mips-elf with newlib

[Michael N. Moran]

Kai Ruottu wrote:
> Michael N. Moran kirjoitti:
> 
>> Other than some issues with makeinfo, I had no problem building and
>> installing binutils-2.16.91.0.7:
>> 
>> 
>> # mips-elf-ld -V GNU ld version 2.16.91.0.7 20060317 Supported
>> emulations: elf32ebmip
>> 
>> However, building GCC 4.1.1 gave me the same error:
>> 
>> mips-elf-ranlib ./libgcc.a mips-elf-ranlib: ./libgcc.a: Malformed
>> archive
> 
> Really odd...  My build GCC for the binutils, GCC etc. Linux/x86
> stuff has been for a while a self-built (cross) gcc-3.2.3 for the Red
> Hat 7.3 target, although running it now on a Pentium-III with SuSE
> 10.1/i586 distro.  So the "compiling GCC" should be quite near yours
> own...
> 
> But tools like 'bison', 'flex' etc. on the $build system could have
> their influence....  I have attached the big-endian-hard-float
> 'libgcc.a' I just got so you could compare yours with it, for
> instance see what your 'ranlib' could say about it....

# /tools/gnu/gcc/4.1.1/mips-elf/bin/mips-elf-ranlib libgcc.a
/tools/gnu/gcc/4.1.1/mips-elf/bin/mips-elf-ranlib: libgcc.a: Malformed
archive

That makes it look suspiciously like my binutils build may have
a problem.

Indeed, I just tried your libgcc.a using the mips-elf-ranlib
produced on my Fedora Core 2 machine with an identical configure
script and binutils version (2.16.91.0.7), and the ranlib worked
find (no Malformed archive).

GCC 4.1.1 is still building on that (rather slow) machine. I'll
post the results tomorrow, but I'm guessing the results will
be different. Hmmm.

> Your build system sounds being something like the 'Red Hat Linux 9'
> and if so, it expected to be stable and working.

It *is* RH9, modulo some manual package "upgrades" over the
past 5 years.


-- 
Michael N. Moran           (h) 770 516 7918
5009 Old Field Ct.         (c) 678 521 5460
Kennesaw, GA, USA 30144    http://mnmoran.org

"So often times it happens, that we live our lives in chains
   and we never even know we have the key."
The Eagles, "Already Gone"

The Beatles were wrong: 1 & 1 & 1 is 1




--
For unsubscribe information see http://sourceware.org/lists.html#faq