[AWS] [Template Parser] How to use Cursor_Tag

[Mathieu Benard]

Pascal :

This looks like a matrix, no?
>
> If MSG_NAME is a vector : "v1" & "v2"
>
> and WORD_NAME a vector of vector:
>   "n1.1" & "n1.2"
>   "n2.1" & "n2.2"
>
> Then:
>
>   @@TABLE@@
>   <message name="@_MSG_NAME_@">....
>       @@TABLE@@
>       <word name="@_WORD_NAME_@">....
>      @@END_TABLE@@
>    @@END_TABLE@@
>
> will be transformed as:
>
>   <message name="v1">....
>       <word name="n1.1">....
>       <word name="n1.2">....
>   <message name="v2">....
>       <word name="n2.1">....
>       <word name="n2.2">....
>
> Isn't this what you want?
>

Yes, actually it's what I wanted. I tried and it works perfectly. However,
the complete structure is a bit more complexe and my example was not
exhaustive. The structure is actually a tree. Thus, messages contain words,
but words contain fields and so on. So, there are a lot of composition
levels to deal with.

> Actually, I don't know how I could find the good words matching a given
> > message. Moreover, a given word could belong to many different messages.
>
> Not sure I understand this.
>

You're right, it was not clear at all. Actually, I just wanted to say that
one instance of word may appear in several messages.
Ex :
      <message name="m1">
            <word name="w1">
      <message name="m2">
             <word name="w1">

Thus, in the vector tag, word w1 will be duplicated and I 'd like to find a
way to avoid this redundancy. I don't know if it's clearer now, but anyway,
with your example it works well.

Thomas :

If you intend to create very complex XML structures, I'll advice you
> to use XML/Ada instead:
>
> http://libre.adacore.com/libre/tools/xmlada/
>
> templates_parser is a great template engine, but for very complex
> output you might find it's simple vector/matrix tags and @@TABLE@@
> construction a bit "lacking".
>

Actually, the goals of my project  are  not limited to generate xml files
but also ada source code, test code and even documentation based on
templates.

To achieve this, I have two inputs : an xml file which contains data and
templates files.
I already use XMLada (SAX api) to retrieve data from the xml input, but I
have to use my own internal data structure in order to instanciate
templates, this reason make the use of the Templates_Parser adequated.

Finally, I can achieve my goals with vector/matrix tags, but I think that
cursor_tag would be more efficient, if they allow to iterate on data
structure (ada record).

Thanks for the answers,
Regards,

-- 
Mathieu Bénard
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/aws/attachments/20100507/ab9fa427/attachment.html