Commits on Jun 22, 2003

            * New optimization to do.
            * Fix problem description for the bug 188 (thanks to cliini on
    Alexey Dejneka committed Jun 22, 2003
Commits on Jun 20, 2003

    	YA instance of insufficient attention to DESCRIBE detail
    William Harold Newman committed Jun 20, 2003

            Bug 258 is not a new bug - it is a longly wanted example of
            bug 148, working under recent SBCL. Thanks, Bill!
    Alexey Dejneka committed Jun 20, 2003
Commits on Jun 19, 2003

    	new bug 258
    	also added some text to try to clarify bug 257 (and hopefully
    		got it right)
    William Harold Newman committed Jun 19, 2003

    	output to proper stream in DESCRIBE-OBJECT (HASH-TABLE T)
    William Harold Newman committed Jun 19, 2003

    	PRINT-UNREADABLE-OBJECT isn't specified to do anything nice
    		with pprint logical blocks, and it's hard to do
    		anything nice without surprising the user, and (as
    		pointed out by Antonio Martinez) it's specifically
    		not supposed to do some of the whitespace stuff it was
    		doing. So just ignore any pretty-streamness and do
    		entirely physical output.
    	tweaked ENCAPSULATE logic so that BACKTRACE reporting of
    		TRACEd functions will be slightly less obscure:
    			1: (.... )
    		instead of
    			1: ("varargs entry for #'(LAMBDA (&REST SB!INT:ARG-LIST) ...)"
    	ruthlessly plundered CMU CL CVS and Gerd's emailed
    		expertise for %NO-PRIMARY-METHOD .ARGS. code
    William Harold Newman committed Jun 19, 2003
Commits on Jun 18, 2003

    	Cosmetic fix for almost invisible bug
    	... %CHECK-VECTOR-SEQUENCE-BOUNDS is always transformed, so
    	use that in its "interpreter" definition; also, that way we
    	don't refer to an undefined variable any more.
    csrhodes committed Jun 18, 2003
Commits on Jun 17, 2003

    	Fix SLOT-MISSING/SLOT-UNBOUND bugs found by Paul Dietz' test
    	... return a single value for SLOT-VALUE, the new value for
    		(SETF SLOT-VALUE), a boolean equivalent for SLOT-BOUNDP
    		and the object for SLOT-MAKUNBOUND
    `	... adjust a bogus test in our regression test suite :-/
    csrhodes committed Jun 17, 2003
  2. (put the lost line)

    Alexey Dejneka committed Jun 17, 2003

            * Fixed DATA-VECTOR-SET-C/SIMPLE-BIT-VECTOR with index 0 on
              X86 (found by Paul Dietz).
    Alexey Dejneka committed Jun 17, 2003

            * Fix bug 15: enable emitting a style warning for redefining
              FTYPE in PROCLAIM;
            * fix bug 46c: uncomment the corresponding checks in COERCE;
            * NOTINLINE does not prevent using function type;
            * write SIMPLE-= method for functions;
            * signal STYLE-WARNING on IR1 transform redefinition;
            * combine conflicting tansformers for %CHECK-BOUND.
    Alexey Dejneka committed Jun 17, 2003

    	merged  Nikodemus Siivola "make slightly less braindead"
    		patch (sbcl-devel 14 Jun 2003)
    	text/comment tweaking
    William Harold Newman committed Jun 17, 2003
Commits on Jun 16, 2003

    	Conditionize COMPILER-NOTE
    	... s/COMPILER-NOTE/COMPILER-NOTIFY/ for the function
    	... a couple of exports in package-data-list
    	(also move a DECLAIM from the start of a MACROLET)
    csrhodes committed Jun 16, 2003

    	Restore type derivation for known/declared functions, even
    	when their use is declared NOTINLINE.
    csrhodes committed Jun 16, 2003
Commits on Jun 15, 2003

    	Reading of symbols in a package named by "" (reported by hmot on
    	#lisp IRC)
    	... the only way one can get one of these beasts is by multiple
    		escapes with nothing in them, so remember if we've seen
    		multiple escape characters
    csrhodes committed Jun 15, 2003
Commits on Jun 14, 2003

            * Merged in Robert E. Brown's patch for the bug 235b;
            * MAKE-ARRAY: in complex case infer a dimension.
    Alexey Dejneka committed Jun 14, 2003

    	A couple of fixes:
    	... SEARCH and test predicate argument ordering: patch from
    		Wolfhard Buss cmucl-imp 2003-06-13
    	... VECTOR-PUSH-EXTEND and type inference: disable MAKE-ARRAY
    		dimension type inferencing for non-simple 1d arrays, as
    		the dimension can change too easily.  Regressions noted
    		by pfdietz' test suite.
    csrhodes committed Jun 14, 2003

            * Add more type checks to the source transforms of equality
              predicates (found under influence of Adam Warner).
    Alexey Dejneka committed Jun 14, 2003
Commits on Jun 13, 2003

    	Merge patch I (+defgenerics.diff) from Andreas Fuchs
    	sbcl-devel 2003-06-10 for sb-bsd-sockets:
    	... add DEFGENERICS (enough to make LOAD-OP quiet, yay!)
    csrhodes committed Jun 13, 2003

    	Grab bag of fixes:
    	... PROFILE on a string (naming a package) shouldn't try to
    		profile macros and special operators any more;
    	... SB-KERNEL, not KERNEL, in "I'm deeply confused" error
    		message strings;
    	... UNBOUND-SLOT is a CELL-ERROR, so use the NAME slot (and
    		delete the SLOT slot :-) (thanks to pfdietz)
    	... delete the INITIALIZE-INFO slot from SLOT-CLASS (observation
    		from Gerd Moellmann)
    	... DESCRIBE on unfinalized classes shouldn't cause an error
    		(reported by kr at molecubotics sbcl-devel 2003-06-13)
    	... fix bug in FORMATTER revealed by shiny new exciting format
    		string for DESCRIBE-OBJECT (CLASS T): original args
    		aren't necessarily available in pretty-printer
    		expansion (specifically, not for "~@< ~:>");
    	... tests for some of the above.
    csrhodes committed Jun 13, 2003

            * Fixed typo on OUTPUT-VECTOR.
    Alexey Dejneka committed Jun 13, 2003

            * SB-CLTL2: first try on VARIABLE-INFORMATION;
            * MAKE-ARRAY: infer array size in complex case;
            * second look at CONCATENATE optimization: create new START
              variable for each sequence. It would be nice to write a
              regression test for
               (time (compile nil
                              '(lambda ()
                                (list (concatenate 'string
    Alexey Dejneka committed Jun 13, 2003
Commits on Jun 12, 2003

    	Patch from Istvan Marko to help make a #+sb-threads runtime work
    	properly: __asm__ wants to be _asm__ __volatile__
    telent committed Jun 12, 2003

            * Declare return type of structure constructors;
            * make some FTYPE declarations to be "exact";
            * GENERATE-TYPE-CHECKS prints too hairy type in the full form;
            * new optimization quality: INSERT-DEBUG-CATCH;
            * avoid using TYPE-SPECIFIER in the VALUES type deriver.
    Alexey Dejneka committed Jun 12, 2003
Commits on Jun 11, 2003

    	Better error reporting for bad types:
    	... signal a simple-error on bare AND, OR, MEMBER or VALUES.
    	... when within the compiler, handle argument count mismatch
    		errors within CAREFUL-[VALUES-]SPECIFIER-TYPE.
    	... (DEFTYPE FOO () 'FIXNUM) (TYPEP 11 'FOO) (TYPEP 11 '(FOO))
    		is legal.
    csrhodes committed Jun 11, 2003

            * bug fix: in macro-like defining macros/special operators the
              implicit block does not enclose lambda list.
    Alexey Dejneka committed Jun 11, 2003
Commits on Jun 10, 2003

    	Fix yet another Dietz bug:
    	... in DEFCLASS, we only redefine an existing class with name
    	CLASS-NAME if said name is the proper name for the class;
    	... ask for a new class (by passing NIL to ENSURE-CLASS-USING-CLASS)
    csrhodes committed Jun 10, 2003

    	Make VALUES derive type optimizer accurate:
    	... since we have this nice interpretation of VALUES types, use
    		it in the VALUES derive-type optimizer -- we will return
    		exactly	as many values as VALUES has arguments;
    	... fix the bugs that this reveals in sbcl; :-)
    	... enables us to detect more bogosity: test for some more type
    		mismatches being caught.
    	... (relatedly) fix one more duplicate definition in
    	Array initializer type warning fix:
    	... don't do (csubtypep (ctype-of x) eltype), because that's wrong
    		for e.g. X = #\a and eltype being STANDARD-CHAR; use
    		CTYPEP instead.
    csrhodes committed Jun 10, 2003

            Minor fixes:
            ... don't define a LENGTH type for disassembly; we're not
            allowed to do that, because LENGTH is in the CL package.  Call
            it DISASSEM-LENGTH instead.  (reported by rtoy on #lisp IRC)
            ... cut'n'pasteo: restart print functions in warm init now refer
    	to variables that are bound, rather than those that aren't.
    csrhodes committed Jun 10, 2003

            * Signal a style warning when DECLAIM is met in a declaration
            * Don't join blocks if the separating continuation's dest is CRETURN;
            * DO-USES: in the restart mode stop iterations when the block
              is deleted under us.
    Alexey Dejneka committed Jun 10, 2003
Commits on Jun 9, 2003

    	another way to break the compiler -- seems to be a bad day
    		 for interaction of my coding style with SBCL
    William Harold Newman committed Jun 9, 2003

    	Contrib fixes for non-x86/non-linux
    	... don't use GET-LISP-OBJ-ADDRESS on fixnums or characters;
    	... don't try to define "posix" linux-specific functions
    	also minor frobs
    csrhodes committed Jun 9, 2003

    	COMPILE-FILE and "static linking"
    	... we are allowed to consider references to functions defined
    	in the same file as such.  At present, we don't inline such
    	references, but merely use previously-derived type information
    	when compiling calls.
    	... also, since the consequences are undefined for multiple
    	definitions in the same file, add a warning for that case (and
    	fix the examples in the codebase itself :-)
    csrhodes committed Jun 9, 2003

    	new BUGS entry
    William Harold Newman committed Jun 9, 2003
Commits on Jun 8, 2003

    	large contrib/simple-streams patch including many CMUCL
    	changes by Paul Foley
    	(Thanks to Rudi Schlatte sbcl-devel 2003.06.08)
    	... melded streams are implemented
    	... removed device-extend; it's deprecated in acl's
    	    simple-streams implementation and will go away there too
    	... prettier print-object methods
            ... various small fixes
    	... Implemented file-position
    	... Removed some sbcl internals redefinition: no need to
    	    rewrite unix-lseek, use sb-posix for mmap, munmap
            ... Can now remove sb-grovel requirement too
    	... Commented out (but not deleted yet) heavily
    	    sbcl-internals-dependent slot access machinery
    	    (def-stream-class, with-stream-class, sm).
            ... Various bug fixes, e.g. read-line now works for lines
    	    longer than 80 characters
    telent committed Jun 8, 2003