[XAR2] New implementation of Blocklayout compiler

["Jason" <judgej (at) xaraya.com>]

"Marcel van der Boom" <marcel (at) hsdev.com> wrote in message 
news:eik034$7ap$1 (at) newton.xaraya.com...
> Jason wrote:
>> "Brian Bain" <xaraya (at) tefen.net> wrote in message
>>> This causes me some worry.  I've been playing with ajax on Xaraya quite 
>>> a bit lately, and I've found that creating my own "ajax" templates seems 
>>> to work pretty well.  The typical template may read like this:
>>>
>>> <screenElement1>
>>>    <item id="123" value="Item 1" />
>>>    <item id="234" value="Item 2" />
>>> </screenElement1>
>>> <screenElement2>
>>>    <foo bar="bar1" />
>>>    <foo bar="bar 2" />
>>> </screenElement2>
>>>
>>> Does this change mean that I won't be able to do such a thing unless I 
>>> define each screenElement and item and foo?
>>
> No problem, they just go through. The default behaviour of the compiler is 
> to copy the input to the output if there is no specific match. We only 
> match tags in the xar: namespace + comments (ignored) + 
> processing-instructions (ignored). Everything else is copied (almost) 
> verbatim. (there will be some whitespace normalisation etc., but this 
> should not change behaviour nor structure. Unfortunately, we've already 
> found 1 case where it does in (X)HTML rendering )

There are certainly browser inconsistencies (and bugs) where white space can 
have an effect. Internnet Explorer and lists formatted into menus comes to 
mind.

>> I'd guess this would not be a problem. The above looks like well-formed 
>> XML, so should pass through okay. It may not pass a validating parser, 
>> without a matching DTD, but I think Marcel is talking about well-formed 
>> XML, rather than validated XML, at this stage at least.
>>
> Both actually. Well formed XML is not a choice anymore, but a must. 
> Without it, nothing works. Validation can be built up step by step. We 
> have no structural requirements at this point. i.e. any tag can appear 
> anywhere with any attribute (note: this is _input_ , validation of 
> _output_ is a whole other game )
>
> At least for part of the xar: namespace it makes sense to add some DTD 
> rules (xar:else only makes sense as child of xar:if and xar:sec for 
> example )
>
> Previously it was useless to define it, because no part of the input chain 
> would pay attention to it. This has changed now, so we can (and probably 
> should) define a DTD for xar: now.
>
> marcel 


_______________________________________________
Xaraya_devel mailing list
Xaraya_devel (at) xaraya.com
http://xaraya.com/mailman/listinfo/xaraya_devel