cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Daniel Fagerstrom <>
Subject Re: jx:text
Date Fri, 20 Jan 2006 15:59:38 GMT
Joerg Heinicke wrote:
> Daniel Fagerstrom <danielf <at>> writes:
>>> <xyz>
>>>   <jx:strip-whitespace>
>>>   ${}
>>>   <nested-tag><jx:preserve-whitespace>     ${abc.description}
>>>             </jx:preserve-whitespace></nested-tag>
>>>   </jx:strip-whitespace>
>>> </xyz>
> This looks very awful. Not that I need this feature, but ...
And that is the issue, do we want to complicate the template framework 
with ad hoc constructions that few needs?

After have spending much more effort than I would have liked in 
re-architecturing JXTG I'm very reluctant in allowing features that 
complicates the framework. It is built for having elements as 
instructions not attributes. We could add the possibility to have 
attributes as instructions as well, but in that it should be a general 
purpose addition, not a hack for a single attribute instruction.

>> Might be that it is awful, but that goes for the rest of JXTG as well 
>> (or any XML language), it is still not a reason to complicate life for 
>> users by mixing attribute and element syntax. Furthermore, I doubt that 
>> what you do above is a common use case.
> I don't think that the above syntax complicates users' life in comparison to
> attribute syntax. Just compare it to namespace declaration in XML:
> <element xmlns="myNamespace">
>   <element xmlns="yourNamespace">
>     <element xmlns="someonesElseNamespace">
>       <element/>
>     </element>
>   </element>
> </element>
> Now imagine this one with namespace elements ...
> If there is a clear separation between concerns (i.e. mapping elements to
> instructions and attributes to what Leszek called processing directives (btw,
> XML also knows 'processing instructions')) I see no problems using attributes.
I don't see any clear separation of concern in this case. Processing 
instructions in XML is info about how the whole document should be 
processed, not something at the element level.


View raw message