If the tpl doesn’t change, you could just put it in a chunk and do something like this instead of the function call
foreach($records as $record) {
$output .= $modx->getChunk('TplChunk',$record->toArray());
}
Then you wouldn’t have the overhead of creating all those chunk objects that you don’t really need.
It would be faster yet to just define the tpl string once and use str_replace on it each time through the loop.
-
☆ A M B ☆
- 2,475 Posts
The $tpl changes each call... or almost every call. I have my own version of this function (a la Evo’s parse function) that’s based on str_replace, and it is MUCH faster. But I miss out on the opportunities of having output filters that way.
Quote from: Everett at Jan 22, 2011, 12:12 AM
The $tpl changes each call... or almost every call. I have my own version of this function (a la Evo’s parse function) that’s based on str_replace, and it is MUCH faster. But I miss out on the opportunities of having output filters that way.
If there’s a finite number of $tpls, you’d still be much better off putting them in an array and just selecting the appropriate one. You can still use toPlaceholders() and output modifiers, so they will be processed when the page is rendered.
-
☆ A M B ☆
- 2,475 Posts
Ok, I tried the above (or something close to it), and it was only marginally faster... my page took 6.7 seconds to load. Without me re-usng the Chunks, it takes about 7.5 seconds to load. The str_replace version of the function causes this particular page to load in less than 0.2 seconds (!!).
Were you thinking of a different approach, Bob?
I was thinking of something that didn’t need chunks at all if speed is really an issue.
I was thinking of something like this if speed is an issue:
<?php
$tpls = array(
'tpl1'=> 'This is my first tpl [[+product]]',
'tpl2' => 'This is my second tpl [[+product]]',
'tpl3' => 'This is my third tpl [[+product]]',
);
foreach ($products as $product) {
$output .= str_replace('[[+product]]', $product->get('name'), $tpls['tpl1']);
}
You’d have to process any output modifiers yourself, but that would also speed thing up a lot.
-
☆ A M B ☆
- 2,475 Posts
Yeah, I see where you’re going with that... hmm.... unfortunately my architecture is pretty built up at this point, so re-designing the structure of my objects/children/etc is probably not viable. Maybe I could get everything I need by doing the str_replace() and then doing a final pass through MODx’s process method -- any lingering unprocessed placeholders would be the ones that had output filters on them. The output filters aren’t a do or die necessity, they’re just really nice to have.
I’ll play around with it.
I can see why you might want to do it the easy way. What I’m suggesting is basically how getResources and Wayfinder work, but if the hit in speed isn’t too bad, I usually use the built-in methods, although I’d probably use getChunk($tpl[$i], $properties), which seems easier to me, though I almost never use output modifiers. They’re painfully slow because before any work gets done, MODx has to parse not only the tags, but (separately) any modifiers, figure out what kind of modifier it is, look for chained modifiers, parse the conditions, etc. You can always do it faster in code because there’s no tag to parse you already know how the output should be modified.
Often, you can speed things up a lot just by pasting the modifier code into your snippet.
