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 ** made %COMPILE set up debugging data more like the way the
9 debugger expects (and maybe even completely
11 *** made (DISASSEMBLE 'PRINT) work
12 * incompatible changes listed in NEWS:
13 ** changed debugger prompt to "5]", "5[2]", "5[3]", etc.
14 ** changed default output representation of *PRINT-ESCAPE*-ed
15 unprintable ASCII characters to #\Nul, #\Soh, etc.
16 * some easy FIXMEs with high disruptive potential:
17 ** Search lists go away.
18 ** Grep for ~D and and change most of them to ~S.
19 * more renaming in global external names:
20 ** used DEFINE-THE-FOO-THING and DEFFOO style consistently (and
21 deprecated supported extensions named in the DEF-FOO
22 style, e.g. SB-ALIEN:DEF-ALIEN-ROUTINE)
23 ** reserved DO-FOO-style names for iteration macros
24 ** finished s/FUNCTION/FUN/
26 ** s/TOP-LEVEL/TOPLEVEL/
27 * global style systematization:
28 ** s/#'(lambda/(lambda/
29 ** four-space indentation in C
30 =======================================================================
33 * building with CLISP (or explaining why not)
34 * faster bootstrapping (both make.sh and slam.sh)
35 ** added mechanisms for automatically finding dead code, and
36 used them to remove dead code
37 ** moved stuff from warm init into cold init where possible
38 (so that slam.sh will run faster and also just because
39 ideally everything would be in cold init)
40 ** profiled and tweaked
41 * more EVAL/EVAL-WHEN/%COMPILE/DEFUN/DEFSTRUCT cleanup:
42 ** made %COMPILE understand magicality of DEFUN FOO
43 w.r.t. e.g. preexisting inlineness of FOO
44 ** used %COMPILE where COMPILE-TOP-LEVEL used to be used
45 ** removed now-redundant COMPILE-TOP-LEVEL and
46 FUNCTIONAL-KIND=:TOP-LEVEL stuff from the compiler
47 ** made FUNCTION-NAME logic work on closures, so that
48 various public functions like CL:PACKAGEP which
49 are now implemented as closures (because
50 they're structure slot accessors) won't be so
52 ** %SLOT-ACCESSOR/%SLOT-ACCESSOR stuff can probably go away,
53 since we inline expand all slot accessors into
54 %INSTANCE-REF and the optimizer knows all it needs
56 * rewrote long-standing confusing error restarts for redefining
58 * outstanding embarrassments
59 ** cut-and-pasted DEF-BOOLEAN-ATTRIBUTE (maybe easier to fix
60 now that EVAL-WHEN does what it should..)
62 ** :IGNORE-ERRORS-P cruft in stems-and-flags.lisp-expr
63 ** weird double-loading (first in GENESIS, then in warm init)
64 of src/assembly/target/*.lisp stuff, and the associated
65 weirdness of the half-baked state (compiler almost but
66 not quite ready for prime time..) of the system after
68 * fixups now feasible because of pre7 changes
69 ** ANSIfied DECLAIM INLINE stuff (deprecating MAYBE-INLINE)
70 * miscellaneous simple refactoring
72 ** renamed %PRIMITIVE to %VOP
73 * These days ANSI C has inline functions, so..
74 ** redid many cpp macros as inline functions:
75 HeaderValue, Pointerp, CEILING, ALIGNED_SIZE,
76 GET_FREE_POINTER, SET_FREE_POINTER,
77 GET_GC_TRIGGER, SET_GC_TRIGGER, GetBSP, SetBSP,
78 os_trunc_foo(), os_round_up_foo()
79 ** removed various avoid-evaluating-C-macro-arg-twice
81 * added mechanisms for automatically finding dead symbols is
82 package-data.lisp-expr (i.e. those symbols not bound,
83 fbound, defined as types, or whatever), and used them
84 to remove dead symbols
85 * made system handle stack overflow safely unless SAFETY is dominated
87 * Probably get rid of or at least rework the fdefinition/encapsulation
88 system so that (SYMBOL-FUNCTION 'FOO) = (FDEFINITION 'FOO).
89 =======================================================================
92 * refactored in preparation for moving CLOS into cold init and merging
93 SB-PCL:FOO with CL:FOO (for FOO=CLASS, FOO=CLASS-OF, etc.)
94 ** systematized support for MOP (new regression tests, maybe
95 new SB-MOP package..) to try to make sure things don't
96 get mislaid in the upcoming CLOS restructuring
97 ** extracted type system from SB-KERNEL into new SB-TYPE
99 ** reimplemented GENERIC-FUNCTION as a primitive object (or
100 maybe made SB-MOP:FUNCALLABLE-STANDARD-OBJECT the
101 primitive object, and then let GENERIC-FUNCTIONs
102 inherit from that) instead of structures with
103 :ALTERNATE-METACLASS and funcallableness. Now
104 FUNCALLABLE-INSTANCE can go away. (And now the new
105 funcallable primitive objects need to go into
106 collections like *FUN-HEADER-WIDETAGS* where
107 FUNCALLABLE-INSTANCE objects used to be.)
108 ** reimplemented CONDITIONs as primitive objects instead of
109 structures with :ALTERNATE-METACLASS. Now (between
110 this and the change to GENERIC-FUNCTIONs)
111 DEFSTRUCT :ALTERNATE-METACLASS can go away.
112 ** (maybe) Now INSTANCE_POINTER_LOWTAG can become just
113 STRUCTURE_POINTER_LOWTAG, and the concept of
114 SB-KERNEL:INSTANCE (including INSTANCEP,
115 (SPECIFIER-TYPE 'INSTANCE), etc.) can go away.
116 * moved CLOS into cold init, in order to allow CLOS to be used in the
117 implementation of the core system (e.g. the type system and the
118 compiler) and in order to support merger of CL:CLASS with
120 * (maybe) eliminated warm init altogether in favor of cold init
121 * (maybe, especially if warm init can be eliminated) rationalized
122 the build process, fixing miscellaneous pre-0.5.0 stuff that's
123 transparently not the right thing
124 ** removed separate build directories, now just building in
125 place with .sbclcoldfasl extensions
126 * (maybe) more refactoring in preparation for merging SB-PCL:FOO
127 into CL:FOO: reimplemented type system OO dispatch
128 (!DEFINE-TYPE-METHOD, etc.) in terms of CLOS OO dispatch
129 * merged SB-PCL:FOO into CL:FOO (and similarly CLASS-OF, etc.)
130 * added some automatic tests for basic binary compatibility, in hopes
131 that it might be practical to maintain binary compatibility
132 between minor maintenance releases on the stable branch (but no
133 promises, sorry, since I've never tried to do this before, and
134 have no idea how much of a pain this'll be)
135 =======================================================================
136 other priorities, no particular time:
138 * bug fixes, especially really annoying bugs (ANSI or not) and any
139 ANSI bugs (i.e. not just bugs in extras like the debugger or
140 "declarations are assertions", but violations of the standard)
141 * better communication with the outside world (scratching WHN's
142 personal itch): I don't want socket-level stuff so much as I
143 want RPC-level or higher (CORBA?) interfaces and (possibly
144 through RPC or CORBA) GUI support
145 =======================================================================
146 important but out of scope (for WHN, anyway: Patches from other people
147 are still welcome!) until after 1.0:
149 * sadly deteriorated support for ANSI-style block compilation
150 (static linking of DEFUNs within a single file or
151 WITH-COMPILATION-UNIT)
152 * various GC issues (exuberant cut-and-paste coding,
153 possibly dangerously over-conservative handling
154 of neighbors of function objects, general GC efficiency)
155 * package issues other than SB!TYPE, SB!MOP, and dead exported
157 * Any systematic effort to fix compiler consistency checks is
158 out of scope. (However, it still might be possible to
159 determine that some or all of them are hopelessly stale
161 ===============================================================================
162 other known issues with no particular target date:
164 bugs listed on the man page
166 hundreds of FIXME notes in the sources from WHN
168 various other unfinished business from CMU CL and before, marked with
169 "XX" or "XXX" or "###" or "***" or "???" or "pfw" or "@@@@" or "zzzzz"
170 or probably also other codes that I haven't noticed or have forgotten.
172 (Things marked as KLUDGE are in general things which are ugly or
173 confusing, but that, for whatever reason, may stay that way
175 =======================================================================
176 "There's nothing an agnostic can't do as long as he doesn't know
177 whether he believes in anything or not."
180 "God grant me serenity to accept the code I cannot change, courage to
181 change the code I can, and wisdom to know the difference."
184 "Accumulation of half-understood design decisions eventually chokes a
185 program as a water weed chokes a canal. By refactoring you can ensure
186 that your full understanding of how the program should be designed is
187 always reflected in the program. As a water weed quickly spreads its
188 tendrils, partially understood design decisions quickly spread their
189 effects throughout your program. No one or two or even ten individual
190 actions will be enough to eradicate the problem."
191 -- Martin Fowler, in _Refactoring: Improving the Design of Existing
194 "I wish I didn't know now what I didn't know then."