[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#341882: gcc-4.0: [mips] support for tri-arch on mips & mipsel



On Sat, 3 Dec 2005, Daniel Jacobowitz wrote:

+! MULTILIB_DIRNAMES = n32 32 64
+! MULTILIB_DIRNAMES = 32 . 64

Why bother?

Stuff in rules.d (binary-libstdcxx.mk) assume that $biarch means put it
in a subdir name "64", and that $biarch32 means put it in a subdir named
"32".

        : # remove precompiled headers
        -find $(d) -type d -name '*.gch' | xargs rm -rf

ifeq ($(biarch),yes)
        mv $(d)/$(PF)/lib64/lib*c++*.{a,so} $(d)/$(gcc_lib_dir)/64/.
endif
ifeq ($(biarch32),yes)
        mv $(d)/$(PF)/lib32/lib*c++*.{a,so} $(d)/$(gcc_lib_dir)/32/.
endif

but, we end up with subdirs names "n32" and "64", which causes

: # remove precompiled headers
find debian/tmp -type d -name '*.gch' | xargs rm -rf
mv debian/tmp/usr/lib64/lib*c++*.{a,so} debian/tmp/usr/lib/gcc/mipsel-linux-gnu/4.0.3/64/.
mv debian/tmp/usr/lib32/lib*c++*.{a,so} debian/tmp/usr/lib/gcc/mipsel-linux-gnu/4.0.3/32/.
mv: when moving multiple files, last argument must be a directory
Try `mv --help' for more information.
make[1]: *** [stamps/08-binary-stamp-libstdcxx-dev] Error 1

Changing MULTILIB_DIRNAMES to match that assumption seemed less
intrusive.


+ #ifdef GPLUSPLUS_TOOL_INCLUDE_DIR
+     /* Pick up GNU C++ target-dependent include files.  */
+-    { GPLUSPLUS_TOOL_INCLUDE_DIR, "G++", 1, 1, 0 },
++    { GPLUSPLUS_INCLUDE_DIR "/" TARGET32_MACHINE, "G++", 1, 1, 0, 32 },
++    { GPLUSPLUS_INCLUDE_DIR "/" TARGET64_MACHINE, "G++", 1, 1, 0, 64 },
++#endif

Do you have any evidence that this is necessary?

No, in fact, when I went back to look at this part again, I realized
that I am not applying this patch, but neglected to remove it from my
tree, so it inadvertantly ended up in my diff. Please consider
biarch-include-mips.dpatch to be removed from the propsed patch (I'll
send an updated patch when we're done poking at the first one).



                                Stuart

Stuart R. Anderson                               anderson@netsweng.com
Network & Software Engineering                   http://www.netsweng.com/
1024D/37A79149:                                  0791 D3B8 9A4C 2CDC A31F
                                                 BD03 0A62 E534 37A7 9149



Reply to: