Serialization
- clean a bit
NodeBaseclass. - wip serialization of
NodeBasethat exemplify the 'intrusive' serialization (Serializerclass member) + test. - wip serialization of Eigen types (binary only) + test.
- serialization
TimeStamp+ test. - serialization entry-point object is referred by typeid.
Merge request reports
Activity
@jsola Can you please review and eventually merge this ?
mentioned in commit 09abd7fc
72 unsigned int nodeId() const; 75 unsigned int nodeId() const; 73 76 std::string getClass() const; 74 std::string getType() const {return node_type_;} 75 std::string getName() const; 77 std::string getType() const; 78 std::string getName() const; 76 79 77 void setType(const std::string& _name){node_type_ = _name;}; 80 void setType(const std::string& _type); 78 81 void setName(const std::string& _name); 79 82 80 virtual ProblemPtr getProblem(){return problem_ptr_.lock();} 81 void setProblem(ProblemPtr _prob_ptr) {problem_ptr_ = _prob_ptr;} 82 83 virtual ProblemPtr getProblem() const; -
Jeremie: This
constkeyword ingetProblem()prevents this method from being overloaded properly. In overloads, theproblem_ptr_in the class is sometimes written, and therefore this method cannot beconst.This could have been designed differently. But what I want to say is that, by merging this MR, I could not compile the
masterbranch. And therefore I wish to warn you about MR's not being properly tested.Did you compile this MR before doing the request?
I did compiled and tested, 2 weeks ago when I opened it :s . Sorry for that.
I have to insist on asking you to ask whomever is in charge of gitlab maintenance at IRI to setup a simple CI (continuous integration). This would automatically and constantly try to compile (possibly run tests) on every pull request / change, therefor save of lot of time and efforts to everyone and especially you.
-
I am about to push it to master. EDIT: pushed.
Edited by Joan Solà Ortega
