Re: Are we on?

I haven't tried these patches,
but it looks like dwarf support
is longing.  On my architecture,
if I try to ups itself, I get the following -

Fatal internal error: attempt to reference nonexistent object (aborting)

But on small simple apps, it works.

I'm running

uname -a

Linux myhost.com 2.6.31.6-166.fc12.x86_64 #1 SMP Wed Dec 9 10:46:22 EST 
2009 x86_64 x86_64 x86_64 GNU/Linux

gcc -v
Using built-in specs.
Target: x86_64-unknown-linux-gnu
Configured with: ../gcc-4.3.4/configure --prefix=/usr/local 
--enable-languages=c,c++ --with-tune=generic --with-arch_32=i686 
--enable-shared --enable-threads=posix --enable-checking=release
Thread model: posix
gcc version 4.3.4 (GCC)


-- 
Stephen


On 12/21/2009 08:54 PM, Callum Gibson wrote:
> On 18Dec09 15:34, Bakul Shah wrote:
> }I'd love to see the FreeBSD port of ups get resurrected.  Any
> }chance to get the patches? Thanks.
>
> I received word back from Tom Hughes. He still makes commits to sourceforge.
>
> Since there has been some interest, here are my patches to get things
> going under FreeBSD/i386 RELENG_7. Should also work on 8.X.
>
> I've attached 2 separate diffs. The first diffs.freebsd, was to get it
> to compile cleanly on FreeBSD, however I believe some of the diffs are
> not FreeBSD specific. In particular, the cleanup around the definition
> of some of the Elf_* variables to avoid some of the weirdness (like Elf32
> functions being defined to Elf64 functions and vice versa) depending on
> the architecture. In any case, it should be harmless to apply the diffs
> under any operating system.
>
> I hasten to add, I have not tried this on FreeBSD/amd64 as yet, but will
> get around to that.
>
> The second file, diffs.fix, addresses some bugs with the new dwarf code,
> mostly in the form of uninitialised memory, but also completing the refactoring
> of a new internal type that was introduced. I was getting crashes when
> expanding struct members until I tracked these down.
>
> Under FreeBSD, you need to use "configure --enable-elf" to make it
> work properly. Most BSD based systems have moved to elf from a.out format
> now whereas ups assumes a.out still.
>
> After running configure you need to edit ifdefs.h and change HAVE_GLOBAL_IOB
> to 0 because FreeBSD's stdio doesn't have an accessible IOB structure for
> ups to frob around in, in spite of what configure thinks. The test for
> "__sF.*" in configure.in is too general for FreeBSD's header. Probably
> "\<__sF\>" would be better.
>
> Finally, you'll need to use "patch -p1" to use these diffs.
>
>      C
>
>