Re: [Bug 514579] Re: gcc-4.4 cross build for m68k on i486 fails
On Thu, 17 Jun 2010, Thorsten Glaser wrote:
> On Wed, 16 Jun 2010, Matthias Klose wrote:
>
> > - multiarch shouldn't be disabled
> > - the selection which multilibs to build should be done by patching gcc/config/m68k/*
>
> Hrm. He didn’t really answer my questions about the c++ symbols
> problem, but this would help us further along at least.
>
> Does anyone have an idea about the symbols thing?
> • LP bug: https://bugs.launchpad.net/ubuntu/+source/gcc-4.4/+bug/514579
> • build log: http://launchpadlibrarian.net/44949657/gcc44m68k.log.gz
> (338K gzip’d, look at the end)
> • Thread: http://lists.debian.org/debian-68k/2010/04/msg00025.html
>
> What multiarch targets do we want to build?
Hopefully, all those targets available to the etch-m68k compilers. Not
sure whether fidoa was one of those?
This thread is interesting:
http://gcc.gnu.org/ml/gcc-patches/2007-03/msg01069.html
Maybe email the CodeSourcery people who implemented fido support, and ask
them what to do about libgfortran?
>
> I can see use for m68040, dunno about m68060, don’t think mcpu32
> or mfidoa. What about Coldfire – is that a gcc-4.5 thing (I’m
> strictly concentrating on gcc-4{3,4} with libc-2.7 at the moment)?
>
> Does anyone have an idea about the c++ symbols thing?
Not just C++ apparently...
$ grep symbols.doesn.t gcc44m68k.log
dpkg-gensymbols: warning: debian/libgcc2/DEBIAN/symbols doesn't match
completely debian/libgcc2.symbols.m68k
dpkg-gensymbols: warning: debian/libobjc2/DEBIAN/symbols doesn't match
completely debian/libobjc2.symbols
dpkg-gensymbols: warning: debian/libstdc++6/DEBIAN/symbols doesn't match
completely debian/libstdc++6.symbols.m68k
In your discussion with Matthias, I see that !defined(_GLIBCXX_HAVE_TLS)
is given as the cause. As it happens, I've recently opened a couple of bug
reports about backporting TLS [1, 2].
You'll see that I've requested that the patch "m68k-allow-gnu99.diff" be
dropped from both 4.4 and 4.5. I hope this does not harm your work on gcc?
4.3 at least, won't be affected.
Regards,
Finn
[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=586005
[2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=586060
>
> bye,
> //mirabilos
>
Reply to: