Great! I tried to update the patch myself, but there were too much conflicts. Thank you very much Antoine for taking this!

Seems there are bugs in frame_setlineno() related to code duplications. And deduplicating the 'finally' blocks doesn't solve all of them. See comments on GitHub.

I don't like the new END_ITER instruction. This complicates (and slows down) the evaluation loop and the peepholer. This also adds a limitation on the peepholer (END_ITER can't be optimized out). We can use other way for determaining the end of the 'for' block. The argument of the FOR_ITER instruction points to the end of the 'for' block. Just scan all addresses from 0 to max_addr and count all FOR_ITER instructions and their targets in the range between min_addr and max_addr.