Bernhard Scholz

Results 168 comments of Bernhard Scholz

I believe we need further investigation. What is wrong with the current behaviour? Ideally, we enclose OpenMP operations with `#ifdef`s, checking whether OpenMP is enabled in the compilation environment. If...

Can you run `clang-format -i src/include/souffle/swig/SwigInterface.h` to ensure that the code changes are uniformly formatted. See below, ``` 10 diff --git a/src/include/souffle/swig/SwigInterface.h b/src/include/souffle/swig/SwigInterface.h 11 index 41aeab1..e175f37 100644 12 --- a/src/include/souffle/swig/SwigInterface.h...

Can you provide some examples in `tests/interface/swig` that access the interface?

Instead of an `#if` in the code checking for 32bit or 64bit, please could you use types as shown in this functor implementation: https://github.com/souffle-lang/souffle/blob/master/tests/interface/functors/functors.cpp

Also, can you add python/java code that accesses the tuples in a relation.

We could introduce a new label beside the old one that conforms to the latest standard and slowly transit to the new label over the next few versions (using the...

This is a nice observation! But note that we cannot assume that all programs with ungrounded variables will become programs with grounded variables after applying the magic-set transform. It is...

Thinking more about your example, there are two aspects in your example that would need to be addressed separately. (1) Ungrounded variables: They are caused by functors (in your example...

Indeed, I did not list the special cases for the zero and one arity case.