This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

RE: [rfa/rfc] Build libcommon.a for gdb and gdbserver


  Hi all,

  this is not enough for my
--host=x86_64-w64-mingw32
configuration to be able to compile libcommon.a:

  CC is now set correctly to x86_64-w64-mingw32-gcc,
but the compilation uses ALL_CFLAGS which is not in FLAGS_TO_PASS
which means that CFLAGS is not added to the compilation options.
  This means or instance that -D__USE_MINGW_ACCESS
is not passed to common subdirectory (no idea if
this has any influence yet, but it might in the future...)

  The locally compiled intl library is not found,
and on mingw target you often don't have a system libintl.h
so this is the reason of my failure for x86_64-w64-mingw32

  On my mingw32 native build, I have a compiled libintl.a
in ../../intl directory, but the compilation reads
the system installed libintl.h..

  INTL_CFLAGS should thus also be handled to supports
targets that have no system libintl and thus 
create one in ../intl directory.
intl source directory should probably be handled just like bfd 
library is if a library has been compiled.

  I was also wondering about what happens to gdbserver if
you use --target= x86_64-w64-mingw32
only, but the gdb/configure script refuses
to build gdbserver if ${host} is different from ${target},
so that this is not an issue.
  I was wondering if there was a real impossibility to do this
kind of cross-compilation, but that is another topic,
by far less urging as the problems I encounter currently
which prevent successful generation of gdb from cvs HEAD.

  Sorry for more claiming than helping out ...

Pierre



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]