[PD] Building externals on OSX

[Jonathan Wilkes]

On 10/22/2013 05:34 PM, Dan Wilcox wrote:
> I'm not going to argue OS politics, again use what works best for you. 
> I can understand the frustration with how the build system works on 
> OSX. It is actually really nice in a lot of ways but it took me a 
> while to get the hang of it after switching from Linux.
>
> Without counting Debian, I still think there's really no need at this 
> point to support a PPC build for OSX. When I wrote "drop PPC support" 
> in earlier emails, I was referring only to OSX.

I understand.  It just seemed you were explaining why PPC is no longer 
supported as the natural outgrowth of software development progress.  I 
mentioned Debian both because I'm curious if anyone runs it on PPC, and 
also as a counter-example to the type of development model Apple uses.  
(One which is able to support a much wider variety of architectures at a 
fraction of the cost.)

A longer discussion of OS politics is probably an unnecessary digression 
here: both you and I are building things in OSX, and that won't change 
by pointing out differences in our philosophy for why we do it.  But I 
just wanted to point out that "use what works for you" should expand to 
"use what works for you without creating a moral hazard for all of us".  
One doesn't need to use the language of free software advocates in order 
to do that, but free software is undoubtedly part of any viable solution.

If the internet that connects all these devices weren't actively being 
weaponized I probably wouldn't quibble over this.

Well, I might, but it'd feel more like a pasttime and less like a 
necessity. :)

> The code as it is now should compile just fine on Debian PPC since the 
> only architecture differences as far as I know would be compiling for 
> little endian versus big endian on Intel. I don't think there are any 
> architecture specific assembly / function calls in Pd.
>
> So in the end, dropping OSX PPC support helps in simplifying the build 
> scripts at least. Again, I think there's really no need to host newer 
> Pd OSX PPC binaries, just leave the last one there since anyone using 
> it will be on a much older version of OSX anyway.

Ok, that sounds like the way to go then.

-Jonathan

>
> On Oct 22, 2013, at 5:05 PM, pd-list-request at iem.at 
> <mailto:pd-list-request at iem.at> wrote:
>
>>>>
>>>> Also, do you have any references for the claim that the vast 
>>>> majority of OSX
>>>> users have moved away from PPC?
>>>
>>> http://update.omnigroup.com/ (Hardware / CPU type): Intel 97.8% PPC 2.2%
>>>
>>> https://www.adium.im/sparkle/ (CPU type): Intel 97.83% PPC 2.71%
>>
>> Thanks.  Those are low numbers, but I'd imagine the number of PPC 
>> users is
>> still fairly high:
>> http://www.statisticbrain.com/apple-computer-company-statistics/
>>
>>>
>>>> I find Jobs' claim that Apple doesn't ship
>>>> junk to generally be true, and combined with their development 
>>>> model the
>>>> unfortunate result would seem to be that poor people still using 
>>>> their once
>>>> sleek and sexy devices are ignored along with their now ugly, 
>>>> unprofitable
>>>> devices.
>>>
>>> Well, those "sleek and sexy" PPC devices were last made & sold in 
>>> 2005, so it's not a surprise the vast majority of people using OSX 
>>> have Intel machines mainly because software developers (& the OS) 
>>> have moved on to 32 bit and now 64 bit intel years ago.
>>
>> Debian supports PPC, no?  Anyone know how it does on the old 
>> machines?  I suppose since Pd is in the repos one could say it still 
>> supports PPC. :)
>>
>>>
>>> Your political bias notwithstanding (I say use what works for you),
>>
>> Well, I'd call it a political stance.  And where it seemed quirky and 
>> deeply
>> personal when I first adopted it, it now seems simply to be a restatement
>> of the scientific method for computer security, at a time when there have
>> been revelations that show our computers really need to be as secure as
>> possible against attacks.
>>
>> I'd also point out that yours is a political stance.  While I understand
>> it, I must disagree with it because in terms of security it is much more
>> difficult to use the scientific method to check whether the specs 
>> actually
>> fit the implementation.  In some cases on proprietary OSes neither are
>> known so you're forced to reverse engineer the software, and for
>> complex systems that's too time consuming and expensive to do.
>
> --------
> Dan Wilcox
> @danomatika
> danomatika.com <http://danomatika.com>
> robotcowboy.com <http://robotcowboy.com>
>
>
>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20131024/60a4205f/attachment-0001.htm>