Open arorashu opened 4 years ago
The failing tests for CPISync are probably due to Issue#74. I will need to take a look why the CuckooSyncTest fails.
What I see in the CI build report is very weird. For example, it says that CuckooSyncTest.cpp:61:Assertion
fails and that assertion is still commented out in your code, should not be executed at all (you seem not to pull from master recently, see #71).
Could you please merge master
to your branch and make sure why these tests fail?
Additionally, consider supporting the very basic test set generation function that would just simply return two std::multiset<shared_ptr<DataObject>>
, as DataObject
is our primary abstraction for the set elements. Set A and the set B, just to keep it simple.
What I see in the CI build report is very weird. For example, it says that
CuckooSyncTest.cpp:61:Assertion
fails and that assertion is still commented out in your code, should not be executed at all (you seem not to pull from master recently, see #71).
So this is because Github merges the two branches itself, and runs the CI on the merge. I validated it from the CI output.
Could you please merge
master
to your branch and make sure why these tests fail?
Yeah, in effect the test failures are resultant of merging with master
@arorashu I see... Can you please look a bit closer into that to find the root cause of those failures?
So, CPISync fails for the case where server is empty. I have currently narrowed it down to the CPISync::set_reconcile
function. For the case where server is empty, it does not populate the delta_self
and delta_other
entries. I will update when I have fixed that.
Increasing Cuckoo filter size from 2^8
to 2^17
, gives better sync performance at the expense of running time
Addresses Issue#62
Create DatasetGenerator class to create different standard types of datasets. Accepts isMultiset and isLargeSync as boolean flags to generate these specific datasets as well.