MODx Recursive Parser - Beta 2a

Quote from: xwisdom at Jul 11, 2006, 08:19 AM

As for version 1.0 I think Jason is working on a recursive parser for the new core. I’m not sure if he will be using this version of the parser.

Right, I implemented said parser back in December, and that will be the parser in 1.0/Tattoo. No worries though; I believe most if not all of the tag processing features being introduced here can be accomodated in one way or another, though I’ll need to go through all the new features one by one to make sure

@David, I just checked and reply notification works for me. Can you please double check ?

Just i case: "More replies may be posted, but you won’t receive any more notifications until you read the topic."

regards.

zi

Yeah OK Zi I’ll check, might be me, I have been so overworked lately... my moving to a new appartment doesn’t help !

Quote from: davidm at Jul 11, 2006, 12:00 PM

And thanks for the patch You mean you implemented what you meant to do for Beta2 in such short time ? Or is it transitionnal fix ?
Anyway : that’s great !

Works like a charm

Glad you got it working David. This was just a quick fix to assist you with the html entities and as it turns out it only helped to mark the parser smarter.

As for Beta2 I’ve thus far implemented 30% for the new features (mainly the removal of eval() when executing snippets and debugger). A lot more needs to be tweaked and added. You can download what I’ve done thus far from my branch of the SVN. Look inside raymond/greenToken.

Thanks Raymond, but I am not sure I have this kind of SVN access... I’ll stick with the release you do for now

Just had a thought about the new parser which would eliminate a TON of support posts:

Since we can now escape things in the parameter calls, we could all people to wrap params in ’ and " as well, no?

Combined with cache control in the snippets, this might be a great support request eliminator!

Yea, great idea Ryan!

Maybe use " but ’ I don’t know about that, for instance in french we use the ’ in some strings and words like after a d’ or a l’, wouldn’t that in turn require we escape those if in parameter (like in php we have to d\’ to escape the ’ ) ?

The point of using ’ or " is that they’re both what PHP uses and people often confuse ` with ’ ... daily. People do it all the time.

I’d still advocate using backtics as you’ll have to escape fewer things, so this would be an optional addition to the parser, not a change to the parameter wrapping characters. However, if it has even a marginal negative impact in parser performance, I say just skip it.

if it is enabled, that there can also be " and ’ around snippet parameter value, I think that the replacing those chars just around snippet parameter value could be then done automatically, so user’s wouldn’t need to escape anything manually (other than backticks)..

so it would automatically turn this:

[[Snippet? &param="test string with "quotes" in value" &param=’test’]]

to this:

[[Snippet? &param=`test string with "quotes" in value` &param=`test`]]

Maybe even possible with some genious regular expression pattern? (my brain doesn’t like those, so I’m not even trying) But don’t know how hard it would be to implement this to work in all cases, because there can now be multi-line parameter values and complex structural snippet calls etc...