3 ;;;; tags which are set during the build process and which end up in
4 ;;;; CL:*FEATURES* in the target SBCL, plus some comments about other
5 ;;;; CL:*FEATURES* tags which have special meaning to SBCL or which
6 ;;;; have a special conventional meaning
8 ;;;; Note that the recommended way to customize the features of a
9 ;;;; local build of SBCL is not to edit this file, but instead to
10 ;;;; tweak customize-target-features.lisp. If you define a function
11 ;;;; in customize-target-features.lisp, it will be used to transform
12 ;;;; the target features list after it's read and before it's used.
13 ;;;; E.g. you can use code like this:
15 ;;;; (flet ((enable (x) (pushnew x list))
16 ;;;; (disable (x) (setf list (remove x list))))
17 ;;;; #+nil (enable :sb-show)
18 ;;;; (enable :sb-after-xc-core)
19 ;;;; #+nil (disable :sb-doc)
21 ;;;; By thus editing a local file (one which is not in the source
22 ;;;; distribution, and which is in .cvsignore) your customizations
23 ;;;; will remain local even if you do things like "cvs update",
24 ;;;; will not show up if you try to submit a patch with "cvs diff",
25 ;;;; and might even stay out of the way if you use other non-CVS-based
26 ;;;; methods to upgrade the files or store your configuration.
28 ;;;; This software is part of the SBCL system. See the README file for
29 ;;;; more information.
31 ;;;; This software is derived from the CMU CL system, which was
32 ;;;; written at Carnegie Mellon University and released into the
33 ;;;; public domain. The software is in the public domain and is
34 ;;;; provided with absolutely no warranty. See the COPYING and CREDITS
35 ;;;; files for more information.
39 ;; features present in all builds
44 ;; FIXME: Isn't there a :x3jsomething feature which we should set too?
45 ;; No. CLHS says ":x3j13 [...] A conforming implementation might or
46 ;; might not contain such a feature." -- CSR, 2002-02-21
51 ;; Douglas Thomas Crosher's conservative generational GC (the only one
52 ;; we currently support for X86).
53 ;; :gencgc used to be here; CSR moved it into
54 ;; local-target-features.lisp-expr via make-config.sh, as alpha,
55 ;; sparc and ppc ports don't currently support it. -- CSR, 2002-02-21
57 ;; We're running under a UNIX. This is sort of redundant, and it was also
58 ;; sort of redundant under CMU CL, which we inherited it from: neither SBCL
59 ;; nor CMU CL supports anything but UNIX (and "technically not UNIX"es
60 ;; such as *BSD and Linux). But someday, maybe we might, and in that case
61 ;; we'd presumably remove this, so its presence conveys the information
62 ;; that the system isn't one which follows such a change.
66 ;; features present in this particular build
69 ;; Setting this enables the compilation of documentation strings
70 ;; from the system sources into the target Lisp executable.
71 ;; Traditional Common Lisp folk will want this option set.
72 ;; I (WHN) made it optional because I came to Common Lisp from
73 ;; C++ through Scheme, so I'm accustomed to asking
74 ;; Emacs about things that I'm curious about instead of asking
75 ;; the executable I'm running.
78 ;; Do regression and other tests when building the system. You might
79 ;; or might not want this if you're not a developer, depending on how
80 ;; paranoid you are. You probably do want it if you are a developer.
81 ;; This test does not affect the target system (in much the same way
82 ;; as :sb-after-xc-core, below).
85 ;; Make more debugging information available (for debugging SBCL
86 ;; itself). If you aren't hacking or troubleshooting SBCL itself,
87 ;; you probably don't want this set.
89 ;; At least two varieties of debugging information are enabled by this
91 ;; * SBCL is compiled with a higher level of OPTIMIZE DEBUG, so that
92 ;; the debugger can tell more about the state of the system.
93 ;; * Various code to print debugging messages, and similar debugging code,
94 ;; is compiled only when this feature is present.
96 ;; Note that the extra information recorded by the compiler at
97 ;; this higher level of OPTIMIZE DEBUG includes the source location
98 ;; forms. In order for the debugger to use this information, it has to
99 ;; re-READ the source file. In an ordinary installation of SBCL, this
100 ;; re-READing may not work very well, for either of two reasons:
101 ;; * The sources aren't present on the system in the same location that
102 ;; they were on the system where SBCL was compiled.
103 ;; * SBCL is using the standard readtable, without the added hackage
104 ;; which allows it to handle things like target features.
105 ;; If you want to be able to use the extra debugging information,
106 ;; therefore, be sure to keep the sources around, and run with the
107 ;; readtable configured so that the system sources can be read.
110 ;; Build SBCL with the old CMU CL low level debugger, "ldb". In the
111 ;; ideal world you would not need this unless you are messing with
112 ;; SBCL at a very low level (e.g., trying to diagnose GC problems, or
113 ;; trying to debug assembly code for a port to a new CPU). However,
114 ;; experience shows that sooner or later everyone lose()'s, in which
115 ;; case SB-LDB can at least provide an informative backtrace.
118 ;; This isn't really a target Lisp feature at all, but controls
119 ;; whether the build process produces an after-xc.core file. This
120 ;; can be useful for shortening the edit/compile/debug cycle when
121 ;; you modify SBCL's own source code, as in slam.sh. Otherwise
122 ;; you don't need it.
125 ;; Enable extra debugging output in the assem.lisp assembler/scheduler
126 ;; code. (This is the feature which was called :DEBUG in the
127 ;; original CMU CL code.)
130 ;; Setting this makes SBCL more "fluid", i.e. more amenable to
131 ;; modification at runtime, by suppressing various INLINE declarations,
132 ;; compiler macro definitions, FREEZE-TYPE declarations; and by
133 ;; suppressing various burning-our-ships-behind-us actions after
134 ;; initialization is complete; and so forth. This tends to clobber the
135 ;; performance of the system, so unless you have some special need for
136 ;; this when hacking SBCL itself, you don't want this set.
139 ;; Enable code for collecting statistics on usage of various operations,
140 ;; useful for performance tuning of the SBCL system itself. This code
141 ;; is probably pretty stale (having not been tested since the fork from
142 ;; base CMU CL) but might nonetheless be a useful starting point for
143 ;; anyone who wants to collect such statistics in the future.
146 ;; Peter Van Eynde's increase-bulletproofness code for CMU CL
148 ;; Some of the code which was #+high-security before the fork has now
149 ;; been either made unconditional, deleted, or rewritten into
150 ;; unrecognizability, but some remains. What remains is not maintained
151 ;; or tested in current SBCL, but I haven't gone out of my way to
155 ; :high-security-support
157 ;; low-level thread primitives support
159 ;; As of SBCL 0.8, this is only supposed to work in x86 Linux with
160 ;; NPTL support (usually kernel 2.6, though sme Red Hat distributions
161 ;; with older kernels also have it) and is implemented using clone(2)
162 ;; and the %fs segment register. Note that no consistent effort to
163 ;; audit the SBCL library code for thread safety has been performed,
164 ;; so caveat executor.
169 ;; While on linux we are able to use futexes for our locking
170 ;; primitive, on other platforms we don't have this luxury. NJF's
171 ;; lutexes present a locking API similar to the futex-based API that
172 ;; allows for sb-thread support on x86 OS X, Solaris and
177 ;; On some operating systems the FS segment register (used for SBCL's
178 ;; thread local storage) is not reliably preserved in signal
179 ;; handlers, so we need to restore its value from the pthread thread
181 ; :restore-tls-segment-register-from-tls
183 ;; Support for detection of unportable code (when applied to the
184 ;; COMMON-LISP package, or SBCL-internal pacakges) or bad-neighbourly
185 ;; code (when applied to user-level packages), relating to material
186 ;; alteration to packages or to bindings in symbols in packages.
189 ;; Support for the entirety of the 21-bit character space defined by
190 ;; the Unicode consortium, rather than the classical 8-bit ISO-8859-1
194 ;; Record source location information for variables, classes, conditions,
195 ;; packages, etc. Gives much better information on M-. in Slime, but
196 ;; increases core size by about 100kB.
199 ;; This affects the definition of a lot of things in bignum.lisp. It
200 ;; doesn't seem to be documented anywhere what systems it might apply
201 ;; to. It doesn't seem to be needed for X86 systems anyway.
204 ;; This is set in classic CMU CL, and presumably there it means
205 ;; that the floating point arithmetic implementation
206 ;; conforms to IEEE's standard. Here it definitely means that the
207 ;; floating point arithmetic implementation conforms to IEEE's standard.
208 ;; I (WHN 19990702) haven't tried to verify
209 ;; that it does conform, but it should at least mostly conform (because
210 ;; the underlying x86 hardware tries).
213 ;; CMU CL had, and we inherited, code to support 80-bit LONG-FLOAT on the x86
214 ;; architecture. Nothing has been done to actively destroy the long float
215 ;; support, but it hasn't been thoroughly maintained, and needs at least
216 ;; some maintenance before it will work. (E.g. the LONG-FLOAT-only parts of
217 ;; genesis are still implemented in terms of unportable CMU CL functions
218 ;; which are not longer available at genesis time in SBCL.) A deeper
219 ;; problem is SBCL's bootstrap process implicitly assumes that the
220 ;; cross-compilation host will be able to make the same distinctions
221 ;; between floating point types that it does. This assumption is
222 ;; fundamentally sleazy, even though in practice it's unlikely to break down
223 ;; w.r.t. distinguishing SINGLE-FLOAT from DOUBLE-FLOAT; it's much more
224 ;; likely to break down w.r.t. distinguishing DOUBLE-FLOAT from LONG-FLOAT.
225 ;; Still it's likely to be quite doable to get LONG-FLOAT support working
226 ;; again, if anyone's sufficiently motivated.
230 ;; miscellaneous notes on other things which could have special significance
231 ;; in the *FEATURES* list
234 ;; Any target feature which affects binary compatibility of fasl files
235 ;; needs to be recorded in *FEATURES-POTENTIALLY-AFFECTING-FASL-FORMAT*
238 ;; notes on the :NIL and :IGNORE features:
240 ;; #+NIL is used to comment out forms. Occasionally #+IGNORE is used
241 ;; for this too. So don't use :NIL or :IGNORE as the names of features..
243 ;; notes on :SB-XC and :SB-XC-HOST features (which aren't controlled by this
244 ;; file, but are instead temporarily pushed onto *FEATURES* or
245 ;; *TARGET-FEATURES* during some phases of cross-compilation):
247 ;; :SB-XC-HOST stands for "cross-compilation host" and is in *FEATURES*
248 ;; during the first phase of cross-compilation bootstrapping, when the
249 ;; host Lisp is being used to compile the cross-compiler.
251 ;; :SB-XC stands for "cross compiler", and is in *FEATURES* during the second
252 ;; phase of cross-compilation bootstrapping, when the cross-compiler is
253 ;; being used to create the first target Lisp.
255 ;; notes on the :SB-ASSEMBLING feature (which isn't controlled by
258 ;; This is a flag for whether we're in the assembler. It's
259 ;; temporarily pushed onto the *FEATURES* list in the setup for
260 ;; the ASSEMBLE-FILE function. It would be a bad idea
261 ;; to use it as a name for a permanent feature.
263 ;; notes on local features (which are set automatically by the
264 ;; configuration script, and should not be set here unless you
265 ;; really, really know what you're doing):
267 ;; machine architecture features:
269 ;; any Intel 386 or better, or compatibles like the AMD K6 or K7
271 ;; any x86-64 CPU running in 64-bit mode
273 ;; DEC/Compaq Alpha CPU
275 ;; any Sun UltraSPARC (possibly also non-Ultras -- currently untested)
281 ;; any MIPS CPU (in little-endian mode with :little-endian -- currently
284 ;; (CMU CL also had a :pentium feature, which affected the definition
285 ;; of some floating point vops. It was present but not enabled or
286 ;; documented in the CMU CL code that SBCL is derived from, and has
287 ;; now been moved to the backend-subfeatures mechanism.)
289 ;; properties derived from the machine architecture
290 ;; :control-stack-grows-downward-not-upward
291 ;; On the X86, the Lisp control stack grows downward. On the
292 ;; other supported CPU architectures as of sbcl-0.7.1.40, the
293 ;; system stack grows upward.
294 ;; Note that there are other stack-related differences between the
295 ;; X86 port and the other ports. E.g. on the X86, the Lisp control
296 ;; stack coincides with the C stack, meaning that on the X86 there's
297 ;; stuff on the control stack that the Lisp-level debugger doesn't
298 ;; understand very well. As of sbcl-0.7.1.40 things like that are
299 ;; just parameterized by #!+X86, but it'd probably be better to
300 ;; use new flags like :CONTROL-STACK-CONTAINS-C-STACK.
302 ;; :stack-allocatable-closures
303 ;; The compiler can allocate dynamic-extent closures on stack.
306 ;; Alien callbacks have been implemented for this platform.
308 ;; operating system features:
309 ;; :linux = We're intended to run under some version of Linux.
310 ;; :bsd = We're intended to run under some version of BSD Unix. (This
311 ;; is not exclusive with the features which indicate which
312 ;; particular version of BSD we're intended to run under.)
313 ;; :freebsd = We're intended to run under FreeBSD.
314 ;; :openbsd = We're intended to run under OpenBSD.
315 ;; :netbsd = We're intended to run under NetBSD.
316 ;; :darwin = We're intended to run under Darwin (including MacOS X).
317 ;; :sunos = We're intended to run under Solaris user environment
318 ;; with the SunOS kernel.
319 ;; :osf1 = We're intended to run under Tru64 (aka Digital Unix
321 ;; (No others are supported by SBCL as of 0.9.6, but :hpux or :irix
322 ;; support could be ported from CMU CL if anyone is sufficiently
323 ;; motivated to do so, and it'd even be possible, though harder, to
324 ;; port the system to Microsoft Windows.)