coapp-developers team mailing list archive

[Garrett Serack <garretts@xxxxxxxxxxxxx>]

It looks like IA64 is going to have an end-of-life on Windows, so that certainly has a lower priority.

And I really can't argue much with either of your points, and I don't see a compelling reason why we should, other than "wouldn't it be nice if..." 

G

Garrett Serack | Open Source Software Developer | Microsoft Corporation 
I don't make the software you use; I make the software you use better on Windows.


-----Original Message-----
From: William A. Rowe Jr. [mailto:wmrowe@xxxxxxxxx] 
Sent: Friday, April 16, 2010 9:31 AM
To: Garrett Serack
Cc: Olaf van der Spek; coapp-developers@xxxxxxxxxxxxxxxxxxx
Subject: Re: [Coapp-developers] Bundling x64 and x86 binaries in the same library package?

On 4/16/2010 11:16 AM, Garrett Serack wrote:
> Yes, that's pretty much it.
> 
> Well, really:
> 
> C:\program files\common files\CoApp\...
> 
> And
> 
> C:\program files (x86)\common files\CoApp\...
> 
> See http://coapp.org/Blueprints/Packages/4.Shared_Library_Packages
> 
> Although, I just saw a problem with that.
> *******************************
> 
> A developer on a 32 bit system should be able to cross-compile for 64 bit platform. 
> 
> The current spec doesn't account for 64 bit libraries on a 32 bit system. Damn..

Strongly encourage that this *not* become an initial design requirement.

 1) 64 bit is sufficiently cheap and available now that anyone interested in
    targeting 64 bit should be able to acquire a 64 bit box and OS for nothing
    more than investing in a 32 bit box and OS.

 2) various packages create generated sources, using compiled code.  Which leads
    to two build specifications of the HOST and the TARGET architecture, and
    in some cases dual builds of binaries common to the host and target.

The amount of pain to implement is simply not worth it, unless Windows CE, IA64 from amd build environments, or other oddball architectures are being similarly considered.  But for the first go-around?