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 * more renaming (esp. for global as opposed to lexical names):
8 ** Rename "cold" stuff (e.g. SB-COLD and src/cold/, though
9 not e.g. #'COLD-INIT) to "boot".
10 * fixed CREDITS, since it's gone pretty stale
12 * pending patches and bug reports that go in (or else get handled
13 somehow, rejected/logged/whatever) before 0.7.0:
14 =======================================================================
17 * patches postponed until after 0.7.0:
18 ** CSR "rough patch to fix bug 106" 2001-10-28
19 ** Alexey Dejneka "bug 111" 2001-12-30
20 * building with CLISP (or explaining why not). This will likely involve
21 a rearrangement of the build system so that it never renames
22 the output from COMPILE-FILE, because CLISP's COMPILE-FILE
23 outputs two (!) files and as far as I can tell LOAD uses both
24 of them. Since I have other motivations for this rearrangement
25 besides CLISPiosyncrasies, I'm reasonably motivated to do it.
26 * urgent EVAL/EVAL-WHEN/%COMPILE/DEFUN/DEFSTRUCT cleanup:
27 ** made inlining DEFUN inside MACROLET work again
28 ** (also, while working on INLINE anyway, it should be easy
29 to flush the old MAYBE-INLINE cruft entirely,
30 including e.g. on the man page)
31 ** fixed bug 137 (more)
32 * faster bootstrapping (both make.sh and slam.sh)
33 ** added mechanisms for automatically finding dead code, and
34 used them to remove dead code
35 ** moved stuff from warm init into cold init where possible
36 (so that slam.sh will run faster and also just because
37 ideally everything would be in cold init)
38 ** profiled and tweaked
39 * more EVAL/EVAL-WHEN/%COMPILE/DEFUN/DEFSTRUCT cleanup:
40 ** made %COMPILE understand magicality of DEFUN FOO
41 w.r.t. e.g. preexisting inlineness of FOO
42 ** used %COMPILE where COMPILE-TOP-LEVEL used to be used
43 ** removed now-redundant COMPILE-TOP-LEVEL and
44 FUNCTIONAL-KIND=:TOP-LEVEL stuff from the compiler
45 ** made FUNCTION-NAME logic work on closures, so that
46 various public functions like CL:PACKAGEP which
47 are now implemented as closures (because
48 they're structure slot accessors) won't be so
50 ** %SLOT-ACCESSOR/%SLOT-ACCESSOR stuff can probably go away,
51 since we inline expand all slot accessors into
52 %INSTANCE-REF and the optimizer knows all it needs
54 * rewrote long-standing confusing error restarts for redefining
56 * outstanding embarrassments
57 ** cut-and-pasted DEF-BOOLEAN-ATTRIBUTE (maybe easier to fix
58 now that EVAL-WHEN does what it should..)
60 ** :IGNORE-ERRORS-P cruft in stems-and-flags.lisp-expr
61 ** weird double-loading (first in GENESIS, then in warm init)
62 of src/assembly/target/*.lisp stuff, and the associated
63 weirdness of the half-baked state (compiler almost but
64 not quite ready for prime time..) of the system after
66 * fixups now feasible because of pre7 changes
67 ** ANSIfied DECLAIM INLINE stuff (deprecating MAYBE-INLINE)
68 * miscellaneous simple refactoring
70 ** renamed %PRIMITIVE to %VOP
71 * These days ANSI C has inline functions, so..
72 ** redid many cpp macros as inline functions:
73 HeaderValue, Pointerp, CEILING, ALIGNED_SIZE,
74 GET_FREE_POINTER, SET_FREE_POINTER,
75 GET_GC_TRIGGER, SET_GC_TRIGGER, GetBSP, SetBSP,
76 os_trunc_foo(), os_round_up_foo()
77 ** removed various avoid-evaluating-C-macro-arg-twice
79 * added mechanisms for automatically finding dead symbols is
80 package-data.lisp-expr (i.e. those symbols not bound,
81 fbound, defined as types, or whatever), and used them
82 to remove dead symbols
83 * made system handle stack overflow safely unless SAFETY is dominated
85 * Probably get rid of or at least rework the fdefinition/encapsulation
86 system so that (SYMBOL-FUNCTION 'FOO) = (FDEFINITION 'FOO).
87 =======================================================================
90 * refactored in preparation for moving CLOS into cold init and merging
91 SB-PCL:FOO with CL:FOO (for FOO=CLASS, FOO=CLASS-OF, etc.)
92 ** systematized support for MOP (new regression tests, maybe
93 new SB-MOP package..) to try to make sure things don't
94 get mislaid in the upcoming CLOS restructuring
95 ** extracted type system from SB-KERNEL into new SB-TYPE
97 ** reimplemented GENERIC-FUNCTION as a primitive object (or
98 maybe made SB-MOP:FUNCALLABLE-STANDARD-OBJECT the
99 primitive object, and then let GENERIC-FUNCTIONs
100 inherit from that) instead of structures with
101 :ALTERNATE-METACLASS and funcallableness. Now
102 FUNCALLABLE-INSTANCE can go away. (And now the new
103 funcallable primitive objects need to go into
104 collections like *FUN-HEADER-WIDETAGS* where
105 FUNCALLABLE-INSTANCE objects used to be.)
106 ** reimplemented CONDITIONs as primitive objects instead of
107 structures with :ALTERNATE-METACLASS. Now (between
108 this and the change to GENERIC-FUNCTIONs)
109 DEFSTRUCT :ALTERNATE-METACLASS can go away.
110 ** (maybe) Now INSTANCE_POINTER_LOWTAG can become just
111 STRUCTURE_POINTER_LOWTAG, and the concept of
112 SB-KERNEL:INSTANCE (including INSTANCEP,
113 (SPECIFIER-TYPE 'INSTANCE), etc.) can go away.
114 * moved CLOS into cold init, in order to allow CLOS to be used in the
115 implementation of the core system (e.g. the type system and the
116 compiler) and in order to support merger of CL:CLASS with
118 * (maybe) eliminated warm init altogether in favor of cold init
119 * (maybe, especially if warm init can be eliminated) rationalized
120 the build process, fixing miscellaneous pre-0.5.0 stuff that's
121 transparently not the right thing
122 ** removed separate build directories, now just building in
123 place with .sbclcoldfasl extensions
124 * (maybe) more refactoring in preparation for merging SB-PCL:FOO
125 into CL:FOO: reimplemented type system OO dispatch
126 (!DEFINE-TYPE-METHOD, etc.) in terms of CLOS OO dispatch
127 * merged SB-PCL:FOO into CL:FOO (and similarly CLASS-OF, etc.)
128 * added some automatic tests for basic binary compatibility, in hopes
129 that it might be practical to maintain binary compatibility
130 between minor maintenance releases on the stable branch (but no
131 promises, sorry, since I've never tried to do this before, and
132 have no idea how much of a pain this'll be)
133 ========================================================================
134 for 1.0 (fixes of lower priority which I'd nonetheless be embarrassed
135 to leave unfixed in 1.0):
136 * all too many BUGS entries and FIXMEs
137 =======================================================================
138 other priorities, no particular time:
140 * bug fixes, especially really annoying bugs (ANSI or not) and any
141 ANSI bugs (i.e. not just bugs in extras like the debugger or
142 "declarations are assertions", but violations of the standard)
143 * better communication with the outside world (scratching WHN's
144 personal itch): I don't want socket-level stuff so much as I
145 want RPC-level or higher (CORBA?) interfaces and (possibly
146 through RPC or CORBA) GUI support
147 =======================================================================
148 important but out of scope (for WHN, anyway: Patches from other people
149 are still welcome!) until after 1.0:
151 * sadly deteriorated support for ANSI-style block compilation
152 (static linking of DEFUNs within a single file or
153 WITH-COMPILATION-UNIT)
154 * various GC issues (exuberant cut-and-paste coding,
155 possibly dangerously over-conservative handling
156 of neighbors of function objects, general GC efficiency)
157 * package issues other than SB!TYPE, SB!MOP, and dead exported
159 * Any systematic effort to fix compiler consistency checks is
160 out of scope. (However, it still might be possible to
161 determine that some or all of them are hopelessly stale
163 ===============================================================================
164 other known issues with no particular target date:
166 bugs listed on the man page
168 hundreds of FIXME notes in the sources from WHN
170 various other unfinished business from CMU CL and before, marked with
171 "XX" or "XXX" or "###" or "***" or "???" or "pfw" or "@@@@" or "zzzzz"
172 or probably also other codes that I haven't noticed or have forgotten.
174 (Things marked as KLUDGE are in general things which are ugly or
175 confusing, but that, for whatever reason, may stay that way
177 =======================================================================
178 "There's nothing an agnostic can't do as long as he doesn't know
179 whether he believes in anything or not."
182 "God grant me serenity to accept the code I cannot change, courage to
183 change the code I can, and wisdom to know the difference."
186 "Accumulation of half-understood design decisions eventually chokes a
187 program as a water weed chokes a canal. By refactoring you can ensure
188 that your full understanding of how the program should be designed is
189 always reflected in the program. As a water weed quickly spreads its
190 tendrils, partially understood design decisions quickly spread their
191 effects throughout your program. No one or two or even ten individual
192 actions will be enough to eradicate the problem."
193 -- Martin Fowler, in _Refactoring: Improving the Design of Existing
196 "I wish I didn't know now what I didn't know then."