Re: ereg + performance problem

Hi Mirek, thanks for pointing this out. This looks like a great site -- although, as you say, will require some [longer term] study. I'll have closer look at Smarty as well... "Mirek Novak" <mirek.novak@auto.cz> wrote in message news:3FF54110.3030502@auto.cz... > Hi, > I think that smarty [ http://smarty.php.net ] can do this for > u. It > has tag nesting and, of course, you can define your own tag as a plugin. > > As for performance problem - this is known "feature" of regexp. Solution > is not simple. You, IMHO, have to rethink your approach. For parsing > larger or complex files is much better to use state machine (finite > state automaton). Try to search google for more theory. For ex. > http://en.wikipedia.org/wiki/Finite_state_automaton > > Martin Helie wrote: > > Hello, > > > > I'm writing a routine that recursively reads an HTML document, looking for > > "special tags". It's a template system, but contrary to what I've seen out > > there so far, no template engines allow for any kind of customization from > > within the document; they only seem to be variable replacement systems. > > > > What I have in mind is something along the lines of: > > > > <HTML> > > blah blah > > {MAGICTAG param1=a param2=b}some more {ANOTHERTAG}text{/ANOTHERTAG} here > > {/MAGICTAG} > > blah > > </HTML> > > > > The text between ANOTHERTAG would first be changed and then passed along > > with the other text to MAGICTAG. ie, the data is treated from the inside > > out, and nesting is obviously supported. > > > > I've got everything working, but the problem is I'm using quite a few "ereg" > > statements recursively, so performance is not so good. > > > > Here's a simplified version of the core (it's part of a class, and I > > modified the code for a "standalone" version here). > > > > Thanks in advance for any ideas. Feel free to recycle this code. > > > > function parse( $template ) { > > > > //Look for (before*) {a tag} (after*) anything and store matches in > > $regs > > if( ereg( "(.*)[{]([a-zA-Z0-9]+)[}](.*)", $template, $regs ) ) { > > $before = $regs[1]; > > $tag = $regs[2]; > > > > //Now look for (data*) {/closing tag} (after*) in the "after" > > portion of first match > > if( ereg( "(.*)[{][/]" . $tag. "[}](.*)", $regs[3], $regs ) ) > > { > > $after = $regs[2]; > > $data = $regs[1]; > > $resolvedTag = initHandler( $tag, $data ); //handle this tag and > > data > > } > > //support for "standalone" tags (no {/closing tag} ) > > else { > > ereg( "(.*)", $regs[3], $regs ); > > $resolvedTag = getTagValue( $tag ); > > $after = $regs[1]; > > } > > } > > if( $resolvedTag ) { > > $template = > > $before > > . $resolvedTag > > . $after; > > parse( $template ); > > } > > else { > > return $template; > > } > > } > > > > -- > Mirek Novak > > jabberID: mirek@njs.netlab.cz > ICQ: 119499448