I finished up the message randomization code and was able to integrate my tolerant proto implementation into yours without too much trouble. A lot can be cleaned up, but as we discussed yesterday, it's probably better to get the code together first and worry about the rest afterwards.
A few notes:
I can compile both application.pro and test.pro
all tests pass (though there aren't any tolerant proto tests yet)
One issue: the round seems to start when there are fewer nodes than the minimum. If I run 2 nodes, the thing starts running. This happens no matter what round type I set in the config file, so maybe it's a bug/feature of the new Group stuff.
Hi David,
I finished up the message randomization code and was able to integrate my tolerant proto implementation into yours without too much trouble. A lot can be cleaned up, but as we discussed yesterday, it's probably better to get the code together first and worry about the rest afterwards.
A few notes:
One issue: the round seems to start when there are fewer nodes than the minimum. If I run 2 nodes, the thing starts running. This happens no matter what round type I set in the config file, so maybe it's a bug/feature of the new Group stuff.
Henry