Boy, this has been a ride beyond any of my dreams.
It was a huge pain to get it to work. I contemplated many times (and deeply) about abandoning the whole effort. But some irrepressible part of me kept me working on it.
Somehow, it was resuscitated. The gods smiled on me and it worked.
But there was a twist in the tale. And then it threw a mighty fit. It definitely was the lowest point in my entire saga (so far that is). I had actually abandoned it (formally at least). Only God knows why I invested more efforts into it. I have no recollection what pushed me into chasing it.
Somehow, that was achieved. You are a life saver, valgrind. But the overhead was still too much. 65x more than what it should have been. 6x more thn the useful work itself. Simply unacceptable.
But, finally, today, on the auspicious day of Diwali. I declare it done. It goes without saying that I am very pleased with the overall results.
It's done, completely.
Only thing left is merging it into master.
As for how the overhead was eliminated, have a look here. Funny, isn't it?
Today, I can confidently say that this is the most complex bit of engineering ever attempted by me. And I have accomplished my goals set before beginning it. I spent time and effort even in stuff like API orthogonalization. Definitely, my biggest achievement yet.