potential fix for crd with empty cut #63
Conversation
Pull Request Test Coverage Report for Build 70
💛 - Coveralls |
| double demominator = _a < _b ? _a : _b; | ||
| if (iter->second == 0) { |
There was a problem hiding this comment.
Isn't a denominator of 0 the problem?
(Not being totally familiar with the data stored here, why is it safe to ignore the entry due to i?) What is cond[i] storing ?
There was a problem hiding this comment.
From my understanding, the denominator can be 0 only if the numerator is zero as well.
| @@ -163,17 +164,21 @@ vtype graph<vtype,itype>::capacity_releasing_diffusion(vector<vtype>& ref_node, | |||
| } | |||
| } | |||
| vtype length = 0; | |||
| vtype idx; | |||
| vtype idx = 0; | |||
There was a problem hiding this comment.
Shouldn't this be initialized in a different way? if idx = 0 why is the rest of the code guaranteed to produce a valid output?
|
@dgleich Let me know how you think on my latest commit. |
|
Meng, it stilll seems like we can improve this one :) Keep working on it. |
|
Meng, let's do this -- switch the stages to include two types of tests. The first type is one that impacts coverage. Primarily this should focus on the python library as that's where we get coverage info :) The second type doesn't have any coverage info. This just deeply tests algs like CRD. So we only care if the second type passes or not. So coveralls shoudl only run on the first type. Make sense |
Sounds good. |
No description provided.