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 ** made inlining DEFUN inside MACROLET work again
8 ** whatever bug it is that causes s/#'(lambda/(lambda/ to
9 cause compilation failure in condition.lisp
10 ** perhaps function debug name quick fix, if there's an easy way
11 * more renaming in global external names:
12 ** reserved DO-FOO-style names for iteration macros
13 ** finished s/FUNCTION/FUN/
15 ** s/TOP-LEVEL/TOPLEVEL/
16 * global style systematization:
17 ** s/#'(lambda/(lambda/
18 * pending patches and bug reports that go in (or else get handled
19 somehow, rejected/logged/whatever) before 0.7.0:
20 =======================================================================
23 * patches postponed until after 0.7.0:
24 ** Christophe Rhodes "rough patch to fix bug 106" 2001-10-28
25 * building with CLISP (or explaining why not)
26 * urgent EVAL/EVAL-WHEN/%COMPILE/DEFUN/DEFSTRUCT cleanup:
28 * faster bootstrapping (both make.sh and slam.sh)
29 ** added mechanisms for automatically finding dead code, and
30 used them to remove dead code
31 ** moved stuff from warm init into cold init where possible
32 (so that slam.sh will run faster and also just because
33 ideally everything would be in cold init)
34 ** profiled and tweaked
35 * more EVAL/EVAL-WHEN/%COMPILE/DEFUN/DEFSTRUCT cleanup:
36 ** made %COMPILE understand magicality of DEFUN FOO
37 w.r.t. e.g. preexisting inlineness of FOO
38 ** used %COMPILE where COMPILE-TOP-LEVEL used to be used
39 ** removed now-redundant COMPILE-TOP-LEVEL and
40 FUNCTIONAL-KIND=:TOP-LEVEL stuff from the compiler
41 ** made FUNCTION-NAME logic work on closures, so that
42 various public functions like CL:PACKAGEP which
43 are now implemented as closures (because
44 they're structure slot accessors) won't be so
46 ** %SLOT-ACCESSOR/%SLOT-ACCESSOR stuff can probably go away,
47 since we inline expand all slot accessors into
48 %INSTANCE-REF and the optimizer knows all it needs
50 * rewrote long-standing confusing error restarts for redefining
52 * outstanding embarrassments
53 ** cut-and-pasted DEF-BOOLEAN-ATTRIBUTE (maybe easier to fix
54 now that EVAL-WHEN does what it should..)
56 ** :IGNORE-ERRORS-P cruft in stems-and-flags.lisp-expr
57 ** weird double-loading (first in GENESIS, then in warm init)
58 of src/assembly/target/*.lisp stuff, and the associated
59 weirdness of the half-baked state (compiler almost but
60 not quite ready for prime time..) of the system after
62 * fixups now feasible because of pre7 changes
63 ** ANSIfied DECLAIM INLINE stuff (deprecating MAYBE-INLINE)
64 * miscellaneous simple refactoring
66 ** renamed %PRIMITIVE to %VOP
67 * These days ANSI C has inline functions, so..
68 ** redid many cpp macros as inline functions:
69 HeaderValue, Pointerp, CEILING, ALIGNED_SIZE,
70 GET_FREE_POINTER, SET_FREE_POINTER,
71 GET_GC_TRIGGER, SET_GC_TRIGGER, GetBSP, SetBSP,
72 os_trunc_foo(), os_round_up_foo()
73 ** removed various avoid-evaluating-C-macro-arg-twice
75 * added mechanisms for automatically finding dead symbols is
76 package-data.lisp-expr (i.e. those symbols not bound,
77 fbound, defined as types, or whatever), and used them
78 to remove dead symbols
79 * made system handle stack overflow safely unless SAFETY is dominated
81 * Probably get rid of or at least rework the fdefinition/encapsulation
82 system so that (SYMBOL-FUNCTION 'FOO) = (FDEFINITION 'FOO).
83 =======================================================================
86 * refactored in preparation for moving CLOS into cold init and merging
87 SB-PCL:FOO with CL:FOO (for FOO=CLASS, FOO=CLASS-OF, etc.)
88 ** systematized support for MOP (new regression tests, maybe
89 new SB-MOP package..) to try to make sure things don't
90 get mislaid in the upcoming CLOS restructuring
91 ** extracted type system from SB-KERNEL into new SB-TYPE
93 ** reimplemented GENERIC-FUNCTION as a primitive object (or
94 maybe made SB-MOP:FUNCALLABLE-STANDARD-OBJECT the
95 primitive object, and then let GENERIC-FUNCTIONs
96 inherit from that) instead of structures with
97 :ALTERNATE-METACLASS and funcallableness. Now
98 FUNCALLABLE-INSTANCE can go away. (And now the new
99 funcallable primitive objects need to go into
100 collections like *FUN-HEADER-WIDETAGS* where
101 FUNCALLABLE-INSTANCE objects used to be.)
102 ** reimplemented CONDITIONs as primitive objects instead of
103 structures with :ALTERNATE-METACLASS. Now (between
104 this and the change to GENERIC-FUNCTIONs)
105 DEFSTRUCT :ALTERNATE-METACLASS can go away.
106 ** (maybe) Now INSTANCE_POINTER_LOWTAG can become just
107 STRUCTURE_POINTER_LOWTAG, and the concept of
108 SB-KERNEL:INSTANCE (including INSTANCEP,
109 (SPECIFIER-TYPE 'INSTANCE), etc.) can go away.
110 * moved CLOS into cold init, in order to allow CLOS to be used in the
111 implementation of the core system (e.g. the type system and the
112 compiler) and in order to support merger of CL:CLASS with
114 * (maybe) eliminated warm init altogether in favor of cold init
115 * (maybe, especially if warm init can be eliminated) rationalized
116 the build process, fixing miscellaneous pre-0.5.0 stuff that's
117 transparently not the right thing
118 ** removed separate build directories, now just building in
119 place with .sbclcoldfasl extensions
120 * (maybe) more refactoring in preparation for merging SB-PCL:FOO
121 into CL:FOO: reimplemented type system OO dispatch
122 (!DEFINE-TYPE-METHOD, etc.) in terms of CLOS OO dispatch
123 * merged SB-PCL:FOO into CL:FOO (and similarly CLASS-OF, etc.)
124 * added some automatic tests for basic binary compatibility, in hopes
125 that it might be practical to maintain binary compatibility
126 between minor maintenance releases on the stable branch (but no
127 promises, sorry, since I've never tried to do this before, and
128 have no idea how much of a pain this'll be)
129 ========================================================================
130 for 1.0 (fixes of lower priority which I'd nonetheless be embarrassed
131 to leave unfixed in 1.0):
132 * all too many BUGS entries and FIXMEs
133 =======================================================================
134 other priorities, no particular time:
136 * bug fixes, especially really annoying bugs (ANSI or not) and any
137 ANSI bugs (i.e. not just bugs in extras like the debugger or
138 "declarations are assertions", but violations of the standard)
139 * better communication with the outside world (scratching WHN's
140 personal itch): I don't want socket-level stuff so much as I
141 want RPC-level or higher (CORBA?) interfaces and (possibly
142 through RPC or CORBA) GUI support
143 =======================================================================
144 important but out of scope (for WHN, anyway: Patches from other people
145 are still welcome!) until after 1.0:
147 * sadly deteriorated support for ANSI-style block compilation
148 (static linking of DEFUNs within a single file or
149 WITH-COMPILATION-UNIT)
150 * various GC issues (exuberant cut-and-paste coding,
151 possibly dangerously over-conservative handling
152 of neighbors of function objects, general GC efficiency)
153 * package issues other than SB!TYPE, SB!MOP, and dead exported
155 * Any systematic effort to fix compiler consistency checks is
156 out of scope. (However, it still might be possible to
157 determine that some or all of them are hopelessly stale
159 ===============================================================================
160 other known issues with no particular target date:
162 bugs listed on the man page
164 hundreds of FIXME notes in the sources from WHN
166 various other unfinished business from CMU CL and before, marked with
167 "XX" or "XXX" or "###" or "***" or "???" or "pfw" or "@@@@" or "zzzzz"
168 or probably also other codes that I haven't noticed or have forgotten.
170 (Things marked as KLUDGE are in general things which are ugly or
171 confusing, but that, for whatever reason, may stay that way
173 =======================================================================
174 "There's nothing an agnostic can't do as long as he doesn't know
175 whether he believes in anything or not."
178 "God grant me serenity to accept the code I cannot change, courage to
179 change the code I can, and wisdom to know the difference."
182 "Accumulation of half-understood design decisions eventually chokes a
183 program as a water weed chokes a canal. By refactoring you can ensure
184 that your full understanding of how the program should be designed is
185 always reflected in the program. As a water weed quickly spreads its
186 tendrils, partially understood design decisions quickly spread their
187 effects throughout your program. No one or two or even ten individual
188 actions will be enough to eradicate the problem."
189 -- Martin Fowler, in _Refactoring: Improving the Design of Existing
192 "I wish I didn't know now what I didn't know then."