| blob | 57142af36d85548c95893c01f0b651ef9abed3ff |
1 ;;;; the top level interfaces to the compiler, plus some other
2 ;;;; compiler-related stuff (e.g. CL:CALL-ARGUMENTS-LIMIT) which
3 ;;;; doesn't obviously belong anywhere else
5 ;;;; This software is part of the SBCL system. See the README file for
6 ;;;; more information.
7 ;;;;
8 ;;;; This software is derived from the CMU CL system, which was
9 ;;;; written at Carnegie Mellon University and released into the
10 ;;;; public domain. The software is in the public domain and is
11 ;;;; provided with absolutely no warranty. See the COPYING and CREDITS
12 ;;;; files for more information.
16 ;;; FIXME: Doesn't this belong somewhere else, like early-c.lisp?
33 ;;; Whether reference to a thing which cannot be defined causes a full
34 ;;; warning.
39 ;;; Set to NIL to disable loop analysis for register allocation.
42 ;;; Bind this to a stream to capture various internal debugging output.
45 ;;; The current block compilation state. These are initialized to the
46 ;;; :BLOCK-COMPILE and :ENTRY-POINTS arguments that COMPILE-FILE was
47 ;;; called with.
48 ;;;
49 ;;; *BLOCK-COMPILE-ARG* holds the original value of the :BLOCK-COMPILE
50 ;;; argument, which overrides any internal declarations.
57 ;;; When block compiling, used by PROCESS-FORM to accumulate top level
58 ;;; lambdas resulting from compiling subforms. (In reverse order.)
62 ;;; The current non-macroexpanded toplevel form as printed when
63 ;;; *compile-print* is true.
67 #!+sb-doc
70 #!+sb-doc
73 #!+sb-doc
74 "When this is true, the compiler prints to *STANDARD-OUTPUT* progress
75 information about the phases of compilation of each function. (This
79 #!+sb-doc
80 "The defaulted pathname of the file currently being compiled, or NIL if not
83 #!+sb-doc
84 "The TRUENAME of the file currently being compiled, or NIL if not
91 ;;; the SOURCE-INFO structure for the current compilation. This is
92 ;;; null globally to indicate that we aren't currently in any
93 ;;; identifiable compilation.
96 ;;; This is true if we are within a WITH-COMPILATION-UNIT form (which
97 ;;; normally causes nested uses to be no-ops).
100 ;;; Count of the number of compilation units dynamically enclosed by
101 ;;; the current active WITH-COMPILATION-UNIT that were unwound out of.
104 ;;; Mumble conditional on *COMPILE-PROGRESS*.
121 ;; Used during compilation to map code paths to the matching
122 ;; instrumentation conses.
124 ;; Used during compilation to keep track of with source paths have been
125 ;; instrumented in which blocks.
127 ;; Stores the code coverage instrumentation results. Keys are namestrings,
128 ;; the value is a list of (CONS PATH STATE), where STATE is NIL for
129 ;; a path that has not been visited, and T for one that has.
132 \f
133 ;;;; WITH-COMPILATION-UNIT and WITH-COMPILATION-VALUES
136 #!+sb-doc
137 "Affects compilations that take place within its dynamic extent. It is
138 intended to be eg. wrapped around the compilation of all files in the same system.
140 Following options are defined:
142 :OVERRIDE Boolean-Form
143 One of the effects of this form is to delay undefined warnings until the
144 end of the form, instead of giving them at the end of each compilation.
145 If OVERRIDE is NIL (the default), then the outermost
146 WITH-COMPILATION-UNIT form grabs the undefined warnings. Specifying
147 OVERRIDE true causes that form to grab any enclosed warnings, even if it
148 is enclosed by another WITH-COMPILATION-UNIT.
150 :POLICY Optimize-Declaration-Form
151 Provides dynamic scoping for global compiler optimization qualities and
152 restrictions, limiting effects of subsequent OPTIMIZE proclamations and
153 calls to SB-EXT:RESTRICT-COMPILER-POLICY to the dynamic scope of BODY.
155 If OVERRIDE is false, specified POLICY is merged with current global
156 policy. If OVERRIDE is true, current global policy, including any
157 restrictions, is discarded in favor of the specified POLICY.
159 Supplying POLICY NIL is equivalent to the option not being supplied at
160 all, ie. dynamic scoping of policy does not take place.
162 This option is an SBCL-specific experimental extension: Interface
163 subject to change.
165 :SOURCE-NAMESTRING Namestring-Form
166 Attaches the value returned by the Namestring-Form to the internal
167 debug-source information as the namestring of the source file. Normally
168 the namestring of the input-file for COMPILE-FILE is used: this option
169 can be used to provide source-file information for functions compiled
170 using COMPILE, or to override the input-file of COMPILE-FILE.
172 If both an outer and an inner WITH-COMPILATION-UNIT provide a
173 SOURCE-NAMESTRING, the inner one takes precedence. Unaffected
174 by :OVERRIDE.
176 This is an SBCL-specific extension.
178 :SOURCE-PLIST Plist-Form
179 Attaches the value returned by the Plist-Form to internal debug-source
180 information of functions compiled in within the dynamic extent of BODY.
182 Primarily for use by development environments, in order to eg. associate
183 function definitions with editor-buffers. Can be accessed using
184 SB-INTROSPECT:DEFINITION-SOURCE-PLIST.
186 If an outer WITH-COMPILATION-UNIT form also provide a SOURCE-PLIST, it
187 is appended to the end of the provided SOURCE-PLIST. Unaffected
188 by :OVERRIDE.
190 This is an SBCL-specific extension.
192 Examples:
194 ;; Prevent proclamations from the file leaking, and restrict
195 ;; SAFETY to 3 -- otherwise uses the current global policy.
196 (with-compilation-unit (:policy '(optimize))
197 (restrict-compiler-policy 'safety 3)
200 ;; Using default policy instead of the current global one,
201 ;; except for DEBUG 3.
202 (with-compilation-unit (:policy '(optimize debug)
203 :override t)
206 ;; Same as if :POLICY had not been specified at all: SAFETY 3
207 ;; proclamation leaks out from WITH-COMPILATION-UNIT.
208 (with-compilation-unit (:policy nil)
209 (declaim (optimize safety))
211 "
224 ;; Inside another WITH-COMPILATION-UNIT, a WITH-COMPILATION-UNIT is
225 ;; ordinarily (unless OVERRIDE) basically a no-op.
253 ;;; Is NAME something that no conforming program can rely on
254 ;;; defining?
257 (:function
260 (:type
267 ;;; This is to be called at the end of a compilation unit. It signals
268 ;;; any residual warnings about unknown stuff, then prints the total
269 ;;; error counts. ABORT-P should be true when the compilation unit was
270 ;;; aborted by throwing out. ABORT-COUNT is the number of dynamically
271 ;;; enclosed nested compilation units that were aborted.
298 #+sb-xc-host nil
300 (:function
302 "~@<The function ~S is undefined, and its name is ~
303 reserved by ANSI CL so that even if it were ~
304 defined later, the code doing so would not be ~
306 (:type
309 "~@<Undefined type ~S. The name starts with ~S: ~
310 probably use of a quoted type name in a context ~
311 where the name is not evaluated.~:@>"
314 "~@<Undefined type ~S. Note that name ~S is ~
315 reserved by ANSI CL, so code defining a type with ~
316 that name would not be portable.~:@>" name
317 name))))
326 "~W more use~:P of undefined ~(~A~) ~S"
327 more kind name)
329 "~W more use~:P of undefined ~(~A~) ~S"
330 more kind name))))))))))
340 abort-p)
344 "~& Undefined ~(~A~)~p:~
345 ~% ~{~<~% ~1:;~S~>~^ ~}"
348 ~[~:;~:*~& caught ~W ERROR condition~:P~]~
349 ~[~:;~:*~& caught ~W WARNING condition~:P~]~
350 ~[~:;~:*~& caught ~W STYLE-WARNING condition~:P~]~
351 ~[~:;~:*~& printed ~W note~:P~]"
360 ;;; Evaluate BODY, then return (VALUES BODY-VALUE WARNINGS-P
361 ;;; FAILURE-P), where BODY-VALUE is the first value of the body, and
362 ;;; WARNINGS-P and FAILURE-P are as in CL:COMPILE or CL:COMPILE-FILE.
363 ;;; This also wraps up WITH-IR1-NAMESPACE functionality.
365 ;; These bindings could just as well be in WITH-IR1-NAMESPACE, but
366 ;; since they're primarily debugging tools, it's nicer to have
367 ;; a wider unique scope by ID.
392 \f
393 ;;;; component compilation
396 #!+sb-doc
397 "The upper limit on the number of times that we will consecutively do IR1
398 optimization that doesn't introduce any new code. A finite limit is
404 ;;; Repeatedly optimize COMPONENT until no further optimizations can
405 ;;; be found or we hit our iteration limit. When we hit the limit, we
406 ;;; clear the component and block REOPTIMIZE flags to discourage the
407 ;;; next optimization attempt from pounding on the same code.
451 ;;; KLUDGE: This was bumped from 5 to 10 in a DTC patch ported by MNA
452 ;;; from CMU CL into sbcl-0.6.11.44, the same one which allowed IR1
453 ;;; transforms to be delayed. Either DTC or MNA or both didn't explain
454 ;;; why, and I don't know what the rationale was. -- WHN 2001-04-28
455 ;;;
456 ;;; FIXME: It would be good to document why it's important to have a
457 ;;; large value here, and what the drawbacks of an excessively large
458 ;;; value are; and it might also be good to make it depend on
459 ;;; optimization policy.
465 ;;; Iterate doing FIND-DFO until no new dead code is discovered.
478 ;;; Do all the IR1 phases for a non-top-level component.
483 ;make this 0.
506 ;; We delay the generation of type checks until
507 ;; the type constraints have had time to
508 ;; propagate, else the compiler can confuse itself.
533 #!+inline-constants
535 #!+inline-constants
537 #!+inline-constants
552 ;; Assign BLOCK-NUMBER for any cleanup blocks introduced by
553 ;; stack analysis. There shouldn't be any unreachable code after
554 ;; control, so this won't delete anything.
604 #-sb-xc-host
616 code-length
617 fixup-notes
623 code-length
624 fixup-notes
628 ;; We're done, so don't bother keeping anything around.
633 ;;; Delete components with no external entry points before we try to
634 ;;; generate code. Unreachable closures can cause IR2 conversion to
635 ;;; puke on itself, since it is the reference to the closure which
636 ;;; normally causes the components to be combined.
643 (:external
651 ;; miscellaneous sanity checks
652 ;;
653 ;; FIXME: These are basically pretty wimpy compared to the checks done
654 ;; by the old CHECK-IR1-CONSISTENCY code. It would be really nice to
655 ;; make those internal consistency checks work again and use them.
660 ;; sanity check to prevent weirdness from propagating insidiously as
661 ;; far from its root cause as it did in bug 138: Make sure that
662 ;; thing-to-COMPONENT links are consistent.
668 ;; Record xref information before optimization. This way the
669 ;; stored xref data reflects the real source as closely as
670 ;; possible.
680 #|
692 |#
694 ;; FIXME: What is MAYBE-MUMBLE for? Do we need it any more?
708 \f
709 ;;;; clearing global data structures
710 ;;;;
711 ;;;; FIXME: Is it possible to get rid of this stuff, getting rid of
712 ;;;; global data structures entirely when possible and consing up the
713 ;;;; others from scratch instead of clearing and reusing them?
715 ;;; Clear the INFO in constants in the *FREE-VARS*, etc. In
716 ;;; addition to allowing stuff to be reclaimed, this is required for
717 ;;; correct assignment of constant offsets, since we need to assign a
718 ;;; new offset for each component. We don't clear the FUNCTIONAL-INFO
719 ;;; slots, since they are used to keep track of functions across
720 ;;; component boundaries.
735 ;;; Blow away the REFS for all global variables, and let COMPONENT
736 ;;; be recycled.
748 x))
755 \f
756 ;;;; trace output
758 ;;; Print out some useful info about COMPONENT to STREAM.
779 \f
780 ;;;; file reading
781 ;;;;
782 ;;;; When reading from a file, we have to keep track of some source
783 ;;;; information. We also exploit our ability to back up for printing
784 ;;;; the error context and for recovering from errors.
785 ;;;;
786 ;;;; The interface we provide to this stuff is the stream-oid
787 ;;;; SOURCE-INFO structure. The bookkeeping is done as a side effect
788 ;;;; of getting the next source form.
790 ;;; A FILE-INFO structure holds all the source information for a
791 ;;; given file.
794 #-no-ansi-print-object
798 ;; If a file, the truename of the corresponding source file. If from
799 ;; a Lisp form, :LISP. If from a stream, :STREAM.
801 ;; the external format that we'll call OPEN with, if NAME is a file.
803 ;; the defaulted, but not necessarily absolute file name (i.e. prior
804 ;; to TRUENAME call.) Null if not a file. This is used to set
805 ;; *COMPILE-FILE-PATHNAME*, and if absolute, is dumped in the
806 ;; debug-info.
808 ;; the file's write date (if relevant)
810 ;; the source path root number of the first form in this file (i.e.
811 ;; the total number of forms converted previously in this
812 ;; compilation)
814 ;; parallel vectors containing the forms read out of the file and
815 ;; the file positions that reading of each form started at (i.e. the
816 ;; end of the previous form)
820 ;;; The SOURCE-INFO structure provides a handle on all the source
821 ;;; information for an entire compilation.
823 #-no-ansi-print-object
827 ;; the UT that compilation started at
829 ;; the IRT that compilation started at
831 ;; the FILE-INFO structure for this compilation
833 ;; the stream that we are using to read the FILE-INFO, or NIL if
834 ;; no stream has been opened yet
836 ;; for coalescing DEFINITION-SOURCE-LOCATION of effectively toplevel forms
837 ;; inside one truly toplevel form.
839 ;; if the current compilation is recursive (e.g., due to EVAL-WHEN
840 ;; processing at compile-time), the invoking compilation's
841 ;; source-info.
844 ;;; Given a pathname, return a SOURCE-INFO structure.
849 :external-format external-format
852 ;;; Return a SOURCE-INFO to describe the incremental compilation of FORM.
860 ;;; Walk up the SOURCE-INFO list until we either reach a SOURCE-INFO
861 ;;; with no parent (e.g., from a REPL evaluation) or until we reach a
862 ;;; SOURCE-INFO whose FILE-INFO denotes a file.
870 finfo))))
872 ;;; Return a form read from STREAM; or for EOF use the trick,
873 ;;; popularized by Kent Pitman, of returning STREAM itself. If an
874 ;;; error happens, then convert it to standard abort-the-compilation
875 ;;; error condition (possibly recording some extra location
876 ;;; information). CONDITION-NAME is what to signal on error,
877 ;;; and should be INPUT-ERROR-IN-COMPILE-FILE or a subclass of it.
878 ;;; The signaled condition encapsulates a reader condition.
884 ;; We don't need to supply :POSITION here because
885 ;; READER-ERRORs already know their position in the file.
886 :condition condition
888 ;; ANSI, in its wisdom, says that READ should return END-OF-FILE
889 ;; (and that this is not a READER-ERROR) when it encounters end of
890 ;; file in the middle of something it's trying to read.
893 :condition condition
894 ;; We need to supply :POSITION here because the END-OF-FILE
895 ;; condition doesn't carry the position that the user
896 ;; probably cares about, where the failed READ began.
897 :position position
901 :condition condition
902 :position position
905 ;;; If STREAM is present, return it, otherwise open a stream to the
906 ;;; current file. There must be a current file.
907 ;;;
908 ;;; FIXME: This is probably an unnecessarily roundabout way to do
909 ;;; things now that we process a single file in COMPILE-FILE (unlike
910 ;;; the old CMU CL code, which accepted multiple files). Also, the old
911 ;;; comment said
912 ;;; When we open a new file, we also reset *PACKAGE* and policy.
913 ;;; This gives the effect of rebinding around each file.
914 ;;; which doesn't seem to be true now. Check to make sure that if
915 ;;; such rebinding is necessary, it's still done somewhere.
928 ;;; Close the stream in INFO if it is open.
936 ;;; Loop over FORMS retrieved from INFO. Used by COMPILE-FILE and
937 ;;; LOAD when loading from a FILE-STREAM associated with a source
938 ;;; file. ON-ERROR is the name of a condition class that should
939 ;;; be signaled if anything goes wrong during a READ.
956 file-info))))
959 file-info))
963 ;;; Read and compile the source file.
971 ;; It's easy to get into a situation where cold-init crashes and the only
972 ;; backtrace you get from ldb is TOP-LEVEL-FORM, which means you're anywhere
973 ;; within the 23000 or so blobs of code deferred until cold-init.
974 ;; Seeing each file finish narrows things down without the noise of :sb-show,
975 ;; but this hack messes up form positions, so it's not on unless asked for.
990 ;;; Return the INDEX'th source form read from INFO and the position
991 ;;; where it was read.
997 \f
998 ;;;; processing of top level forms
1000 ;;; This is called by top level form processing when we are ready to
1001 ;;; actually compile something. If *BLOCK-COMPILE* is T, then we still
1002 ;;; convert the form, but delay compilation, pushing the result on
1003 ;;; *TOPLEVEL-LAMBDAS* instead.
1007 ;; Don't bother to compile simple objects that just sit there.
1021 nil))))))
1023 ;;; Macroexpand FORM in the current environment with an error handler.
1024 ;;; We only expand one level, so that we retain all the intervening
1025 ;;; forms in the source path. A compiler-macro takes precedence over
1026 ;;; an ordinary macro as specified in CLHS 3.2.3.1
1027 ;;; Note that this function is _only_ for processing of toplevel forms.
1028 ;;; Non-toplevel forms use IR1-CONVERT-FUNCTOID which considers compiler macros.
1041 condition))))
1043 ;;; Process a PROGN-like portion of a top level form. FORMS is a list of
1044 ;;; the forms, and PATH is the source path of the FORM they came out of.
1045 ;;; COMPILE-TIME-TOO is as in ANSI "3.2.3.1 Processing of Top Level Forms".
1051 ;;; Process a top level use of LOCALLY, or anything else (e.g.
1052 ;;; MACROLET) at top level which has declarations and ordinary forms.
1053 ;;; We parse declarations and then recursively process the body.
1060 ;; FIXME: VALUES declaration
1061 ;;
1062 ;; Binding *POLICY* is pretty much of a hack, since it
1063 ;; causes LOCALLY to "capture" enclosed proclamations. It
1064 ;; is necessary because CONVERT-AND-MAYBE-COMPILE uses the
1065 ;; value of *POLICY* as the policy. The need for this hack
1066 ;; is due to the quirk that there is no way to represent in
1067 ;; a POLICY that an optimize quality came from the default.
1068 ;;
1069 ;; FIXME: Ideally, something should be done so that DECLAIM
1070 ;; inside LOCALLY works OK. Failing that, at least we could
1071 ;; issue a warning instead of silently screwing up.
1073 ;; This is probably also a hack
1075 ;; ditto
1079 ;;; Parse an EVAL-WHEN situations list, returning three flags,
1080 ;;; (VALUES COMPILE-TOPLEVEL LOAD-TOPLEVEL EXECUTE), indicating
1081 ;;; the types of situations present in the list.
1085 '(:compile-toplevel
1086 compile
1087 :load-toplevel
1088 load
1089 :execute
1090 eval)))
1095 deprecated-names)))
1097 situations)
1102 ;;; utilities for extracting COMPONENTs of FUNCTIONALs
1113 result))))
1117 result))))
1120 &key
1121 name
1123 ;; I'd thought NIL should
1124 ;; work, but it doesn't.
1125 ;; -- WHN 2001-09-20
1136 lambda-expression
1138 ;; Convert the XEP using the policy of the real function. Otherwise
1139 ;; the wrong policy will be used for deciding whether to type-check
1140 ;; the parameters of the real function (via CONVERT-CALL /
1141 ;; PROPAGATE-TO-ARGS). -- JES, 2007-02-27
1144 :source-name source-name
1156 ;; Any leftover REFs to FUN outside local calls get replaced with the
1157 ;; XEP.
1164 xep
1165 fun)
1166 xep)))
1168 ;;; Compile LAMBDA-EXPRESSION into *COMPILE-OBJECT*, returning a
1169 ;;; description of the result.
1170 ;;; * If *COMPILE-OBJECT* is a CORE-OBJECT, then write the function
1171 ;;; into core and return the compiled FUNCTION value.
1172 ;;; * If *COMPILE-OBJECT* is a fasl file, then write the function
1173 ;;; into the fasl file and return a dump handle.
1174 ;;;
1175 ;;; If NAME is provided, then we try to use it as the name of the
1176 ;;; function for debugging/diagnostic information.
1179 &key
1180 name
1182 ;; This magical idiom seems to be the appropriate
1183 ;; path for compiling standalone LAMBDAs, judging
1184 ;; from the CMU CL code and experiment, so it's a
1185 ;; nice default for things where we don't have a
1186 ;; real source path (as in e.g. inside CL:COMPILE).
1197 :name name
1200 ;; FIXME: The compile-it code from here on is sort of a
1201 ;; twisted version of the code in COMPILE-TOPLEVEL. It'd be
1202 ;; better to find a way to share the code there; or
1203 ;; alternatively, to use this code to replace the code there.
1204 ;; (The second alternative might be pretty easy if we used
1205 ;; the :LOCALL-ONLY option to IR1-FOR-LAMBDA. Then maybe the
1206 ;; whole FUNCTIONAL-KIND=:TOPLEVEL case could go away..)
1224 result
1225 ;; KLUDGE: This code duplicates some other code in this
1226 ;; file. In the great reorganzation, the flow of program
1227 ;; logic changed from the original CMUCL model, and that
1228 ;; path (as of sbcl-0.7.5 in SUB-COMPILE-FILE) was no
1229 ;; longer followed for CORE-OBJECTS, leading to BUG
1230 ;; 156. This place is transparently not the right one for
1231 ;; this code, but I don't have a clear enough overview of
1232 ;; the compiler to know how to rearrange it all so that
1233 ;; this operation fits in nicely, and it was blocking
1234 ;; reimplementation of (DECLAIM (INLINE FOO)) (MACROLET
1235 ;; ((..)) (DEFUN FOO ...))
1236 ;;
1237 ;; FIXME: This KLUDGE doesn't solve all the problem in an
1238 ;; ideal way, as (1) definitions typed in at the REPL
1239 ;; without an INLINE declaration will give a NULL
1240 ;; FUNCTION-LAMBDA-EXPRESSION (allowable, but not ideal)
1241 ;; and (2) INLINE declarations will yield a
1242 ;; FUNCTION-LAMBDA-EXPRESSION headed by
1243 ;; SB-C:LAMBDA-WITH-LEXENV, even for null LEXENV. -- CSR,
1244 ;; 2002-07-02
1245 ;;
1246 ;; (2) is probably fairly easy to fix -- it is, after all,
1247 ;; a matter of list manipulation (or possibly of teaching
1248 ;; CL:FUNCTION about SB-C:LAMBDA-WITH-LEXENV). (1) is
1249 ;; significantly harder, as the association between
1250 ;; function object and source is a tricky one.
1251 ;;
1252 ;; FUNCTION-LAMBDA-EXPRESSION "works" (i.e. returns a
1253 ;; non-NULL list) when the function in question has been
1254 ;; compiled by (COMPILE <x> '(LAMBDA ...)); it does not
1255 ;; work when it has been compiled as part of the top-level
1256 ;; EVAL strategy of compiling everything inside (LAMBDA ()
1257 ;; ...). -- CSR, 2002-11-02
1274 form)
1283 form)
1287 ;;; Handle the evaluation the a :COMPILE-TOPLEVEL body during
1288 ;;; compilation. Normally just evaluate in the appropriate
1289 ;;; environment, but also compile if outputting a CFASL.
1302 ;; In case a null lexenv is created, it needs to get the newly
1303 ;; effective global policy, not the policy currently in *POLICY*.
1307 ;;; Process a top level FORM with the specified source PATH.
1308 ;;; * If this is a magic top level form, then do stuff.
1309 ;;; * If this is a macro, then expand it.
1310 ;;; * Otherwise, just compile it.
1311 ;;;
1312 ;;; COMPILE-TIME-TOO is as defined in ANSI
1313 ;;; "3.2.3.1 Processing of Top Level Forms".
1324 path)
1329 ;; When we're cross-compiling, consider: what should we
1330 ;; do when we hit e.g.
1331 ;; (EVAL-WHEN (:COMPILE-TOPLEVEL)
1332 ;; (DEFUN FOO (X) (+ 7 X)))?
1333 ;; DEFUN has a macro definition in the cross-compiler,
1334 ;; and a different macro definition in the target
1335 ;; compiler. The only sensible thing is to use the
1336 ;; target compiler's macro definition, since the
1337 ;; cross-compiler's macro is in general into target
1338 ;; functions which can't meaningfully be executed at
1339 ;; cross-compilation time. So make sure we do the EVAL
1340 ;; here, before we macroexpand.
1341 ;;
1342 ;; Then things get even dicier with something like
1343 ;; (DEFCONSTANT-EQX SB!XC:LAMBDA-LIST-KEYWORDS ..)
1344 ;; where we have to make sure that we don't uncross
1345 ;; the SB!XC: prefix before we do EVAL, because otherwise
1346 ;; we'd be trying to redefine the cross-compilation host's
1347 ;; constants.
1348 ;;
1349 ;; (Isn't it fun to cross-compile Common Lisp?:-)
1350 #+sb-xc-host
1355 ;; like SB!XC:DEFCONSTANT and SB!XC:DEFTYPE
1356 ;; should be equivalent to their CL: counterparts
1357 ;; when being compiled as target code. We leave
1358 ;; the rest of the form uncrossed because macros
1359 ;; might yet expand into EVAL-WHEN stuff, and
1360 ;; things inside EVAL-WHEN can't be uncrossed
1361 ;; until after we've EVALed them in the
1362 ;; cross-compilation host.)
1366 slightly-uncrossed)))
1368 ;; (Now that we're no longer processing toplevel
1369 ;; forms, and hence no longer need to worry about
1370 ;; EVAL-WHEN, we can uncross everything.)
1372 ;; (We have to demote COMPILE-TIME-TOO to NIL
1373 ;; here, no matter what it was before, since
1374 ;; otherwise we'd tend to EVAL subforms more than
1375 ;; once, because of WHEN COMPILE-TIME-TOO form
1376 ;; above.)
1378 ;; When we're not cross-compiling, we only need to
1379 ;; macroexpand once, so we can follow the 1-thru-6
1380 ;; sequence of steps in ANSI's "3.2.3.1 Processing of
1381 ;; Top Level Forms".
1382 #-sb-xc-host
1390 path
1391 compile-time-too)))))))
1393 #+sb-xc-host
1394 ;; (There are no xc EVAL-WHEN issues in the ATOM case until
1395 ;; (1) SBCL gets smart enough to handle global
1396 ;; DEFINE-SYMBOL-MACRO or SYMBOL-MACROLET and (2) SBCL
1397 ;; implementors start using symbol macros in a way which
1398 ;; interacts with SB-XC/CL distinction.)
1400 #-sb-xc-host
1406 form))))
1413 ;; CT, LT, and E here are as in Figure 3-7 of ANSI
1414 ;; "3.2.3.1 Processing of Top Level Forms".
1419 e))))
1421 body path new-compile-time-too))
1426 magic
1431 path
1432 compile-time-too))
1436 magic
1440 path
1441 compile-time-too
1451 \f
1452 ;;;; load time value support
1453 ;;;;
1454 ;;;; (See EMIT-MAKE-LOAD-FORM.)
1456 ;;; Return T if we are currently producing a fasl file and hence
1457 ;;; constants need to be dumped carefully.
1461 ;;; Compile FORM and arrange for it to be called at load-time. Return
1462 ;;; the dumper handle and our best guess at the type of the object.
1472 ;;; Compile the FORMS and arrange for them to be called (for effect,
1473 ;;; not value) at load time.
1478 ;;; Do the actual work of COMPILE-LOAD-TIME-VALUE or
1479 ;;; COMPILE-MAKE-LOAD-FORM-INIT-FORMS.
1485 lambda)))
1487 ;;; This is called by COMPILE-TOPLEVEL when it was passed T for
1488 ;;; LOAD-TIME-VALUE-P (which happens in COMPILE-LOAD-TIME-STUFF). We
1489 ;;; don't try to combine this component with anything else and frob
1490 ;;; the name. If not in a :TOPLEVEL component, then don't bother
1491 ;;; compiling, because it was merged with a run-time component.
1500 \f
1501 ;;;; COMPILE-FILE
1511 ;;; Smash LAMBDAS into a single component, compile it, and arrange for
1512 ;;; the resulting function to be called.
1522 ;;; Compile top level code and call the top level lambdas. We pick off
1523 ;;; top level lambdas in non-top-level components here, calling
1524 ;;; SUB-c-t-l-l on each subsequence of normal top level lambdas.
1533 lambdas
1534 ;; this used to read ":start start", but
1535 ;; start can be greater than len, which
1536 ;; is an error according to ANSI - CSR,
1537 ;; 2002-04-25
1539 len)))
1548 ;;; Compile LAMBDAS (a list of CLAMBDAs for top level forms) into the
1549 ;;; object file.
1550 ;;;
1551 ;;; LOAD-TIME-VALUE-P seems to control whether it's MAKE-LOAD-FORM and
1552 ;;; COMPILE-LOAD-TIME-VALUE stuff. -- WHN 20000201
1585 ;;; Actually compile any stuff that has been queued up for block
1586 ;;; compilation.
1606 lexenv))
1610 nil
1626 lexenv))
1636 ;;; Read all forms from INFO and compile them, with output to OBJECT.
1637 ;;; Return (VALUES ABORT-P WARNINGS-P FAILURE-P).
1675 ;; Dump the code coverage records into the fasl.
1684 list))
1685 nil
1686 nil)))
1693 nil))))
1694 ;; Some errors are sufficiently bewildering that we just fail
1695 ;; immediately, without trying to recover and compile more of
1696 ;; the input file.
1702 "~@<~@:_compilation aborted because of fatal error: ~2I~_~A~@:_~:>"
1707 ;;; Return a pathname for the named file. The file must exist.
1716 new
1732 ;;; Print some junk at the beginning and end of compilation.
1740 file-info)
1742 :print-weekday nil
1749 won
1755 ;;; Open some files and call SUB-COMPILE-FILE. If something unwinds
1756 ;;; out of the compile, then abort the writing of the output file, so
1757 ;;; that we don't overwrite it with known garbage.
1760 &key
1762 ;; ANSI options
1764 ;; FIXME: ANSI doesn't seem to say anything about
1765 ;; *COMPILE-VERBOSE* and *COMPILE-PRINT* being rebound by this
1766 ;; function..
1771 ;; extensions
1775 #!+sb-doc
1776 "Compile INPUT-FILE, producing a corresponding fasl file and
1777 returning its filename.
1779 :PRINT
1780 If true, a message per non-macroexpanded top level form is printed
1781 to *STANDARD-OUTPUT*. Top level forms that whose subforms are
1782 processed as top level forms (eg. EVAL-WHEN, MACROLET, PROGN) receive
1783 no such message, but their subforms do.
1785 As an extension to ANSI, if :PRINT is :top-level-forms, a message
1786 per top level form after macroexpansion is printed to *STANDARD-OUTPUT*.
1787 For example, compiling an IN-PACKAGE form will result in a message about
1788 a top level SETQ in addition to the message about the IN-PACKAGE form'
1789 itself.
1791 Both forms of reporting obey the SB-EXT:*COMPILER-PRINT-VARIABLE-ALIST*.
1793 :BLOCK-COMPILE
1794 Though COMPILE-FILE accepts an additional :BLOCK-COMPILE
1795 argument, it is not currently supported. (non-standard)
1797 :TRACE-FILE
1798 If given, internal data structures are dumped to the specified
1799 file, or if a value of T is given, to a file of *.trace type
1800 derived from the input file name. (non-standard)
1802 :EMIT-CFASL
1803 (Experimental). If true, outputs the toplevel compile-time effects
1804 of this file into a separate .cfasl file."
1805 ;;; Block compilation is currently broken.
1806 #|
1807 "Also, as a workaround for vaguely-non-ANSI behavior, the
1808 :BLOCK-COMPILE argument is quasi-supported, to determine whether
1809 multiple functions are compiled together as a unit, resolving function
1810 references at compile time. NIL means that global function names are
1811 never resolved at compilation time. Currently NIL is the default
1813 of the ANSI spec allows this behavior under all circumstances, the
1814 compiler's runtime scales badly when it tries to do this for large
1815 files. If/when this performance problem is fixed, the block
1816 compilation default behavior will probably be made dependent on the
1817 SPEED and COMPILATION-SPEED optimization values, and the
1818 :BLOCK-COMPILE argument will probably become deprecated."
1819 |#
1852 default-trace-file-pathname
1854 default-trace-file-pathname))))
1888 ;; CLHS says that the first value is NIL if the "file could not
1889 ;; be created". We interpret this to mean "a valid fasl could not
1890 ;; be created" -- which can happen if the compilation is aborted
1891 ;; before the whole file has been processed, due to eg. a reader
1892 ;; error.
1894 ;; Hack around filesystem race condition...
1896 warnings-p
1897 failure-p)))
1898 \f
1899 ;;; a helper function for COMPILE-FILE-PATHNAME: the default for
1900 ;;; the OUTPUT-FILE argument
1901 ;;;
1902 ;;; ANSI: The defaults for the OUTPUT-FILE are taken from the pathname
1903 ;;; that results from merging the INPUT-FILE with the value of
1904 ;;; *DEFAULT-PATHNAME-DEFAULTS*, except that the type component should
1905 ;;; default to the appropriate implementation-defined default type for
1906 ;;; compiled files.
1910 retyped))
1912 ;;; KLUDGE: Part of the ANSI spec for this seems contradictory:
1913 ;;; If INPUT-FILE is a logical pathname and OUTPUT-FILE is unsupplied,
1914 ;;; the result is a logical pathname. If INPUT-FILE is a logical
1915 ;;; pathname, it is translated into a physical pathname as if by
1916 ;;; calling TRANSLATE-LOGICAL-PATHNAME.
1917 ;;; So I haven't really tried to make this precisely ANSI-compatible
1918 ;;; at the level of e.g. whether it returns logical pathname or a
1919 ;;; physical pathname. Patches to make it more correct are welcome.
1920 ;;; -- WHN 2000-12-09
1922 &key
1925 #!+sb-doc
1926 "Return a pathname describing what file COMPILE-FILE would write to given
1927 these arguments."
1931 \f
1932 ;;;; MAKE-LOAD-FORM stuff
1934 ;;; The entry point for MAKE-LOAD-FORM support. When IR1 conversion
1935 ;;; finds a constant structure, it invokes this to arrange for proper
1936 ;;; dumping. If it turns out that the constant has already been
1937 ;;; dumped, then we don't need to do anything.
1938 ;;;
1939 ;;; If the constant hasn't been dumped, then we check to see whether
1940 ;;; we are in the process of creating it. We detect this by
1941 ;;; maintaining the special *CONSTANTS-BEING-CREATED* as a list of all
1942 ;;; the constants we are in the process of creating. Actually, each
1943 ;;; entry is a list of the constant and any init forms that need to be
1944 ;;; processed on behalf of that constant.
1945 ;;;
1946 ;;; It's not necessarily an error for this to happen. If we are
1947 ;;; processing the init form for some object that showed up *after*
1948 ;;; the original reference to this constant, then we just need to
1949 ;;; defer the processing of that init form. To detect this, we
1950 ;;; maintain *CONSTANTS-CREATED-SINCE-LAST-INIT* as a list of the
1951 ;;; constants created since the last time we started processing an
1952 ;;; init form. If the constant passed to emit-make-load-form shows up
1953 ;;; in this list, then there is a circular chain through creation
1954 ;;; forms, which is an error.
1955 ;;;
1956 ;;; If there is some intervening init form, then we blow out of
1957 ;;; processing it by throwing to the tag PENDING-INIT. The value we
1958 ;;; throw is the entry from *CONSTANTS-BEING-CREATED*. This is so the
1959 ;;; offending init form can be tacked onto the init forms for the
1960 ;;; circular object.
1961 ;;;
1962 ;;; If the constant doesn't show up in *CONSTANTS-BEING-CREATED*, then
1963 ;;; we have to create it. We call MAKE-LOAD-FORM and check to see
1964 ;;; whether the creation form is the magic value
1965 ;;; :SB-JUST-DUMP-IT-NORMALLY. If it is, then we don't do anything. The
1966 ;;; dumper will eventually get its hands on the object and use the
1967 ;;; normal structure dumping noise on it.
1968 ;;;
1969 ;;; Otherwise, we bind *CONSTANTS-BEING-CREATED* and
1970 ;;; *CONSTANTS-CREATED-SINCE- LAST-INIT* and compile the creation form
1971 ;;; much the way LOAD-TIME-VALUE does. When this finishes, we tell the
1972 ;;; dumper to use that result instead whenever it sees this constant.
1973 ;;;
1974 ;;; Now we try to compile the init form. We bind
1975 ;;; *CONSTANTS-CREATED-SINCE-LAST-INIT* to NIL and compile the init
1976 ;;; form (and any init forms that were added because of circularity
1977 ;;; detection). If this works, great. If not, we add the init forms to
1978 ;;; the init forms for the object that caused the problems and let it
1979 ;;; deal with it.
1982 ;;; FIXME: Shouldn't these^ variables be unbound outside LET forms?
1986 ;; KLUDGE: This special hack is because I was too lazy
1987 ;; to rework DEF!STRUCT so that the MAKE-LOAD-FORM
1988 ;; function of LAYOUT returns nontrivial forms when
1989 ;; building the cross-compiler but :IGNORE-IT when
1990 ;; cross-compiling or running under the target Lisp. --
1991 ;; WHN 19990914
2000 ;; If the constant is a reference to a named constant, we can
2001 ;; just use SYMBOL-VALUE during LOAD.
2007 #-sb-xc-host
2012 constant creation-form init-form))
2015 (:sb-just-dump-it-normally
2017 t)
2018 (:ignore-it
2019 nil)
2032 constant
2036 nil)
2038 constant)))
2044 collect name into names
2045 collect form into forms
2048 nil)))
2052 nil)))))
2054 \f
2055 ;;;; Host compile time definitions
2056 #+sb-xc-host
2061 #+sb-xc-host