[j-devel] 0.20.2.7

[Peter G. <pe...@ar...>]

This morning's development snapshot (j 0.20.2.7, lisp 0.0.3.6) is up:

    http://armedbear.org/j.zip          (source)
    http://armedbear.org/j-jar.zip      (just j.jar)

This snapshot fixes a couple of buglets brought to light by the release
of Java 1.5 Beta 1, the more serious of which was that the configure
script didn't recognize 1.5 as a valid version number for Java.

In addition, by request, there is now a rather primitive mode for
Objective C, used by default for .m files. (To be honest, the request
was not specifically for a rather primitive mode.) There is also a
downright vestigial mode for assembly language, used by default for
=2Easm and .inc files.

There are still a few open bugs and/or feature requests that I haven't
had a chance to do anything about. With one exception, these don't
appear to be particularly serious.

The exception is the disappearing lines bug:

    http://sourceforge.net/tracker/index.php?func=3Ddetail&aid=3D878031&g=
roup_id=3D55057&atid=3D475785

This bug has been known for a long time and is (for me) entirely
unreproducible. I've seen it with my own eyes exactly once, so I know
first-hand that it's real, but I've never been able to get it to happen
when I try to track the problem down. Since it's Linux-only, I suspect
it may be a bug in Java; it would be interesting to know if it's still
a problem with the 1.5 beta.

Armed Bear Lisp in this snapshot fails 159 out of the 15864 tests in
the GCL ANSI test suite (an improvement from 521 failures out of 14271
tests in 0.20.2.3, the last snapshot for which such statistics were
recorded). This represents a nominal compliance of 98.9977% (loosely
speaking, of course, since, among other things, the test suite itself
does not yet cover 100% of ANSI Common Lisp).

When test suite compliance reaches 99%, the name will officially be
changed to "Armed Bear Common Lisp", provided the implementation is
otherwise complete. Since 20 required functions have still not been
implemented, the name change is not quite as close as 98.9977% would
tend to suggest.

Thanks for your support.

-Peter