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

Re: Bug#341884: libc6: [mips] tri-arch support for mips & mipsel

On Thu, 2 Mar 2006, Aurelien Jarno wrote:

Please note also that I used a modified version of this patch (mainly
because the glibc has changed since you reported the bug) to bootstrap
32-bit and 64-bit glibc/gcc/zlib on mips.

I am glad to hear that you were successful!

You can find all the corresponding packages on:

It looks these are big endian. This nicely compliments the (slightly older)
little endian packages at http://www.helix-hpc.com/apt/.

I know I need to play catch up once again. Sometimes it seems like a new
source package comes out before I can finish the builds from re-syncing with
the previous package 8-).

I should be able to get the LE packages caught again up in the next couple
of days.

+Conflicts: libc6-mips32

Why a confict there? Does such a package has existed in the archive?

No, not in the Debian archive. It existed in my local archive, and was the
result of some package renaming that was done in response to some earlier
comments. This particular Conflict should not be needed in the real packages.

+Provides: libn32c-dev

What about providing lib32c-dev instead?

I can't get to my source right this momment to check, but isn't lib32c-dev
provided by the o32 package?

compared to the other ports, and it also have the advantage for other
biarch packages to use a unique name across all the architectures.

If there is a change that makes things more consistant, I would have
no objection.

Note also that the other architectures does not encode the ABI name in
32-bit or 64-bit packages. I mean that the package is not called for
example libi386c-dev and the libgcc package is called lib32gcc1-dev and
not libi386gcc1-dev.

Some of the unusual names are needed to distinguish between the two 32 ABIs,
o32 and n32.

+# This is needed to keep binutils (ar & ranlib) from getting confused about
+# the file format being used.
+export GNUTARGET=elf64-tradbigmips

I have concerned with this. I understand that is is a workaround
necessary to currently build the glibc, however, I would prefer to have
a fixed binutils instead of that.

I have been remiss in not getting a proper bug filed against binutils for this.
I'll try to get this taken care of tomorrow.


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: