[Openmcl-devel] pioneers are the ones with the arrows in their backs

[Gary Byers]

On that pleasant note:

The Linux/x86-64 port of OpenMCL isn't really usable or releasable
yet, but a lot of stuff does seem to work.  So far, what works has
worked for me on a couple of platforms, and it'd be helpful to get
some sense of whether or not the same stuff works on platforms that
I don't have access to.  (Most of the development's been done on an
Athlon 64 X2 running Fedora (FC4 and FC5), and I've also used a
Turion 64 laptop running a prerelease of Ubuntu 6).

So, if you:

  - have access to an x86-64 system running Linux (especially some
    other distribution besides those I've used)

  - don't mind the occasional arrow in the back (backtrace doesn't
    work, there are known thread-safety/GC issues, floating-point
    exceptions aren't handled, GC integrity checking code is enabled
    and slowing things down, other things ...)

  - are fairly comfortable building OpenMCL from sources + a heap
    image + interfaces

  - want to see if it "works" with your hardware/Linux distro (where
    "works" means "prints a Welcome message and a prompt, compiles
    itself with some warnings, and isn't broken in some obvious and
    unknown way")

please let me know.

Most x86-64 Linux distributions seem to be fairly "modern" (a recent
2.6 kernel, thread-local-storage enabled in the toolchain and libraries,
NPTL instead of LinuxThreads for pthreads, etc.) and the general idea
is that things should work if these features are present.  If the port's
somehow depending on things that it doesn't need to depend on, it'd
be helpful to try to catch that sort of thing early.