[openbeos] Re: How not to actively prevent a PPC version of your software and write better code at the same time
- From: "Michael Phipps" <mphipps1@xxxxxxxxxxxxxxxx>
- To: openbeos@xxxxxxxxxxxxx
- Date: Tue, 09 Apr 2002 21:58:02 -0400
>> Exactly! That is the thing. This is not PPC or no PPC. This is "PPC
>> with an R5 kernel
>> as a development environment" vs "PPC when the new kernel is ready".
>> I don't think that
>> the added work on everyone's part to add _EXPORT, along with whatever
>> other
>> issues may or may not come up are worth it, RIGHT NOW, to support 5
>> year old HW
>> (which is all that will run R5's kernel).
>
>I named all the issues. I have been taking BeOS x86 GCC software to PPC
>for *years*. I think I am competent to state the issues. Nor is it
>purely an mwcc thing, as I mentioned in my other mail.
Right. Neither is mwcc mentioned. It is a PEF thing. Which is a tiny place to
live in,
right now.
>> IF newer Macs were supported on the R5 kernel OR
>> the change wasn't all of source control wide OR
>> it would be necessary anyway,
>> I would have a different opinion. But to put everyone to extra work
>> for the, what, 5 PPC users on the
>> project who MIGHT write code in the next N months (where N is a
>> single digit), I don't think it is worthwhile.
>
>Hmmm, guess we shouldn't port anything to Alpha, ever, since there are
>no BeOS Alpha users...
And, of course, I never said that. What might make sense is that there is no
impetus for
us to include things that an Alpha port MIGHT need. But, then again, since no
one
can tell what every possible port might need, this makes no sense.
>> Yes. One thing that I would really like to do, someday, is build on a
>> couple of different compilers to see if it works
>> and if any errors fall out. But having one compiler across all
>> platforms is a good thing.
>
>Trust me. You have lots and lots and lots of errors. There is an
>enormous amount of code that has been written with proprietary GCC
>extensions (like not using _EXPORT). The points I raised remove 99% of
>them.
>-Nathan
And I agreed with your points that were reasonable. Good, standard C++ things
that GCC overlooks in its warnings.
Other related posts:
