Stephen R. van den Berg wrote:
I was then wondering if it is being used successfully anywhere already?
I now know more about the Shuffler code than I ever wanted to know (obviously).
However, armed with that knowledge I am reasonably surprised that anyone managed to use it in the old state without triggering its myriad of lurking segmentation fault and memory corruption pitfalls.
I don't use it myself in Pike 8.0, I use it in Pike 8.1 now. But, in the interest of stabilty I can offer to port the fixes from 8.1 to 8.0 (which, I hope, should be reasonably straightforward, I haven't looked yet).
Should I?