sbcl-1.1.12 released

Christophe Rhodes <[email protected]> Mon, 30 Sep 2013 15:25:43 +0100
Newsgroups gmane.lisp.steel-bank.announce
Message-ID <[email protected]>
Dear all,

I have just released sbcl-1.1.12 to the world; the release notes are
appended to this message.  Of note in this release cycle: I will no
longer pretend that a fix for the type annotation problem referred to in
the last two release messages is imminent, and a launchpad bug has been
filed at <https://bugs.launchpad.net/sbcl/+bug/1229340>, and progress
can be tracked there.  Also, there have been reports of heap corruption
/ garbage collection failures on the 32-bit x86 platform, probably
dating back a few months; if anyone out there is experiencing this and
has a repeatable test case, we would appreciate details.

Best wishes,

Christophe

changes in sbcl-1.1.12 relative to sbcl-1.1.11:
  * enhancement: Add sb-bsd-sockets:socket-shutdown, for calling
    shutdown(3). (thanks to Jan Moringen, lp#1207483)
  * enhancement: document extensible sequences.  (thanks to Jan Moringen,
    lp#994528)
  * optimization: EQUAL and EQUALP transforms are smarter.
    (thanks to Elias Martenson, lp#1220084)
  * optimization: CHAR-EQUAL is faster for constant and base-char arguments.
  * bug fix: probe-file now can access symlinks to pipes and sockets in
    /proc/pid/fd on Linux. (reported by Eric Schulte)
  * bug fix: SBCL can now be built on Solaris x86-64.
  * bug fix: Floating point exceptions do not persist on Solaris anymore.
  * bug fix: (setf . a) is pprinted correctly (reported by Douglas Katzman).
  * bug fix: handle compiler-error in LOAD when it's not run from inside EVAL.
    (lp#1219601)
  * bug fix: SB-GMP:MPZ-POW no longer segfaults given a non-bignum base.
    (thanks to Stephan Frank)
  * bug fix: space allocation of result bignums in SB-GMP is more accurate.
    (thanks to Stephan Frank, lp#1206191)
  * bug fix: sb-safepoint can now reliably handle signal interruptions of
    foreign code. (lp#1133018)
  * bug fix: the compiler-macro for MAKE-INSTANCE when emitting "fallback"
    constructors no longer fails to merge actual and default initargs
    (thanks to Jan Moringen, lp#1179858)
  * bug fix: the compiler-macro for MAKE-INSTANCE when emitting "fallback"
    constructors handles non-KEYWORD initialization arguments more correctly.
  * bug fix: loading the SB-SIMPLE-STREAMS contributed module no longer
    clobbers FILE-NAMESTRING.  (thanks to Anton Kovalenko, lp#884603)
  * bug fix: class definitions with CPLs inconsistent with their metaclasses
    are less likely to destroy the object system's integrity.  (lp#309076)
  * bug fix: restart clause parsing in RESTART-CASE is more in line with the
    standard.  (lp#1203585, thanks to Jan Moringen)
  * bug fix: silence a note from RESTART-CASE under high-SPEED optimization
    settings.  (lp#1023721)
  * bug fix: getting the order of arguments to
    SB-MOP:SET-FUNCALLABLE-INSTANCE-FUNCTION wrong produces a sensible error
    rather than a failed AVER.  (reported by Paul Nathan)
  * bug fix: Parsing of &optional/&key/&rest arguments now never overwrites
    arguments during copying on x86 and x86-64; it may still happen on other
    platforms when there are more fixed arguments than stack slots.
    (reported by Jan Moringen)


------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from 
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk