wlav
wlav
Just adding a couple of remarks; feel free to ignore. :) Full disclosure: I'm the author of cppyy, and saw some of my documents referenced above, which is how I...
Yes, this is known: the JIT in the current version of LLVM underlying Cling uses emulated TLS even if the platform supports it natively. More recent gcc no longer does...
I see. Yes, it does make sense, but there's no easy place to put it if coded like the above example: the semantics of owernship passing are completely encoded in...
Looks like upstream isn't providing any enums (still looking into that one). The function is missing b/.c `__dir__` didn't perform updates. Was b/c it used to leak, but that's no...
Interestingly, the Cling code suggests that this 'Error' is only a warning. I presume this is coming from the dictionary generation, but although I do know that it stores the...
From what I can tell, Cling has hard-coded the enabling of coloring (and is using its own scheme). I've asked upstream whether there's a way of disabling it.
Yes, upstream confirms the hard-codedness. :( I'll have to patch in in cppyy-backend. If you care about it right now, it's in `CIFactory.cpp`, line 1432. Or here, from upstream: https://github.com/root-project/root/blob/master/interpreter/cling/lib/Interpreter/CIFactory.cpp#L1414
cppyy checks the CPU and adds `-mavx` to the Cling CLI arguments if the processor supports it. You can override this behavior by setting the envar `EXTRA_CLING_ARGS` (which defaults to...
> Wouldn't it be better to 'look at' the CPU the process is running on and figure out what it can provide? That's what I meant with "if supported." :)...
Yes, `-march=native` does the trick AFAICT, I'll replace the CPU feature check with it and see whether the math folks notice/complain. :) Completely aside, in the other thread, you mentioned...