[Openmcl-devel] Alpha-testing 0.14

[Gary Byers]

If you're interested in alpha-testing OpenMCL 0.14 (and are comfortable
with the circular bootstrapping process), I put recent Darwin binaries
and a terse README file at:

<http://clozure.com/testing>

which is also reachable as:

<ftp://clozure.com/pub/testing>

You can get sources via CVS, using the "Pre_0_14" tag.

0.14 isn't binary-compatible with previous versions; you'd presumably
want to keep separate "ccl" trees for 0.14 and 0.13.x, and may want to
copy and modify your installed "openmcl" shell script to make it
easier to invoke 0.14 (and still possible to invoke 0.13.x).

One of the things that's missing is a big list of what's missing ...

The most obvious changes are related to the multiprocessing API (which
has never been documented in OpenMCL, but should have been.)  There's
some functionality that's intentionally missing, some that's unintentionally
missing, and some that should be changed/removed but hasn't been.  This
should all be written down, but hasn't been (yet.)

Native threads expose resource-contention issues, some of which can be
quite subtle (and some of which are pretty blatant.)  I've caught a lot
of them, but it's reasonable to assume that more remain.  Sometimes,
the symptoms of these issues are hard to relate to the cause; if you
experience strange behavior (other than missing/broken functionality)
please try to reproduce and/or report it.

In general, 0.14 is still something of a moving target: fairly significant
changes are still made frequently, and those changes are sometimes hard
to bootstrap.  I'll try to keep fairly recent binaries available in the
/testing directory; those binaries may not match the CVS version, but
should usually/ideally be able to compile it.

I've been using 0.14 regularly for several months now; in my experience,
it's not catastrophically buggy.  I use a fairly small subset of the
lisp (the parts exercised by frequent recompilation of everything ...);
if you exercise a different subset, your experience may be different.

Please report bugs, and please let me know if there's anything I can
do to make this process smoother.  Thanks.

Gary Byers
gb at clozure.com


_______________________________________________
Openmcl-devel mailing list
Openmcl-devel at clozure.com
http://clozure.com/cgi-bin/mailman/listinfo/openmcl-devel