Permalink
Commits on Nov 28, 2004
  1. 0.8.17:

    	release, tagged as sbcl_0_8_17
    William Harold Newman committed Nov 28, 2004
Commits on Nov 20, 2004
  1. 0.8.16.45:

    	made the system stop guessing whether a fooCASE normal-clause
    		might have been intended as an otherwise-clause (and
    		stop issuing STYLE-WARNINGs); this fixes an unreasonable
    		guess reported by Tony Martinez sbcl-devel 2004-11-09
    William Harold Newman committed Nov 20, 2004
Commits on Nov 19, 2004
  1. 0.8.16.44: direct-subclass update protocol bugfix

                * Fixes the issue noted by David Morse of superclasses
                   direct-subclass lists not being correctly updated.
    nikodemus committed Nov 19, 2004
  2. 0.8.16.43: Fixes for various CLOS/MOP bugs

                * Correct canonization of DEFCLASS slot options (also move
                   checking to macroexpansion time of DEFCLASS, and do
                   some trivial reorganization of the defclass.lisp to
                   make the diff bigger then it really is.) (reported
                   by Bruno Haible)
                * (SETF FIND-CLASS) with new-value that is a
                   FORWARD-REFERENCED-CLASS. (reported by Bruno Haible)
                * CLASS-PROTOTYPE signals an error if the class is not yet
                   finalized; also clean up some of the class-prototype
                   machinery a bit. (reported by Bruno Haible)
                * ALLOCATE-INSTANCE for BUILT-IN-CLASS now a) exists b)
                   signals an error.
                * Add tests & record an as-of-yet-unfixed bug.
    nikodemus committed Nov 19, 2004
Commits on Nov 17, 2004
  1. 0.8.16.42:

    	Commit patch from vja (sbcl-help 2004-11-17) for unit enumerations
    csrhodes committed Nov 17, 2004
Commits on Nov 16, 2004
  1. 0.8.16.41:

            * Partial workaround for the bug 262: inline expansion of a
              local function is canceled by conversion of RETURN-FROM,
              referring a deleted CTRAN. (Fixes bug reported by Peter
              Denno on sbcl-devel.)
    Alexey Dejneka committed Nov 16, 2004
  2. 0.8.16.40:

    	Fix for "~VR" nil 5 problem (from PFD ansi-tests)
    csrhodes committed Nov 16, 2004
Commits on Nov 15, 2004
  1. 0.8.16.39:

            * Fix bug in ~^: parameter equal to NIL should mean
              "unsupplied" (found by PFD's test suite).
    Alexey Dejneka committed Nov 15, 2004
  2. 0.8.16.38: Duplicate LOOP bindings

                * Check duplication of all, not just iteration
                   variables; remove needless special-casing
                   for iteration variable binding creation.
    nikodemus committed Nov 15, 2004
Commits on Nov 8, 2004
  1. 0.8.16.37: fixed #351

                * Better error-handling and reporting for malformed LET
                  and LET* forms.
    nikodemus committed Nov 8, 2004
  2. 0.8.16.36: fixed #350

                * Use a bisection algorithm based on suggestion by by Paul
                   F. Dietz for printing bignums, and clean up integer
                   printing a bit while at it.
    nikodemus committed Nov 8, 2004
Commits on Nov 7, 2004
  1. 0.8.16.35:

            Fix FASL loader in the presence of foreign-datarefs on unicode
    
            * Make fop-foreign-dataref-fixup treat the foreign symbol as a
              BASE-CHAR string.
            * Make foreign.test.sh use a FASL file instead of a LOADed .lisp
              file.
            * Incidentally, also fix tests/foreign.test.sh, too: it returned
              success even when there was a failure in the main test.
    antifuchs committed Nov 7, 2004
Commits on Nov 6, 2004
  1. 0.8.16.34:

            * Fix MISC.437: differ necessary and unnecessary component
              reoptimizations; unused code flushing is necassary (for
              variable references).
              ... disable forward optimization pass after running out of
                  reoptimization limit.
    Alexey Dejneka committed Nov 6, 2004
  2. 0.8.16.33:

            * Fix MISC.320: forbid intercomponent reference of a
              DEFINED-FUN-FUNCTIONAL.
    Alexey Dejneka committed Nov 6, 2004
Commits on Nov 4, 2004
  1. 0.8.16.32: Fix #302

                * PRIMITIVE-TYPE used to return (any) for all intersection
                   types. Make it smarter.
    nikodemus committed Nov 4, 2004
Commits on Nov 3, 2004
  1. 0.8.16.31:

    	Oops. Remove bogus :print-function from CLOOP.
    jsnell committed Nov 3, 2004
  2. 0.8.16.30:

    	Fix PFD ansi-tests REAL.1 and REAL.2.
    	... Ow my eyes!  Assorted horribleness.
    csrhodes committed Nov 3, 2004
  3. 0.8.16.29:

    	Fix FORMAT.PAREN PFD ansi-tests for #!+sb-unicode.
    csrhodes committed Nov 3, 2004
  4. 0.8.16.28:

    	Fix CERROR when given a condition datum.
    	... arguments are simply passed to the continue format control.
    csrhodes committed Nov 3, 2004
Commits on Nov 2, 2004
  1. 0.8.16.27:

    	made SXHASH values for CLOS instances be well-distributed
    William Harold Newman committed Nov 2, 2004
  2. 0.8.16.26:

    	THE END (sort of)
    	... add new (enabled-by-default) :sb-unicode feature to
    		base-target-features.lisp-expr;
    	... make the system aware that loading fasls with the wrong feature
    		is a bad idea;
    	... one compiler fix for CHAR-FOO functions (the transforms have
    		to be on portably-unparseable types)
    csrhodes committed Nov 2, 2004
  3. 0.8.16.25:

    	Merge the rest of character_branch under #!+sb-unicode
    	... untested with #!+sb-unicode, but it seems to work OK without.
    		One more build/test cycle to go.
    
    	This patch brought to you by --ifdef
    csrhodes committed Nov 2, 2004
Commits on Nov 1, 2004
  1. 0.8.16.24:

    	What I think are the final commonalities between the current tree
    	and the soon-to-be-added #!+sb-unicode build option.
    
    	This patch was brought to you by character_branch and much puzzling
    	over diffs.
    csrhodes committed Nov 1, 2004
  2. 0.8.16.23:

    	Merge in character_branch runtime changes, protected by #ifdefs
    
    	This patch brought to you by character_branch
    csrhodes committed Nov 1, 2004
  3. 0.8.16.22:

    	Some more changes that can be common to both a widechar sbcl
    	and a narrow one.
    
    	This patch brought to you by character_branch and shell brace
    	expansion
    csrhodes committed Nov 1, 2004
  4. 0.8.16.21:

            * Keep the principle "reanalysis implies reoptimization" on
              the last reoptimization pass (fix bug MISC.436, 438).
    Alexey Dejneka committed Nov 1, 2004
Commits on Oct 31, 2004
  1. 0.8.16.20:

    	A couple more BASE-CHAR-CODE-LIMIT-related fixes:
    	... legal format directives are all base-chars;
    	... define the BASE-CHAR type in terms of the limit constant.
    	Also add remaining test cases to the test suite, just to get them
    	out of the way.
    csrhodes committed Oct 31, 2004
  2. 0.8.16.19:

            * Fix bug MISC.435: disable SUBSTITUTE-SINGLE-USE-LVAR when a
              ref is inside a block to be deleted.
    Alexey Dejneka committed Oct 31, 2004
Commits on Oct 30, 2004
  1. 0.8.16.18:

    	Introduce BASE-CHAR-CODE-LIMIT.  Also make various printer/reader
    	structures scalable.
    
    	This patch brought to you by character_branch
    csrhodes committed Oct 30, 2004
  2. 0.8.16.17

            * Fix bug MISC.434: constraining REF type can terminate a
              block;
              ... block termination immediately marks the successor for
                  deletion.
    Alexey Dejneka committed Oct 30, 2004
  3. 0.8.16.16:

    	Various string-related fixes and harmless changes, mostly
    	apparently cosmetic (but necessary for base-char not equalling
    	character).
    	... tests for dumper behaviour
    
    	This patch brought to you by character_branch
    csrhodes committed Oct 30, 2004
Commits on Oct 29, 2004
  1. 0.8.16.15:

    	Fix BUG #308
    	... use ucd.dat to generate a database for characters with information
    		about graphicness, caseness and the like;
    	... use the database in the ANSI character operators;
    	... (frob the compiler transforms to work with latin-1 characters)
    
    	This patch was brought to you by character_branch
    csrhodes committed Oct 29, 2004
  2. 0.8.16.14:

    	External format support
    	... not latin9, though -- need to think about how that might work
    		in a character-poor sbcl.
    	... delete accented characters in comments from
    		package-locks.impure.lisp -- Something Will Have To Be Done
    
    	This patch brought to you by the letters U, T, F and the number 8.
    csrhodes committed Oct 29, 2004
  3. 0.8.16.13:

    	Add FAST-READ-CHAR optimization, thanks to Teemu Kalvas
    	... except for :io streams.
    
    	This patch was brought to you by character_branch
    csrhodes committed Oct 29, 2004
Commits on Oct 28, 2004
  1. 0.8.16.12:

    	Deal with oversight of ~~ directive in compile-time checking.
    	(Bruno Haible cmucl-imp 2004-10-28)
    csrhodes committed Oct 28, 2004