[Lldb-commits] [PATCH 3/3] Rearrange #include's, initialization/termination.

[Stephen Wilson]

Hi Greg,

Greg Clayton <gclayton at apple.com> writes:
> Stephen, let me know what issue you run into when tryingo t build these?

OK.  I definitely jumped the gun on this one -- I thought the code in
Process/Utility was a bit more interdependent than it is :)

Does the following patch look OK to you?  This is the only issue I found
among UnwindAssemblyProfiler-x86, ArchDefaultUnwindPlan-x86, and
ArchVolatileRegs-x86.

The only remaining problem is how to get the Makefiles working for this
directory.  I will take a closer look at the LLVM Makefiles, but I do
not remember seeing a way to exclude files within a directory from the
build.

Would it be unreasonable to split up Process/Utility with a few more
logical subdirectories in which we place platform-specific components if
it turns out the current Makefile system cannot be pushed into dealing
with the layout as is?


Thanks!


diff --git a/source/Plugins/Process/Utility/UnwindAssemblyProfiler-x86.cpp b/source/Plugins/Process/Utility/UnwindAssemblyProfiler-x86.cpp
index 6b64886..61e3c70 100644
--- a/source/Plugins/Process/Utility/UnwindAssemblyProfiler-x86.cpp
+++ b/source/Plugins/Process/Utility/UnwindAssemblyProfiler-x86.cpp
@@ -844,11 +844,11 @@ UnwindAssemblyProfiler_x86::FirstNonPrologueInsn (AddressRange& func, Target& ta
 lldb_private::UnwindAssemblyProfiler *
 UnwindAssemblyProfiler_x86::CreateInstance (const lldb_private::ArchSpec &arch)
 {
-   uint32_t cpu = arch.GetCPUType ();
-   if (cpu != CPU_TYPE_X86_64 && cpu != CPU_TYPE_I386)
+    ArchSpec::CPU cpu = arch.GetGenericCPUType ();
+    if (cpu != ArchSpec::eCPU_x86_64 && cpu != ArchSpec::eCPU_i386)
        return NULL;
 
-   return new UnwindAssemblyProfiler_x86 (cpu == CPU_TYPE_X86_64 ? k_x86_64 : k_i386);
+   return new UnwindAssemblyProfiler_x86 (cpu == ArchSpec::eCPU_x86_64 ? k_x86_64 : k_i386);
 }