optimize by using a fast module - why not? and by the way, JIT doesn't have many features, while HT::Compiled has the dot syntax. use the examples/bench.pl included in the distribution to see some benchmark results. for example i have used HTC in a cronjob which generated XML - instead of LibXML. before the cronjob took so long that we could just run it 4 times a day. after that it could be run 6 times a day. keep in mind, i'm the author. i'm a fan =) just my 2 cents tina
