On Mar 22, 2006, at 12:00 , Dennis Peterson wrote:
Dale Walsh wrote:
On Mar 21, 2006, at 21:32 , Dennis Peterson wrote:
Dale Walsh wrote:
On Mar 20, 2006, at 09:03 , Dennis Peterson wrote:
Torbjorn Granlund, the author of gmp, is aware of the problem
and has stated he's not interested in a fix.
If you google the error "tmp-dive_1.s:98..." you will see core
duo owners are having the same problem.
Yes, I ran into this as well, but it does build if you disable
shared.
If you want to generate a proper fix, I'm sure that it would be
appreciated but one thing to note, I've been unable to build it
as a dual architecture and this would be good to achieve as well.
Building it without shared was a slam-dunk but not especially
helpful. This did the trick:
configure --host=none-apple-darwin
Building for generic x86, reduces functionality but does provide a
working solution, now you just need to build using some additional
flags to get it built right.
It wasn't my requirement to build it for general release - we're
committed to Mac on Intel only going forward and so this works fine
for our datacenters.
What you have generated is marginal at best, doesn't include key
functions required for proper digital signature support but you can
choose to settle for what you have generated in the hopes that you
don't experience any problems.
You have a partial functioning GMP/MP/MPFR library which makes the
digital signature support unreliable, the configure you used doesn't
enable specific features and functions required for full digital
support so if you're really committed to the Intel Mac and providing
proper software support, I'd look into correcting this.
Myself, I'm committed to finding a solution for Macs in general, I am
not narrow minded or biased and only support intel or ppc, I'm
ecstatic if my solutions/improvements work for other platforms but my
goal is ultimately to provide a better build for Macs.
Since the time of this writing I've successfully generated a
darwinbuild project that builds the libraries in the proper format
and includes all of the required functions for proper use.
I'd be able to go one step further and provide 100% functionality and
build compatibility for Intel Macs and other platforms would also
benefit if I could find a solution for the 9 tiny assembler files (by
converting these to regular source) but suffice it to say I have been
able to at least get the basic key functions required to build and
that full digital support is now possible for ClamAV software on
Intel Macs.
dp
-- Dale
_______________________________________________
http://lurker.clamav.net/list/clamav-users.html
