[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [MiNT] Cross-compiling and RPMs (was Re: g++ bug or is my build enviroment not sane?)
Hello Vincent, rest,
If the only problem is cross-compiling, I can make RPMs on aranym on a quad4
pc, if someone makes the .spec files for me!
Wednesday, October 28, 2009, 12:16:48 AM, you wrote:
> Patrice Mandin wrote:
>> Speaking of RPMs, I would like to cross-build RPMs of uptodate packages
>> for MiNT, but the RPM format does not make things easy for cross
>> compilation. Or maybe I did not find the right infos to do so.
> Hello, Patrice.
> One day I tried to cross-build an RPM, but I went into trouble. I tried to
> use the latest RPM tool, but it produces RPMs incompatible with the (old)
> SpareMiNT one. I also went to the conclusion that RPM is not designed for
> cross-compiling.
> I managed to partially cross-build an RPM of binutils 2.18 last year.
> Basically, I cross-built the binaries, then I switched to ARAnyM in order to
> use the SpareMiNT RPM utility more or less like zip to make an archive. This
> is a very bad method, but it was good enough to build a working binary RPM.
> This experimental work (including build script) is available here:
> http://vincent.riviere.free.fr/soft/m68k-atari-mint/archives/unsupported/
> Note that I made a (small) special effort to make the new binutils 2.20
> buildable with GCC 2.95. So they should be buildable from MiNT, by using the
> standard, good RPM method. In order to do that, a recent version of texinfo
> must be installed, I believe that the RPM built by Mark should be correct.
> It can be found here:
> http://storage.atari-source.org:8000/atari/personal/package_staging/
> About GCC, the patch is perfectly correct, but I never found a clean way to
> build it, because of its interdependencies with the MiNTLib. I never looked
> at how the RPMs of GCC 2.95 were built.
>> Also, all libraries should be built for multilib usage (i.e.
>> 000/020/020-60/coldfire builds).
> I remember you sent to me long ago an additional patch for GCC for enabling
> a multilib for 68020 without FPU. I didn't include it in my patch yet. We
> should decide what set of multilibs should be "officially" supported by MiNT.
> Well, I think that multilib problem is secondary, it could be treated after
> making some a first set of RPMs.
--
Best regards,
George mailto:ggn@hol.gr