Fwd: [RFC, PowerPC, Darwin] Clean up generation of ha16() / lo16() markers

[Ulrich Weigand]

David Fang <fang at csl.cornell.edu> wrote on 17.05.2013 21:34:49:

> > It would be good if you could run some tests that require a native
> > system; in particular, the project/test-suite tests, or the JIT
> > tests, or maybe a clang bootstrap.
>
> I am working towards bootstrapping powerpc-darwin8 (-O0 -fPIC -fno-common

> -no-integrated-as), but am running into miscompilations.  I'm able to
> get 60% into stage 2 without complaints from the system assembler, so
> that's a good sign.
>
> In particular, tblgen is crashing and so is unittests/ADTTests.
> http://llvm.org/bugs/show_bug.cgi?id=15965
> http://llvm.org/bugs/show_bug.cgi?id=15901
>
> Bootstrapping is a long road ahead.

Huh, OK.  I wasn't aware that the target doesn't actually bootstrap
at the moment ...

Well, I guess since there are tests that verify generation of hi/lo
markers at the assembler source level, and those all pass after my
patch, it should probably be OK ...

Bye,
Ulrich