Two more comments below ... Am Mi., 31. Juli 2019 um 01:40 Uhr schrieb Ahmed Essam < theartful...@gmail.com>:
> Hello all, > > Yesterday was the end of week 9 (I think I might've lost count!). Find > what I did in this week in this blogpost > https://medium.com/@theartful.ae/gsoc-19-with-shogun-week-9-adde4e742555 > > I think I've hit a blocker on the SGString PR. It's not possible to define > typemaps for std::vector<SGVector<T>> because "std_vector.i" (swig's > library to support vectors) already defines them. Their typemaps work well > with primitive types, but I couldn't get them to work with custom types. > Maybe we split the refactoring into two steps where we first remove SGString and keep SGStringList (renamed to SGVectorList), and then refactor that in a second step? See comments in PR > > On the spdlog PR, it needs review. Also do we want to support compile-time > loglevels? So for example when shogun is build in release mode, we can get > rid of all SG_DEBUG and SG_GCDEBUG. > I like the idea of not compiling the debug messages in release mode. Same could be done with ASSERT > > This week I will finish up old PRs, and work on subsets. > > Thanks, > Ahmed Essam. >
