On Fri, 21 Apr 2006 12:44:45 -0700, Johannes Graumann  
<johannes_graumann@web.de> wrote:

> nico wrote:
>
>> On Thu, 20 Apr 2006 17:27:50 -0700, Johannes Graumann
>> <johannes_graumann@web.de> wrote:
>>
>>> nico wrote:
>>>
>>>> Why don't you test if the attribute is set before applying it to the  
>>>> TR?
>>>> You see the problem on the last row, but the fact is that you force  
>>>> the
>>>> bottomframe clear for every row (which is dirty IMHO).
>>> It's not that easy, since the bottomframe - once set - is persistent  
>>> ...
>>> so
>>> every row underneath my header (e.g.) will have 'bottomframe=on' set.
>>
>> No, the options are local to the current level where it is set. If
>> bottomframe is on for one row, it won't set bottomframe to the next  
>> rows.
>> In the following example, only two rows (the header and another one in  
>> the
>> body) have local bottomframe on, and the last row bottomframe is done  
>> with
>> table setup:
>
> The persistence issue is XML specific, I believe and therefore does not
> apply to your code example.

It's because you systematically apply the attribute. Doing this should  
work, even if there's maybe a smarter way of doing:

\defineXMLnested
   [tr]
   [bottomframe=,rulethickness=0.5pt]
   {\doifelse{\XMLop{bottomframe}}{}%
    {\expanded{\bTR}}%
    {\expanded{\bTR[bottomframe=\XMLop{bottomframe},rulethickness=\XMLop{rulethickness}]}}}
   {\eTR}

I attach the full example.

Regards,
BG