[sword-devel] osis2mod changes for Words of Christ and container elements

[DM Smith]

Greg Hellings wrote:
> DM, two questions:
>
> On Thu, Sep 11, 2008 at 1:57 PM, DM Smith <dmsmith555 at yahoo.com> wrote:
>
>   
>>   Added #define INCLUDE_TAGS to include all tags.
>>       Currently osis2mod preserves text and most tags.
>>       This would include all tags and their attributes, but it is
>> turned off.
>>     
>
> Why did you turn this off by default?  I would presume it might break
> output with filters which are not currently set to strip out this
> extra tagset - was that your reason?
>   
It preserves prior behavior.
That behavior is that <verse>, </verse>, <chapter>, </chapter> and <div 
type="book"> are not included in the module.
>   
>>    One thing to note, I've used "genX" as the sID/eID value (where X is
>> unique to the pair). This allows for reverse engineering in mod2osis.
>>     
>
> What is the value placed in X?
osis2mod starts with 1 and every time an sID is created, osis2mod use 
the next number.

A stack is used to find the corresponding end tag and when it is found, 
the top element is changed from sID to eID and output.

>   Recreating in mod2osis should be
> relatively straightforward if my formulated thoughts on how you
> described it are accurate.  However, there would need to be some
> indication from the Sword library that we're entering a new
> section/paragraph.  I don't believe mod2osis currently looks for any
> such information.  Finding those locations seems almost like it would
> fall to the filters to have direct access to the module data,
> especially for a section which might be placed before the start
> milestone for version book.1.1 or book.0.0, etc.
>   

I haven't looked at mod2osis. If the module is already osis, then there 
is no need for filters. A simple loop over all keys calling getRawText 
and transforming that should be nearly sufficient.

You might want to have two routines, one for OSIS modules and one for 
other encodings.

-- DM