Yup, this is basically a variation of #1, where you need to know the DOM of the final HTML document.
There are less than ten (like 2-6) pod to ?html things with anything resembling a user base, and only 2 are "popular"...
So there is not a whole lot to know, you make the text, so make it easy, resulting DOM is either one target node <div>##tablestart ... ##tableend</div> or a bunch of siblings <p>##tablestart...<p>##tableend
Both are easily found if you tag the table
with jquery $(':contains("##tablestart")').each...text...replace...
or xpath (jquery also has a plugin) var iterator = document.evaluate('//*[ not(./*) and contains( ., "##tablestart")/ ]', document.body, null, XPathResult.UNORDERED_NODE_ITERATOR_TYPE, null ); ...
Perhaps there should be something like =for !HTML or =begin !HTML ... =end !HTML? :) If you think you need that, you should generate the pod from some other template language ... if it doesn't drive you insane
Yes, you could invent a =begin :format, but then you'd have to convince all those pod module authors to accept your solution, and all the users would have to upgrade .... seems like a nonstarter , pod is already too complex :)
|