Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 1179, Tue Mar 26 03:04:46 2002 UTC	revision 1282, Thu Oct 17 13:07:53 2002 UTC
#	Line 14	Line 14
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16	Name: Matthias Blume	Name: Matthias Blume
17		Date: 2002/10/17 09:10:00 EDT
18		Tag: Release_110_42
19		Description:
20
21		In good old tradition, there has been a slight hiccup so that we have
22		to patch 110.42 after the fact.   The old release tag has been replaced
23		(see below).
24
25		The change solves a problem with two competing approaches the
26		configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
27		each other's way.
28
29		This change only affects the runtime system code and the installer script.
30		(No new bootfiles.)
31
32		----------------------------------------------------------------------
33		Name: Matthias Blume
34		Date: 2002/10/16 12:00:00 EDT
35		Tag: Release_110_42_removed
36		Description:
37
38		New working release.  New bootfiles.
39
40		----------------------------------------------------------------------
41		Name: Matthias Blume
42		Date: 2002/10/10 13:10:00 EDT
43		Tag: blume-20021010-ppc-divs
44		Description:
45
46		The mltree operator DIVS must be implemented with an overflow check on
47		the PPC because the hardware indicates divide-by-zero using "overflow" as
48		well.
49
50		----------------------------------------------------------------------
51		Name: Matthias Blume
52		Date: 2002/07/23 11:45:00
53		Tag: blume-20020723-smlnj-home
54		Description:
55
56		Sml now senses the SMLNJ_HOME environment variable.  If this is set,
57		then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
58		CM_PATHCONFIG is also set), the path configuration file is assumed
59		to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
60		the entire tree to some other place and everything will "just work".
61
62		(Companion commands such as ml-build and ml-makedepend also sense this
63		variable.)
64
65		----------------------------------------------------------------------
66		Name: Matthias Blume
67		Date: 2002/07/12 21:19:00 EDT
68		Tag: blume-20020712-liveness
69		Description:
70
71		Exported two useful "step" functions from liveness module (MLRISC).
72
73		----------------------------------------------------------------------
74		Name: Matthias Blume
75		Date: 2002/07/05 16:00 EDT
76		Tag: Release_110_41
77		Description:
78
79		New working release.  New bootfiles.
80
81		----------------------------------------------------------------------
82		Name: Matthias Blume
83		Date: 2002/07/05 10:25:00 EDT
84		Tag: blume-20020705-btimp
85		Description:
86
87		Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
88		other clients can set up backtracing support.
89
90		----------------------------------------------------------------------
91		Name: Matthias Blume
92		Date: 2002/06/25 17:23:00 EDT
93		Tag: blume-20020625-fpmax
94		Description:
95
96		Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
97		case.  (The sense of the isNaN test was reversed -- which made min and
98		max always return their first argument.)
99
100		----------------------------------------------------------------------
101		Name: Matthias Blume
102		Date: 2002/06/11
103		Tag: blume-20020611-unixpath
104		Description:
105
106		Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
107
108		----------------------------------------------------------------------
109		Name: Matthias Blume
110		Date: 2002/06/10 16:35:00 EDT
111		Tag: blume-20020610-ieeereal
112		Description:
113
114		I back-ported my implementation of IEEEReal.fromString from the
115		idlbasis-devel branch so that we can test it.
116
117		Another small change is that ppDec tries to give more information
118		than just "<sig>" in the case of functors.  However, this code is
119		broken in some mysterious way if the functor's body's signature
120		has not been declared by ascription but gets inferred from the
121		implementation.  This needs fixing...
122
123		----------------------------------------------------------------------
124		Name: Matthias Blume
125		Date: 2002/05/31
126		Tag: blume-20020531-btrace-mode
127		Description:
128
129		Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
130		the wayside when I switched over to using Controls everywhere.)
131
132		----------------------------------------------------------------------
133		Name: Lal George
134		Date: 2002/05/23 12:21:40 EDT
135		Tag: george-20020523-visual-labels
136		Description:
137
138		Labels are now displayed in the graphical output to make
139		the fall-through and target blocks obvious.
140
141		----------------------------------------------------------------------
142		Name: Matthias Blume
143		Date: 2002/05/22 11:03:00 EDT
144		Tag: blume-20020522-shrink
145		Description:
146
147		John tweaked yesterday's fix for 1131 to handle an out-of-memory
148		situation that comes up when allocating huge arrays.
149
150		----------------------------------------------------------------------
151		Name: Matthias Blume
152		Date: 2002/05/21 16:00:00 EDT
153		Tag: Release_110_40
154		Description:
155
156		New working release (110.40).  New bootfiles.
157
158		[Also: John Reppy fixed GC bug 1131.]
159
160		----------------------------------------------------------------------
161		Name: Matthias Blume
162		Date: 2002/05/21 12:35:00 EDT
163		Tag: blume-20020521-cmdoc
164		Description:
165
166		CM documentation update.
167
168		----------------------------------------------------------------------
169		Name: Matthias Blume
170		Date: 2002/05/21 10:55:00 EDT
171		Tag: blume-20020521-misc
172		Description:
173
174		- John tweaked runtime to be silent on heap export (except when
175		GC messages are on).
176		- I added a few more things (cross-compiling versions of CMB) to
177		config/preloads (as suggestions).
178
179		----------------------------------------------------------------------
180		Name: Matthias Blume
181		Date: 2002/05/20 22:25:00 EDT
182		Tag: blume-20020520-controls
183		Description:
184
185		- Added ControlUtil structure to control-lib.cm.
186		- Use it throughout.
187		- Used Controls facility to define MLRISC controls (as opposed to
188		registering MLRISC control ref cells with Controls after the fact)
189		- Fixed messed-up controls priorities.
190
191		* Removed again all the stuff from config/preloads that one wouldn't
192		be able to preload at the time the initial heap image is built.
193		(Many libraries, e.g., CML, do not exist yet at this time.  The
194		only libraries that can be preloaded via config/preloads are those
195		that come bundled with the bootfiles.)
196
197		----------------------------------------------------------------------
198		Name: Matthias Blume
199		Date: 2002/05/20 10:59:00 EDT
200		Tag: blume-20020520-preloads
201		Description:
202
203		Added a lot of commented-out suggestions for things to be included
204		in config/preloads.
205
206		----------------------------------------------------------------------
207		Name: Allen Leung
208		Date: 2002/05/18 14:20:00 EDT
209		Tag: leunga-20020518-mdl
210		Description:
211
212		o Made the mdl tool stuff compile and run again.
213		o I've disabled all the stuff that depends on RTL specifications; they
214		are all badly broken anyway.
215
216		----------------------------------------------------------------------
217		Name: Matthias Blume
218		Date: 2002/05/17 16:49:00 EDT
219		Tag: blume-20020517-controls
220		Description:
221
222		0. John Reppy made several modifications to the SML/NJ library.
223		In particular, there is a shiny new controls-lib.cm.
224
225		1. Pushed new controls interface through compiler so that everything
226		compiles again.
227
228		2. Added FormatComb and FORMAT_COMB to the CML version of the
229		SML/NJ library (so that CML compiles again).
230
231		3. Modified init scripts because XXX_DEFAULT environment variables
232		are no longer with us.  (Boot-time initialization is now done
233		using the same environment variables that are also used for
234		startup-time initialization of controls.)
235
236		----------------------------------------------------------------------
237		Name: Lal George
238		Date: 2002/05/15 09:20:10 EDT
239		Tag: george-20020515-pseudo-op-decls
240		Description:
241
242		All pseudo-ops emitted before the first segment declaration
243		such as TEXT, DATA, and BSS directives are assumed to be global
244		declarations and are emitted first in the assembly file. This is
245		useful in a number of situations where one has pseudo-ops that are not
246		specific to any segment, and also works around the constraint that one
247		cannot have client pseudo-ops in the TEXT segment.
248
249		Because no segment is associated with these declarations it is
250		an error to allocate any space or objects before the first segment
251		directive and an exception will be raised. However, we cannot make
252		this check for client pseudo-ops.
253
254		These top level declarations are a field in the CFG graph_info.
255		In theory you can continue to add to this field after the CFG has been
256		built -- provided you know what you are doing;-)
257
258		----------------------------------------------------------------------
259		Name: Matthias Blume
260		Date: 2002/05/13 16:40:00 EDT
261		Tag: blume-20020513-pp-etc
262		Description:
263
264		A few minor bugfixes:
265
266		- Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
267		(Bogus printouts for redefined bindings still occur.  Compiler
268		bug should no longer occur now.  We need to redo the prettyprinter
269		from scratch.)
270
271		- CM pathname printer now also adds escape sequences for ( and )
272
273		- commend and docu fixes for ml-nlffi
274
275		----------------------------------------------------------------------
276		Name: Matthias Blume
277		Date: 2002/05/10 16:40:00 EDT
278		Tag: blume-20020510-erg-textio
279		Description:
280
281		Applied the following bugfix provided by Emden Gansner:
282
283		Output is corrupted when outputSubstr is used rather than output.
284
285		The problem occurs when a substring
286
287		ss = (s, dataStart, dataLen)
288
289		where dataStart > 0, fills a stream buffer with avail bytes left.
290		avail bytes of s, starting at index dataStart, are copied into the
291		buffer, the buffer is flushed, and then the remaining dataLen-avail
292		bytes of ss are copied into the beginning of the buffer. Instead of
293		starting this copy at index dataStart+avail in s, the current code
294		starts the copy at index avail.
295
296		Fix:
297		In text-io-fn.sml, change line 695 from
298		val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
299		to
300		val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
301
302		----------------------------------------------------------------------
303		Name: Matthias Blume
304		Date: 2002/04/12 13:55:00 EDT
305		Tag: blume-20020412-assyntax
306		Description:
307
308		1. Grabbed newer assyntax.h from the XFree86 project.
309		2. Fiddled with how to compile X86.prim.asm without warnings.
310		3. (Very) Minor cleanup in CM.
311
312		----------------------------------------------------------------------
313		Name: Matthias Blume
314		Date: 2002/04/01 (no joke!) 17:07:00 EST
315		Tag: blume-20020401-x86div
316		Description:
317
318		Added full support for div/mod/rem/quot on the x86, using the machine
319		instruction's two results (without clumsily recomputing the remainder)
320		directly where appropriate.
321
322		Some more extensive power-of-two support was added to the x86 instruction
323		selector (avoiding expensive divs, mods, and muls where they can be
324		replaced with cheaper shifts and masks).  However, this sort of thing
325		ought to be done earlier, e.g., within the CPS optimizer so that
326		all architectures benefit from it.
327
328		The compiler compiles to a fixed point, but changes might be somewhat
329		fragile nevertheless.  Please, report any strange things that you might
330		see wrt. div/mod/quot/rem...
331
332		----------------------------------------------------------------------
333		Name: Matthias Blume
334		Date: 2002/03/29 17:22:00
335		Tag: blume-20020329-div
336		Description:
337
338		Fixed my broken div/mod logic.  Unfortunately, this means that the
339		inline code for div/mod now has one more comparison than before.
340		Fast paths (quotient > 0 or remainder = 0) are not affected, though.
341		The problem was with quotient = 0, because that alone does not tell
342		us which way the rounding went.  One then has to look at whether
343		remainder and divisor have the same sign...  :(
344
345		Anyway, I replaced the bootfiles with fresh ones...
346
347		----------------------------------------------------------------------
348		Name: Matthias Blume
349		Date: 2002/03/29 14:10:00 EST
350		Tag: blume-20020329-inlprims
351		Description:
352
353		NEW BOOTFILES!!!    Version number bumped to 110.39.3.
354
355		Primops have changed. This means that the bin/boot-file formats have
356		changed as well.
357
358		To make sure that there is no confusion, I made a new version.
359
360
361		CHANGES:
362
363		* removed REMT from mltree (remainder should never overflow).
364
365		* added primops to deal with divisions of all flavors to the frontend
366
367		* handled these primops all the way through so they map to their respective
368		MLRISC support
369
370		* used these primops in the implementation of Int, Int32, Word, Word32
371
372		* removed INLDIV, INLMOD, and INLREM as they are no longer necessary
373
374		* parameterized INLMIN, INLMAX, and INLABS by a numkind
375
376		* translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
377		floating point
378
379		* used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
380		and Real (but Real.abs maps to a separate floating-point-only primop)
381
382
383		TODO items:
384
385		* Hacked Alpha32 instruction selection, disabling the selection of REMx
386		instructions because the machine instruction encoder cannot handle
387		them.  (Hppa, PPC, and Sparc instruction selection did not handle
388		REM in the first place, and REM is supported by the x86 machine coder.)
389
390		* Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
391		selection phase.  (The two can be streamlined because the hardware
392		delivers both quotient and remainder at the same time anyway.)
393
394		* Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
395		(Currently the behavior is inconsistent both across architectures and
396		wrt. the draft Basis spec.)
397
398		* Word8 should eventually be handled natively, too.
399
400		* There seems to be one serious bug in mltree-gen.sml.  It appears, though,
401		as if there currently is no execution path that could trigger it in
402		SML/NJ.  (The assumptions underlying functions arith and promotable do not
403		hold for things like multiplication and division.)
404
405		----------------------------------------------------------------------
406		Name: Matthias Blume
407		Date: 2002/03/27 16:27:00 EST
408		Tag: blume-20020327-mlrisc-divisions
409		Description:
410
411		Added support for all four division operations (ML's div, mod, quot,
412		and rem) to MLRISC.  In the course of doing so, I also rationalized
413		the naming (no more annoying switch-around of DIV and QUOT), by
414		parameterizing the operation by div_rounding_mode (which can be either
415		DIV_TO_ZERO or DIV_TO_NEGINF).
416
417		The generic MLTreeGen functor takes care of compiling all four
418		operations down to only round-to-zero div.
419
420		Missing pieces:
421
422		* Doing something smarter than relying on MLTreeGen on architectures
423		like, e.g., the x86 where hardware division delivers both quotient and
424		remainder at the same time.  With this, the implementation of the
425		round-to-neginf operations could be further streamlined.
426
427		* Remove inlining support for div/mod/rem from the frontend and replace it
428		with primops that get carried through to the backend.  Do this for all
429		int and word types.
430
431		----------------------------------------------------------------------
432		Name: Matthias Blume
433		Date: 2002/03/25 17:25:00 EST
434		Tag: blume-20020325-divmod
435		Description:
436
437		I improved (hopefully without breaking them) the implementation of Int.div,
438		Int.mod, and Int.rem.   For this, the code in translate.sml now takes
439		advantage of the following observations:
440
441		Let  q = x quot y      r = x rem y
442		d = x div  y      m = x mod y
443
444		where "quot" is the round-to-zero version of integer division that
445		hardware usually provides.  Then we have:
446
447		r = x - q * y        where neither the * nor the - will overflow
448		d = if q >= 0 orelse x = q * y then q else q - 1
449		where neither the * nor the - will overflow
450		m = if q >= 0 orelse r = 0 then r else r + y
451		where the + will not overflow
452
453		This results in substantial simplification of the generated code.
454		The following table shows the number of CFG nodes and edges generated
455		for
456		fun f (x, y) = x OPER y
457		(* with OPER \in div, mod, quot, rem *)
458
459
460		OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
461		--------------------------------------------------------
462		div |         24 |         39 |         12 |         16
463		mod |         41 |         71 |         12 |         16
464		quot |          8 |         10 |          8 |         10
465		rem |         10 |         14 |          8 |         10
466
467
468		----------------------------------------------------------------------
469		Name: Matthias Blume
470	Date: 2002/03/25 22:06:00 EST	Date: 2002/03/25 22:06:00 EST
471	Tag: blume-20020325-cprotobug	Tag: blume-20020325-cprotobug
472	Description:	Description:

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.1179   changed lines   Added in v.1282

---

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