> I noticed that in fresh clone with fixes integrated to Runtime for emulator > in Linux environment ... in emulator.mk file there are still uses of > -std=c++98 instead of c++11. Good obsevation. However, I think that usage is ok, at least for now, as those are used in substitution patterns that apparently don't match. One could try removing those outer substitutions altogether; the inner substitutions for removing -mcpu=cortex-m0 are needed, I believe. But the build system has changed so much that it is possible that even them are unnecessary by now. > Moreover, the tests cases got compiled but I got same error as Eero Hakala: > > test_SPI_begin.o:(.rodata._ZL5SPI_2+0x4): undefined reference to `spimap2' > test_SPI_begin.o:(.rodata._ZL3SPI+0x4): undefined reference to `spimap1' I wouldn't worry about these at the moment, Ivan and I will be working on this, in the feature-coap branch. > But, other test cases are running for me. Excellent. Now you can proceed to thinking about and experimenting with the test cases for real. Please consider the ones in tests only as early and quick models, not something to be retained. I guess it would be best to compile the libraries into shared libraries for the emulator, and load those shared libraries from python. But it may be easier to start with just parsing the textual output of the current tests. --Pekka