Re: DarwinPorts (MacPorts) Advisory: Stay with at least Xcode 2.2.1 on Intel (if not PPC)
Tabitha McNerney <tabithamc <at> gmail.com>
2006-08-27 10:49:11 GMT
On 8/27/06, Ryan Schmidt <subversion-2006c <at> ryandesign.com> wrote:
>
> On Aug 27, 2006, at 11:38, Tabitha McNerney wrote:
>
> > I allow for the possibility that there might be a problem with UFS
> > volumes
> > on an Intel machine. I will also toggle back to DP 1.3.1 (with
> > Xcode 2.2.1)
> > to see if that makes a difference or not.
>
> What do you think you are gaining by using UFS? Apple technical
> support has always actively recommended against the use of UFS. I do
> not understand why Apple engineering ever made the option available.
Ryan, UFS has a rich history with *BSD platforms including NeXTSTEP.
Darwin's implementation of UFS is similar to FreeBSD's with the difference
being that Darwin UFS is big endian regardless if a disc volume formatted
with Darwin UFS is big or little endian (e.g., PPC or Intel). Thus, I can
not in any way believe that MacPorts on Mac Intel hardware running Mac OS X
10.4.5 or higher should have any problem building ports on a volume that is
formatted with the Darwin UFS. Darwin UFS has always been case sensitive and
if I recall correctly, HFSX is still quite new (introduced a year or two
ago?) so when you think about it, go back to when Apple acquired NeXT and
they made Mac OS X 10.0 ... and 10.1 and Jaguar 10.2 -- there was no other
case sensitive file system option besides Darwin UFS. I'm not even sure if
Panther 10.3 had HFSX or not.
Also, HFS+ does not support sparse files (Darwin UFS does!) and if you look
at Apple's evolution of the Disk Utility app, you can see how sparse files
and disc images have come along as a new option when imaging the file
(Continue reading)