[Dev] Tiger switch test?

[Andi Vajda]

Getting python 2.4 built on 10.4 with gcc 4.0.0 seems to be quite another 
matter. After fixing up more bogus links in /usr/lib this time, I got a 
limping python built but it's complaining a lot about missing sem_ 
implementations and there are bunch of compile errors in extension modules 
that I didn't look into.
At this point, I don't know if it's a gcc 4.0.0 issue or a Mac OS X 10.4 issue 
(or both) but a fair amount of debugging is necessary before python can build.

Andi..

On Tue, 3 May 2005, Andi Vajda wrote:

>
>> Well, I tried a full build, but didn't get past persistence/db:
>> 
>> /usr/bin/libtool: for architecture: cputype (16777234) cpusubtype (0) file: 
>> -lSystem is not an object file (not allowed in a library)
>> make[2]: *** [libdb-4.3.la] Error 1
>> make[1]: *** [build] Error 2
>> make: *** [persistence/db] Error 2
>> 
>
> Seems to build fine for me. Here is what I did:
>  - installed Tiger on a firewire drive, with a 'scratch' install
>  - booted from it
>  - installed Tiger dev tools included on the DVD, omitting gcc 3.3 (only gcc
>    4.0 got installed)
>  - fixed the links in /usr/bin that make gcc/g++/cc/c++ point to an
>    inexistant gcc 3.3 toolchain to point to the gcc 4.0 one
>  - checked out a new osaf tree
>  - cd external/persistence/db
>  - make
>    it built fine, without anything unusual
>
> I didn't do a full chandler build yet. I'm still building gcj from the most 
> recent gcc 4.1 snapshot. For that, I also had to fix another bogus link in 
> /usr/include: /usr/include/stdint.h needs to point at gcc/darwin/4.0/stdint.h
> instead of gcc/darwin/default/stdint.h
>
> So far, 10.4 seems allright. Even Virtual PC 6.0 appears to work just fine.
>
> Andi..
> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>
> Open Source Applications Foundation "Dev" mailing list
> http://lists.osafoundation.org/mailman/listinfo/dev
>