On Friday September 30 2016 10:10:40 Lawrence Velázquez wrote:
I coined this admittedly a bit wild idea the other day that all of ${prefix}
could be hosted on a R/W dynamic disk image, aka a sparse image or sparse
bundle. That would make it possible to install the more "unixy" ports to a
case-sensitive filesystem without need for guiding the user through the steps
required to create a case-sensitive volume. It would also provide an
alternative to those of us who want to use the default /opt/local prefix but
still install MacPorts on a different volume. Rather than replacing /opt/local
with a symlink after the initial install, we'd simply give the desired location
for the image, and keep /opt/local as its mount point.
I also like the idea of being able to take the whole MacPorts tree offline with
a simple command.
I do something comparable with my "LinuxPorts" adaptation; it resides on a
series of nested ZFS datasets.
> >> This sound convoluted. Also remember that MacPorts is not confined to
> >> installing files in ${prefix}.
> >
> > A tad, maybe. Anything that gets installed outside of ${prefix} is
> > largely out of control, but it's probably also safe to say that those
> > files are where they are because they're somehow specific to the OS
> > and thus don't make assumptions about filename case.
Creation:
%> hdiutil create -type [SPARSE|SPARSEBUNDLE] -imagekey sparse-band-size=16384
-fs JHFS+X -layout NONE -volname MacPorts -uid 0 -gid 0 -o ${imgFileName}
Mount:
%> hdiutil attach ${imgFileName} [-kernel] -mountpoint ${prefix} -nobrowse
-owners on -noautoopen
-kernel may or may not have advantages (no helper process) but only works with
SPARSE (UDSP), not SPARSEBUNDLES. Contrary to what I expected, a SPARSEBUNDLE
isn't slower than a sparse image, on the contrary even *). It also has the
advantage that it can be compressed with afsctool, periodically (and when
unmounted, evidently).
The attach command can be put in a "RunAtLoad" launchd plist in
/Library/LaunchAgents, to avoid having to mess with /etc/fstab .
*) In fact, the XBench-1.3 disk-only score was about 2x on mounted the sparse
bundle than on the volume hosting the bundle. Go figure ...
_______________________________________________
macports-dev mailing list
[email protected]
https://lists.macosforge.org/mailman/listinfo/macports-dev
