How difficult would it be to make a single step python converter to take a minecraft 1.x world file, extract only the block data, and convert it into an empty minetest world (or place it at eg y=500, so it doesn't get corrupted by regen). Then our tools like worldedit would work, and the toolchain would be drastically reduced.
Worldedit is totally unstable attempting to import any large volumes. The python version of Minetestmapper-numpy can read an entire minetest world in seconds (or worst case minutes). Apparently the c version is even faster, but python is so much easier for most of us to code and we don't have to recompile for every different computer. tomtsschem.py is so incredibly slow (presumably because it's relying on mcedit to access the file, but also possibly because it doesn't follow the minecraft block order, so extraction is very inefficient.
I'm thinking it would be far better to just bypass mcedit entirely.
The minecraft anvil file format is very well documented, so shouldn't be that difficult to extract from. The difficult bit I don't understand it the files seem to have compressed sections within them, and I'm not sure how you'd go about extracting that. Actually mineblend:
http://randomsamples.info/project/mineblend/ has python code to extract direct to blender, so would probably be a good start, and there are probably others
Then you could have a simple text-based lookup table to convert from a minecraft block to a minetest one (just as tomtsschem.py does) and resort the blocks into chunks for minetest.
Finally put it directly into a minetest world sqlite file with any header code.
Even if it took all night for a world to convert it would be worth it as it give minetest users access to the thousands of amazing worlds and objects built by minecraft users...
EDIT: I've started a new post on this, as the more I think about it the better the idea seems:
https://forum.minetest.net/viewtopic.php?f=5&t=11146