It's time, once again, for a new SBCL release. SBCL is released once a month, rain or shine, so a release means "it's the end of the month", not "important new features merit a release." That doesn't mean the features aren't important, just that you don't necessarily have to run out and get a new SBCL just because there's a new one.
The NEWS file in the SBCL release chronicles what's new from release to release. Here are the NEWS entries for 1.0.14:
changes in sbcl-1.0.14 relative to sbcl-1.0.13:
- new feature: SB-EXT:*EXIT-HOOKS* are called when the process exits (see documentation for details.)
- revived support for OpenBSD (contributed by Josh Elsasser)
- partially fixed bug #108: ROOM no longer suffers from occasional (AVER (SAP= CURRENT END)) failures .
- fixed bug #402: proclaimed non-standard declarations in DEFMETHOD bodies no longer cause a WARNING to be signalled. (reported by Vincent Arkesteijn)
- bug fix: (TRUNCATE X 0) when X is a bignum now correctly signals DIVISION-BY-ZERO. Similarly for MOD and REM (which suffered due to the bug in TRUNCATE.) (reported by Michael Weber)
- bug fix: SB-SPROF:REPORT no longer signals an error if there are no samples. (reported by Andy Hefner)
- bug fix: functions compiled using (COMPILE NIL '(LAMBDA ...)) no longer appear as (NIL ...) frames in backtraces.
- bug fix: RESOLVE-CONFLICT (and the other name conflict machinery) is now actually exported from SB-EXT as documented. (reported by Maciej Katafiasz)
- bug fix: sb-aclrepl now correctly understands how to inspect single-floats on 64-bit platforms where single-floats are not boxed.
- bug fix: SB-MOP:CLASS-SLOTS now signals an error if the class has not yet been finalized. (reported by Levente Meszaros)
- bug fix: CLOSE :ABORT T behaves more correctly on Windows.
- DESCRIBE and (DOCUMENTATION ... 'OPTIMIZE) describe meaning of SBCL-specific optimize qualities.
You can also watch SBCL's development progress by reading Planet SBCL.