| blob | 2283291f6e47d8a5eb9265fb5ea80d5eac6bec17 |
1 ;;;; stuff which is not specific to any particular build phase, but
2 ;;;; used by most of them
3 ;;;;
4 ;;;; Note: It's specifically not used when bootstrapping PCL, because
5 ;;;; we do SAVE-LISP after that, and we don't want to save extraneous
6 ;;;; bootstrapping machinery into the frozen image which will
7 ;;;; subsequently be used as the mother of all Lisp sessions.
9 ;;;; This software is part of the SBCL system. See the README file for
10 ;;;; more information.
11 ;;;;
12 ;;;; This software is derived from the CMU CL system, which was
13 ;;;; written at Carnegie Mellon University and released into the
14 ;;;; public domain. The software is in the public domain and is
15 ;;;; provided with absolutely no warranty. See the COPYING and CREDITS
16 ;;;; files for more information.
18 ;;; SB-COLD holds stuff used to build the initial SBCL core file
19 ;;; (including not only the final construction of the core file, but
20 ;;; also the preliminary steps like e.g. building the cross-compiler
21 ;;; and running the cross-compiler to produce target FASL files).
33 ;; If only 1 integer, assume same parallelism for both passes.
36 ;; 0 means no parallelism. 1 means use at most one subjob,
37 ;; just in case you want to test the controlling loop.
40 ;; Parallelism on pass 1 works only if LOAD does not compile.
41 ;; Otherwise it's slower than compiling serially.
42 ;; (And this has only been tested with sb-fasteval, not sb-eval.)
44 value2)))))
55 ;;; If TRUE, then COMPILE-FILE is being invoked only to process
56 ;;; :COMPILE-TOPLEVEL forms, not to produce an output file.
57 ;;; This is part of the implementation of parallelized make-host-2.
60 ;;; prefixes for filename stems when cross-compiling. These are quite arbitrary
61 ;;; (although of course they shouldn't collide with anything we don't want to
62 ;;; write over). In particular, they can be either relative path names (e.g.
63 ;;; "host-objects/" or absolute pathnames (e.g. "/tmp/sbcl-xc-host-objects/").
64 ;;;
65 ;;; The cross-compilation process will force the creation of these directories
66 ;;; by executing CL:ENSURE-DIRECTORIES-EXIST (on the xc host Common Lisp).
71 ;; Target fasl files are LOADed (actually only quasi-LOADed, in
72 ;; GENESIS) only by SBCL code, and it doesn't care about particular
73 ;; extensions, so we can use something arbitrary.
76 ;; Target fasl files from SB!C:ASSEMBLE-FILE are LOADed via GENESIS.
77 ;; The source files are compiled once as assembly files and once as
78 ;; normal lisp files. In the past, they were kept separate by
79 ;; clever symlinking in the source tree, but that became less clean
80 ;; as ports to host environments without symlinks started appearing.
81 ;; In order to keep them separate, we have the assembled versions
82 ;; with a separate suffix.
85 ;;; a function of one functional argument, which calls its functional argument
86 ;;; in an environment suitable for compiling the target. (This environment
87 ;;; includes e.g. a suitable *FEATURES* value.)
91 ;;; a function with the same calling convention as CL:COMPILE-FILE, to be
92 ;;; used to translate ordinary Lisp source files into target object files
96 ;;; designator for a function with the same calling convention as
97 ;;; SB-C:ASSEMBLE-FILE, to be used to translate assembly files into target
98 ;;; object files
100 \f
101 ;;;; some tools
103 ;;; Take the file named X and make it into a file named Y. Sorta like
104 ;;; UNIX, and unlike Common Lisp's bare RENAME-FILE, we don't allow
105 ;;; information from the original filename to influence the final
106 ;;; filename. (The reason that it's only sorta like UNIX is that in
107 ;;; UNIX "mv foo bar/" will work, but the analogous
108 ;;; (RENAME-FILE-A-LA-UNIX "foo" "bar/") should fail.)
109 ;;;
110 ;;; (This is a workaround for the weird behavior of Debian CMU CL
111 ;;; 2.4.6, where (RENAME-FILE "dir/x" "dir/y") tries to create a file
112 ;;; called "dir/dir/y". If that behavior goes away, then we should be
113 ;;; able to get rid of this function and use plain RENAME-FILE in the
114 ;;; COMPILE-STEM function above. -- WHN 19990321
118 ;; example in the ANSI spec for TRUENAME.)
121 ;; From the ANSI spec: "In this case, the file is closed
122 ;; when the truename is tried, so the truename
123 ;; information is reliable."
129 ;;; other miscellaneous tools
134 ;;; Try to minimize/conceal any non-standardness of the host Common Lisp.
136 \f
137 ;;;; special read-macros for building the cold system (and even for
138 ;;;; building some of our tools for building the cold system)
142 ;;; When cross-compiling, the *FEATURES* set for the target Lisp is
143 ;;; not in general the same as the *FEATURES* set for the host Lisp.
144 ;;; In order to refer to target features specifically, we refer to
145 ;;; *SHEBANG-FEATURES* instead of *FEATURES*, and use the #!+ and #!-
146 ;;; readmacros instead of the ordinary #+ and #- readmacros.
150 nil
162 "target features *SHEBANG-FEATURES*=~%~@<~S~:>~%"
176 "target backend-subfeatures *SHEBANG-BACKEND-FEATURES*=~@<~S~:>~%"
179 ;;; Call for effect of signaling an error if no target picked.
182 ;;; You can get all the way through make-host-1 without either one of these
183 ;;; features, but then 'bit-bash' will fail to cross-compile.
187 ;;; Some feature combinations simply don't work, and sometimes don't
188 ;;; fail until quite a ways into the build. Pick off the more obvious
189 ;;; combinations now, and provide a description of what the actual
190 ;;; failure is (not always obvious from when the build fails).
205 sb-safepoint sb-thruption sb-wtimer
206 sb-dynamic-core)))"
209 ;; Subtle memory corruption follows when sb-dynamic-core is
210 ;; active, and non-threaded allocation routines have not been
211 ;; updated to take the additional indirection into account.
212 ;; Let's avoid this unusual combination.
215 ;; It sorta kinda works to have both, but there should be no need,
216 ;; and it's not really supported.
218 ;; There is still hope to make multithreading on DragonFly x86-64
228 failed-test-descriptions)))
229 \f
230 ;;;; cold-init-related PACKAGE and SYMBOL tools
232 ;;; Once we're done with possibly ANSIfying the COMMON-LISP package,
233 ;;; it's probably a mistake if we change it (beyond changing the
234 ;;; values of special variables such as *** and +, anyway). Set up
235 ;;; machinery to warn us when/if we change it.
236 ;;;
237 ;;; All code depending on this is itself dependent on #!+SB-SHOW.
238 #!+sb-show
242 \f
243 ;;;; master list of source files and their properties
245 ;;; flags which can be used to describe properties of source files
249 ;; cross-compiler which runs on the host ANSI Lisp. ("not host
250 ;; code", i.e. does not execute on host -- but may still be
251 ;; cross-compiled by the host, so that it executes on the target)
252 :not-host
253 ;; meaning: This file is not to be compiled as part of the target
254 ;; SBCL. ("not target code" -- but still presumably host code,
255 ;; used to support the cross-compilation process)
256 :not-target
257 ;; meaning: This file must always be compiled by 'slam.lisp' even if
258 ;; the object is not out of date with respect to its source.
259 ;; Necessary if there are compile-time-too effects that are not
260 ;; reflected into make-host-2 by load-time actions of make-host-1.
261 :slam-forcibly
262 ;; meaning: The #'COMPILE-STEM argument :TRACE-FILE should be T.
263 ;; When the compiler is SBCL's COMPILE-FILE or something like it,
264 ;; compiling "foo.lisp" will generate "foo.trace" which contains lots
265 ;; of exciting low-level information about representation selection,
266 ;; VOPs used by the compiler, and bits of assembly.
267 :trace-file
268 ;; meaning: This file is to be processed with the SBCL assembler,
269 ;; not COMPILE-FILE. (Note that this doesn't make sense unless
270 ;; :NOT-HOST is also set, since the SBCL assembler doesn't exist
271 ;; while the cross-compiler is being built in the host ANSI Lisp.)
272 :assem
273 ;; meaning: The #'COMPILE-STEM argument called :IGNORE-FAILURE-P
274 ;; should be true. (This is a KLUDGE: I'd like to get rid of it.
275 ;; For now, it exists so that compilation can proceed through the
276 ;; legacy warnings in src/compiler/x86/array.lisp, which I've
277 ;; never figured out but which were apparently acceptable in CMU
278 ;; CL. Eventually, it would be great to just get rid of all
279 ;; warnings and remove support for this flag. -- WHN 19990323)
291 ,@body
294 ;;; Given a STEM, remap the path component "/target/" to a suitable
295 ;;; target directory.
303 stem)))
306 ;;; Determine the source path for a stem.
311 ;;; Determine the object path for a stem/flags/mode combination.
316 (:host-compile
317 ;; On some xc hosts, it's impossible to LOAD a fasl file unless it
318 ;; has the same extension that the host uses for COMPILE-FILE
319 ;; output, so we have to be careful to use the xc host's preferred
320 ;; extension.
331 ;;; Check for stupid typos in FLAGS list keywords.
334 ;; We do duplicate stem comparison based on the object path in
335 ;; order to cover the case of stems with an :assem flag, which
336 ;; have two entries but separate object paths for each. KLUDGE:
337 ;; We have to bind *target-obj-prefix* here because it's normally
338 ;; set up later in the build process and we don't actually care
339 ;; what it is so long as it doesn't change while we're checking
340 ;; for duplicate stems.
346 ;; FIXME: We should make sure that the :assem flag is only used
347 ;; when paired with :not-host.
351 set-difference)))))
352 \f
353 ;;;; tools to compile SBCL sources to create the cross-compiler
355 ;;; a wrapper for compilation/assembly, used mostly to centralize
356 ;;; the procedure for finding full filenames from "stems"
357 ;;;
358 ;;; Compile the source file whose basic name is STEM, using some
359 ;;; standard-for-the-SBCL-build-process procedures to generate the
360 ;;; full pathnames of source file and object file. Return the pathname
361 ;;; of the object file for STEM.
362 ;;;
363 ;;; STEM and FLAGS are as per DO-STEMS-AND-FLAGS. MODE is one of
364 ;;; :HOST-COMPILE and :TARGET-COMPILE.
368 ;; Lisp Way, although it works just fine for common UNIX environments.
369 ;; Should it come to pass that the system is ported to environments
370 ;; where version numbers and so forth become an issue, it might become
371 ;; urgent to rewrite this using the fancy Common Lisp PATHNAME
372 ;; machinery instead of just using strings. In the absence of such a
373 ;; port, it might or might be a good idea to do the rewrite.
374 ;; -- WHN 19990815
377 ;; Compile-for-effect happens simultaneously with a forked compile,
378 ;; so we need the for-effect output not to stomp on the real output.
394 ;; We're about to set about building a new object file. First, we
395 ;; delete any preexisting object file in order to avoid confusing
396 ;; ourselves later should we happen to bail out of compilation
397 ;; with an error.
401 ;; Original comment:
402 ;;
403 ;; Work around a bug in CLISP 1999-01-08 #'COMPILE-FILE: CLISP
404 ;; mangles relative pathnames passed as :OUTPUT-FILE arguments,
405 ;; but works OK with absolute pathnames.
406 ;;
407 ;; following discussion on cmucl-imp 2002-07
408 ;; "COMPILE-FILE-PATHNAME", it would seem safer to deal with
409 ;; absolute pathnames all the time; it is no longer clear that the
410 ;; original behaviour in CLISP was wrong or that the current
411 ;; behaviour is right; and in any case absolutifying the pathname
412 ;; insulates us against changes of behaviour. -- CSR, 2002-08-09
414 ;; (Note that this idiom is taken from the ANSI
415 ;; documentation for TRUENAME.)
418 ;; Compilation would overwrite the
419 ;; temporary object anyway and overly
420 ;; strict implementations default
421 ;; to :ERROR.
425 ;; and some compilers (e.g. OpenMCL) will complain if they're
426 ;; asked to write over a file that exists already (and isn't
427 ;; recognizeably a fasl file), so
431 ;; Try to use the compiler to generate a new temporary object file.
437 retry-compile-file
449 src)
453 obj)
455 :report report-recompile-restart
458 :report report-continue-restart
460 ;; Don't leave failed object files lying around.
464 ;; Otherwise: success, just fall through.
467 ;; If we get to here, compilation succeeded, so it's OK to rename
468 ;; the temporary output file to the permanent object file.
474 ;; nice friendly traditional return value
478 ;;; Execute function FN in an environment appropriate for compiling the
479 ;;; cross-compiler's source code in the cross-compilation host.
483 ;; the CROSS-FLOAT-INFINITY-KLUDGE, as documented in
484 ;; base-target-features.lisp-expr:
486 '(:sb-propagate-float-type
492 ;;; Process a file as source code for the cross-compiler, compiling it
493 ;;; (if necessary) in the appropriate environment, then loading it
494 ;;; into the cross-compilation host Common lisp.
502 ;;; like HOST-CLOAD-STEM, except that we don't bother to compile
506 \f
507 ;;;; tools to compile SBCL sources to create object files which will
508 ;;;; be used to create the target SBCL .core file
510 ;;; Run the cross-compiler on a file in the source directory tree to
511 ;;; produce a corresponding file in the target object directory tree.
518 ;;; (This function is not used by the build process, but is intended
519 ;;; for interactive use when experimenting with the system. It runs
520 ;;; the cross-compiler on test files with arbitrary filenames, not
521 ;;; necessarily in the source tree, e.g. in "/tmp".)
534 ;; Dependence of the assembler on the compiler
535 ;; feels a bit backwards, but assembly needs
536 ;; TN-SC, TN-OFFSET, etc. because the compiler
537 ;; doesn't speak the assembler's language.
538 ;; Rather vice-versa.
540 ;; Both SB-ASSEM and SB-DISASSEM export these two symbols.
541 ;; Neither is shadowing-imported. If you need one, package-qualify it.
544 pkg))