Darren J Moffat wrote:
> John Plocher wrote:
>> There is a drafty opinion in the case dir that Joe, Glenn and Garrett
>> have
>> kindly looked at, but since Roland (the project team) has not yet
>> provided
>> feedback on either "the spec" or "the opinion", and I expect some
>> changes/
>> additions based on his feedback, I am reluctant to push this case to
>> a formal
>> vote tomorrow.
>>
>> I will try to bring things to closure before COB Thursday so this can
>> pop up
>> on the ARC's radar screen in time for next week's meeting.
>
> "32-bit-only processor architectures obviously can not run
> 64-bit code, and should not install 64-bit anything."
>
> That is counter to the current architecture and I think unwise. There
> is no reason why an install done on a 32 bit kernel should not install
> 64 bit binaries. This would actually be a step backwards to where we
> were in Solaris 7.
Yes. But maybe the idea is that there could be some processor
architecture (say ARM) that we wind up supporting where there is no
definition for 64-bits. That's quite different from i386, where 64-bit
support is conditional based upon the specific processor.
>
> The issue is the word "install". Installation isn't IMO part of this
> case so I recommend just dropping that whole sentence.
Probably a good idea.
- Garrett
