Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 641, Thu May 11 01:16:45 2000 UTC	revision 684, Tue Jul 11 04:57:48 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: 2000/07/11 13:58:00 JST
17		Tag: blume-20000711-doctypo
18		Description:
19
20		Fixed a few typos in CM manual.
21
22		----------------------------------------------------------------------
23		Name: Allen Leung
24		Date: 2000/06/15 00:38:00
25		Tag: leunga-20000704-sparc-x86
26
27		1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
28		2. fix to x86 RET bug [found by Dan Grossman]
29		3. sparc assembly bug fix for ticc instructions [found by Fermin]
30
31		Affects c-- and moby only
32
33		----------------------------------------------------------------------
34		Name: Matthias Blume
35		Date: 2000/07/04 15:26:00
36		Tag: blume-20000704-trigger
37		Description:
38
39		1. Improvements to CM manual.
40		2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
41		of getting a back-trace.  The function, when called, raises an
42		internal exception which explicitly carries the full back-trace history,
43		so it is unaffected by any intervening handle-raise pairs ("trivial"
44		or not).  The interactive loop will print that history once it arrives
45		at top level.
46		Short of having all exceptions implicitly carry the full history, the
47		recommended way of using this facility is:
48		- compile your program with instrumentation "on"
49		- run it, when it raises an exception, look at the history
50		- if the history is "cut off" because of some handler, go and modify
51		your program so that it explicitly calls BTrace.trigger
52		- recompile (still instrumented), and rerun; look at the full history
53
54		----------------------------------------------------------------------
55		Name: Matthias Blume
56		Date: 2000/07/03 15:36:00 JST
57		Tag: blume-20000702-manual
58		Description:
59
60		Small corrections and updates to CM manual.
61
62		----------------------------------------------------------------------
63		Name: Matthias Blume
64		Date: 2000/06/29 16:04:00 JST
65		Tag: blume-20000629-yacctool
66		Description:
67
68		Changes:
69
70		1. Class "mlyacc" now takes separate arguments to pass options to
71		generated .sml- and .sig-files independently.
72		2. Corresponding CM manual updates.
73		3. BTrace module now also reports call sites.  (However, for loop clusters
74		it only shows from where the cluster was entered.)  There are associated
75		modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
76
77		----------------------------------------------------------------------
78		Name: Matthias Blume
79		Date: 2000/06/27 16:51:00 JST
80		Tag: blume-20000627-noweb
81		Description:
82
83		Changes:
84
85		1. Implemented "subdir" and "witness" options for noweb tool.
86		This caused some slight internal changes in CM's tool implementation.
87		2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
88		some remaining issues from earlier path anchor changes.
89		3. Updated CM manual accordingly.
90
91		4. Changed implementation of back-tracing so that I now consider it
92		ready for prime-time.
93
94		In particular, you don't have to explicitly trigger the back-trace
95		anymore.  Instead, if you are running BTrace-instrumented code and
96		there is an uncaught exception (regardless of whether or not it was
97		raised in instrumented code), the top-level evalloop will print
98		the back-trace.
99
100		Features:
101
102		- Instrumented and uninstrumented code work together seemlessly.
103		(Of course, uninstrumented code is never mentioned in actual
104		back-traces.)
105
106		- Asymptotic time- and space-complexity of instrumented code is
107		equal to that of uninstrumented code.  (This means that
108		tail-recursion is preserved by the instrumentation phase.)
109
110		- Modules whose code has been instrumented in different sessions
111		work together without problem.
112
113		- There is no penalty whatsoever on uninstrumented code.
114
115		- There is no penalty on "raise" expressions, even in
116		instrumented code.
117
118		A potential bug (or perhaps it is a feature, too):
119
120		A back-trace reaches no further than the outermost instrumented
121		non-trivial "raise".  Here, a "trivial" raise is one that is the
122		sole RHS of a "handle" rule.  Thus, back-traces reach trough
123
124		<exp> handle e => raise e
125
126		and even
127
128		<exp> handle Foo => raise Bar
129
130		and, of course, through
131
132		<exp> handle Foo => ...
133
134		if the exception was not Foo.
135
136		Back-traces always reach right through any un-instrumented code
137		including any of its "handle" expressions, trivial or not.
138
139		To try this out, do the following:
140
141		- Erase all existing binfiles for your program.
142		(You may keep binfiles for those modules where you think you
143		definitely don't need back-tracing.)
144		- Turn on back-trace instrumentation:
145		SMLofNJ.Internals.BTrace.mode (SOME true);
146		- Recompile your program.  (I.e., run "CM.make" or "use".)
147		- You may now turn instrumentation off again (if you want):
148		SMLofNJ.Internals.BTrace.mode (SOME false);
149		- Run your program as usual.  If it raises an exception that
150		reaches the interactive toplevel, then a back-trace will
151		automatically be printed.  After that, the toplevel loop
152		will print the exception history as usual.
153
154		----------------------------------------------------------------------
155		Name: Matthias Blume
156		Date: 2000/06/26 09:56:46 JST
157		Tag: blume-20000626-setup
158		Description:
159
160		CM: - setup-parameter to "sml" added; this can be used to run arbitrary
161		ML code before and after compiling a file (e.g., to set compiler
162		flags)
163
164		Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
165		- associated changes to btrace.sml (BTrace instrumentation pass)
166		- cleaner implementation of btimp.sml (BTrace tracing and report
167		module)
168
169		CM manual: * new path encoding documented
170		* description of setup-parameter to "sml" added
171
172		The biggest user-visible change to back-tracing is that it is no
173		longer necessary to compile all traced modules within the same
174		session.  (This was a real limitation.)
175
176		----------------------------------------------------------------------
177		Name: Matthias Blume
178		Date: 2000/06/24 12:40:00 JST
179		Tag: blume-20000624-startup
180		Description:
181
182		Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
183		bit too often -- to put it mildly. :)
184
185		----------------------------------------------------------------------
186		Name: Matthias Blume
187		Date: 2000/06/23 18:20:00 JST
188		Tag: blume-20000623-btrace
189		Description:
190
191		This updates adds a backtrace facility to aid programmers in debugging
192		their programs.  This involves the following changes:
193
194		1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
195		keeping track of the current call stack.  When programs are compiled
196		in a special mode, the compiler will insert calls to these hooks
197		into the user program.
198		"Hook" means that it is possible for different implementations of
199		back-tracing to register themselves (at different times).
200
201		2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
202		as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
203		It can be turned on using this call:
204		SMLofNJ.Internals.BTrace.mode (SOME true);
205		Turning it off again:
206		SMLofNJ.Internals.BTrace.mode (SOME false);
207		Querying the current status:
208		SMLofNJ.Internals.BTrace.mode NONE;
209		Annotated programs are about twice as big as normal ones, and they
210		run a factor of 2 to 4 slower with a dummy back-trace plugin (one
211		where all hooks do nothing).  The slowdown with a plugin that is
212		actually useful (such as the one supplied by default) is even greater,
213		but in the case of the default plugin it is still only an constant
214		factor (amortized).
215
216		3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
217		with a sub-structure BTrace for controlling back-tracing.  In particular,
218		the above-mentioned function "mode" controls whether the annotation
219		phase is invoked by the compiler.  Another important function is
220		"trigger": when called it aborts the current execution and causes
221		the top-level loop to print a full back-trace.
222
223		4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
224		for back-tracing.  It keeps track of the dynamic call stack and in
225		addition to that it keeps a partial history at each "level" of that
226		stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
227		d and b (at separate times, dynamically), then the report will show:
228
229		GOTO   d
230		/c
231		GOTO  \b
232		CALL   a
233
234		This shows that there was an initial non-tail call of a, then a
235		tail-call to b or c, looping behavior in a cluster of functions that
236		consist of b and c, and then a goto from that cluster (i.e., either from
237		b or from c) to d.
238
239		Note that (depending on the user program) the amount of information
240		that the back-trace module has to keep track of at each level is bounded
241		by a constant.  Thus, the whole implementation has the same asymptotical
242		complexity as the original program (both in space and in time).
243
244		5. compiler/TopLevel/interact/evalloop.sml has been modified to
245		handle the special exception SMLofNJ.Internals.BTrace.BTrace
246		which is raised by the "trigger" function mentioned above.
247
248		Notes on usage:
249
250		- Annotated code works well together with unannotated code:
251		Unannotated calls simply do not show up at all in the backtrace.
252
253		- It is not a good idea to let modules that were annotated during
254		different sessions run at the same time.  This is because the compiler
255		chooses small integers to identify individual functions, and there
256		will be clashes if different modules were compiled in separate sessions.
257		(Nothing will crash, and you will even be told about the clashes, but
258		back-trace information will in general not be useful.)
259
260		- Back-tracing can be confused by callcc and capture.
261
262		- The only way of getting a back-trace right now is to explicitly
263		invoke the "trigger" function from your user program.  Eventually, we
264		should make every exception carry back-trace information (if
265		available).  But since this creates more overhead at "raise"-time
266		(similar to the current exnHistory overhead), I have not yet
267		implemented this.  (The implementation will be rather easy.)  With
268		exceptions carrying back-trace information, this facility will be even
269		more useful because users don't need to modify their programs...
270
271		- While it is possible to compile the compiler with back-trace
272		annotations turned on (I did it to get some confidence in
273		correctness), you must make absolutely sure that core.sml and
274		btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
275		be compiled with annotation because there is no core access yet, but
276		if you compile btimp.sml with annotation, then the system will go into
277		an infinite recursion and crash.)
278		Since CM currently does not know about BTrace, the only way to turn
279		annotations on and off for different modules of the compiler is to
280		interrupt CMB.make, change the settings, and re-invoke it.  Of course,
281		this is awkward and clumsy.
282
283		Sample sessions:
284
285		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
286		- SMLofNJ.Internals.BTrace.mode (SOME true);
287		[autoloading]
288		[autoloading done]
289		val it = false : bool
290		- structure X = struct
291		-     fun main n = let
292		-         fun a (x, 0) = d x
293		-           | a (x, n) = b (x, n - 1)
294		-         and b (x, n) = c (x, n)
295		-         and c (x, n) = a (x, n)
296		-         and d x = e (x, 3)
297		-         and e (x, 0) = f x
298		-           | e (x, n) = e (x, n - 1)
299		-         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
300		-           | f n = n * g (n - 1)
301		-         and g n = a (n, 3)
302		-     in
303		-         f n
304		-     end
305		- end;
306		structure X : sig val main : int -> int end
307		- X.main 3;
308		*** BACK-TRACE ***
309		GOTO   stdIn:4.2-13.20: X.main[2].f
310		GOTO-( stdIn:4.2-13.20: X.main[2].e
311		GOTO   stdIn:4.2-13.20: X.main[2].d
312		/ stdIn:4.2-13.20: X.main[2].a
313		| stdIn:4.2-13.20: X.main[2].b
314		GOTO-\ stdIn:4.2-13.20: X.main[2].c
315		CALL   stdIn:4.2-13.20: X.main[2].g
316		GOTO   stdIn:4.2-13.20: X.main[2].f
317		GOTO-( stdIn:4.2-13.20: X.main[2].e
318		GOTO   stdIn:4.2-13.20: X.main[2].d
319		/ stdIn:4.2-13.20: X.main[2].a
320		| stdIn:4.2-13.20: X.main[2].b
321		GOTO-\ stdIn:4.2-13.20: X.main[2].c
322		CALL   stdIn:4.2-13.20: X.main[2].g
323		GOTO   stdIn:4.2-13.20: X.main[2].f
324		GOTO-( stdIn:4.2-13.20: X.main[2].e
325		GOTO   stdIn:4.2-13.20: X.main[2].d
326		/ stdIn:4.2-13.20: X.main[2].a
327		| stdIn:4.2-13.20: X.main[2].b
328		GOTO-\ stdIn:4.2-13.20: X.main[2].c
329		CALL   stdIn:4.2-13.20: X.main[2].g
330		GOTO   stdIn:4.2-13.20: X.main[2].f
331		CALL   stdIn:2.15-17.4: X.main[2]
332		-
333
334		(Note that because of a FLINt bug the above code currently does not
335		compile without BTrace turned on.)
336
337		Here is another example, using my modified Tiger compiler:
338
339		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
340		- SMLofNJ.Internals.BTrace.mode (SOME true);
341		[autoloading]
342		[autoloading done]
343		val it = false : bool
344		- CM.make "sources.cm";
345		[autoloading]
346		...
347		[autoloading done]
348		[scanning sources.cm]
349		[parsing (sources.cm):parse.sml]
350		[creating directory CM/SKEL ...]
351		[parsing (sources.cm):tiger.lex.sml]
352		...
353		[wrote CM/sparc-unix/semant.sml]
354		[compiling (sources.cm):main.sml]
355		[wrote CM/sparc-unix/main.sml]
356		[New bindings added.]
357		val it = true : bool
358		- Main.compile ("../testcases/merge.tig", "foo.out");
359		*** BACK-TRACE ***
360		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
361		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
362		CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
363		GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
364		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
365		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
366		CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
367		/ lib/semant.sml:411.65-543.8: SemantFun[2].transDec
368		CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
369		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
370		CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
371		CALL   main.sml:1.18-118.4: Main.compile[2]
372		-
373
374		----------------------------------------------------------------------
375		Name: Matthias Blumen
376		Date: 2000/06/21 18:00:00 JST
377		Tag: blume-20000621-manual
378		Description:
379
380		CM manual update: Path environments documented.
381
382		----------------------------------------------------------------------
383		Name: Matthias Blume
384		Date: 2000/06/19 13:40:00
385		Tag: blume-20000619-manual
386		Description:
387
388		CM manual and system/README update.  This only covers the fact that
389		there are no more implicit anchors.  (Path environments and the "bind"
390		option to "cm" have yet to be documented.)
391
392		----------------------------------------------------------------------
393		Name: Matthias Blume
394		Date: 2000/06/19 11:05:00 JST
395		Tag: blume-20000619-chdir-bugfix
396		Description:
397
398		Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
399
400		----------------------------------------------------------------------
401		Name: Matthias Blume
402		Date: 2000/06/18 22:00:10 JST
403		Tag: blume-20000618-implicit-anchors-really-gone
404		Description:
405
406		I updates the previous HISTORY entry where I forgot to mention that
407		implicit anchors are no longer with us.
408
409		The current update also gets rid of the (now useless) controller
410		CM.Control.implicit_anchors.
411
412		----------------------------------------------------------------------
413		Name: Matthias Blume
414		Date: 2000/06/16 17:30:00 JST
415		Tag: blume-20000616-anchorenv
416		Description:
417
418		This patch implements the long anticipated (just kidding :) "anchor
419		environment" mechanism.  In the course of doing this, I also
420		re-implemented CM's internal "SrcPath" module from scratch.  The new
421		one should be more robust in certain boundary cases.  In any case, it
422		is a lot cleaner than its predecessor (IMHO).
423
424		This time, although there is yet another boot file format change, I
425		kept the unpickler backward-compatible.  As a result, no new bootfiles
426		are necessary and bootstrapping is straightforward.  (You cannot read
427		new bootfiles into an old system, but the other way around is no
428		problem.)
429
430		Visible changes:
431
432		** 0. Implicit path anchors (without the leading $-symbol) are no
433		longer recognized at all. This means that such path names are not
434		illegal either.  For example, the name basis.cm simply refers to a
435		local file called "basis.cm" (i.e, the name is an ordinary path
436		relative to .cm-files directory).  Or, to put it differently, only
437		names that start with $ are anchored paths.
438
439		** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
440		vanished.
441
442		John (Reppy) had critizised this as soon as I originally proposed and
443		implemented it, but at that time I did not really deeply believe
444		him. :) Now I came full-circle because I need the $<singlearc> syntax
445		in another place where it cannot be seen as an abbreviation for
446		$/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
447		seems to mean (i.e., it "expands" into the corresponding anchor
448		value).
449
450		However, when paths are used as members in CM description files, it
451		continues to be true that there must be at least another arc after the
452		anchor.  This is now enforced separately during semantic analysis
453		(i.e., from a lexical/syntactical point of view, the notation is ok.)
454
455		** 2. The "cm" class now accepts an option "bind".  The option's value
456		is a sub-option list of precisely two items -- one labeled "anchor"
457		and the other one labeled "value".  As you might expect, "anchor" is
458		used to specify an anchor name to be bound, and "value" specifies what
459		the anchor is being bound to.
460
461		The value must be a directory name and can be given in either standard
462		syntax (including the possibility that it is itself an anchored path)
463		or native syntax.
464
465		Examples:
466
467		foo.cm (bind:(anchor:bar value:$mystuff/bar))
468		lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
469
470		and so on.
471
472		The meaning of this is that the .cm-file will be processed with an
473		augmented anchor environment where the given anchor(s) is/are bound to
474		the given values(s).
475
476		The rationale for having this feature is this: Suppose you are trying
477		to use two different (already stable) libraries a.cm and b.cm (that
478		you perhaps didn't write yourself).  Further, suppose each of these
479		two libraries internally uses its own auxiliary library $aux/lib.cm.
480		Normally you would now have a problem because the anchor "lib" can not
481		be bound to more than one value globally.  Therefore, the project that
482		uses both a.cm and b.cm must locally redirect the anchor to some other
483		place:
484
485		a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
486		b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
487
488		This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
489		/usr/lib/smlnj/b-stuff/aux.cm, respectively.
490
491		Hard-wiring path names is a bit inflexible (and CM will verbosely warn
492		you when you do so at the time of CM.stabilize).  Therefore, you can
493		also use an anchored path as the value:
494
495		a.cm (bind:(anchor:lib value:$a-lib))
496		b.cm (bind:(anchor:lib value:$b-lib))
497
498		Now you can globally configure (using the usual CM.Anchor.anchor or
499		pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
500		itself is always locally bound, setting it globally is no longer
501		meaningful or necessary (but it does not hurt either).  In fact, "lib"
502		can still be used as a global anchor for separate purposes.  As a
503		matter of fact, one can locally define "lib" in terms of a global
504		"lib":
505
506		a.cm (bind:(anchor:lib value:$lib/a))
507		b.cm (bind:(anchor:lib value:$lib/b))
508
509		** 3: The encoding of path names has changed.  This affects the way
510		path names are shown in CM's progress report and also the internal
511		protocol encoding used for parallel make.
512
513		The encoding now uses one or more ':'-separated segments.  Each
514		segments corresponds to a file that has been specified relative to the
515		file given by its preceding segment.  The first segment is either
516		relative to the CWD, absolute, or anchored.  Each segment itself is
517		basically a Unix pathname; all segments but the first are relative.
518
519		Example:
520
521		$foo/bar/baz.cm:a/b/c.sml
522
523		This path denotes the file bar/a/b/c.sml relative to the directory
524		denoted by anchor "foo".  Notice that the encoding also includes
525		baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
526		paths are resolved relative to the .cm-files directory, so baz.cm must
527		be ignored to get the "real" pathname.
528
529		To make this fact more obvious, CM puts the names of such "virtual
530		arcs" into parentheses when they appear in progress reports. (No
531		parentheses will appear in the internal protocol encoding.)  Thus,
532		what you really see is:
533
534		$foo/bar/(baz.cm):a/b/c.sml
535
536		I find this notation to be much more informative than before.
537
538		Another new feature of the encoding is that special characters
539		including parentheses, colons, (back)slashes, and white space are
540		written as \ddd (where ddd is the decimal encoding of the character).
541
542		*** The CM manual still needs to be updated.
543
544		----------------------------------------------------------------------
545		Name: Allen Leung
546		Date: 2000/06/15 00:38:00
547		Tag: leunga-20000615-x86-peephole
548
549		x86 Peephole fix by Fermin.  Affects c-- and moby only.
550
551		----------------------------------------------------------------------
552		Name: Matthias Blume
553		Date: 2000/06/12 11:40:00
554		Tag: blume-20000612-parmakefix
555		Description:
556
557		More cleanup after changing the file naming scheme: This time I
558		repaired the parallel make mechanism for CMB.make which I broke earlier.
559
560		----------------------------------------------------------------------
561		Name: Allen Leung
562		Date: 2000/06/09 01:25:00
563		Tag: leunga-20000609-various
564
565		None of these things should affect normal SML/NJ operations
566
567		1. Peephole improvements provided by Fermin (c--)
568		2. New annotation DEFUSE for adding extra dependence (moby)
569		3. New X86 LOCK instructions (moby)
570		4. New machine description language for reservation tables (scheduling)
571		5. Fixes to various optimization/analysis modules (branch chaining, dominator
572		trees etc.)
573		6. I've changed the CM files so that they can work with versions
574		110.0.6, 110.25 and 110.28
575
576		----------------------------------------------------------------------
577		Name: Matthias Blume
578		Date: 2000/06/09 12:40:00
579		Tag: blume-20000609-log
580		Description:
581
582		- Removed all(?) remaining RCS Log entries from sources.
583
584		- Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
585		anchored paths).
586
587		----------------------------------------------------------------------
588		Name: Matthias Blume
589		Date: 2000/06/07 17:00:00 JST
590		Tag: blume-20000607-no-implicit-anchors
591		Description:
592
593		1. This update changes the default setting for
594		CM.Control.implicit_anchors from true to false.  This means that
595		implicit anchors are no longer permitted by default.  I also tried to
596		make sure that nothing else still relies on implicit anchors.
597		(This is the next step on the schedule towards a CM that does not even
598		have the notion of implicit anchors anymore.)
599
600		2. More CM manual updates.
601
602		3. I managed to track down and fix the pickling bug I mentioned last
603		time.  Because of the previously existing workaround, this entails no
604		immediate practical changes.
605
606		----------------------------------------------------------------------
607		Name: Matthias Blume
608		Date: 2000/06/06 11:15:00 JST
609		Tag: blume-20000606-lazierpickle
610		Description:
611
612		!!!! NEW BOOT FILES !!!!
613
614		* The main purpose of this update is to make library pickles lazier in
615		order to reduce the initial space penalty for autoloading a library.
616		As a result, it is now possible to have $smlnj/compiler.cm
617		pre-registered.  This should take care of the many complaints or
618		inquiries about missing structure Compiler.  This required changes to
619		CM's internal data structures and small tweaks to some algorithms.
620
621		As a neat additional effect, it is no longer necessary (for the sake
622		of lean heap image files) to distinguish between a "minimal" CM and a
623		"full" CM.  Now, there is only one CM (i.e., the "full" version:
624		$smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
625		interactive top level. ($smlnj/cm/minimal.cm is gone.)
626
627		To make the life of compiler-hackers easier, "makeml" now also
628		pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
629		words, after you bootstrap a new sml for the first time, you will not
630		have to autoload $smlnj/cmb.cm again afterwards.  (The first time
631		around you will still have to do it, though.)
632
633		* A second change consists of major updates to the CM manual.  There
634		are now several appendices with summary information and also a full
635		specification of the CM description file syntax.
636
637		* In directory src/system I added the script "allcross".  This script
638		invokes sml and cross-compiles the compiler for all supported
639		architectures.  (Useful when providing a new set of boot files.)
640
641		* There seems to be a latent bug in my "lazy pickles" mechanism.  I
642		added a small tweak to pickle-util.sml to work around this problem,
643		but it is not a proper fix yet.  I will investigate further.  (The
644		effect of the bug was an inflation of library pickle size.)
645
646		* Version number increased to 110.28.1 (to avoid compatibility problems).
647
648		----------------------------------------------------------------------
649		Name: Allen Leung
650		Date: 2000/05/25 17:28 EDT
651		Tag: leunga-20000525-ra
652		Description:
653
654		Fixed a bug in freezing phase of the register allocator.
655
656		----------------------------------------------------------------------
657		Name: Allen Leung
658		Date: 2000/05/15 22:53 EDT
659		Tag: leunga-20000515-alpha-x86-ra
660		Description:
661
662		1. Alpha
663
664		Slight cleanup.  Removed the instruction SGNXL
665
666		2. X86
667
668		Added the following instructions to the instruction set:
669
670		ROLx, RORx,
671		BTx, BTSx, BTLx, BTRx,
672		XCHGx, and variants with the LOCK prefix
673
674		3. Register Allocation
675
676		The module ra-rewrite-with-renaming has been improved.
677
678		These have no effect on SML/NJ.
679
680		----------------------------------------------------------------------
681		Name: Matthias Blume
682		Date: 2000/05/15 16:20:00 JST
683		Tag: blume-20000515-lightrebuild
684		Description:
685
686		1. I added an alternative to "-rebuild" to "makeml".  The difference is
687		that prior to calling CMB.make' the CM-variable "LIGHT" will be
688		defined.  In effect, the command will not build any cross-compiler
689		backends and therefore finish more quickly.
690
691		The "fixpt" script also takes a "-light" switch to be able to use
692		this new facility while compiling for a fixpoint.
693
694		2. I replaced all mentions of anchored paths in group owner specifications
695		with simple relative paths (usually starting with "..").
696		The rationale is that a library's internal workings should not be
697		compromised by the lack of some anchor.  (An anchor is necessary
698		for someone who wants to refer to the library by an anchored path,
699		but it should not be necessary to build the same library in the first
700		place.)
701
702		3. I changed the way CM's tool mechanism determines the shell command
703		string used for things like ml-yacc etc. so that it does not break
704		when CM.Control.implicit_anchors is turned off.
705
706		----------------------------------------------------------------------
707		Name: Matthias Blume
708		Date: 2000/05/12 18:20:00 JST
709		Tag: blume-20000512-ml-build
710		Description:
711
712		Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
713		getting installed properly (by config/install.sh).
714
715		----------------------------------------------------------------------
716		Name: Matthias Blume
717		Date: 2000/05/12 17:30:00 JST
718		Tag: blume-20000512-anchors
719		Description:
720
721		!!! NEW BOOT FILES !!!
722
723		This change is in preparation of fading out support for "implicitly
724		anchored path names".  I went through all sources and used the
725		explicit (and relatively new) $-notation.  See system/README and the
726		CM manual for more info on this.
727
728		I also modified the anchoring scheme for some things such as "smlnj",
729		"MLRISC", "cm", etc. to take advantage of the fact that explicit
730		anchors are more expressive: anchor name and first arc do not have to
731		coincide.  This entails the following user-visible change:
732
733		You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
734		particular, when you fire up sml with a command-line argument, say,
735		e.g.:
736
737		sml '$smlnj/cmb.cm'
738
739		At the ML toplevel prompt:
740
741		CM.autoload "$smlnj/cmb.cm";
742
743		There is also a new controller in CM.Control that can be used to turn
744		off all remaining support for implicit anchors by saying:
745
746		CM.autoload "$smlnj/
747		#set CM.Control.implicit_anchors false;
748
749		This causes CM to reject implicitly anchored paths.  This is (for the
750		time being) less permissive than the "final" version where there will
751		be no more such implicit anchors and relative paths will be just that:
752		relative.
753
754		The next step (version after next version?) will be to make the
755		default for CM.Control.implicit_anchors false.  After the dust has
756		settled, I can then produce the "final" version of this...
757
758		Note: Since bootstrapping is a bit tricky, I provided new boot files.
759
760		----------------------------------------------------------------------
761		Name: Matthias Blume
762		Date: 2000/05/11 16:30:00 JST
763		Tag: blume-20000511-sources
764		Description:
765
766		The main change is that I added function CM.sources as a generalized
767		version of the earlier CM.makedepend.  This entails the following
768		additional changes:
769
770		- CM.makedepend has been dropped.
771
772		- CM manual has been updated.
773
774		- TOOLS signature and API have been changed.
775
776		----------------------------------------------------------------------
777	Name: Allen Leung	Name: Allen Leung
778	Date: 2000/05/10 21:17 EDT	Date: 2000/05/10 21:17 EDT
779	Tag: leunga-20000510-moby-c--ssa	Tag: leunga-20000510-moby-c--ssa
#	Line 374	Line 1136
1136
1137	3.  Assembly	3.  Assembly
1138
1139	When generating assemby, resolve the value of client defined constants,	When generating assembly, resolve the value of client defined constants,
1140	instead of generating symbolic values.  This is controlled by the	instead of generating symbolic values.  This is controlled by the
1141	new flag "asm-resolve-constants", which is default to true.	new flag "asm-resolve-constants", which is default to true.
1142
#	Line 397	Line 1159
1159
1160	To this end, I arranged that instead of "structure Core" as "structure	To this end, I arranged that instead of "structure Core" as "structure
1161	_Core" is bound in the pervasive environment.  Core access is done via	_Core" is bound in the pervasive environment.  Core access is done via
1162	_Core (which can never be accidentially rebound because _Core is not a	_Core (which can never be accidentally rebound because _Core is not a
1163	legal surface-syntax symbol).	legal surface-syntax symbol).
1164
1165	The current solution is much cleaner because the core environment is	The current solution is much cleaner because the core environment is
#	Line 407	Line 1169
1169	with dynamic and symbolic parts of the core environment.	with dynamic and symbolic parts of the core environment.
1170
1171	Remaining hackery (to bind the "magic" symbol _Core) is localized in the	Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1172	compilation mananger's bootstrap compiler (actually: in the "init group"	compilation manager's bootstrap compiler (actually: in the "init group"
1173	handling).  See the comments in src/system/smlnj/init/init.cmi for	handling).  See the comments in src/system/smlnj/init/init.cmi for
1174	more details.	more details.
1175
#	Line 522	Line 1284
1284	(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
1285	used).	used).
1286
1287	This was done to accomodate the new "make" and "shell" tools which	This was done to accommodate the new "make" and "shell" tools which
1288	facilitate fairly seemless hookup to portions of code managed using	facilitate fairly seamless hookup to portions of code managed using
1289	Makefiles or Shell scripts.	Makefiles or Shell scripts.
1290
1291	There are no classes "shared" or "private" anymore.  Instead, the	There are no classes "shared" or "private" anymore.  Instead, the
#	Line 535	Line 1297
1297
1298	All existing tools are described in the CM manual.	All existing tools are described in the CM manual.
1299
1300	- Slightly better error handling.  (CM now surpresses many followup	- Slightly better error handling.  (CM now suppresses many followup
1301	error messages that tended to be more annoying than helpful.)	error messages that tended to be more annoying than helpful.)
1302
1303	2. Major changes to the compiler's static environment data structures.	2. Major changes to the compiler's static environment data structures.
#	Line 669	Line 1431
1431
1432	I've changed andl to testl in the floating point test sequence	I've changed andl to testl in the floating point test sequence
1433	whenever appropriate.  The Intel optimization guide states that	whenever appropriate.  The Intel optimization guide states that
1434	testl is perferable to andl.	testl is preferable to andl.
1435
1436	7. RA (x86 only)	7. RA (x86 only)
1437
#	Line 851	Line 1613
1613
1614	1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an	1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1615	additional argument called "template" which is an optional string that	additional argument called "template" which is an optional string that
1616	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
1617	explanation.	explanation.
1618
1619	2. A special-purpose tool can be "regisitered" by simply dropping the	2. A special-purpose tool can be "registered" by simply dropping the
1620	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1621	directory where the .cm file lives that uses this tool.  (The	directory where the .cm file lives that uses this tool.  (The
1622	behavior/misfeature until now was to look for the tool description	behavior/misfeature until now was to look for the tool description
#	Line 898	Line 1660
1660	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
1661	<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,
1662	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
1663	perhaps got accidentially killed).  In this case, fixpt will quickly	perhaps got accidentally killed).  In this case, fixpt will quickly
1664	move through what exists before continuing where it left off earlier,	move through what exists before continuing where it left off earlier,
1665	and, thus, saves a lot of time.	and, thus, saves a lot of time.
1666
#	Line 948	Line 1710
1710	it from that remote directory.	it from that remote directory.
1711	This should simplify installation further:  For machines that have	This should simplify installation further:  For machines that have
1712	access to the internet, just fetch <version>-config.tgz, unpack it,	access to the internet, just fetch <version>-config.tgz, unpack it,
1713	edit config/targets, and go (run config/install.sh).  The scipt will	edit config/targets, and go (run config/install.sh).  The script will
1714	fetch everything else that it might need all by itself.	fetch everything else that it might need all by itself.
1715
1716	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.641   changed lines   Added in v.684

---

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