3 * filed off the roughest edges (or, perhaps, at least hammered down the
4 protruding rusty nails and snipped off the trailing razor wire,
5 leaving some filing for later:-) from the monster
6 EVAL/EVAL-WHEN/%COMPILE/DEFUN/DEFSTRUCT cleanup:
7 ** substantially rewrote DEFSTRUCT implementation to work
8 cleanly with EVAL-WHEN, not sleazily use DEFUN for
9 structure functions, implement out-of-line structure
10 accessors as closures, reduce or eliminate non-ANSI
11 magicality of structure functions
12 *** made structure type tests work again
13 *** got rid of bogus warnings about "redefinition" of
15 ** made inlining DEFUN inside MACROLET work again
16 ** made %COMPILE set up debugging data more like the way the
17 debugger expects (and maybe even completely
19 * incompatible changes listed in NEWS:
20 ** changed debugger prompt to "5]", "5[2]", "5[3]", etc.
21 ** changed default output representation of *PRINT-ESCAPE*-ed
22 unprintable ASCII characters to #\Nul, #\Soh, etc.
23 * some easy FIXMEs with high disruptive potential:
24 ** Search lists go away.
25 ** Grep for ~D and and change most of them to ~S.
26 * more renaming in global external names:
27 ** used DEFINE-THE-FOO-THING and DEFFOO style consistently (and
28 deprecated supported extensions named in the DEF-FOO
29 style, e.g. SB-ALIEN:DEF-ALIEN-ROUTINE)
30 ** reserved DO-FOO-style names for iteration macros
31 ** finished s/FUNCTION/FUN/
33 ** s/TOPLEVEL/TOP-LEVEL/
34 * global style systematization:
35 ** s/#'(lambda/(lambda/
36 ** four-space indentation in C
37 =======================================================================
40 * building with CLISP (or explaining why not)
41 * faster bootstrapping (both make.sh and slam.sh)
42 ** added mechanisms for automatically finding dead code, and
43 used them to remove dead code
44 ** moved stuff from warm init into cold init where possible
45 (so that slam.sh will run faster and also just because
46 ideally everything would be in cold init)
47 ** profiled and tweaked
48 * more EVAL/EVAL-WHEN/%COMPILE/DEFUN/DEFSTRUCT cleanup:
49 ** made %COMPILE understand magicality of DEFUN FOO
50 w.r.t. e.g. preexisting inlineness of FOO
51 ** used %COMPILE where COMPILE-TOP-LEVEL used to be used
52 ** removed now-redundant COMPILE-TOP-LEVEL and
53 FUNCTIONAL-KIND=:TOP-LEVEL stuff from the compiler
54 ** made FUNCTION-NAME logic work on closures, so that
55 various public functions like CL:PACKAGEP which
56 are now implemented as closures (because
57 they're structure slot accessors) won't be so
59 ** %SLOT-ACCESSOR/%SLOT-ACCESSOR stuff can probably go away,
60 since we inline expand all slot accessors into
61 %INSTANCE-REF and the optimizer knows all it needs
63 * rewrote long-standing confusing error restarts for redefining
65 * outstanding embarrassments
66 ** cut-and-pasted DEF-BOOLEAN-ATTRIBUTE (maybe easier to fix
67 now that EVAL-WHEN does what it should..)
69 ** :IGNORE-ERRORS-P cruft in stems-and-flags.lisp-expr
70 ** weird double-loading (first in GENESIS, then in warm init)
71 of src/assembly/target/*.lisp stuff, and the associated
72 weirdness of the half-baked state (compiler almost but
73 not quite ready for prime time..) of the system after
75 * fixups now feasible because of pre7 changes
76 ** ANSIfied DECLAIM INLINE stuff (deprecating MAYBE-INLINE)
77 * miscellaneous simple refactoring
79 ** renamed %PRIMITIVE to %VOP
80 * These days ANSI C has inline functions, so..
81 ** redid many cpp macros as inline functions:
82 HeaderValue, Pointerp, CEILING, ALIGNED_SIZE,
83 GET_FREE_POINTER, SET_FREE_POINTER,
84 GET_GC_TRIGGER, SET_GC_TRIGGER, GetBSP, SetBSP,
85 os_trunc_foo(), os_round_up_foo()
86 ** removed various avoid-evaluating-C-macro-arg-twice
88 * added mechanisms for automatically finding dead symbols is
89 package-data.lisp-expr (i.e. those symbols not bound,
90 fbound, defined as types, or whatever), and used them
91 to remove dead symbols
92 * made system handle stack overflow safely unless SAFETY is dominated
94 =======================================================================
97 * refactored in preparation for moving CLOS into cold init and merging
98 SB-PCL:FOO with CL:FOO (for FOO=CLASS, FOO=CLASS-OF, etc.)
99 ** systematized support for MOP (new regression tests, maybe
100 new SB-MOP package..) to try to make sure things don't
101 get mislaid in the upcoming CLOS restructuring
102 ** extracted type system from SB-KERNEL into new SB-TYPE
104 ** reimplemented GENERIC-FUNCTION as a primitive object (or
105 maybe made SB-MOP:FUNCALLABLE-STANDARD-OBJECT the
106 primitive object, and then let GENERIC-FUNCTIONs
107 inherit from that) instead of structures with
108 :ALTERNATE-METACLASS and funcallableness. Now
109 FUNCALLABLE-INSTANCE can go away. (And now the new
110 funcallable primitive objects need to go into
111 collections like *FUN-HEADER-WIDETAGS* where
112 FUNCALLABLE-INSTANCE objects used to be.)
113 ** reimplemented CONDITIONs as primitive objects instead of
114 structures with :ALTERNATE-METACLASS. Now (between
115 this and the change to GENERIC-FUNCTIONs)
116 DEFSTRUCT :ALTERNATE-METACLASS can go away.
117 ** (maybe) Now INSTANCE_POINTER_LOWTAG can become just
118 STRUCTURE_POINTER_LOWTAG, and the concept of
119 SB-KERNEL:INSTANCE (including INSTANCEP,
120 (SPECIFIER-TYPE 'INSTANCE), etc.) can go away.
121 * moved CLOS into cold init, in order to allow CLOS to be used in the
122 implementation of the core system (e.g. the type system and the
123 compiler) and in order to support merger of CL:CLASS with
125 * (maybe) eliminated warm init altogether in favor of cold init
126 * (maybe, especially if warm init can be eliminated) rationalized
127 the build process, fixing miscellaneous pre-0.5.0 stuff that's
128 transparently not the right thing
129 ** removed separate build directories, now just building in
130 place with .sbclcoldfasl extensions
131 * (maybe) more refactoring in preparation for merging SB-PCL:FOO
132 into CL:FOO: reimplemented type system OO dispatch
133 (!DEFINE-TYPE-METHOD, etc.) in terms of CLOS OO dispatch
134 * merged SB-PCL:FOO into CL:FOO (and similarly CLASS-OF, etc.)
135 * added some automatic tests for basic binary compatibility, in hopes
136 that it might be practical to maintain binary compatibility
137 between minor maintenance releases on the stable branch (but no
138 promises, sorry, since I've never tried to do this before, and
139 have no idea how much of a pain this'll be)
140 =======================================================================
141 other priorities, no particular time:
143 * bug fixes, especially really annoying bugs (ANSI or not) and any
144 ANSI bugs (i.e. not just bugs in extras like the debugger or
145 "declarations are assertions", but violations of the standard)
146 * better communication with the outside world (scratching WHN's
147 personal itch): I don't want socket-level stuff so much as I
148 want RPC-level or higher (CORBA?) interfaces and (possibly
149 through RPC or CORBA) GUI support
150 =======================================================================
151 important but out of scope (for WHN, anyway: Patches from other people
152 are still welcome!) until after 1.0:
154 * sadly deteriorated support for ANSI-style block compilation
155 (static linking of DEFUNs within a single file or
156 WITH-COMPILATION-UNIT)
157 * various GC issues (exuberant cut-and-paste coding,
158 possibly dangerously over-conservative handling
159 of neighbors of function objects, general GC efficiency)
160 * package issues other than SB!TYPE, SB!MOP, and dead exported
162 * Any systematic effort to fix compiler consistency checks is
163 out of scope. (However, it still might be possible to
164 determine that some or all of them are hopelessly stale
166 ===============================================================================
167 other known issues with no particular target date:
169 bugs listed on the man page
171 hundreds of FIXME notes in the sources from WHN
173 various other unfinished business from CMU CL and before, marked with
174 "XX" or "XXX" or "###" or "***" or "???" or "pfw" or "@@@@" or "zzzzz"
175 or probably also other codes that I haven't noticed or have forgotten.
177 (Things marked as KLUDGE are in general things which are ugly or
178 confusing, but that, for whatever reason, may stay that way
180 =======================================================================
181 "There's nothing an agnostic can't do as long as he doesn't know
182 whether he believes in anything or not."
185 "God grant me serenity to accept the code I cannot change, courage to
186 change the code I can, and wisdom to know the difference."
189 "Accumulation of half-understood design decisions eventually chokes a
190 program as a water weed chokes a canal. By refactoring you can ensure
191 that your full understanding of how the program should be designed is
192 always reflected in the program. As a water weed quickly spreads its
193 tendrils, partially understood design decisions quickly spread their
194 effects throughout your program. No one or two or even ten individual
195 actions will be enough to eradicate the problem."
196 -- Martin Fowler, in _Refactoring: Improving the Design of Existing
199 "I wish I didn't know now what I didn't know then."