how to change encfs installation script to skip macfuse

[Dan Ports]

On Tue, Mar 15, 2011 at 07:18:26PM +0000, Bayard Bell wrote:
> I'm not sure what the general rules are with this, but I've noticed that <port>-devel allows people who are so inclined to get closer to the bleeding edge, presumably forewarned that what they are running is of a markedly different status. Let people work with the -devel branch and decide from there when the 2.1 branch is sufficiently stable to replace 2.0 in the standard port.

Yes, that's an option. It's worth noting that there's also an official
beta release of MacFUSE (v. 2.1.5). I'm not sure offhand what its
benefits over 2.0 are. It doesn't have 64-bit kernel support, which
seems to be what most people are (understandably!) looking for.

We haven't done a -devel version of macfuse before, in part because I
hadn't heard any demand for it, and in part because it wasn't clear
what version that would be. Development has been (and still is,
somewhat) fragmented, with a bunch of random unofficial releases
appearing on forums.

On Tue, Mar 15, 2011 at 03:16:30PM -0500, Ryan Schmidt wrote:
> That's a good idea, to start with anyway. (Later, if 2.1.x is deemed to work well enough, and the original developer shows no signs of updating the "official" version, we could decide to use 2.1.x in the regular macfuse port anyway.)
> 
> My own research on the topic suggests that 2.1.9 is currently an ok version, including 64-bit kernel support, but it seemed that there were several different versions built by different people, but that Tuxera seemed to be taking the lead with macfuse development these days. I have not yet tried it myself; though I greatly want 64-bit kernel support, I also greatly want not to cause kernel panics, which is what some earlier versions would cause.

Yes, there were a bunch of versions floating around for a while that
purported to support 64-bit kernels. Many of them actually worked by
disabling various thread-safety checks without actually being
thread-safe, and thereby wound up causing frequent kernel panics.

It looks like the Tuxera code is better in that respect, although I
believe it uses fairly coarse-grained locking which can limit
concurrency.

Much of the problem here is that it's hard to keep track of these
different releases, which are coming from different people, some of
which work, some of which don't, and some of which are poorly
documented. It would certainly be better for the MacFUSE community
(such as it is) to straighten this out, rather than us...

Dan

-- 
Dan R. K. Ports              MIT CSAIL                http://drkp.net/