Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 1183, Fri Mar 29 19:09:48 2002 UTC	revision 1353, Mon Sep 8 16:57:39 2003 UTC
#	Line 13	Line 13
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: Matthias Blume (blume@tti-c.org)
17		Date: 2003/09/08 11:55:00 CDT
18		Tag: blume-20030908-fullpath
19		Description:
20
21		Made Win32 version of OS.FileSys.fullPath return current directory
22		when given an empty string.  This is what the spec says, and incidentally,
23		CM depends on it.  (CM otherwise goes into an infinite loop in certain
24		cases when presented with the name of a non-existing .cm file.)
25
26		----------------------------------------------------------------------
27		Name: Matthias Blume (blume@tti-c.org)
28		Date: 2003/09/04 16:30:00 CDT
29		Tag: blume-20030905-slices-etc
30		Description:
31
32		1. Changed interface to vectors and arrays in Basis to match
33		(draft) Basis spec.
34		2. Added signatures and implementations of slices according to
35		Basis spec.
36		3. Edited source code throughout the system to make it compile again
37		under 1. and 2.  (In some cases code had to be added to have it
38		match the new signatures.)
39		4. MLRISC should be backward-compatible: the copies of the originals
40		of files that needed to change under 3. were retained, the .cm files
41		check the compiler version number and use old versions when
42		appropriate.
43		5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
44		dirstream -> string option (in accordance with Basis spec).
45		6. When generating code that counts lines, ml-lex used function
46		CharVector.foldli, taking advantage of its old interface.
47		This has been replaced with the corresponding code from
48		CharVectorSlice.  (html-lex must be re-lexed!)
49		7. BitArray in smlnj-lib/Util has been extended/modified to match the
50		new MONO_ARRAY signature.  (Do we need BitArraySlice?)
51		8. Removed temporary additions (fromInternal, toInternal) from the
52		(now obsolete) IntInf in smlnj-lib/Util.
53		9. Cleaned up structure Byte.
54		10. Added localOffset, scan, and fromString to Date (according to spec).
55		Cleaned/corrected implementation of Date.
56		(Still need to check for correctness; implement better canonicalizeDate.)
57		11. Added "scan" to signature IEEE_REAL.
58		12. Some improvements to IntInf [in particular: efficiency-hack for
59		mod and rem when second operand is 2 (for parity checks).]
60		13. Changed representation of type Time.time, using a single IntInf.int
61		value counting microseconds.  This considerably simplified the
62		implementation of structure Time.  We now support negative time
63		values; scan and fromString handle signs.
64		14. Functor PrimIO now takes two additional arguments (VectorSlice and
65		ArraySlice).
66
67		----------------------------------------------------------------------
68		Name: Matthias Blume (blume@tti-c.org)
69		Date: 2003/08/28 17:00:00 CDT
70		Tag: blume-20030828-intinf
71		Description:
72
73		This is a major update which comes with a version number bump
74		(110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
75		and an implementation of IntInf in the Basis.
76
77		There are a fairly large number of related changes and updates throughout
78		the system:
79
80		Basis:
81		- Implemented IntInf.
82		- Made LargeInt a projection of IntInf (by filtering through INTEGER).
83		- Added some missing Real64 operations, most notably Real.toLargeInt.
84		- Added FixedInt as a synonym for Int32.
85
86		compiler:
87		* Added support for a built-in intinf type.
88		- literals
89		- pattern matching
90		- conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
91		- overloading on literals and operations
92
93		This required adding a primitive type intinf, some additional
94		primops, and implementations for several non-trivial intinf
95		operations in Core.  (The intinf type is completely abstract
96		to the compiler; all operations get delegated back to the Core.)
97
98		* Intinf equality is handled by polyequal.  However, the compiler
99		does not print its usual warning in this case (since polyequal
100		is the right thing to do there).
101
102		* Improved the organization of structure InlineT.
103
104		* A word about conversion primops:
105		If conversions involving intinf do not cancel out during
106		CPS contract, then the compiler must insert calls to Core functions.
107		Since all core access must be resolved already during the FLINT
108		translate phase, it would be too late a the time of CPS contract
109		to add new Core calls.  For this reason, conversion primops
110		for intinf carry two arguments:  1. the numeric argument that
111		they are supposed to convert, and 2. the Core function that
112		can help with this conversion if necessary.  If CPS contract
113		eliminates a primop, then the associated Core function becomes
114		dead and goes away.  Intinf conversion primops that do not get
115		eliminated by CPS contract get rewritten into calls of their
116		core functions by a separate, new phase.
117
118		interactive system:
119		- Control.Print.intinfDepth controls max length of intinf constants
120		being printed. (Analogous to Control.Print.stringDepth.)
121		- Cleanup in printutil and pputil: got rid of unused stuff and
122		duplicates; replaced some of the code with code that makes better
123		use of library functionality.
124
125		CM:
126		Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
127		are no longer silent.
128
129		CKIT:
130		Fixed mismatched uses of Int32 and LargeInt.  I always decided
131		in favor of LargeInt -- which is now the same as IntInf.
132		CKIT-knowledgable people should check whether this is what's
133		intended and otherwise change things back to using Int32 or
134		FixedInt.
135
136		Throughout the code:
137		Started using IntInf.int literals and built-in operations
138		(e.g., comparison with 0) where this seems appropriate.
139
140
141		----------------------------------------------------------------------
142		Name: Dave MacQueen (dbm@cs.uchicago.edu)
143		Date: 2003/08/13 11:36:00 CDT
144		Tag: dbm-20030813-mcz-merge1
145		Description:
146
147		Merging changes from the mcz-branch development branch into trunk.
148		These changes involve replacement of the emulated old prettyprinter
149		interface with direct use of the SML/NJ Lib PP library, and fixing
150		of a couple of bugs (895, 1186) relating to error messages.  A new
151		prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
152		has been added.
153
154		----------------------------------------------------------------------
155		Name: Matthias Blume (blume@tti-c.org)
156		Date: 2003/08/11 15:45:00 CDT
157		Tag: blume-20030811-windows
158		Description:
159
160		Version number bumped to 110.42.9.  NEW BOOTFILES!!!
161
162		http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
163
164		This patch restores SML/NJ's ability to run under win32.  There are a
165		number of changes, including fixes for several bugs that had gone
166		unnoticed until now:
167
168		- uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
169		trying to run the win32 version from within cygwin.)
170
171		- There are a number of simple .bat scripts that substitute for their
172		corresponding Unix shell-scripts.  (See below.)
173
174		- The internals of ml-build have been modified slightly.  The main
175		difference is that instead of calling ".link-sml" (or link-sml.bat)
176		using OS.Process.system, the ML process delegates this task back
177		to the script.  Otherwise problems arise in mixed environments such
178		as Cygwin where scripts look and work like Unix scripts, but
179		where OS.Process.system cannot run them.
180
181		- In CM, the srcpath pickler used native pathname syntax -- which
182		is incorrect in the case of cross-compilation.  The new pickle format
183		is independent of platform-specific naming conventions.
184
185		- Path configuration files (such as lib/pathconfig) can now choose
186		between native and standard syntax.  Placing a line of the form
187
188		standard!
189
190		into the file causes all subsequent paths to be interpreted using
191		CM standard pathname syntax (= Unix conventions); a line
192
193		native!
194
195		switches back to native style.  This was needed so that
196		path config files can be written portably, see src/system/pathconfig.
197
198		- Runtime system:
199
200		- win32-filesys.c:  get_file_time and set_file_time now
201		access modification time, not creation time.
202
203		- I/O code made aware of new array representation.
204
205		- Bug fixes in X86.prim.masm.
206
207		- src/system/makeml made aware of win32. (For use under cygwin
208		and other Unix-environments for windows.)
209
210		- In Basis, fixed off-by-one error in win32-io.sml (function vecF)
211		which caused BinIO.inputAll to fail consistently.
212
213		.bat scripts:
214
215		Windows .bat scripts assume that SMLNJ_HOME is defined.
216
217		- sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
218		applications (sml, ml-yacc, ml-lex).
219		- ml-build.bat: analogous to ml-build.
220		- config\install.bat:  Analogous to config/install.sh.  This requires
221		that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
222		(nmake etc. must be on the path, and vcvars32 must have been run.)
223		Moreover, sources for ml-lex and ml-yacc need to exist under src,
224		and the bootfile hierarchy must have been unpacked under
225		sml.boot.x86-win32.
226		The script is very primitive and does a poor job at error checking.
227		It only installs the base system, ml-lex, and ml-yacc.  No other
228		libraries are being installed (i.e., you get only those that
229		are part of the compiler.)
230		- link-sml.bat:  analogous to .link-sml, but not currently used
231
232		Unrelated bug fixes:
233
234		- ml-nlffigen now exports structures ST_* corresponding to incomplete
235		types.
236		- Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
237		otherwise.)
238
239		----------------------------------------------------------------------
240		Name: Dave MacQueen (macqueen@cs.uchicago.edu)
241		Date: 2003/06/17
242		Tag: macqueen-20030617-bug895
243		Description:
244
245		Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
246		Tag will be used for new development branch (mcz-branch) for
247		use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
248
249		----------------------------------------------------------------------
250		Name: Matthias Blume (blume@tti-c.org)
251		Date: 2003/05/27 16:55:00 CDT
252		Tag: blume-20030527-polyeq
253		Description:
254
255		Tried to eliminated most cases of polymorphic equality.
256
257		----------------------------------------------------------------------
258		Name: Matthias Blume (blume@tti-c.org)
259		Date: 2003/05/21 17:45:00 CDT
260		Tag: blume-20030517-complete
261		Description:
262
263		Two changes:
264
265		1. Added a flag for controlling whether non-exhaustive bindings will
266		be treated as errors (default is false).
267		2. Cleaned up the *entire* source tree so that CMB.make goes through
268		without a single non-exhaustive match- or bind warning.
269
270		----------------------------------------------------------------------
271		Name: Matthias Blume (blume@tti-c.org)
272		Date: 2003/05/17 10:20:00 CDT
273		Tag: blume-20030517-absyn
274		Description:
275
276		1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
277
278		This mainly affects the quality of error messages.  However, some
279		of the code is now more straightforward than before. (Treatment of
280		the above four constructs in translate.sml is much simpler than
281		the "macro-expansion" that was going on before.  Plus, the mach-
282		compiler no longer gets invoked just to be able to compile an
283		if-expression.)
284
285		2. The ErrorMsg.Error exception is now caught and absorbed by the
286		interactive loop.
287
288		----------------------------------------------------------------------
289		Name: Allen Leung
290		Date: 2003/05/16 13:05:00 CDT
291		Tag: leunga-20030516-cygwin-runtime
292		Description:
293
294		Ported the runtime system to cygwin, which uses the unix
295		x86-unix bin files.  Missing/buggy features:
296
297		o getnetbyname, getnetbyaddr: these functions seem to be missing in
298		the Cygwin library.
299		o Ctrl-C handling may be flaky.
300		o Windows system calls and Windows I/O are not supported.
301
302		A new set of binfiles is located at:
303
304		http://www.dorsai.org/~leunga/boot.x86-unix.tgz
305
306		This is only needed for bootstrapping the cygwin version of smlnj.
307		Other x86 versions can use the existing binfiles.
308
309		----------------------------------------------------------------------
310		Name: Matthias Blume
311		Date: 2003/04/08 15:42:00 CDT
312		Tag: blume-20030408-listpair
313		Description:
314
315		1. Added a target 'mlrisc' to installer.
316
317		2. Added missing elements to structure ListPair.
318
319		----------------------------------------------------------------------
320		Name: Allen Leung
321		Date: 2003/01/07 10:40:00 EST
322		Tag: leunga-20030107-int-rem
323		Description:
324
325		Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
326		The arguments to the SUBL instruction were swapped.
327
328		----------------------------------------------------------------------
329		Name: Matthias Blume
330		Date: 2002/12/12 16:25:00 EST
331		Tag: blume-20021212-risc-ra
332		Description:
333
334		Fixed a serious bug in the rewrite code for FP spilling/reloading that
335		sent the RA into an infinite loop when floating point registers get
336		spilled.  (Because of this bug, e.g., nucleic stopped compiling between
337		110.37 and 110.38.)
338		There was another set of potential problems related to the handling of
339		MLRISC annotations (but those did not yet cause real problems, apparently).
340
341		----------------------------------------------------------------------
342		Name: Matthias Blume
343		Date: 2002/12/06 22:40:00 EST
344		Tag: blume-20021206-cm-fileid
345		Description:
346
347		Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
348		This fixes the problem of CM getting confused by files that suddenly
349		change their identity (e.g., by getting unlinked and recreated by some
350		text editor such as vi).  There might be a better/cheaper/cleaner way
351		of doing this, but for now this will have to do.
352
353		----------------------------------------------------------------------
354		Name: Matthias Blume
355		Date: 2002/10/28 09:50:00 EST
356		Tag: blume-20021028-typecheck
357		Description:
358
359		Exported structure Typecheck from $smlnj/viscomp/core.cm.
360
361		----------------------------------------------------------------------
362		Name: Matthias Blume
363		Date: 2002/10/17 09:10:00 EDT
364		Tag: Release_110_42
365		Description:
366
367		In good old tradition, there has been a slight hiccup so that we have
368		to patch 110.42 after the fact.   The old release tag has been replaced
369		(see below).
370
371		The change solves a problem with two competing approaches the
372		configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
373		each other's way.
374
375		This change only affects the runtime system code and the installer script.
376		(No new bootfiles.)
377
378		----------------------------------------------------------------------
379		Name: Matthias Blume
380		Date: 2002/10/16 12:00:00 EDT
381		Tag: Release_110_42_removed
382		Description:
383
384		New working release.  New bootfiles.
385
386		----------------------------------------------------------------------
387		Name: Matthias Blume
388		Date: 2002/10/10 13:10:00 EDT
389		Tag: blume-20021010-ppc-divs
390		Description:
391
392		The mltree operator DIVS must be implemented with an overflow check on
393		the PPC because the hardware indicates divide-by-zero using "overflow" as
394		well.
395
396		----------------------------------------------------------------------
397		Name: Matthias Blume
398		Date: 2002/07/23 11:45:00
399		Tag: blume-20020723-smlnj-home
400		Description:
401
402		Sml now senses the SMLNJ_HOME environment variable.  If this is set,
403		then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
404		CM_PATHCONFIG is also set), the path configuration file is assumed
405		to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
406		the entire tree to some other place and everything will "just work".
407
408		(Companion commands such as ml-build and ml-makedepend also sense this
409		variable.)
410
411		----------------------------------------------------------------------
412		Name: Matthias Blume
413		Date: 2002/07/12 21:19:00 EDT
414		Tag: blume-20020712-liveness
415		Description:
416
417		Exported two useful "step" functions from liveness module (MLRISC).
418
419		----------------------------------------------------------------------
420		Name: Matthias Blume
421		Date: 2002/07/05 16:00 EDT
422		Tag: Release_110_41
423		Description:
424
425		New working release.  New bootfiles.
426
427		----------------------------------------------------------------------
428		Name: Matthias Blume
429		Date: 2002/07/05 10:25:00 EDT
430		Tag: blume-20020705-btimp
431		Description:
432
433		Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
434		other clients can set up backtracing support.
435
436		----------------------------------------------------------------------
437		Name: Matthias Blume
438		Date: 2002/06/25 17:23:00 EDT
439		Tag: blume-20020625-fpmax
440		Description:
441
442		Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
443		case.  (The sense of the isNaN test was reversed -- which made min and
444		max always return their first argument.)
445
446		----------------------------------------------------------------------
447		Name: Matthias Blume
448		Date: 2002/06/11
449		Tag: blume-20020611-unixpath
450		Description:
451
452		Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
453
454		----------------------------------------------------------------------
455		Name: Matthias Blume
456		Date: 2002/06/10 16:35:00 EDT
457		Tag: blume-20020610-ieeereal
458		Description:
459
460		I back-ported my implementation of IEEEReal.fromString from the
461		idlbasis-devel branch so that we can test it.
462
463		Another small change is that ppDec tries to give more information
464		than just "<sig>" in the case of functors.  However, this code is
465		broken in some mysterious way if the functor's body's signature
466		has not been declared by ascription but gets inferred from the
467		implementation.  This needs fixing...
468
469		----------------------------------------------------------------------
470		Name: Matthias Blume
471		Date: 2002/05/31
472		Tag: blume-20020531-btrace-mode
473		Description:
474
475		Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
476		the wayside when I switched over to using Controls everywhere.)
477
478		----------------------------------------------------------------------
479		Name: Lal George
480		Date: 2002/05/23 12:21:40 EDT
481		Tag: george-20020523-visual-labels
482		Description:
483
484		Labels are now displayed in the graphical output to make
485		the fall-through and target blocks obvious.
486
487		----------------------------------------------------------------------
488		Name: Matthias Blume
489		Date: 2002/05/22 11:03:00 EDT
490		Tag: blume-20020522-shrink
491		Description:
492
493		John tweaked yesterday's fix for 1131 to handle an out-of-memory
494		situation that comes up when allocating huge arrays.
495
496		----------------------------------------------------------------------
497		Name: Matthias Blume
498		Date: 2002/05/21 16:00:00 EDT
499		Tag: Release_110_40
500		Description:
501
502		New working release (110.40).  New bootfiles.
503
504		[Also: John Reppy fixed GC bug 1131.]
505
506		----------------------------------------------------------------------
507		Name: Matthias Blume
508		Date: 2002/05/21 12:35:00 EDT
509		Tag: blume-20020521-cmdoc
510		Description:
511
512		CM documentation update.
513
514		----------------------------------------------------------------------
515		Name: Matthias Blume
516		Date: 2002/05/21 10:55:00 EDT
517		Tag: blume-20020521-misc
518		Description:
519
520		- John tweaked runtime to be silent on heap export (except when
521		GC messages are on).
522		- I added a few more things (cross-compiling versions of CMB) to
523		config/preloads (as suggestions).
524
525		----------------------------------------------------------------------
526		Name: Matthias Blume
527		Date: 2002/05/20 22:25:00 EDT
528		Tag: blume-20020520-controls
529		Description:
530
531		- Added ControlUtil structure to control-lib.cm.
532		- Use it throughout.
533		- Used Controls facility to define MLRISC controls (as opposed to
534		registering MLRISC control ref cells with Controls after the fact)
535		- Fixed messed-up controls priorities.
536
537		* Removed again all the stuff from config/preloads that one wouldn't
538		be able to preload at the time the initial heap image is built.
539		(Many libraries, e.g., CML, do not exist yet at this time.  The
540		only libraries that can be preloaded via config/preloads are those
541		that come bundled with the bootfiles.)
542
543		----------------------------------------------------------------------
544		Name: Matthias Blume
545		Date: 2002/05/20 10:59:00 EDT
546		Tag: blume-20020520-preloads
547		Description:
548
549		Added a lot of commented-out suggestions for things to be included
550		in config/preloads.
551
552		----------------------------------------------------------------------
553		Name: Allen Leung
554		Date: 2002/05/18 14:20:00 EDT
555		Tag: leunga-20020518-mdl
556		Description:
557
558		o Made the mdl tool stuff compile and run again.
559		o I've disabled all the stuff that depends on RTL specifications; they
560		are all badly broken anyway.
561
562		----------------------------------------------------------------------
563		Name: Matthias Blume
564		Date: 2002/05/17 16:49:00 EDT
565		Tag: blume-20020517-controls
566		Description:
567
568		0. John Reppy made several modifications to the SML/NJ library.
569		In particular, there is a shiny new controls-lib.cm.
570
571		1. Pushed new controls interface through compiler so that everything
572		compiles again.
573
574		2. Added FormatComb and FORMAT_COMB to the CML version of the
575		SML/NJ library (so that CML compiles again).
576
577		3. Modified init scripts because XXX_DEFAULT environment variables
578		are no longer with us.  (Boot-time initialization is now done
579		using the same environment variables that are also used for
580		startup-time initialization of controls.)
581
582		----------------------------------------------------------------------
583		Name: Lal George
584		Date: 2002/05/15 09:20:10 EDT
585		Tag: george-20020515-pseudo-op-decls
586		Description:
587
588		All pseudo-ops emitted before the first segment declaration
589		such as TEXT, DATA, and BSS directives are assumed to be global
590		declarations and are emitted first in the assembly file. This is
591		useful in a number of situations where one has pseudo-ops that are not
592		specific to any segment, and also works around the constraint that one
593		cannot have client pseudo-ops in the TEXT segment.
594
595		Because no segment is associated with these declarations it is
596		an error to allocate any space or objects before the first segment
597		directive and an exception will be raised. However, we cannot make
598		this check for client pseudo-ops.
599
600		These top level declarations are a field in the CFG graph_info.
601		In theory you can continue to add to this field after the CFG has been
602		built -- provided you know what you are doing;-)
603
604		----------------------------------------------------------------------
605		Name: Matthias Blume
606		Date: 2002/05/13 16:40:00 EDT
607		Tag: blume-20020513-pp-etc
608		Description:
609
610		A few minor bugfixes:
611
612		- Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
613		(Bogus printouts for redefined bindings still occur.  Compiler
614		bug should no longer occur now.  We need to redo the prettyprinter
615		from scratch.)
616
617		- CM pathname printer now also adds escape sequences for ( and )
618
619		- commend and docu fixes for ml-nlffi
620
621		----------------------------------------------------------------------
622		Name: Matthias Blume
623		Date: 2002/05/10 16:40:00 EDT
624		Tag: blume-20020510-erg-textio
625		Description:
626
627		Applied the following bugfix provided by Emden Gansner:
628
629		Output is corrupted when outputSubstr is used rather than output.
630
631		The problem occurs when a substring
632
633		ss = (s, dataStart, dataLen)
634
635		where dataStart > 0, fills a stream buffer with avail bytes left.
636		avail bytes of s, starting at index dataStart, are copied into the
637		buffer, the buffer is flushed, and then the remaining dataLen-avail
638		bytes of ss are copied into the beginning of the buffer. Instead of
639		starting this copy at index dataStart+avail in s, the current code
640		starts the copy at index avail.
641
642		Fix:
643		In text-io-fn.sml, change line 695 from
644		val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
645		to
646		val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
647
648		----------------------------------------------------------------------
649		Name: Matthias Blume
650		Date: 2002/04/12 13:55:00 EDT
651		Tag: blume-20020412-assyntax
652		Description:
653
654		1. Grabbed newer assyntax.h from the XFree86 project.
655		2. Fiddled with how to compile X86.prim.asm without warnings.
656		3. (Very) Minor cleanup in CM.
657
658		----------------------------------------------------------------------
659		Name: Matthias Blume
660		Date: 2002/04/01 (no joke!) 17:07:00 EST
661		Tag: blume-20020401-x86div
662		Description:
663
664		Added full support for div/mod/rem/quot on the x86, using the machine
665		instruction's two results (without clumsily recomputing the remainder)
666		directly where appropriate.
667
668		Some more extensive power-of-two support was added to the x86 instruction
669		selector (avoiding expensive divs, mods, and muls where they can be
670		replaced with cheaper shifts and masks).  However, this sort of thing
671		ought to be done earlier, e.g., within the CPS optimizer so that
672		all architectures benefit from it.
673
674		The compiler compiles to a fixed point, but changes might be somewhat
675		fragile nevertheless.  Please, report any strange things that you might
676		see wrt. div/mod/quot/rem...
677
678		----------------------------------------------------------------------
679		Name: Matthias Blume
680		Date: 2002/03/29 17:22:00
681		Tag: blume-20020329-div
682		Description:
683
684		Fixed my broken div/mod logic.  Unfortunately, this means that the
685		inline code for div/mod now has one more comparison than before.
686		Fast paths (quotient > 0 or remainder = 0) are not affected, though.
687		The problem was with quotient = 0, because that alone does not tell
688		us which way the rounding went.  One then has to look at whether
689		remainder and divisor have the same sign...  :(
690
691		Anyway, I replaced the bootfiles with fresh ones...
692
693		----------------------------------------------------------------------
694	Name: Matthias Blume	Name: Matthias Blume
695	Date: 2002/03/29 14:10:00 EST	Date: 2002/03/29 14:10:00 EST
696	Tag: blume-20020329-inlprims	Tag: blume-20020329-inlprims

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.1183   changed lines   Added in v.1353

---

root@smlnj-gforge.cs.uchicago.edu	ViewVC Help
Powered by ViewVC 1.0.0