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