I would have highly preferred if you had forked this on Bitbucket, making
it possible for me to integrate your changes. I see no good reason why you
decided to eliminate history (including authorship) by creating a new
repository from scratch.

I'm also unhappy with the fact that you have applied a license of your
choice to the files without consulting with their original author.

I'm uncertain why you picked 'gcc' as the compiler of choice under OS X
when for a couple of years now Apple has deprecated GCC.

Overall, the intent is appreciated, although the way you went about it all
is odd.


On Mon, Mar 31, 2014 at 9:00 PM, Enry the Ermit <[email protected]> wrote:

> I made a start on a macos brew script based on yours. You can fetch it
> here :
>
> https://github.com/enrytheermit/gnustep/tree/master/gnustep-android
>
>
>
> Regards,
>
> Enry
>
>
>
> Dnia Poniedziałek, 31 Marca 2014 16:26 Mathias Bauer <
> [email protected]> napisał(a)
>
> Hi Ivan,
>
>
>
> first of all, thanks for sharing your work. I think I understand what
>
> you did, nevertheless following your instructions with just two changes
>
> (see below), I get an error at the very last step that until now I don't
>
> understand (mainly because I don't see where it comes from).
>
>
>
> Am 30.03.14 23:58, schrieb Ivan Vučica:
>
> > On Sun, Mar 30, 2014 at 10:50 PM, Mathias Bauer <[email protected]
>
> > <mailto:[email protected]>> wrote:
>
> >
>
> >
>
> >     Currently I'm trying to convince the "configure" of gnustep base to
> use
>
> >     the Android toolchain and find the freshly built libobjc2 for
> linking.
>
> >
>
> >
>
> > This *should* work out of the box on a clean Ubuntu 12.04 machine:
>
> > https://bitbucket.org/ivucica/gnustep-android
>
> > I don't know how well it'll work on a non-clean machine, and you should
>
> > definitely open the scripts and see what they're doing. For example, you
>
> > may want to modify them to avoid downloading another copy of Android SDK
>
> > and NDK (which is, IIRC, what these scripts will try to do).
>
>
>
> I made two small changes: I added the following parameters to the
>
> configure call of gnustep make:
>
>
>
> --enable-objc-nonfragile-abi --with-layout=gnustep
>
>
>
> and the following one to the configure call of gnustep base:
>
>
>
> --disable-mixedabi
>
>
>
> I don't know if they are still needed nowadays, but at least (especially
>
> because of the old compiler used - clang 3.3) I wanted to make sure that
>
> this won't create problems. Nevertheless the error in the last step
>
> happened also when I omitted these changes.
>
>
>
> I consider the results a bit "creepy", as "make install" will put
>
> Android binaries to places where X86 binaries belong, but for the time
>
> being I can ignore this. I assume that for a proper cross compiling we
>
> need to set up a kind of "sysroot" where e.g. GNUstep and libobjc2 get
>
> installed to and where other libraries can find them.
>
>
>
> When I build gnustep base, it seems that the exception handling is not
>
> detected properly:
>
>
>
> > ../../Headers/Foundation/NSException.h:48:2: error: "gnustep-base is
> configured to use 'traditional' exceptions, but you are
>
> >       building for 'native' exceptions."
>
> > #error  "gnustep-base is configured to use 'traditional' exceptions, but
> you are building for 'native' exceptions."
>
>
>
> Having a look into the configure output, I detected several scary
>
> warnings, so all in all I don't think that this is usable now. I will
>
> continue to investigate this.
>
>
>
> Regards,
>
> Mathias
>
>
>
> _______________________________________________
>
> Discuss-gnustep mailing list
>
> [email protected]
>
> https://lists.gnu.org/mailman/listinfo/discuss-gnustep
>
>
>
>
>
>
>
>
> _______________________________________________
> Discuss-gnustep mailing list
> [email protected]
> https://lists.gnu.org/mailman/listinfo/discuss-gnustep
>
>


-- 
Ivan Vučica
[email protected]
_______________________________________________
Discuss-gnustep mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/discuss-gnustep

Reply via email to