grpc: unfuck abseil once again

srcpkgs/grpc/patches/google-sucks.patch

@@ -5,6 +5,7 @@ An all-in-one patch that fixes several issues:
 3) powerpc stacktrace implementation only works on glibc (disabled on musl)
 4) powerpc stacktrace implementation has ppc64 assumptions (fixed)
 5) examine_stack.cpp makes glibc assumptions on powerpc (fixed)
+6) randen_hwaes does not build on all ppc (fixed)
 
 --- third_party/abseil-cpp/absl/base/internal/unscaledcycleclock.cc
 +++ third_party/abseil-cpp/absl/base/internal/unscaledcycleclock.cc
@@ -109,3 +110,13 @@ An all-in-one patch that fixes several issues:
        // Check that alleged sp before signal is nonnull and is reasonably
        // aligned.
        if (sp_before_signal != nullptr &&
+--- third_party/abseil-cpp/absl/random/internal/randen_hwaes.cc
++++ third_party/abseil-cpp/absl/random/internal/randen_hwaes.cc
+@@ -141,6 +141,7 @@ ABSL_NAMESPACE_END
+ #include <altivec.h>
+ // <altivec.h> #defines vector __vector; in C++, this is bad form.
+ #undef vector
++#undef bool
+ 
+ // Rely on the PowerPC AltiVec vector operations for accelerated AES
+ // instructions. GCC support of the PPC vector types is described in: