On Thu, Oct 31, 2013 at 1:47 PM, Jonathan Wakely <jwakely.gcc@xxxxxxxxx> wrote: > On 31 October 2013 18:13, Cyd Haselton wrote: >> On Thu, Oct 31, 2013 at 8:25 AM, Jonathan Wakely <jwakely.gcc@xxxxxxxxx> wrote: >>> On 31 October 2013 13:22, Cyd Haselton wrote: >>>> Bug submitted, thanks. >>> >>> Again, please don't top-post on this list. >> >> Again? I don't remember seeing a first request... > > There have been a few requests on this list recently, but not to you, > so the "again" is out of context, sorry. > >>>> Not sure if this is relevant but the symbol table from objdump -x <4.8 >>>> cc1> after adding explicit cast to files.c shows odd characters >>>> >>> >>> The difference is that GCC 4.8 is compiled with a C++ compiler so the >>> symbol names are mangled to encode the function signatures, as >>> required by C++ in order to support function overloading and other >>> features. >> >> Could this have an effect on cross-building a native GCC cc1 binary >> for Android...i.e. make it more problematic or impossible? > > No, it shouldn't be a problem. Hopefully people don't mind bottom posting. Just wanted to mention that, as a test, I replaced the following lines in libcpp/Makefile: # The name of the compiler to use. COMPILER = $(CXX) COMPILER_FLAGS = $(ALL_CXXFLAGS) DEPMODE = $(CXXDEPMODE) with this (from the gcc 4.7 sources): # The name of the compiler to use. ENABLE_BUILD_WITH_CXX = @ENABLE_BUILD_WITH_CXX@ ifneq ($(ENABLE_BUILD_WITH_CXX),yes) COMPILER = $(CC) COMPILER_FLAGS = $(ALL_CFLAGS) DEPMODE = $(CCDEPMODE) else COMPILER = $(CXX) COMPILER_FLAGS = $(ALL_CXXFLAGS) DEPMODE = $(CXXDEPMODE) endif While the cc1 build failed...and I'm not sure if the two are related but I wasn't surprised...the libcpp build completed. I'm not trying to contradict what was said earlier...that the change to compilation with a c++ compiler shouldn't change anything...but I am wondering if there's a c++/g++ option I'm missing that is preventing the build from completing.
