1 ;;;; tags which are set during the build process and which end up in
2 ;;;; CL:*FEATURES* in the target SBCL, plus some comments about other
3 ;;;; CL:*FEATURES* tags which have special meaning to SBCL or which
4 ;;;; have a special conventional meaning
6 ;;;; This software is part of the SBCL system. See the README file for
9 ;;;; This software is derived from the CMU CL system, which was
10 ;;;; written at Carnegie Mellon University and released into the
11 ;;;; public domain. The software is in the public domain and is
12 ;;;; provided with absolutely no warranty. See the COPYING and CREDITS
13 ;;;; files for more information.
17 ;; features present in all builds
22 ;; FIXME: Isn't there a :x3jsomething feature which we should set too?
27 ;; Douglas Thomas Crosher's conservative generational GC (the only one
28 ;; we currently support)
31 ;; We're running under a UNIX. This is sort of redundant, and it was also
32 ;; sort of redundant under CMU CL, which we inherited it from: neither SBCL
33 ;; nor CMU CL supports anything but UNIX (and "technically not UNIX"es
34 ;; such as *BSD and Linux). But someday, maybe we might, and in that case
35 ;; we'd presumably remove this, so its presence conveys the information
36 ;; that the system isn't one which follows such a change.
40 ;; features present in this particular build
43 ;; Setting this enables the compilation of documentation strings
44 ;; from the system sources into the target Lisp executable.
45 ;; Traditional Common Lisp folk will want this option set.
46 ;; I (WHN) made it optional because I came to Common Lisp from
47 ;; C++ through Scheme, so I'm accustomed to asking
48 ;; Emacs about things that I'm curious about instead of asking
49 ;; the executable I'm running.
52 ;; When this is set, EVAL is implemented as an "IR1 interpreter":
53 ;; code is compiled into the compiler's first internal representation,
54 ;; then the IR1 is interpreted. When this is not set, EVAL is implemented
55 ;; as a little bit of hackery wrapped around a call to COMPILE, i.e.
56 ;; the system becomes a "compiler-only implementation" of Common Lisp.
57 ;; As of sbcl-0.6.7, the compiler-only implementation is prototype code,
58 ;; and much less mature than the old IR1 interpreter. Thus, the safe
59 ;; thing is to leave :SB-INTERPRETER set. However, the compiler-only
60 ;; system is noticeably smaller, so you might want to omit
61 ;; :SB-INTERPRETER if you have a small machine.
63 ;; Probably, the compiler-only implementation will become more
64 ;; stable someday, and support for the IR1 interpreter will then be
65 ;; dropped. This will make the system smaller and easier to maintain
66 ;; not only because we no longer need to support the interpreter,
67 ;; but because code elsewhere in the system (the dumper, the debugger,
68 ;; etc.) no longer needs special cases for interpreted code.
71 ;; Do regression and other tests when building the system. You
72 ;; might or might not want this if you're not a developer,
73 ;; depending on how paranoid you are. You probably do want it if
74 ;; you are a developer.
77 ;; Setting this makes more debugging information available.
78 ;; If you aren't hacking or troubleshooting SBCL itself, you
79 ;; probably don't want this set.
81 ;; At least two varieties of debugging information are enabled by this
83 ;; * SBCL is compiled with a higher level of OPTIMIZE DEBUG, so that
84 ;; the debugger can tell more about the state of the system.
85 ;; * Various code to print debugging messages, and similar debugging code,
86 ;; is compiled only when this feature is present.
88 ;; Note that the extra information recorded by the compiler at
89 ;; this higher level of OPTIMIZE DEBUG includes the source location
90 ;; forms. In order for the debugger to use this information, it has to
91 ;; re-READ the source file. In an ordinary installation of SBCL, this
92 ;; re-READing may not work very well, for either of two reasons:
93 ;; * The sources aren't present on the system in the same location that
94 ;; they were on the system where SBCL was compiled.
95 ;; * SBCL is using the standard readtable, without the added hackage
96 ;; which allows it to handle things like target features.
97 ;; If you want to be able to use the extra debugging information,
98 ;; therefore, be sure to keep the sources around, and run with the
99 ;; readtable configured so that the system sources can be read.
102 ;; Enable extra debugging output in the assem.lisp assembler/scheduler
103 ;; code. (This is the feature which was called :DEBUG in the
104 ;; original CMU CL code.)
107 ;; Setting this makes SBCL more "fluid", i.e. more amenable to
108 ;; modification at runtime, by suppressing various INLINE declarations,
109 ;; compiler macro definitions, FREEZE-TYPE declarations; and by
110 ;; suppressing various burning-our-ships-behind-us actions after
111 ;; initialization is complete; and so forth. This tends to clobber the
112 ;; performance of the system, so unless you have some special need for
113 ;; this when hacking SBCL itself, you don't want this set.
116 ;; Enable code for collecting statistics on usage of various operations,
117 ;; useful for performance tuning of the SBCL system itself. This code
118 ;; is probably pretty stale (having not been tested since the fork from
119 ;; base CMU CL) but might nonetheless be a useful starting point for
120 ;; anyone who wants to collect such statistics in the future.
123 ;; Peter Van Eynde's increase-bulletproofness code
125 ;; This is not maintained or tested in current SBCL, but I haven't
126 ;; gone out of my way to remove or break it, either.
129 ; :high-security-support
131 ;; multiprocessing support
133 ;; This is not maintained or tested in current SBCL. I haven't gone out
134 ;; of my way to break it, but since it's derived from an old version of
135 ;; CMU CL where multiprocessing was pretty shaky, it's likely to be very
137 ;; :MP enables multiprocessing
138 ;; :MP-I486 is used, only within the multiprocessing code, to control
139 ;; what seems to control processor-version-specific code. It's
140 ;; probably for 486 or later, i.e. could be set as long as
141 ;; you know you're not running on a 386, but it doesn't seem
142 ;; to be documented anywhere, so that's just a guess.
146 ;; KLUDGE: used to suppress stale code related to floating point infinities.
147 ;; I intend to delete this code completely some day, since it was a pain
148 ;; for me to try to work with and since all benefits it provides are
149 ;; non-portable. Until I actually pull the trigger, though, I've left
150 ;; various stale code in place protected with #!-SB-INFINITIES.
153 ;; This affects the definition of a lot of things in bignum.lisp. It
154 ;; doesn't seem to be documented anywhere what systems it might apply to.
155 ;; It doesn't seem to be needed for X86 systems anyway.
158 ;; This is probably true for some processor types, but not X86. It affects
159 ;; a lot of floating point code.
160 ; :negative-zero-is-not-zero
162 ;; This is mentioned in cmu-user.tex, which says that it enables
163 ;; the compiler to reason about integer arithmetic. It also seems to
164 ;; control other fancy numeric reasoning, e.g. knowing the result type of
165 ;; a remainder calculation given the type of its inputs.
167 ;; CROSS-FLOAT-INFINITY-KLUDGE: The :PROPAGATE-FLOAT-TYPE and
168 ;; :PROPAGATE-FUN-TYPE features are problematic when building
169 ;; the cross-compiler itself. Their implementation depends on
170 ;; floating point infinities, which might not be supported in the
171 ;; cross-compilation host. In order to avoid this problem, while
172 ;; still supporting these features in the target Lisp compiler,
173 ;; we use the :WILL-PROPAGATE-FLOAT-TYPE feature when building
174 ;; the cross-compiler, and munge it into :PROPAGATE-FLOAT-TYPE
175 ;; only when building the target compiler; and similarly for
176 ;; :WILL-PROPAGATE-FUN-TYPE.
177 ;:will-propagate-float-type ; (becomes :PROPAGATE-FLOAT-TYPE)
179 ;; According to cmu-user.tex, this enables the compiler to infer result
180 ;; types for mathematical functions like SQRT, EXPT, and LOG, allowing
181 ;; it to e.g. eliminate the possibility that a complex result will be
182 ;; generated. This applies only to the target compiler, not the
183 ;; cross-compiler: see CROSS-FLOAT-INFINITY-KLUDGE.
184 ;:will-propagate-fun-type ; (becomes :PROPAGATE-FUN-TYPE)
186 ;; It's unclear to me what this does (but it was enabled in the code that I
187 ;; picked up from Peter Van Eynde). -- WHN 19990224
188 :constrain-float-type
190 ;; This is set in classic CMU CL, and presumably there it means
191 ;; that the floating point arithmetic implementation
192 ;; conforms to IEEE's standard. Here it definitely means that the
193 ;; floating point arithmetic implementation conforms to IEEE's standard.
194 ;; I (WHN 19990702) haven't tried to verify
195 ;; that it does conform, but it should at least mostly conform (because
196 ;; the underlying x86 hardware tries).
199 ;; This seems to be the pre-GENCGC garbage collector for CMU CL, which was
200 ;; AFAIK never supported for the X86.
203 ;; CMU CL had, and we inherited, code to support 80-bit LONG-FLOAT on the x86
204 ;; architecture. Nothing has been done to actively destroy the long float
205 ;; support, but it hasn't been thoroughly maintained, and needs at least
206 ;; some maintenance before it will work. (E.g. the LONG-FLOAT-only parts of
207 ;; genesis are still implemented in terms of unportable CMU CL functions
208 ;; which are not longer available at genesis time in SBCL.) A deeper
209 ;; problem is SBCL's bootstrap process implicitly assumes that the
210 ;; cross-compilation host will be able to make the same distinctions
211 ;; between floating point types that it does. This assumption is
212 ;; fundamentally sleazy, even though in practice it's unlikely to break down
213 ;; w.r.t. distinguishing SINGLE-FLOAT from DOUBLE-FLOAT; it's much more
214 ;; likely to break down w.r.t. distinguishing DOUBLE-FLOAT from LONG-FLOAT.
215 ;; Still it's likely to be quite doable to get LONG-FLOAT support working
216 ;; again, if anyone's sufficiently motivated.
220 ;; miscellaneous notes on other things which could have special significance
221 ;; in the *FEATURES* list
224 ;; notes on the :NIL and :IGNORE features:
226 ;; #+NIL is used to comment out forms. Occasionally #+IGNORE is used
227 ;; for this too. So don't use :NIL or :IGNORE as the names of features..
229 ;; notes on :SB-XC and :SB-XC-HOST features (which aren't controlled by this
230 ;; file, but are instead temporarily pushed onto *FEATURES* or
231 ;; *TARGET-FEATURES* during some phases of cross-compilation):
233 ;; :SB-XC-HOST stands for "cross-compilation host" and is in *FEATURES*
234 ;; during the first phase of cross-compilation bootstrapping, when the
235 ;; host Lisp is being used to compile the cross-compiler.
237 ;; :SB-XC stands for "cross compiler", and is in *FEATURES* during the second
238 ;; phase of cross-compilation bootstrapping, when the cross-compiler is
239 ;; being used to create the first target Lisp.
241 ;; notes on the :PROPAGATE-FLOAT-TYPE and :PROPAGATE-FUN-TYPE
242 ;; features: See the comments on CROSS-FLOAT-INFINITY-KLUDGE.
244 ;; notes on the :SB-ASSEMBLING feature (which isn't controlled by
247 ;; This is a flag for whether we're in the assembler. It's
248 ;; temporarily pushed onto the *FEATURES* list in the setup for
249 ;; the ASSEMBLE-FILE function. It would be a bad idea
250 ;; to use it as a name for a permanent feature.
252 ;; notes on local features (which are set automatically by the
253 ;; configuration script, and should not be set here unless you
254 ;; really, really know what you're doing):
256 ;; machine architecture features:
257 ;; :x86 ; any Intel 386 or better, or compatibles like the AMD K6 or K7
258 ;; (No others are supported by SBCL as of 0.6.7, but :alpha or
259 ;; :sparc support could be ported from CMU CL if anyone is
260 ;; sufficiently motivated to do so.)
261 ;; (CMU CL also had a :pentium feature, which affected the definition
262 ;; of some floating point vops. It was present but not enabled in the
263 ;; CMU CL code that SBCL is derived from, and is present but stale
264 ;; in SBCL as of 0.6.7.)
266 ;; operating system features:
267 ;; :linux = We're intended to run under some version of Linux.
268 ;; :bsd = We're intended to run under some version of BSD Unix. (This
269 ;; is not exclusive with the features which indicate which
270 ;; particular version of BSD we're intended to run under.)
271 ;; :freebsd = We're intended to run under FreeBSD.
272 ;; :openbsd = We're intended to run under FreeBSD.
273 ;; (No others are supported by SBCL as of 0.6.7, but :hpux or
274 ;; :solaris support could be ported from CMU CL if anyone is
275 ;; sufficiently motivated to do so.)