Re: Patch for config.guess for "cross-compiling" x86_64 -> powerpc on Mac OS X <= 10.6

[Earnie]

On 4/18/2017 2:06 PM, Mojca Miklavec wrote:
> Hi,
>
> On 18 April 2017 at 18:47, Jan Engelhardt wrote:
>> On Tuesday 2017-04-18 17:44, Mojca Miklavec wrote:
>>> On 18 April 2017 at 17:27, Jan Engelhardt wrote:
>>>>
>>>> That will invoke /usr/bin/powerpc-unknown-linux-gnu-gcc etc., which
can either
>>>> be the PPC compiler, or a shell script invoking an existing
native-by-default
>>>> gcc with more arguments.
>>>
>>> The problem of this approach is that the configure scripts etc. then
>>> assume that the computer where you run the configuration is not able
>>> to run the resulting binaries.
>>
>> Yes, that is a bit unfortunate -- but usually points to a problem
>> *elsewhere*. If a software package needs, say, an utility to generate
>> more source code from an IDL [think protobuf], the package in
>> question should ensure it uses the native compiler for these things.
>> I also acknowledge that autotools does not expose the native CC very
>> well.
>
> No, the idea is not to use the native compiler.
>
> From a random example using AC_TRY_RUN([...]) the configure script
> compiles the code for the host platform, runs it and checks whether
> the resulting binary crashes. The idea is to check the binary for
> *host* platform, not the guest.
>
> The autotools need to know whether the will actually run the code or
> not. When cross-compiling, the autotools cannot simply run the
> resulting binary and the result has to be guessed or a database of all
> features of all platforms needs to be present. When compiling
> natively, the code can easily run.
>
> If I use --host=powerpc-<whatever> on x86_64, how can autotools decide
> whether or not powerpc binaries can run or not?
>
>>> Building for i386 or ppc is not really cross-compiling in the
>>> traditional sense. An x86_64 machine can still run the ppc binaries
>>> without problems
>>
>> Last time I tried, x86_64 chips only support the i686 and AMD64 ISAs
>> natively. For something like ppc, one would need a binfmt_misc(5)
>> entry with some qemu magic attached.
>
> On Macs (10.4 - 10.5) it works automatically without any qemu magic
> (there is some rosetta-something in the background, but that's fully
> transparent for users).
>

Forgive the brief interlude without any trial of the idea but if you
specify the --build the same as the --host for this scenario does it
work for you?

If I understand correctly you're looking for something similar to
executing a 32bit binary in a 64bit world where the OS adjusts for the
32bits in its 64bitness.  The idea is that specifying the --host and
--build to be the same triplet the configure would execute the resulting
binary in tests and the compiler choice would be based on the --host
triplet.

>
> But in order not to get too off-topic: is anything about my proposed
> patch actually problematic?
>

I haven't looked at the patch, I don't know.

-- 
Earnie