[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] Other Sparemint Decisions

Mark Duckworth wrote:
1. Sparemint repositories. Are we ok with a strict 68000 repository
(m68kmint.rpm), 68020-60 repository (m68020mint.rpm) and a coldfire
repository (v4emint.rpm)?

This looks very correct, but something is missing.
Patrice Mandin noticed that in GCC 4.x, the -m68020-60 option produces code for the FPU. If I'm not wrong, there are some Falcons with a 68030 but not a 6888x FPU. The -m68020-60 binaries will not work on them.

Do we have to add another multilib set for FPUless Falcons ?

2. I will be tearing through packages and fixing them to build with gcc
4.4.2, incrementing their version numbers, upgrading as necessary, etc.
Are we all ok with me taking the lead on this? I don't want to hear
later that people don't think I am qualified or that the work can't be

I totally agree.
It is a shame that your work and Keith's on RPMs have never been integrated into the official SpareMiNT archive.
If someone is not happy with your RPMs, he will patch them and provide new ones.

3. In order to do this effectively we really need gcc 4.4.2, binutils
2.20 and a new mintlib release rpms that can be trusted. Keith is
starting to work on this.

I really believe my latest binutils+GCC patches are correct.
Until now there has been no clue that the MiNT builds from Keith or MiKRO would be unreliable.

However, as I said on the other thread, the most important thing is to have an automatic build procedure for the whole distribution. So when the compiler, the MiNTLib or some fundamental component is updated, we just have to click one button to get all the new binaries.

If all of these things are answered, I should be able to have much of
sparemint built for coldfire in a week or two, and thus we will be able
to get easymint running on coldfire boards with no emulated code.

You are definitely right, let's go on !

Vincent Rivière