Commits on Apr 29, 2003
  1. @telent

    0.pre8.118

    	Fix WITH-MUTEX bug: don't release the mutex if we didn't
    	manage to acquire it in the first place
    telent committed Apr 29, 2003
Commits on Apr 28, 2003
  1. @csrhodes

    0.pre8.117:

    	:ALLOW-OTHER-KEYS NIL is perfectly valid in an initarg list, so
    	don't flame out when it's present (detected by Paul Dietz'
    	suite: test CLASS-07.10)
    csrhodes committed Apr 28, 2003
  2. @csrhodes

    0.pre8.116:

    	Fix for :DEFAULT-INITARGS with side-effects (detected by Paul
    	Dietz' tests, fix from Gerd Moellmann)
    	... ctor needed to be smarter about the separation between
    		locations and initarg equality
    	Also really add the hyperobject tests (logically part of
    	sbcl-0.pre8.115)
    csrhodes committed Apr 28, 2003
  3. @csrhodes

    0.pre8.115:

    	Fix for SLOT-VALUE inside (SETF SLOT-VALUE-USING-CLASS), from
    	Gerd Moellmann (test case a simplified version of KMR's
    	hyperobject)
    	... don't pass a NIL wrapper to the relevant PCL functions
    	... one or two code cleanups and comments
    csrhodes committed Apr 28, 2003
  4. @csrhodes

    0.pre8.114:

    	CHAR, SCHAR and SVREF act on vectors, not multidimensional
    	arrays
    	... so make the source-transform aware of that, for early error
    		reporting (observation from Gerd Moellmann)
    csrhodes committed Apr 28, 2003
Commits on Apr 27, 2003
  1. @kevinrosenberg

    0.pre8.113:

          * src/pcl/slots.lisp: change return value of slot-makunbound-using-class
          to OBJECT rather than NIL to conform to AMOP
    kevinrosenberg committed Apr 27, 2003
  2. @kevinrosenberg

    0.pre8.112:

          - toplevel.lisp: implements toplevel of recursive repl.
    kevinrosenberg committed Apr 27, 2003
  3. @kevinrosenberg

    0.pre8.112:

          - src/code/toplevel.lisp: Remove changes to REPL and rename
            REPL to REPL-FUN and add hook.
          - sb-aclrepl/tests.lisp: Add display tests.
          - sb-aclrepl/toplevel.lisp: New file. Toplevel REPL with support
            for catching signals
          - sb-aclrepl/README: state that sb-aclrepl must be loaded in
            ~/.sbclrc.
    kevinrosenberg committed Apr 27, 2003
  4. 0.pre8.111:

            * fast version of CONSTANT-FOLD-CALL flushes dead combination.
    Alexey Dejneka committed Apr 27, 2003
Commits on Apr 26, 2003
  1. @telent

    0.pre8.110

    	Syntax error in sparc/linux C runtime
    telent committed Apr 26, 2003
  2. @telent

    0.pre8.109

    	#+sb-threads check for Linux 2.4, because it won't work in 2.2
    	In the process, rename early_kernel to linux_sparc_siginfo_bug,
    	just to make its purpose a little more obvious
    
    	Export WITH-TIMEOUT and TIMEOUT from SB-EXT.  TIMEOUT is now a
    	SERIOUS-CONDITION not an ERROR
    telent committed Apr 26, 2003
  3. 0.pre8.108:

    	Check *FEATURES* compatibility in fasl load.
    William Harold Newman committed Apr 26, 2003
  4. 0.pre8.107:

            * Factored out code shared between MAP and MAP-INTO
              optimizers;
            * use one iteration variable for all arrays in the MAP-INTO
              optimizer.
    Alexey Dejneka committed Apr 26, 2003
  5. 0.pre8.106:

    	doc tweaking
    	refactoring in preparation for checking *FEATURES* compatibility
    		in fasl load
    William Harold Newman committed Apr 26, 2003
Commits on Apr 25, 2003
  1. @telent

    0.pre8.105

    	Add #include <errno.h>, for +sb-threads builds on some Linux
    	distros in which the existing include files don't seem to have
    	dragged it in already
    
    	Change the PSEUDO_ATOMIC_ATOMIC assert in alloc to print more
    	detail.  If you run into this bug, sbcl-devel wants to know.
    telent committed Apr 25, 2003
  2. @kevinrosenberg

    0.pre8.104:

        * src/code/toplevel.lisp: Add special variables to convert SB-IMPL::REPL
          into a recursively invokable funcion
        * src/code/debug.lisp: Add hook for SB-DEBUG::DEBUG-LOOP
        * contrib/sb-aclrepl/tests.lisp: add tests for bignum inspection
        * contrib/sb-aclrepl/repl.lisp: convert to use new SB-IMPL::REPL function,
          add some debugger commands
        * contrib/sb-aclrepl/debug.lisp: use SB-DEBUG::*DEBUG-LOOP-FUN* hook.
          however, hook is not yet enabled by default  while debugger function
          continues development.
    kevinrosenberg committed Apr 25, 2003
  3. 0.pre8.103:

            * Added open coding of MAP-INTO for a vector destination
              (reported by Brian Downing on c.l.l)
    Alexey Dejneka committed Apr 25, 2003
  4. 0.pre8.102:

            Compiler issues a full WARNING on calling of an undefined
            function with a name from the CL package.
    Alexey Dejneka committed Apr 25, 2003
  5. @kevinrosenberg

    0.pre8.100:

        * sb-aclrepl changes
           - Update README with examples and contact information
           - Strip out break-stack concept from repl.lisp
           - Fix bug in trimming *history* when *max-history* is reached
           - Add display of single-float, double-float, bignum hexidecimal contents
           - Fix reporting of object addresses (mask lowtag bits)
           - Add ":i slot <id>" command
           - Some non-active experimental code is in toplevel.lisp and debug.lisp,
             but this is #+ignore'd while in development.
    kevinrosenberg committed Apr 25, 2003
Commits on Apr 24, 2003
  1. @csrhodes

    0.pre8.100:

    	As suggested by Fufie (Stig Sandoe) on #lisp IRC, emit a
    	STYLE-WARNING for slots with names that are STRING= but not EQ.
    	... test that we can nonetheless still make them, given the
    		insecurity I feel about my format string wizardry.
    csrhodes committed Apr 24, 2003
  2. @csrhodes

    0.pre8.99:

    	One more -0.0/0.0 fix:
    	... numeric type ranges should be compared with EQUALP, not
    		EQUAL, to make (FLOAT 0.0) be the same as (FLOAT -0.0)
    csrhodes committed Apr 24, 2003
Commits on Apr 23, 2003
  1. @csrhodes

    0.pre8.98:

    	Fix for MEMBER type (as per cmucl-imp 2003-04-23)
    	... (MEMBER 0.0) is not the same as (SINGLE-FLOAT 0.0 0.0);
    	... (MEMBER 0.0 -0.0) is the same as (SINGLE-FLOAT 0.0 0.0)
    	... (NOT (MEMBER 0.0)) needs to be
    		(OR (NOT SINGLE-FLOAT)
    		    (SINGLE-FLOAT * (0.0))
    		    (MEMBER -0.0)
    		    (SINGLE-FLOAT (0.0)));
    	... add some tests for this one.
    	In the process of this fix, make -0.0 and -0.0d0 dumpable by the
    	cross-compiler:
    	... more special cases in src/code/cross-float.lisp.
    	And also let an :SB-SHOW build proceed to the end
    	... don't try to print the slots of PCL objects.
    csrhodes committed Apr 23, 2003
  2. @kevinrosenberg

    0.pre8.97:

          * contrib/sb-rt/Makefile: fixed to install sb-rt as an ASDF system
          * contrib/sb-rt/rt.lisp: renamed from sb-rt.lisp
          * contrib/sb-aclrepl/sb-aclrepl.asd: Reworked to create a separate
            sb-aclrepl-tests ASDF system
    kevinrosenberg committed Apr 23, 2003
  3. @csrhodes

    0.pre8.96:

    	Make ROOM work on cheneygc systems more than half of the time
    	... ensure that we map over allocated objects from the start
    		of the current dynamic (semi-)space, not from the start
    		of the whole dynamic space region.
    csrhodes committed Apr 23, 2003
  4. @kevinrosenberg

    0.pre8.95:

           - Rework sb-aclrepl.asd file to for sb-rt package
           - Rename aclrepl-tests.lisp to tests.lisp
    kevinrosenberg committed Apr 23, 2003
Commits on Apr 22, 2003
  1. @kevinrosenberg

    0.pre8.94:

           - Add missing Makefile so that sb-rt installs
    kevinrosenberg committed Apr 22, 2003
  2. 0.pre8.93:

    	Antonio Martinez pointed out that an EQUAL test is needed for
    		generalized function name lookup.
    	(FORMAT NIL "~S ~S" '1W '1WW) => "|1W| 1WW" seems to be
    		correct after all.
    William Harold Newman committed Apr 22, 2003
  3. @csrhodes

    0.pre8.92:

    	Add SB-RT (a regression tester framework) as a contrib
    	... not with a great fanfare, though, as philosophically it's for
    		contrib use, not general public use;
    	... make sb-aclrepl and sb-bsd-sockets use sb-rt, not their own
    		local copies of rt.
    csrhodes committed Apr 22, 2003
  4. @csrhodes

    0.pre8.91:

    	Fix for ctor MAKE-INSTANCE optimization
    	... don't overzealously assert that a given location is unset;
    		users can supply duplicate initargs.
    csrhodes committed Apr 22, 2003
  5. @csrhodes

    0.pre8.90:

    	Make COPY-SYMBOL work on threaded builds
    	... FAST-SYMBOL-VALUE need not be fast, but must not signal an error
    		on unbound-symbol
    	... add a smoke test
    csrhodes committed Apr 22, 2003
  6. 0.pre8.89:

            * CAST branch seems to take too long time to complete, so:
            ** fixed ENDP.* tests from Paul Dietz' test suit by disabling
              "constant folding" and flushing of IF;
            * fixed type declaration for ENDP as recommended by CSR;
            * ... and ENDP is FLUSHABLE.
    Alexey Dejneka committed Apr 22, 2003
  7. @kevinrosenberg

    0.pre8.88:

          - contrib/sb-aclrepl/rt.lisp: Add pfdietz's eq test to equalp-with-case
          - contrib/sb-aclrepl/aclrepl-tests.lisp: Take advantage of new
            equalp-with-case patch
    kevinrosenberg committed Apr 22, 2003
Commits on Apr 21, 2003
  1. @telent

    0.pre8.87

    	Bug 247 was fixed in 0.pre8.71: update tests and BUGS file to
    	match
    telent committed Apr 21, 2003
  2. @kevinrosenberg

    0.pre8.86:

           - contrib/sb-aclrepl/repl.lisp: improve ld-cmd to allow loads
             such as ":ld ~/foo/bar".
           - contrib/sb-aclrepl/inspect.lisp: Handle cyclic lists.
           - contrib/sb-aclrepl/aclrepl-tests.lisp: add cyclic lists tests.
             Use macros to handle common test patterns greatly reducing
             code duplication.
    kevinrosenberg committed Apr 21, 2003
  3. @telent

    0.pre8.85

    	Merge the definitely-maybe-gc branch:
    
    	Remove all the lispland logic that checks bytes-consed
    	thresholds to see if it's worth GCing: C does these checks
    	anyway.  SUB-GC now always GCs unless *GC-INHIBIT* says not to
    
    	Delete the notify before/after calls and notify-stream.
    	(Planned future change: pass the same information to
    	*{BEFORE,AFTER}-GC-HOOKS* instead)
    
    	*SOFT-HEAP-LIMIT* and MAYBE-GC are gone (unused).
    	*ALREADY-MAYBE-GCING* is also dead, replaced with a proper
    	mutex that will also prevent simultaneous GCs from multiple
    	threads.  This entails moving gc.lisp and purify.lisp after
    	target-thread in the build-order
    
    	(Unithread WITH-MUTEX now honours its WAIT-P argument, so
    	actually works in this situation)
    telent committed Apr 21, 2003