Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 633, Sun Apr 30 16:44:09 2000 UTC	revision 678, Tue Jun 27 07:51:09 2000 UTC
#	Line 12	Line 12
12	Tag: <post-commit CVS tag>	Tag: <post-commit CVS tag>
13	Description:	Description:
14	----------------------------------------------------------------------	----------------------------------------------------------------------
15		Name: Matthias Blume
16		Date: 2000/06/27 16:51:00 JST
17		Tag: blume-20000627-noweb
18		Description:
19
20		Changes:
21
22		1. Implemented "subdir" and "witness" options for noweb tool.
23		This caused some slight internal changes in CM's tool implementation.
24		2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
25		some remaining issues from earlier path anchor changes.
26		3. Updated CM manual accordingly.
27
28		4. Changed implementation of back-tracing so that I now consider it
29		ready for prime-time.
30
31		In particular, you don't have to explicitly trigger the back-trace
32		anymore.  Instead, if you are running BTrace-instrumented code and
33		there is an uncaught exception (regardless of whether or not it was
34		raised in instrumented code), the top-level evalloop will print
35		the back-trace.
36
37		Features:
38
39		- Instrumented and uninstrumented code work together seemlessly.
40		(Of course, uninstrumented code is never mentioned in actual
41		back-traces.)
42
43		- Asymptotic time- and space-complexity of instrumented code is
44		equal to that of uninstrumented code.  (This means that
45		tail-recursion is preserved by the instrumentation phase.)
46
47		- Modules whose code has been instrumented in different sessions
48		work together without problem.
49
50		- There is no penalty whatsoever on uninstrumented code.
51
52		- There is no penalty on "raise" expressions, even in
53		instrumented code.
54
55		A potential bug (or perhaps it is a feature, too):
56
57		A back-trace reaches no further than the outermost instrumented
58		non-trivial "raise".  Here, a "trivial" raise is one that is the
59		sole RHS of a "handle" rule.  Thus, back-traces reach trough
60
61		<exp> handle e => raise e
62
63		and even
64
65		<exp> handle Foo => raise Bar
66
67		and, of course, through
68
69		<exp> handle Foo => ...
70
71		if the exception was not Foo.
72
73		Back-traces always reach right through any un-instrumented code
74		including any of its "handle" expressions, trivial or not.
75
76		To try this out, do the following:
77
78		- Erase all existing binfiles for your program.
79		(You may keep binfiles for those modules where you think you
80		definitely don't need back-tracing.)
81		- Turn on back-trace instrumentation:
82		SMLofNJ.Internals.BTrace.mode (SOME true);
83		- Recompile your program.  (I.e., run "CM.make" or "use".)
84		- You may now turn instrumentation off again (if you want):
85		SMLofNJ.Internals.BTrace.mode (SOME false);
86		- Run your program as usual.  If it raises an exception that
87		reaches the interactive toplevel, then a back-trace will
88		automatically be printed.  After that, the toplevel loop
89		will print the exception history as usual.
90
91		----------------------------------------------------------------------
92		Name: Matthias Blume
93		Date: 2000/06/26 09:56:46 JST
94		Tag: blume-20000626-setup
95		Description:
96
97		CM: - setup-parameter to "sml" added; this can be used to run arbitrary
98		ML code before and after compiling a file (e.g., to set compiler
99		flags)
100
101		Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
102		- associated changes to btrace.sml (BTrace instrumentation pass)
103		- cleaner implementation of btimp.sml (BTrace tracing and report
104		module)
105
106		CM manual: * new path encoding documented
107		* description of setup-parameter to "sml" added
108
109		The biggest user-visible change to back-tracing is that it is no
110		longer necessary to compile all traced modules within the same
111		session.  (This was a real limitation.)
112
113		----------------------------------------------------------------------
114		Name: Matthias Blume
115		Date: 2000/06/24 12:40:00 JST
116		Tag: blume-20000624-startup
117		Description:
118
119		Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
120		bit too often -- to put it mildly. :)
121
122		----------------------------------------------------------------------
123		Name: Matthias Blume
124		Date: 2000/06/23 18:20:00 JST
125		Tag: blume-20000623-btrace
126		Description:
127
128		This updates adds a backtrace facility to aid programmers in debugging
129		their programs.  This involves the following changes:
130
131		1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
132		keeping track of the current call stack.  When programs are compiled
133		in a special mode, the compiler will insert calls to these hooks
134		into the user program.
135		"Hook" means that it is possible for different implementations of
136		back-tracing to register themselves (at different times).
137
138		2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
139		as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
140		It can be turned on using this call:
141		SMLofNJ.Internals.BTrace.mode (SOME true);
142		Turning it off again:
143		SMLofNJ.Internals.BTrace.mode (SOME false);
144		Querying the current status:
145		SMLofNJ.Internals.BTrace.mode NONE;
146		Annotated programs are about twice as big as normal ones, and they
147		run a factor of 2 to 4 slower with a dummy back-trace plugin (one
148		where all hooks do nothing).  The slowdown with a plugin that is
149		actually useful (such as the one supplied by default) is even greater,
150		but in the case of the default plugin it is still only an constant
151		factor (amortized).
152
153		3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
154		with a sub-structure BTrace for controlling back-tracing.  In particular,
155		the above-mentioned function "mode" controls whether the annotation
156		phase is invoked by the compiler.  Another important function is
157		"trigger": when called it aborts the current execution and causes
158		the top-level loop to print a full back-trace.
159
160		4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
161		for back-tracing.  It keeps track of the dynamic call stack and in
162		addition to that it keeps a partial history at each "level" of that
163		stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
164		d and b (at separate times, dynamically), then the report will show:
165
166		GOTO   d
167		/c
168		GOTO  \b
169		CALL   a
170
171		This shows that there was an initial non-tail call of a, then a
172		tail-call to b or c, looping behavior in a cluster of functions that
173		consist of b and c, and then a goto from that cluster (i.e., either from
174		b or from c) to d.
175
176		Note that (depending on the user program) the amount of information
177		that the back-trace module has to keep track of at each level is bounded
178		by a constant.  Thus, the whole implementation has the same asymptotical
179		complexity as the original program (both in space and in time).
180
181		5. compiler/TopLevel/interact/evalloop.sml has been modified to
182		handle the special exception SMLofNJ.Internals.BTrace.BTrace
183		which is raised by the "trigger" function mentioned above.
184
185		Notes on usage:
186
187		- Annotated code works well together with unannotated code:
188		Unannotated calls simply do not show up at all in the backtrace.
189
190		- It is not a good idea to let modules that were annotated during
191		different sessions run at the same time.  This is because the compiler
192		chooses small integers to identify individual functions, and there
193		will be clashes if different modules were compiled in separate sessions.
194		(Nothing will crash, and you will even be told about the clashes, but
195		back-trace information will in general not be useful.)
196
197		- Back-tracing can be confused by callcc and capture.
198
199		- The only way of getting a back-trace right now is to explicitly
200		invoke the "trigger" function from your user program.  Eventually, we
201		should make every exception carry back-trace information (if
202		available).  But since this creates more overhead at "raise"-time
203		(similar to the current exnHistory overhead), I have not yet
204		implemented this.  (The implementation will be rather easy.)  With
205		exceptions carrying back-trace information, this facility will be even
206		more useful because users don't need to modify their programs...
207
208		- While it is possible to compile the compiler with back-trace
209		annotations turned on (I did it to get some confidence in
210		correctness), you must make absolutely sure that core.sml and
211		btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
212		be compiled with annotation because there is no core access yet, but
213		if you compile btimp.sml with annotation, then the system will go into
214		an infinite recursion and crash.)
215		Since CM currently does not know about BTrace, the only way to turn
216		annotations on and off for different modules of the compiler is to
217		interrupt CMB.make, change the settings, and re-invoke it.  Of course,
218		this is awkward and clumsy.
219
220		Sample sessions:
221
222		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
223		- SMLofNJ.Internals.BTrace.mode (SOME true);
224		[autoloading]
225		[autoloading done]
226		val it = false : bool
227		- structure X = struct
228		-     fun main n = let
229		-         fun a (x, 0) = d x
230		-           | a (x, n) = b (x, n - 1)
231		-         and b (x, n) = c (x, n)
232		-         and c (x, n) = a (x, n)
233		-         and d x = e (x, 3)
234		-         and e (x, 0) = f x
235		-           | e (x, n) = e (x, n - 1)
236		-         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
237		-           | f n = n * g (n - 1)
238		-         and g n = a (n, 3)
239		-     in
240		-         f n
241		-     end
242		- end;
243		structure X : sig val main : int -> int end
244		- X.main 3;
245		*** BACK-TRACE ***
246		GOTO   stdIn:4.2-13.20: X.main[2].f
247		GOTO-( stdIn:4.2-13.20: X.main[2].e
248		GOTO   stdIn:4.2-13.20: X.main[2].d
249		/ stdIn:4.2-13.20: X.main[2].a
250		| stdIn:4.2-13.20: X.main[2].b
251		GOTO-\ stdIn:4.2-13.20: X.main[2].c
252		CALL   stdIn:4.2-13.20: X.main[2].g
253		GOTO   stdIn:4.2-13.20: X.main[2].f
254		GOTO-( stdIn:4.2-13.20: X.main[2].e
255		GOTO   stdIn:4.2-13.20: X.main[2].d
256		/ stdIn:4.2-13.20: X.main[2].a
257		| stdIn:4.2-13.20: X.main[2].b
258		GOTO-\ stdIn:4.2-13.20: X.main[2].c
259		CALL   stdIn:4.2-13.20: X.main[2].g
260		GOTO   stdIn:4.2-13.20: X.main[2].f
261		GOTO-( stdIn:4.2-13.20: X.main[2].e
262		GOTO   stdIn:4.2-13.20: X.main[2].d
263		/ stdIn:4.2-13.20: X.main[2].a
264		| stdIn:4.2-13.20: X.main[2].b
265		GOTO-\ stdIn:4.2-13.20: X.main[2].c
266		CALL   stdIn:4.2-13.20: X.main[2].g
267		GOTO   stdIn:4.2-13.20: X.main[2].f
268		CALL   stdIn:2.15-17.4: X.main[2]
269		-
270
271		(Note that because of a FLINt bug the above code currently does not
272		compile without BTrace turned on.)
273
274		Here is another example, using my modified Tiger compiler:
275
276		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
277		- SMLofNJ.Internals.BTrace.mode (SOME true);
278		[autoloading]
279		[autoloading done]
280		val it = false : bool
281		- CM.make "sources.cm";
282		[autoloading]
283		...
284		[autoloading done]
285		[scanning sources.cm]
286		[parsing (sources.cm):parse.sml]
287		[creating directory CM/SKEL ...]
288		[parsing (sources.cm):tiger.lex.sml]
289		...
290		[wrote CM/sparc-unix/semant.sml]
291		[compiling (sources.cm):main.sml]
292		[wrote CM/sparc-unix/main.sml]
293		[New bindings added.]
294		val it = true : bool
295		- Main.compile ("../testcases/merge.tig", "foo.out");
296		*** BACK-TRACE ***
297		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
298		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
299		CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
300		GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
301		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
302		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
303		CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
304		/ lib/semant.sml:411.65-543.8: SemantFun[2].transDec
305		CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
306		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
307		CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
308		CALL   main.sml:1.18-118.4: Main.compile[2]
309		-
310
311		----------------------------------------------------------------------
312		Name: Matthias Blumen
313		Date: 2000/06/21 18:00:00 JST
314		Tag: blume-20000621-manual
315		Description:
316
317		CM manual update: Path environments documented.
318
319		----------------------------------------------------------------------
320		Name: Matthias Blume
321		Date: 2000/06/19 13:40:00
322		Tag: blume-20000619-manual
323		Description:
324
325		CM manual and system/README update.  This only covers the fact that
326		there are no more implicit anchors.  (Path environments and the "bind"
327		option to "cm" have yet to be documented.)
328
329		----------------------------------------------------------------------
330		Name: Matthias Blume
331		Date: 2000/06/19 11:05:00 JST
332		Tag: blume-20000619-chdir-bugfix
333		Description:
334
335		Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
336
337		----------------------------------------------------------------------
338		Name: Matthias Blume
339		Date: 2000/06/18 22:00:10 JST
340		Tag: blume-20000618-implicit-anchors-really-gone
341		Description:
342
343		I updates the previous HISTORY entry where I forgot to mention that
344		implicit anchors are no longer with us.
345
346		The current update also gets rid of the (now useless) controller
347		CM.Control.implicit_anchors.
348
349		----------------------------------------------------------------------
350		Name: Matthias Blume
351		Date: 2000/06/16 17:30:00 JST
352		Tag: blume-20000616-anchorenv
353		Description:
354
355		This patch implements the long anticipated (just kidding :) "anchor
356		environment" mechanism.  In the course of doing this, I also
357		re-implemented CM's internal "SrcPath" module from scratch.  The new
358		one should be more robust in certain boundary cases.  In any case, it
359		is a lot cleaner than its predecessor (IMHO).
360
361		This time, although there is yet another boot file format change, I
362		kept the unpickler backward-compatible.  As a result, no new bootfiles
363		are necessary and bootstrapping is straightforward.  (You cannot read
364		new bootfiles into an old system, but the other way around is no
365		problem.)
366
367		Visible changes:
368
369		** 0. Implicit path anchors (without the leading $-symbol) are no
370		longer recognized at all. This means that such path names are not
371		illegal either.  For example, the name basis.cm simply refers to a
372		local file called "basis.cm" (i.e, the name is an ordinary path
373		relative to .cm-files directory).  Or, to put it differently, only
374		names that start with $ are anchored paths.
375
376		** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
377		vanished.
378
379		John (Reppy) had critizised this as soon as I originally proposed and
380		implemented it, but at that time I did not really deeply believe
381		him. :) Now I came full-circle because I need the $<singlearc> syntax
382		in another place where it cannot be seen as an abbreviation for
383		$/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
384		seems to mean (i.e., it "expands" into the corresponding anchor
385		value).
386
387		However, when paths are used as members in CM description files, it
388		continues to be true that there must be at least another arc after the
389		anchor.  This is now enforced separately during semantic analysis
390		(i.e., from a lexical/syntactical point of view, the notation is ok.)
391
392		** 2. The "cm" class now accepts an option "bind".  The option's value
393		is a sub-option list of precisely two items -- one labeled "anchor"
394		and the other one labeled "value".  As you might expect, "anchor" is
395		used to specify an anchor name to be bound, and "value" specifies what
396		the anchor is being bound to.
397
398		The value must be a directory name and can be given in either standard
399		syntax (including the possibility that it is itself an anchored path)
400		or native syntax.
401
402		Examples:
403
404		foo.cm (bind:(anchor:bar value:$mystuff/bar))
405		lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
406
407		and so on.
408
409		The meaning of this is that the .cm-file will be processed with an
410		augmented anchor environment where the given anchor(s) is/are bound to
411		the given values(s).
412
413		The rationale for having this feature is this: Suppose you are trying
414		to use two different (already stable) libraries a.cm and b.cm (that
415		you perhaps didn't write yourself).  Further, suppose each of these
416		two libraries internally uses its own auxiliary library $aux/lib.cm.
417		Normally you would now have a problem because the anchor "lib" can not
418		be bound to more than one value globally.  Therefore, the project that
419		uses both a.cm and b.cm must locally redirect the anchor to some other
420		place:
421
422		a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
423		b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
424
425		This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
426		/usr/lib/smlnj/b-stuff/aux.cm, respectively.
427
428		Hard-wiring path names is a bit inflexible (and CM will verbosely warn
429		you when you do so at the time of CM.stabilize).  Therefore, you can
430		also use an anchored path as the value:
431
432		a.cm (bind:(anchor:lib value:$a-lib))
433		b.cm (bind:(anchor:lib value:$b-lib))
434
435		Now you can globally configure (using the usual CM.Anchor.anchor or
436		pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
437		itself is always locally bound, setting it globally is no longer
438		meaningful or necessary (but it does not hurt either).  In fact, "lib"
439		can still be used as a global anchor for separate purposes.  As a
440		matter of fact, one can locally define "lib" in terms of a global
441		"lib":
442
443		a.cm (bind:(anchor:lib value:$lib/a))
444		b.cm (bind:(anchor:lib value:$lib/b))
445
446		** 3: The encoding of path names has changed.  This affects the way
447		path names are shown in CM's progress report and also the internal
448		protocol encoding used for parallel make.
449
450		The encoding now uses one or more ':'-separated segments.  Each
451		segments corresponds to a file that has been specified relative to the
452		file given by its preceding segment.  The first segment is either
453		relative to the CWD, absolute, or anchored.  Each segment itself is
454		basically a Unix pathname; all segments but the first are relative.
455
456		Example:
457
458		$foo/bar/baz.cm:a/b/c.sml
459
460		This path denotes the file bar/a/b/c.sml relative to the directory
461		denoted by anchor "foo".  Notice that the encoding also includes
462		baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
463		paths are resolved relative to the .cm-files directory, so baz.cm must
464		be ignored to get the "real" pathname.
465
466		To make this fact more obvious, CM puts the names of such "virtual
467		arcs" into parentheses when they appear in progress reports. (No
468		parentheses will appear in the internal protocol encoding.)  Thus,
469		what you really see is:
470
471		$foo/bar/(baz.cm):a/b/c.sml
472
473		I find this notation to be much more informative than before.
474
475		Another new feature of the encoding is that special characters
476		including parentheses, colons, (back)slashes, and white space are
477		written as \ddd (where ddd is the decimal encoding of the character).
478
479		*** The CM manual still needs to be updated.
480
481		----------------------------------------------------------------------
482		Name: Allen Leung
483		Date: 2000/06/15 00:38:00
484		Tag: leunga-20000615-x86-peephole
485
486		x86 Peephole fix by Fermin.  Affects c-- and moby only.
487
488		----------------------------------------------------------------------
489		Name: Matthias Blume
490		Date: 2000/06/12 11:40:00
491		Tag: blume-20000612-parmakefix
492		Description:
493
494		More cleanup after changing the file naming scheme: This time I
495		repaired the parallel make mechanism for CMB.make which I broke earlier.
496
497		----------------------------------------------------------------------
498		Name: Allen Leung
499		Date: 2000/06/09 01:25:00
500		Tag: leunga-20000609-various
501
502		None of these things should affect normal SML/NJ operations
503
504		1. Peephole improvements provided by Fermin (c--)
505		2. New annotation DEFUSE for adding extra dependence (moby)
506		3. New X86 LOCK instructions (moby)
507		4. New machine description language for reservation tables (scheduling)
508		5. Fixes to various optimization/analysis modules (branch chaining, dominator
509		trees etc.)
510		6. I've changed the CM files so that they can work with versions
511		110.0.6, 110.25 and 110.28
512
513		----------------------------------------------------------------------
514		Name: Matthias Blume
515		Date: 2000/06/09 12:40:00
516		Tag: blume-20000609-log
517		Description:
518
519		- Removed all(?) remaining RCS Log entries from sources.
520
521		- Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
522		anchored paths).
523
524		----------------------------------------------------------------------
525		Name: Matthias Blume
526		Date: 2000/06/07 17:00:00 JST
527		Tag: blume-20000607-no-implicit-anchors
528		Description:
529
530		1. This update changes the default setting for
531		CM.Control.implicit_anchors from true to false.  This means that
532		implicit anchors are no longer permitted by default.  I also tried to
533		make sure that nothing else still relies on implicit anchors.
534		(This is the next step on the schedule towards a CM that does not even
535		have the notion of implicit anchors anymore.)
536
537		2. More CM manual updates.
538
539		3. I managed to track down and fix the pickling bug I mentioned last
540		time.  Because of the previously existing workaround, this entails no
541		immediate practical changes.
542
543		----------------------------------------------------------------------
544		Name: Matthias Blume
545		Date: 2000/06/06 11:15:00 JST
546		Tag: blume-20000606-lazierpickle
547		Description:
548
549		!!!! NEW BOOT FILES !!!!
550
551		* The main purpose of this update is to make library pickles lazier in
552		order to reduce the initial space penalty for autoloading a library.
553		As a result, it is now possible to have $smlnj/compiler.cm
554		pre-registered.  This should take care of the many complaints or
555		inquiries about missing structure Compiler.  This required changes to
556		CM's internal data structures and small tweaks to some algorithms.
557
558		As a neat additional effect, it is no longer necessary (for the sake
559		of lean heap image files) to distinguish between a "minimal" CM and a
560		"full" CM.  Now, there is only one CM (i.e., the "full" version:
561		$smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
562		interactive top level. ($smlnj/cm/minimal.cm is gone.)
563
564		To make the life of compiler-hackers easier, "makeml" now also
565		pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
566		words, after you bootstrap a new sml for the first time, you will not
567		have to autoload $smlnj/cmb.cm again afterwards.  (The first time
568		around you will still have to do it, though.)
569
570		* A second change consists of major updates to the CM manual.  There
571		are now several appendices with summary information and also a full
572		specification of the CM description file syntax.
573
574		* In directory src/system I added the script "allcross".  This script
575		invokes sml and cross-compiles the compiler for all supported
576		architectures.  (Useful when providing a new set of boot files.)
577
578		* There seems to be a latent bug in my "lazy pickles" mechanism.  I
579		added a small tweak to pickle-util.sml to work around this problem,
580		but it is not a proper fix yet.  I will investigate further.  (The
581		effect of the bug was an inflation of library pickle size.)
582
583		* Version number increased to 110.28.1 (to avoid compatibility problems).
584
585		----------------------------------------------------------------------
586		Name: Allen Leung
587		Date: 2000/05/25 17:28 EDT
588		Tag: leunga-20000525-ra
589		Description:
590
591		Fixed a bug in freezing phase of the register allocator.
592
593		----------------------------------------------------------------------
594		Name: Allen Leung
595		Date: 2000/05/15 22:53 EDT
596		Tag: leunga-20000515-alpha-x86-ra
597		Description:
598
599		1. Alpha
600
601		Slight cleanup.  Removed the instruction SGNXL
602
603		2. X86
604
605		Added the following instructions to the instruction set:
606
607		ROLx, RORx,
608		BTx, BTSx, BTLx, BTRx,
609		XCHGx, and variants with the LOCK prefix
610
611		3. Register Allocation
612
613		The module ra-rewrite-with-renaming has been improved.
614
615		These have no effect on SML/NJ.
616
617		----------------------------------------------------------------------
618		Name: Matthias Blume
619		Date: 2000/05/15 16:20:00 JST
620		Tag: blume-20000515-lightrebuild
621		Description:
622
623		1. I added an alternative to "-rebuild" to "makeml".  The difference is
624		that prior to calling CMB.make' the CM-variable "LIGHT" will be
625		defined.  In effect, the command will not build any cross-compiler
626		backends and therefore finish more quickly.
627
628		The "fixpt" script also takes a "-light" switch to be able to use
629		this new facility while compiling for a fixpoint.
630
631		2. I replaced all mentions of anchored paths in group owner specifications
632		with simple relative paths (usually starting with "..").
633		The rationale is that a library's internal workings should not be
634		compromised by the lack of some anchor.  (An anchor is necessary
635		for someone who wants to refer to the library by an anchored path,
636		but it should not be necessary to build the same library in the first
637		place.)
638
639		3. I changed the way CM's tool mechanism determines the shell command
640		string used for things like ml-yacc etc. so that it does not break
641		when CM.Control.implicit_anchors is turned off.
642
643		----------------------------------------------------------------------
644		Name: Matthias Blume
645		Date: 2000/05/12 18:20:00 JST
646		Tag: blume-20000512-ml-build
647		Description:
648
649		Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
650		getting installed properly (by config/install.sh).
651
652		----------------------------------------------------------------------
653		Name: Matthias Blume
654		Date: 2000/05/12 17:30:00 JST
655		Tag: blume-20000512-anchors
656		Description:
657
658		!!! NEW BOOT FILES !!!
659
660		This change is in preparation of fading out support for "implicitly
661		anchored path names".  I went through all sources and used the
662		explicit (and relatively new) $-notation.  See system/README and the
663		CM manual for more info on this.
664
665		I also modified the anchoring scheme for some things such as "smlnj",
666		"MLRISC", "cm", etc. to take advantage of the fact that explicit
667		anchors are more expressive: anchor name and first arc do not have to
668		coincide.  This entails the following user-visible change:
669
670		You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
671		particular, when you fire up sml with a command-line argument, say,
672		e.g.:
673
674		sml '$smlnj/cmb.cm'
675
676		At the ML toplevel prompt:
677
678		CM.autoload "$smlnj/cmb.cm";
679
680		There is also a new controller in CM.Control that can be used to turn
681		off all remaining support for implicit anchors by saying:
682
683		CM.autoload "$smlnj/
684		#set CM.Control.implicit_anchors false;
685
686		This causes CM to reject implicitly anchored paths.  This is (for the
687		time being) less permissive than the "final" version where there will
688		be no more such implicit anchors and relative paths will be just that:
689		relative.
690
691		The next step (version after next version?) will be to make the
692		default for CM.Control.implicit_anchors false.  After the dust has
693		settled, I can then produce the "final" version of this...
694
695		Note: Since bootstrapping is a bit tricky, I provided new boot files.
696
697		----------------------------------------------------------------------
698		Name: Matthias Blume
699		Date: 2000/05/11 16:30:00 JST
700		Tag: blume-20000511-sources
701		Description:
702
703		The main change is that I added function CM.sources as a generalized
704		version of the earlier CM.makedepend.  This entails the following
705		additional changes:
706
707		- CM.makedepend has been dropped.
708
709		- CM manual has been updated.
710
711		- TOOLS signature and API have been changed.
712
713		----------------------------------------------------------------------
714		Name: Allen Leung
715		Date: 2000/05/10 21:17 EDT
716		Tag: leunga-20000510-moby-c--ssa
717		Description:
718
719		Various bug fixes and new features for C--, Moby and MLRISC optimizations.
720		None of these affect SML/NJ.
721
722		1. Register Allocation
723
724		a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
725		This module tries to remove local (i.e. basic block level) redundancies
726		during spilling.
727
728		b. A new framework for performing region based register allocation.
729		Not yet entirely functional.
730
731		2. X86
732
733		a. DefUse for POP was missing the stack pointer [found by Lal]
734		b. Reload for CALL was incorrect in X86Spill [found by John]
735		c. Various fixes in X86Spill so that it can be used correctly for
736		the new spilling module.
737
738		3. SSA/IR
739
740		a. New module ir/dj-dataflow.sml implements elimination based
741		data flow analysis.
742
743		4. MLRiscGen
744
745		a. Fix for gc type annotation
746
747		5. MDGen
748
749		Various fixes for machine description -> ml code translation.  For ssa
750		only.
751
752		----------------------------------------------------------------------
753		Name: Allen Leung
754		Date: 2000/05/08 22:17 EDT
755		Tag: leunga-20000508-labexp
756		Description:
757
758		Fermin has found a few assembly problems with constant expressions
759		generated in LabelExp.  Mostly, the problems involve extra parentheses,
760		which choke on dumb assemblers.  This is his fix.
761
762		----------------------------------------------------------------------
763		Name: Dave MacQueen
764		Date: 2000/04/09 14:00 EDT
765		Tag: dbm-20000502-Version_110_28
766		Description:
767
768		1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
769
770		2. Updated config/version to 110.28
771
772		3. Updated config/srcarchiveurl
773
774		3. New boot files!
775		ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
776
777		----------------------------------------------------------------------
778		Name: Matthias Blume
779		Date: 2000/05/01 19:05:00 JST
780		Tag: blume-20000501-noweb
781		Description:
782
783		A new noweb tool has been added.  The existing system is entirely
784		unaffected by this, but some CM users have asked for renewed noweb
785		support.  Everything is documented in the CM manual.
786
787		New (plugin) libraries:
788
789		noweb-tool.cm
790		nw-ext.cm
791
792		----------------------------------------------------------------------
793	Name: Dave MacQueen	Name: Dave MacQueen
794	Date: 2000/04/30 12:40PM EDT	Date: 2000/04/30 12:40PM EDT
795	Tag: dbm-20000430-bug_fixes	Tag: dbm-20000430-bug_fixes
#	Line 295	Line 1073
1073
1074	3.  Assembly	3.  Assembly
1075
1076	When generating assemby, resolve the value of client defined constants,	When generating assembly, resolve the value of client defined constants,
1077	instead of generating symbolic values.  This is controlled by the	instead of generating symbolic values.  This is controlled by the
1078	new flag "asm-resolve-constants", which is default to true.	new flag "asm-resolve-constants", which is default to true.
1079
#	Line 318	Line 1096
1096
1097	To this end, I arranged that instead of "structure Core" as "structure	To this end, I arranged that instead of "structure Core" as "structure
1098	_Core" is bound in the pervasive environment.  Core access is done via	_Core" is bound in the pervasive environment.  Core access is done via
1099	_Core (which can never be accidentially rebound because _Core is not a	_Core (which can never be accidentally rebound because _Core is not a
1100	legal surface-syntax symbol).	legal surface-syntax symbol).
1101
1102	The current solution is much cleaner because the core environment is	The current solution is much cleaner because the core environment is
#	Line 328	Line 1106
1106	with dynamic and symbolic parts of the core environment.	with dynamic and symbolic parts of the core environment.
1107
1108	Remaining hackery (to bind the "magic" symbol _Core) is localized in the	Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1109	compilation mananger's bootstrap compiler (actually: in the "init group"	compilation manager's bootstrap compiler (actually: in the "init group"
1110	handling).  See the comments in src/system/smlnj/init/init.cmi for	handling).  See the comments in src/system/smlnj/init/init.cmi for
1111	more details.	more details.
1112
#	Line 443	Line 1221
1221	(specified in the .cm file at each instance where the tool's class is	(specified in the .cm file at each instance where the tool's class is
1222	used).	used).
1223
1224	This was done to accomodate the new "make" and "shell" tools which	This was done to accommodate the new "make" and "shell" tools which
1225	facilitate fairly seemless hookup to portions of code managed using	facilitate fairly seamless hookup to portions of code managed using
1226	Makefiles or Shell scripts.	Makefiles or Shell scripts.
1227
1228	There are no classes "shared" or "private" anymore.  Instead, the	There are no classes "shared" or "private" anymore.  Instead, the
#	Line 456	Line 1234
1234
1235	All existing tools are described in the CM manual.	All existing tools are described in the CM manual.
1236
1237	- Slightly better error handling.  (CM now surpresses many followup	- Slightly better error handling.  (CM now suppresses many followup
1238	error messages that tended to be more annoying than helpful.)	error messages that tended to be more annoying than helpful.)
1239
1240	2. Major changes to the compiler's static environment data structures.	2. Major changes to the compiler's static environment data structures.
#	Line 590	Line 1368
1368
1369	I've changed andl to testl in the floating point test sequence	I've changed andl to testl in the floating point test sequence
1370	whenever appropriate.  The Intel optimization guide states that	whenever appropriate.  The Intel optimization guide states that
1371	testl is perferable to andl.	testl is preferable to andl.
1372
1373	7. RA (x86 only)	7. RA (x86 only)
1374
#	Line 772	Line 1550
1550
1551	1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an	1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1552	additional argument called "template" which is an optional string that	additional argument called "template" which is an optional string that
1553	specifiel the layout of the tool command line.  See the CM manual for	specifies the layout of the tool command line.  See the CM manual for
1554	explanation.	explanation.
1555
1556	2. A special-purpose tool can be "regisitered" by simply dropping the	2. A special-purpose tool can be "registered" by simply dropping the
1557	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1558	directory where the .cm file lives that uses this tool.  (The	directory where the .cm file lives that uses this tool.  (The
1559	behavior/misfeature until now was to look for the tool description	behavior/misfeature until now was to look for the tool description
#	Line 819	Line 1597
1597	rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for	rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
1598	<k>=1,2,... is normally not a good idea when invoking fixpt.  However,	<k>=1,2,... is normally not a good idea when invoking fixpt.  However,
1599	they might be the result of an earlier partial run of fixpt (which	they might be the result of an earlier partial run of fixpt (which
1600	perhaps got accidentially killed).  In this case, fixpt will quickly	perhaps got accidentally killed).  In this case, fixpt will quickly
1601	move through what exists before continuing where it left off earlier,	move through what exists before continuing where it left off earlier,
1602	and, thus, saves a lot of time.	and, thus, saves a lot of time.
1603
#	Line 869	Line 1647
1647	it from that remote directory.	it from that remote directory.
1648	This should simplify installation further:  For machines that have	This should simplify installation further:  For machines that have
1649	access to the internet, just fetch <version>-config.tgz, unpack it,	access to the internet, just fetch <version>-config.tgz, unpack it,
1650	edit config/targets, and go (run config/install.sh).  The scipt will	edit config/targets, and go (run config/install.sh).  The script will
1651	fetch everything else that it might need all by itself.	fetch everything else that it might need all by itself.
1652
1653	For CVS users, this mechanism is not relevant for source archives, but	For CVS users, this mechanism is not relevant for source archives, but

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.633   changed lines   Added in v.678

---

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