Recursive parser

680 Posts

aNoble Reply #21, 20 years, 10 months ago

This is great news. A faster and more powerful parser. Thanks for all your hard work on this Remon.

Adam Crownoble: obleDesign.com | [email protected] | [email protected] (Google Talk)

1,732 Posts

xwisdom Reply #22, 20 years, 10 months ago

Cool! That’s looking good so far.

The good news is that it’s a little faster and also supports recursive parsing

Can you say how many snippets where processed?

xWisdom
www.xwisdomhtml.com
The fear of the Lord is the beginning of wisdom:
MODx Co-Founder - Create and do more with less.

1,732 Posts

xwisdom Reply #23, 20 years, 10 months ago

It might also be good to add some debug code to show number of tags parsed and execution order. This only occurs when debug is enabled

xWisdom
www.xwisdomhtml.com
The fear of the Lord is the beginning of wisdom:
MODx Co-Founder - Create and do more with less.

125 Posts

R.S. Reply #24, 20 years, 10 months ago

Sure, I’ll add some debugging stuff, at least, I’ll try to do so

I would call the speed increase rather significant, since it’s just comming from the parser’s find and replace snippets/chunck/etc routines!

Ehm, the home page of the modx install has a:
- DropMenu
- FlexSearchForm
- ListRelated
- Most Recent

Which add significant to the processing time! (the evaluating of those snippets I mean)

Also a number of [(yada)] and [*yada*] which is handled by the processTags funtion now too, but these don’t add much to the processing time.

Some parts of the processTags() can be unified/simplified if the calling convention would be like this:

[[T CODE ]]

Where T is a specific tag, and CODE the name of a snippet/chunk/etc with additional parameters(only for snippets).

So, this:

[*#content*]

will become this:

[[*#content]]

A snippet remains the same, a chunk would be:

[[{chunkname]]

A placeholder:

[[+placeholdername]]

So in processTags we check for first occurences of:
[[*
[[(
[[+
etc
[[

on a match, we substring the template until a ’]]’ is found, and feed that to the correct method.
and continue our recursive processTags();
In case of a snippet, the snippet parsing code checks for nested snippets, just like it is now.

What do you think?

Greetz,

Remon

1 Posts

Zombie Reply #25, 20 years, 10 months ago

Very nice and very logical!

Good work!

1,732 Posts

xwisdom Reply #26, 20 years, 10 months ago

Hi Remon,

You’re right on target. We have been discussing the very same thing over the past few days. I’ll soon post those new tags here for us to review.

xWisdom
www.xwisdomhtml.com
The fear of the Lord is the beginning of wisdom:
MODx Co-Founder - Create and do more with less.

1,732 Posts

xwisdom Reply #27, 20 years, 10 months ago

I would also suggest that we can go ahead and implement the once you have suggested:

[[*
[[(
[[+
etc
[[

And we can later fine tune the tags to match the new recommendations.

xWisdom
www.xwisdomhtml.com
The fear of the Lord is the beginning of wisdom:
MODx Co-Founder - Create and do more with less.

1,732 Posts

xwisdom Reply #28, 20 years, 10 months ago

Note: This part (line 727) of the parser only collects the arguments passed into the system e.g. [[Snippet? &param1=`value`]]

$snippetProperties = '';
if($et>$st) {
	$tag = substr($source,$st+2,$et-$st-2);	// get params
	$parameterpos = strpos($tag,'?',0);
	if ($parameterpos) {
		$snippetname = substr($tag, 0, $parameterpos);
		$snippetProperties = substr($tag, $parameterpos, $et-$st-2);
	}
	else {
		$snippetname = $tag;
	}

// HELP HELP HELP
// EEHHMMMM, can someone explain me why we have a parseProperties function called above, and then
// do (ALMOST) the exact same stuff again below, overwriting the values assigned in parseProperties?????

the parseProperties() function is works on the parameters assigned from within the manager (default params). We then merge the parameters entered by the user as shown above in an effort to allow a developer to override the default params.

xWisdom
www.xwisdomhtml.com
The fear of the Lord is the beginning of wisdom:
MODx Co-Founder - Create and do more with less.

MODX Staff
12,272 Posts

Ryan Thrash Reply #29, 20 years, 10 months ago

Wow... I go away for a few hours and great things happen... think I’ll do that more often! laugh

Thanks Remon/Raymond!

Ryan Thrash, MODX Co-Founder
Follow me on Twitter at @rthrash or catch my occasional unofficial thoughts at thrash.me

125 Posts

R.S. Reply #30, 20 years, 10 months ago

Quote from: rthrash at Nov 15, 2005, 03:21 PM

Wow... I go away for a few hours and great things happen... think I’ll do that more often!

Thanks Remon/Raymond!

Hehe, you write if different, but pronounce it (almost?) the same ;-)

After some headaches (oh, I had one allready, never mind) again a new version of the parser, now with:

[[*yada]]
[[(yada]]
etc.

calls, for those interested, I could provide an url to show it works, and oh why not just giving it, here it is:
http://vt.shuis.tudelft.nl/~remon/test/

Have funn!

Remon

(P.S. I suppose there are a number of bugs left, but thats also part of the fun, no?
For example, friendly urls have a problem somehow)