SVN r364352

@craig.topper, hmm, what happens in terms of CG when LAHF/SAHF are not available? I assume its just worse CG as you could spill AH onto the stack and do a load/store. This actually results in library calls which may not be possible to fulfill.

Move test case around

add additional context and test case

@lebedev.ri - sure, I will add a driver test to ensure that the feature is set on the command line when invoked from the driver, however, I don't think that there is really much in terms of testing that you can do for this type of stuff other than throw a large corpus at it.

Could you please add a test to ensure that Darwin defaults to the old behaviour?

SVN r363440

What I am imagining is basically having a variable which specifies the distribution targets, and having variables which will be constructed based upon those names (similar to how variables for the runtimes are passed today).

Would be nice to get someone like @clayborg to chime in, but, I think that @labath also seems to think that this is fine.

SVN r362425

SVN r362424

This should get the build working again, so lets get this fixed, we can improve it later

Generally, clang-format the changes, it will catch the formatting things.

Oops! Nice catch. This matches the specification.

Actually, I think that we should extend CompilerType and TypeSystem to expose Clang's knowledge of whether a type is passed in a register by means of using clang::RecordDecl::isPassInRegisters

Thanks for adding the test case!

Thanks, this looks much better.

Sounds safe enough to me. I think that it would be nice to add a test that checks the hexagon flags as well.

This looks good

SVN r361512

I think that this makes sense - the one case that I had in mind is pretty convoluted and if it breaks, we can fix it then.

SVN r361443

SVN r361307

@smeenai - not sure what you mean; you mean a clean configuration with the same set of parameters? I'll certainly do that test and expect no differences there. @hintonda - yeah, CACHE ... FORCE should do the trick for that.

SVN r361210

The only thing that Im worried about is cross-compilation accidentally picking up the libc++ headers that are built here (e.g. building on Linux for Linux with the host libc++ headers and building libc++ as a runtime). The rest looks pretty much a substitution and addition of runtime-libcxx-headers target to the runtimes build which is fine.

Fix inclusion