I did use the exact pool software as used by others, it does not work. Not on any of the available branches either. I have no reason to believe stevebrush was not aware of these pool breaking changes, as per the previous issue, which he explicitly quoted the fix to me after I already fixed it.
You very likely did
not use the "exact" same "pool software". The reason I used your pool and recommended it was because your pool allowed a fixed diff share setting, while others didn't have that feature.
When I investigated this something like four months ago I noticed that you ran a different fork of the pool software than all others, which was obvious already by the different source reference and versioning in the footer of your pool page.
I don't know how many forks of that "cryptonote-universal-pool" exist but there must be quite a few and very likely you ran a different one than the other pools. Which as I mentioned previously wasn't a bad thing at all.
To make it clear exactly what I did, here are the following versions I tried.
clintar/cryptonote-xmr-pool
clintar/cryptonote-universal-pool
zone117x/node-cryptonote-pool <--used by others, broken by default
fancoder/cryptonote-universal-pool <--works
in combination with the following node-cryptonote-util
clintar/node-cryptonote-util
clintar/node-cryptonote-util#Nan-2.0
clintar/node-cryptonote-util#xdn-Nan-2.0
clintar/node-cryptonote-util#xmr-Nan-2.0
zone117x/node-cryptonote-util
Snipa22/node-cryptonote-util <--used by others, broken by default
fancoder/node-cryptonote-util <--works
I have dug into the fork history of all of this during my troubleshooting, read every issue involving messages, modified and recompiled node-cryptonote-util. Out of all available combinations and options, only one thing worked and that was fancoder's fork. It's working now and I'm not going to dig any deeper.
In my mind there may be one last difference. I compile my node, and last night there was a new commit, during my troubleshooting I did recompile the node. The possibility exists that the binary released days ago was compiled from code that was not available on the git at time of release. I don't believe this is the case, though.