Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 594, Tue Apr 4 23:41:47 2000 UTC	revision 755, Thu Dec 14 07:57:55 2000 UTC
#	Line 13	Line 13
13	Description:	Description:
14	----------------------------------------------------------------------	----------------------------------------------------------------------
15	Name: Allen Leung	Name: Allen Leung
16		Date: Thu Dec 14 03:45:24 EST 2000
17		Description:
18		Tag:  leunga-20001214-int-inf
19		Description:
20
21		In IntInf, added these standard functions, which are missing from our
22		implementation:
23
24		andb : int * int -> int
25		xorb : int * int -> int
26		orb  : int * int -> int
27		notb : int -> int
28		<<   : int * word -> int
29		~>>  : int * word -> int
30
31		Not tested, I hope they are correct.
32
33		----------------------------------------------------------------------
34		Name: Allen Leung
35		Date: Fri Dec  8 19:23:26 EST 2000
36		Description:
37		Tag:  leunga-20001208-nowhere
38		Description:
39
40		Slight improvements to the 'nowhere' tool to handle OR-patterns,
41		to generate better error messages etc.  Plus a brief manual.
42
43		----------------------------------------------------------------------
44		Name: Lal George
45		Date: 2000/12/08 09:54:02 EST 2000
46		Tag: Release_110_31
47		Description:
48
49		- Version 110.31
50		----------------------------------------------------------------------
51		Name: Allen Leung
52		Date: Thu Dec  7 22:01:04 EST 2000
53		Tag:  leunga-20001207-cell-monster-hack
54		Description:
55
56		Major MLRISC internal changes.  Affect all clients.
57		Summary:
58
59		1.  Type CELLS.cell = int is now replaced by a datatype.
60		As a result, the old regmap is now gone.  Almost all interfaces
61		in MLRISC change as a consequence.
62
63		2.  A new brand version of machine description tool (v3.0) that generates
64		modules expecting the new interface.  The old version is removed.
65
66		3.  The RA interface has been further abstracted into two new functors.
67		RISC_RA and X86RA.  These functors have much simpler interfaces.
68		[See also directory MLRISC/demo.]
69
70		4.  Some other new source->source code generation tools are available:
71
72		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
73		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
74		I use this tool to generate the peephole optimizers---with the new
75		cell type changes, peephole rules are becoming difficult to write
76		without conditional pattern matching.
77
78		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
79		cover the entire MLRISC source tree so many things broke.
80
81		6.  CM files have been moved to the subdirectory MLRISC/cm.
82		They are moved because there are a lot of them and they clutter up the
83		root dir.
84
85		7.  More detailed documentation to come...
86
87		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
88		a makeml -rebuild first.  This is because of other other
89		changes that Matthias has made (see below).
90
91
92		----------------------------------------------------------------------
93		Name: Matthias Blume
94		Date: 2000/11/30 23:12:00 JST
95		Tag: blume-20001130-filereorg
96		Description:
97
98		Some manual updates and some file reorganizations in CM.
99
100		----------------------------------------------------------------------
101		Name: Matthias Blume
102		Date: 2000/11/24 17:45:00 JST
103		Tag: blume-20001124-link
104		Description:
105
106		Drastically improved link traversal code for the case that the dynamic
107		value was already loaded at bootstrap time.  As a result, CM and CMB
108		now both load blazingly fast -- even on a very slow machine.  Also,
109		memory consumption has been further reduced by this.
110
111		Warning: The format of the PIDMAP file has changed.  THerefore, to
112		bootstrap you have to do this:
113
114		1. Run CMB.make
115		2. Make a symbolic link for the boot directory:
116		ln -s sml.boot.ARCH-OS xxx
117		3. "Rebuild" the boot directory:
118		./makeml -boot xxx -rebuild sml ; rm xxx
119		4. Boot normally:
120		./makeml
121
122		----------------------------------------------------------------------
123		Name: Matthias Blume
124		Date: 2000/11/21 21:20:00 JST
125		Tag: blume-20001121-tools
126		Description:
127
128		Continued hacking on autoloading problem -- with success this time.
129		Also changed tool-plugin mechanism.  See new CM manual.
130
131		----------------------------------------------------------------------
132		Name: Matthias Blume
133		Date: 2000/11/19 14:30:00 JST
134		Tag:  blume-20001119-autoload
135		Description:
136
137		Some hacking to make autoloading faster.  Success for CMB, no success
138		so far for CM.  There is a reduced structure CM' that autoloads faster.
139		(This is a temporary, non-documented hack to be eliminated again when
140		the general problem is solved.)
141
142		----------------------------------------------------------------------
143		Name: Matthias Blume
144		Date: 2000/11/17 14:10:00 JST
145		Tag: blume-20001117-pickle-lib
146		Description:
147
148		1. Eliminated comp-lib.cm
149		2. Made pickle-lib.cm
150		3. Eliminated all uses of intset.sml (from comp-lib.cm)
151		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
152		equivalent constructs from smlnj-lib.cm (INtHashTable).
153		5. Point 4. also goes for those uses of intmap.* in MLRISC.
154		Duplicated intmap modules thrown out.
155		6. Hunted down all duplicated SCC code and replaced it with
156		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
157		7. Rewrote Feedback module.
158		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
159		should be thrown out and equivalent modules from smlnj-lib.cm
160		should be used (IntRedBlackSet, IntListSet, ...).
161
162		Confirmed that compiler compiles to fixpoint.
163
164		----------------------------------------------------------------------
165		Name: Allen Leung
166		Date: 2000/11/10 18:00:00
167		Tag: leunga-20001110-new-x86-fp
168
169		A new x86 floating point code generator has been added.
170		By default this is turned off.  To turn this on, do:
171
172		CM.autoload "$smlnj/compiler.cm";
173		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
174
175		Changes:
176
177		1.  Changed FTAN to FPTAN so that the assembly output is correct.
178		2.  Changed the extension callback for FTANGENT to generate:
179
180		fptan
181		fstp  %st(0)
182		instead of
183		fptan
184		fstpl ftempmem
185
186		3.  Numerous assembly fixes for x86.
187
188		5.  Cleaned up the machine code output module x86/x86MC.sml and added
189		support for a whole bunch of instructions and addressing modes:
190
191		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
192		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
193		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
194		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
195		fxch %st(n)
196		fld %st(n)
197		fst %st(n)
198		fst mem
199		fstp %st(n)
200		fucom %st(n)
201		fucomp %st(n)
202
203		All these are now generated when the fast fp mode is turned on.
204
205		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
206
207		----------------------------------------------------------------------
208		Name: Matthias Blume
209		Date: 2000/11/09 11:20:00 JST
210		Tag: blume-20001109-scc
211		Description:
212
213		Eliminated some code duplication:
214
215		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
216		(Otherwise the functor is useless.)
217		2. Used GraphSCCFn where SCCUtilFun was used previously.
218		3. Got rid of SCCUtilFun (in comp-lib.cm).
219
220		----------------------------------------------------------------------
221		Name: Lal George
222		Date: 2000/11/06 09:02:21 EST 2000
223		Tag: Release_110_30
224		Description:
225
226		- Version 110.30
227		----------------------------------------------------------------------
228		Name: Matthias Blume
229		Date: 2000/11/04 14:45:00
230		Tag: blume-20001104-mlbuild
231		Description:
232
233		- Made ml-build faster on startup.
234		- Documentation fixes.
235
236		----------------------------------------------------------------------
237		Name: Matthias Blume
238		Date: 2000/11/02 17:00:00 JST
239		Tag: blume-20001102-condcomp
240		Description:
241
242		- Small tweaks to pickler -- new BOOTFILES!
243		- Version bumped to 110.29.2.
244		- Added conditional compilation facility to init.cmi (see comment there).
245		----------------------------------------------------------------------
246		Name: Allen Leung
247		Date: 2000/10/23 19:31:00
248		Tag: leunga-20001023-demo-ra
249
250		1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
251		2. Test programs for the graph library updated
252		3. Some new MLRISC demo programs added
253
254		----------------------------------------------------------------------
255		Name: Matthias Blume
256		Date: 2000/08/31 22:15:00 JST
257		Tag: blume-20001017-errmsg
258		Description:
259
260		More error message grief: Where there used to be no messages, there
261		now were some that had bogus error regions.  Fixed.
262
263		----------------------------------------------------------------------
264		Name: Matthias Blume
265		Date: 2000/08/31 17:30:00 JST
266		Tag: blume-20001017-v110p29p1
267		Description:
268
269		I made a version 110.29.1 with new bootfiles.
270
271		Changes:  Modified pickler/unpickler for faster and leaner unpickling.
272		CM documentation changes and a small bugfix in CM's error reporting.
273
274		----------------------------------------------------------------------
275		Name: Lal George
276		Date: 2000/09/27 14:42:35 EDT
277		Tag: george-20000927-nodestatus
278		Description:
279
280		Changed the type of the nodestatus, so that:
281
282		SPILLED(~1)             is now SPILLED
283		SPILLED(m) where m>=0   is now MEMREG(m)
284		SPILLED(s) where s<~1   is now SPILL_LOC(~s)
285
286		----------------------------------------------------------------------
287		Name: Matthias Blume
288		Date: 2000/09/07 14:45:00 JST
289		Tag: blume-20000907-cmerrmsg
290		Description:
291
292		Small tweak to CM to avoid getting ML syntax error messages twice.
293
294		----------------------------------------------------------------------
295		Name: Matthias Blume
296		Date: 2000/08/31 18:00:00 JST
297		Tag: blume-20000831-cvsbootfiles
298		Description:
299
300		New URL for boot files (because the 110.29 files on the BL server do
301		now work correctly with my updated install scripts for yacc and lex).
302
303		----------------------------------------------------------------------
304		Name: Matthias Blume
305		Date: 2000/08/08 12:33:00 JST
306		Tag: blume-20000808-manual
307		Description:
308
309		Tiny update to CM manual.
310
311		----------------------------------------------------------------------
312		Name: Allen Leung
313		Date: 2000/08/7 19:31:00
314		Tag: leunga-20000807-a-whole-bunch-of-stuff
315
316		Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
317		and Mobdy.
318
319		1.  x86
320
321		a.  Fixes to peephole module by John and Dan.
322		b.  Assembly fix to SETcc by Allen.
323		c.  Fix to c-call by John.
324		d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
325		e.  Instruction selection optimization to SETcc as suggested by John.
326
327		For example,
328
329		MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
330
331		should generate:
332
333		MOVL a, x
334		SUBL b, x
335		SHRL 31, x
336
337		2.  IR stuff
338
339		A bunch of new DJ-graph related algorithms added.  These
340		speed up SSA construction.
341
342		3.  SSA + Scheduling
343
344		Added code for SSA and scheduling to the repository
345
346		----------------------------------------------------------------------
347		Name: Lal George
348		Date: 2000/07/27 11:53:14 EDT
349
350		Tag: lal-20000727-linux-ppc
351		Description:
352
353		Made changes to support Linux PPC.
354		p.s. I have confirmation that the 110.29 boot files work fine.
355
356		----------------------------------------------------------------------
357		Name: Matthias Blume
358		Date: 2000/07/27 17:40:00 JST
359		Tag: blume-20000727-scripts
360		Description:
361
362		!!!! WARNING !!!!
363		You must recompile the runtime system!
364		!!!! WARNING !!!!
365
366		This is basically another round of script-enhancements:
367
368		1. sml, ml-build, and ml-makedepend accept options -D and -U to define
369		and undefine CM preprocessor symbols.
370
371		2. ml-build avoids generating a new heap image if it finds that the
372		existing one is still ok.  (The condition is that no ML file had to
373		be recompiled and all ML files are found to be older that the heap
374		file.)
375
376		To make this work smoothly, I also hacked the runtime system as
377		well as SMLofNJ.SysInfo to get access to the heap image suffix
378		(.sparc-solaris, ...) that is currently being used.
379
380		Moreover, the signature of CM.mk_standalone has changed.  See the
381		CM manual.
382
383		3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
384		CM manual for details.)
385
386		4. More CM manual updates:
387		- all of the above has been documented.
388		- there is now a section describing the (CM-related) command line
389		arguments that are accepted by the "sml" command
390
391		----------------------------------------------------------------------
392		Name: Matthias Blume
393		Date: 2000/07/25 16:20:00 JST
394		Tag: blume-20000725-makedepend
395		Description:
396
397		Added a script called ml-makedepend.  This can be used in makefiles
398		for Unix' make in a way very similar to the "makedepend" command for
399		C.
400
401		The script internally uses function CM.sources.
402
403		Synopsis:
404
405		ml-makedepend [-f makefile] cmfile targetname
406
407		The default for the makefile is "makefile" (or "Makefile" should
408		"makefile" not exist).
409
410		ml-makedepend adds a cmfile/targetname-specific section to this
411		makefile (after removing the previous version of this section).  The
412		section contains a single dependency specification with targetname on
413		the LHS (targetname is an arbitrary name), and a list of files derived
414		from the cmfile on the RHS.  Some of the files on the RHS are
415		ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
416		"make" variables $(ARCH) and $(OPSYS) in place of the corresponding
417		path names.  The makefile writer is responsible for making sure that
418		these variables have correct at the time "make" is invoked.
419
420		----------------------------------------------------------------------
421		Name: Matthias Blume
422		Date: 2000/07/22 23:30:00 JST
423		Tag: blume-20000722-urlupdate
424		Description:
425
426		Changed BOOT and config/srcarchiveurl to point to BL server:
427
428		ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
429
430		----------------------------------------------------------------------
431		Name: Matthias Blume
432		Date: 2000/07/18 18:00:00 JST
433		Tag: blume-20000718-Version_110_29
434		Description:
435
436		1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
437
438		2. Updated config/version to 110.29
439
440		3. Updated config/srcarchiveurl
441
442		3. New boot files!
443		ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
444
445		----------------------------------------------------------------------
446		Name: Matthias Blume
447		Date: 2000/07/11 13:58:00 JST
448		Tag: blume-20000711-doctypo
449		Description:
450
451		Fixed a few typos in CM manual.
452
453		----------------------------------------------------------------------
454		Name: Allen Leung
455		Date: 2000/06/15 00:38:00
456		Tag: leunga-20000704-sparc-x86
457
458		1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
459		2. fix to x86 RET bug [found by Dan Grossman]
460		3. sparc assembly bug fix for ticc instructions [found by Fermin]
461
462		Affects c-- and moby only
463
464		----------------------------------------------------------------------
465		Name: Matthias Blume
466		Date: 2000/07/04 15:26:00
467		Tag: blume-20000704-trigger
468		Description:
469
470		1. Improvements to CM manual.
471		2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
472		of getting a back-trace.  The function, when called, raises an
473		internal exception which explicitly carries the full back-trace history,
474		so it is unaffected by any intervening handle-raise pairs ("trivial"
475		or not).  The interactive loop will print that history once it arrives
476		at top level.
477		Short of having all exceptions implicitly carry the full history, the
478		recommended way of using this facility is:
479		- compile your program with instrumentation "on"
480		- run it, when it raises an exception, look at the history
481		- if the history is "cut off" because of some handler, go and modify
482		your program so that it explicitly calls BTrace.trigger
483		- recompile (still instrumented), and rerun; look at the full history
484
485		----------------------------------------------------------------------
486		Name: Matthias Blume
487		Date: 2000/07/03 15:36:00 JST
488		Tag: blume-20000702-manual
489		Description:
490
491		Small corrections and updates to CM manual.
492
493		----------------------------------------------------------------------
494		Name: Matthias Blume
495		Date: 2000/06/29 16:04:00 JST
496		Tag: blume-20000629-yacctool
497		Description:
498
499		Changes:
500
501		1. Class "mlyacc" now takes separate arguments to pass options to
502		generated .sml- and .sig-files independently.
503		2. Corresponding CM manual updates.
504		3. BTrace module now also reports call sites.  (However, for loop clusters
505		it only shows from where the cluster was entered.)  There are associated
506		modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
507
508		----------------------------------------------------------------------
509		Name: Matthias Blume
510		Date: 2000/06/27 16:51:00 JST
511		Tag: blume-20000627-noweb
512		Description:
513
514		Changes:
515
516		1. Implemented "subdir" and "witness" options for noweb tool.
517		This caused some slight internal changes in CM's tool implementation.
518		2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
519		some remaining issues from earlier path anchor changes.
520		3. Updated CM manual accordingly.
521
522		4. Changed implementation of back-tracing so that I now consider it
523		ready for prime-time.
524
525		In particular, you don't have to explicitly trigger the back-trace
526		anymore.  Instead, if you are running BTrace-instrumented code and
527		there is an uncaught exception (regardless of whether or not it was
528		raised in instrumented code), the top-level evalloop will print
529		the back-trace.
530
531		Features:
532
533		- Instrumented and uninstrumented code work together seemlessly.
534		(Of course, uninstrumented code is never mentioned in actual
535		back-traces.)
536
537		- Asymptotic time- and space-complexity of instrumented code is
538		equal to that of uninstrumented code.  (This means that
539		tail-recursion is preserved by the instrumentation phase.)
540
541		- Modules whose code has been instrumented in different sessions
542		work together without problem.
543
544		- There is no penalty whatsoever on uninstrumented code.
545
546		- There is no penalty on "raise" expressions, even in
547		instrumented code.
548
549		A potential bug (or perhaps it is a feature, too):
550
551		A back-trace reaches no further than the outermost instrumented
552		non-trivial "raise".  Here, a "trivial" raise is one that is the
553		sole RHS of a "handle" rule.  Thus, back-traces reach trough
554
555		<exp> handle e => raise e
556
557		and even
558
559		<exp> handle Foo => raise Bar
560
561		and, of course, through
562
563		<exp> handle Foo => ...
564
565		if the exception was not Foo.
566
567		Back-traces always reach right through any un-instrumented code
568		including any of its "handle" expressions, trivial or not.
569
570		To try this out, do the following:
571
572		- Erase all existing binfiles for your program.
573		(You may keep binfiles for those modules where you think you
574		definitely don't need back-tracing.)
575		- Turn on back-trace instrumentation:
576		SMLofNJ.Internals.BTrace.mode (SOME true);
577		- Recompile your program.  (I.e., run "CM.make" or "use".)
578		- You may now turn instrumentation off again (if you want):
579		SMLofNJ.Internals.BTrace.mode (SOME false);
580		- Run your program as usual.  If it raises an exception that
581		reaches the interactive toplevel, then a back-trace will
582		automatically be printed.  After that, the toplevel loop
583		will print the exception history as usual.
584
585		----------------------------------------------------------------------
586		Name: Matthias Blume
587		Date: 2000/06/26 09:56:46 JST
588		Tag: blume-20000626-setup
589		Description:
590
591		CM: - setup-parameter to "sml" added; this can be used to run arbitrary
592		ML code before and after compiling a file (e.g., to set compiler
593		flags)
594
595		Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
596		- associated changes to btrace.sml (BTrace instrumentation pass)
597		- cleaner implementation of btimp.sml (BTrace tracing and report
598		module)
599
600		CM manual: * new path encoding documented
601		* description of setup-parameter to "sml" added
602
603		The biggest user-visible change to back-tracing is that it is no
604		longer necessary to compile all traced modules within the same
605		session.  (This was a real limitation.)
606
607		----------------------------------------------------------------------
608		Name: Matthias Blume
609		Date: 2000/06/24 12:40:00 JST
610		Tag: blume-20000624-startup
611		Description:
612
613		Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
614		bit too often -- to put it mildly. :)
615
616		----------------------------------------------------------------------
617		Name: Matthias Blume
618		Date: 2000/06/23 18:20:00 JST
619		Tag: blume-20000623-btrace
620		Description:
621
622		This updates adds a backtrace facility to aid programmers in debugging
623		their programs.  This involves the following changes:
624
625		1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
626		keeping track of the current call stack.  When programs are compiled
627		in a special mode, the compiler will insert calls to these hooks
628		into the user program.
629		"Hook" means that it is possible for different implementations of
630		back-tracing to register themselves (at different times).
631
632		2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
633		as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
634		It can be turned on using this call:
635		SMLofNJ.Internals.BTrace.mode (SOME true);
636		Turning it off again:
637		SMLofNJ.Internals.BTrace.mode (SOME false);
638		Querying the current status:
639		SMLofNJ.Internals.BTrace.mode NONE;
640		Annotated programs are about twice as big as normal ones, and they
641		run a factor of 2 to 4 slower with a dummy back-trace plugin (one
642		where all hooks do nothing).  The slowdown with a plugin that is
643		actually useful (such as the one supplied by default) is even greater,
644		but in the case of the default plugin it is still only an constant
645		factor (amortized).
646
647		3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
648		with a sub-structure BTrace for controlling back-tracing.  In particular,
649		the above-mentioned function "mode" controls whether the annotation
650		phase is invoked by the compiler.  Another important function is
651		"trigger": when called it aborts the current execution and causes
652		the top-level loop to print a full back-trace.
653
654		4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
655		for back-tracing.  It keeps track of the dynamic call stack and in
656		addition to that it keeps a partial history at each "level" of that
657		stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
658		d and b (at separate times, dynamically), then the report will show:
659
660		GOTO   d
661		/c
662		GOTO  \b
663		CALL   a
664
665		This shows that there was an initial non-tail call of a, then a
666		tail-call to b or c, looping behavior in a cluster of functions that
667		consist of b and c, and then a goto from that cluster (i.e., either from
668		b or from c) to d.
669
670		Note that (depending on the user program) the amount of information
671		that the back-trace module has to keep track of at each level is bounded
672		by a constant.  Thus, the whole implementation has the same asymptotical
673		complexity as the original program (both in space and in time).
674
675		5. compiler/TopLevel/interact/evalloop.sml has been modified to
676		handle the special exception SMLofNJ.Internals.BTrace.BTrace
677		which is raised by the "trigger" function mentioned above.
678
679		Notes on usage:
680
681		- Annotated code works well together with unannotated code:
682		Unannotated calls simply do not show up at all in the backtrace.
683
684		- It is not a good idea to let modules that were annotated during
685		different sessions run at the same time.  This is because the compiler
686		chooses small integers to identify individual functions, and there
687		will be clashes if different modules were compiled in separate sessions.
688		(Nothing will crash, and you will even be told about the clashes, but
689		back-trace information will in general not be useful.)
690
691		- Back-tracing can be confused by callcc and capture.
692
693		- The only way of getting a back-trace right now is to explicitly
694		invoke the "trigger" function from your user program.  Eventually, we
695		should make every exception carry back-trace information (if
696		available).  But since this creates more overhead at "raise"-time
697		(similar to the current exnHistory overhead), I have not yet
698		implemented this.  (The implementation will be rather easy.)  With
699		exceptions carrying back-trace information, this facility will be even
700		more useful because users don't need to modify their programs...
701
702		- While it is possible to compile the compiler with back-trace
703		annotations turned on (I did it to get some confidence in
704		correctness), you must make absolutely sure that core.sml and
705		btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
706		be compiled with annotation because there is no core access yet, but
707		if you compile btimp.sml with annotation, then the system will go into
708		an infinite recursion and crash.)
709		Since CM currently does not know about BTrace, the only way to turn
710		annotations on and off for different modules of the compiler is to
711		interrupt CMB.make, change the settings, and re-invoke it.  Of course,
712		this is awkward and clumsy.
713
714		Sample sessions:
715
716		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
717		- SMLofNJ.Internals.BTrace.mode (SOME true);
718		[autoloading]
719		[autoloading done]
720		val it = false : bool
721		- structure X = struct
722		-     fun main n = let
723		-         fun a (x, 0) = d x
724		-           | a (x, n) = b (x, n - 1)
725		-         and b (x, n) = c (x, n)
726		-         and c (x, n) = a (x, n)
727		-         and d x = e (x, 3)
728		-         and e (x, 0) = f x
729		-           | e (x, n) = e (x, n - 1)
730		-         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
731		-           | f n = n * g (n - 1)
732		-         and g n = a (n, 3)
733		-     in
734		-         f n
735		-     end
736		- end;
737		structure X : sig val main : int -> int end
738		- X.main 3;
739		*** BACK-TRACE ***
740		GOTO   stdIn:4.2-13.20: X.main[2].f
741		GOTO-( stdIn:4.2-13.20: X.main[2].e
742		GOTO   stdIn:4.2-13.20: X.main[2].d
743		/ stdIn:4.2-13.20: X.main[2].a
744		| stdIn:4.2-13.20: X.main[2].b
745		GOTO-\ stdIn:4.2-13.20: X.main[2].c
746		CALL   stdIn:4.2-13.20: X.main[2].g
747		GOTO   stdIn:4.2-13.20: X.main[2].f
748		GOTO-( stdIn:4.2-13.20: X.main[2].e
749		GOTO   stdIn:4.2-13.20: X.main[2].d
750		/ stdIn:4.2-13.20: X.main[2].a
751		| stdIn:4.2-13.20: X.main[2].b
752		GOTO-\ stdIn:4.2-13.20: X.main[2].c
753		CALL   stdIn:4.2-13.20: X.main[2].g
754		GOTO   stdIn:4.2-13.20: X.main[2].f
755		GOTO-( stdIn:4.2-13.20: X.main[2].e
756		GOTO   stdIn:4.2-13.20: X.main[2].d
757		/ stdIn:4.2-13.20: X.main[2].a
758		| stdIn:4.2-13.20: X.main[2].b
759		GOTO-\ stdIn:4.2-13.20: X.main[2].c
760		CALL   stdIn:4.2-13.20: X.main[2].g
761		GOTO   stdIn:4.2-13.20: X.main[2].f
762		CALL   stdIn:2.15-17.4: X.main[2]
763		-
764
765		(Note that because of a FLINt bug the above code currently does not
766		compile without BTrace turned on.)
767
768		Here is another example, using my modified Tiger compiler:
769
770		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
771		- SMLofNJ.Internals.BTrace.mode (SOME true);
772		[autoloading]
773		[autoloading done]
774		val it = false : bool
775		- CM.make "sources.cm";
776		[autoloading]
777		...
778		[autoloading done]
779		[scanning sources.cm]
780		[parsing (sources.cm):parse.sml]
781		[creating directory CM/SKEL ...]
782		[parsing (sources.cm):tiger.lex.sml]
783		...
784		[wrote CM/sparc-unix/semant.sml]
785		[compiling (sources.cm):main.sml]
786		[wrote CM/sparc-unix/main.sml]
787		[New bindings added.]
788		val it = true : bool
789		- Main.compile ("../testcases/merge.tig", "foo.out");
790		*** BACK-TRACE ***
791		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
792		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
793		CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
794		GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
795		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
796		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
797		CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
798		/ lib/semant.sml:411.65-543.8: SemantFun[2].transDec
799		CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
800		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
801		CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
802		CALL   main.sml:1.18-118.4: Main.compile[2]
803		-
804
805		----------------------------------------------------------------------
806		Name: Matthias Blumen
807		Date: 2000/06/21 18:00:00 JST
808		Tag: blume-20000621-manual
809		Description:
810
811		CM manual update: Path environments documented.
812
813		----------------------------------------------------------------------
814		Name: Matthias Blume
815		Date: 2000/06/19 13:40:00
816		Tag: blume-20000619-manual
817		Description:
818
819		CM manual and system/README update.  This only covers the fact that
820		there are no more implicit anchors.  (Path environments and the "bind"
821		option to "cm" have yet to be documented.)
822
823		----------------------------------------------------------------------
824		Name: Matthias Blume
825		Date: 2000/06/19 11:05:00 JST
826		Tag: blume-20000619-chdir-bugfix
827		Description:
828
829		Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
830
831		----------------------------------------------------------------------
832		Name: Matthias Blume
833		Date: 2000/06/18 22:00:10 JST
834		Tag: blume-20000618-implicit-anchors-really-gone
835		Description:
836
837		I updates the previous HISTORY entry where I forgot to mention that
838		implicit anchors are no longer with us.
839
840		The current update also gets rid of the (now useless) controller
841		CM.Control.implicit_anchors.
842
843		----------------------------------------------------------------------
844		Name: Matthias Blume
845		Date: 2000/06/16 17:30:00 JST
846		Tag: blume-20000616-anchorenv
847		Description:
848
849		This patch implements the long anticipated (just kidding :) "anchor
850		environment" mechanism.  In the course of doing this, I also
851		re-implemented CM's internal "SrcPath" module from scratch.  The new
852		one should be more robust in certain boundary cases.  In any case, it
853		is a lot cleaner than its predecessor (IMHO).
854
855		This time, although there is yet another boot file format change, I
856		kept the unpickler backward-compatible.  As a result, no new bootfiles
857		are necessary and bootstrapping is straightforward.  (You cannot read
858		new bootfiles into an old system, but the other way around is no
859		problem.)
860
861		Visible changes:
862
863		** 0. Implicit path anchors (without the leading $-symbol) are no
864		longer recognized at all. This means that such path names are not
865		illegal either.  For example, the name basis.cm simply refers to a
866		local file called "basis.cm" (i.e, the name is an ordinary path
867		relative to .cm-files directory).  Or, to put it differently, only
868		names that start with $ are anchored paths.
869
870		** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
871		vanished.
872
873		John (Reppy) had critizised this as soon as I originally proposed and
874		implemented it, but at that time I did not really deeply believe
875		him. :) Now I came full-circle because I need the $<singlearc> syntax
876		in another place where it cannot be seen as an abbreviation for
877		$/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
878		seems to mean (i.e., it "expands" into the corresponding anchor
879		value).
880
881		However, when paths are used as members in CM description files, it
882		continues to be true that there must be at least another arc after the
883		anchor.  This is now enforced separately during semantic analysis
884		(i.e., from a lexical/syntactical point of view, the notation is ok.)
885
886		** 2. The "cm" class now accepts an option "bind".  The option's value
887		is a sub-option list of precisely two items -- one labeled "anchor"
888		and the other one labeled "value".  As you might expect, "anchor" is
889		used to specify an anchor name to be bound, and "value" specifies what
890		the anchor is being bound to.
891
892		The value must be a directory name and can be given in either standard
893		syntax (including the possibility that it is itself an anchored path)
894		or native syntax.
895
896		Examples:
897
898		foo.cm (bind:(anchor:bar value:$mystuff/bar))
899		lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
900
901		and so on.
902
903		The meaning of this is that the .cm-file will be processed with an
904		augmented anchor environment where the given anchor(s) is/are bound to
905		the given values(s).
906
907		The rationale for having this feature is this: Suppose you are trying
908		to use two different (already stable) libraries a.cm and b.cm (that
909		you perhaps didn't write yourself).  Further, suppose each of these
910		two libraries internally uses its own auxiliary library $aux/lib.cm.
911		Normally you would now have a problem because the anchor "lib" can not
912		be bound to more than one value globally.  Therefore, the project that
913		uses both a.cm and b.cm must locally redirect the anchor to some other
914		place:
915
916		a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
917		b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
918
919		This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
920		/usr/lib/smlnj/b-stuff/aux.cm, respectively.
921
922		Hard-wiring path names is a bit inflexible (and CM will verbosely warn
923		you when you do so at the time of CM.stabilize).  Therefore, you can
924		also use an anchored path as the value:
925
926		a.cm (bind:(anchor:lib value:$a-lib))
927		b.cm (bind:(anchor:lib value:$b-lib))
928
929		Now you can globally configure (using the usual CM.Anchor.anchor or
930		pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
931		itself is always locally bound, setting it globally is no longer
932		meaningful or necessary (but it does not hurt either).  In fact, "lib"
933		can still be used as a global anchor for separate purposes.  As a
934		matter of fact, one can locally define "lib" in terms of a global
935		"lib":
936
937		a.cm (bind:(anchor:lib value:$lib/a))
938		b.cm (bind:(anchor:lib value:$lib/b))
939
940		** 3: The encoding of path names has changed.  This affects the way
941		path names are shown in CM's progress report and also the internal
942		protocol encoding used for parallel make.
943
944		The encoding now uses one or more ':'-separated segments.  Each
945		segments corresponds to a file that has been specified relative to the
946		file given by its preceding segment.  The first segment is either
947		relative to the CWD, absolute, or anchored.  Each segment itself is
948		basically a Unix pathname; all segments but the first are relative.
949
950		Example:
951
952		$foo/bar/baz.cm:a/b/c.sml
953
954		This path denotes the file bar/a/b/c.sml relative to the directory
955		denoted by anchor "foo".  Notice that the encoding also includes
956		baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
957		paths are resolved relative to the .cm-files directory, so baz.cm must
958		be ignored to get the "real" pathname.
959
960		To make this fact more obvious, CM puts the names of such "virtual
961		arcs" into parentheses when they appear in progress reports. (No
962		parentheses will appear in the internal protocol encoding.)  Thus,
963		what you really see is:
964
965		$foo/bar/(baz.cm):a/b/c.sml
966
967		I find this notation to be much more informative than before.
968
969		Another new feature of the encoding is that special characters
970		including parentheses, colons, (back)slashes, and white space are
971		written as \ddd (where ddd is the decimal encoding of the character).
972
973		*** The CM manual still needs to be updated.
974
975		----------------------------------------------------------------------
976		Name: Allen Leung
977		Date: 2000/06/15 00:38:00
978		Tag: leunga-20000615-x86-peephole
979
980		x86 Peephole fix by Fermin.  Affects c-- and moby only.
981
982		----------------------------------------------------------------------
983		Name: Matthias Blume
984		Date: 2000/06/12 11:40:00
985		Tag: blume-20000612-parmakefix
986		Description:
987
988		More cleanup after changing the file naming scheme: This time I
989		repaired the parallel make mechanism for CMB.make which I broke earlier.
990
991		----------------------------------------------------------------------
992		Name: Allen Leung
993		Date: 2000/06/09 01:25:00
994		Tag: leunga-20000609-various
995
996		None of these things should affect normal SML/NJ operations
997
998		1. Peephole improvements provided by Fermin (c--)
999		2. New annotation DEFUSE for adding extra dependence (moby)
1000		3. New X86 LOCK instructions (moby)
1001		4. New machine description language for reservation tables (scheduling)
1002		5. Fixes to various optimization/analysis modules (branch chaining, dominator
1003		trees etc.)
1004		6. I've changed the CM files so that they can work with versions
1005		110.0.6, 110.25 and 110.28
1006
1007		----------------------------------------------------------------------
1008		Name: Matthias Blume
1009		Date: 2000/06/09 12:40:00
1010		Tag: blume-20000609-log
1011		Description:
1012
1013		- Removed all(?) remaining RCS Log entries from sources.
1014
1015		- Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1016		anchored paths).
1017
1018		----------------------------------------------------------------------
1019		Name: Matthias Blume
1020		Date: 2000/06/07 17:00:00 JST
1021		Tag: blume-20000607-no-implicit-anchors
1022		Description:
1023
1024		1. This update changes the default setting for
1025		CM.Control.implicit_anchors from true to false.  This means that
1026		implicit anchors are no longer permitted by default.  I also tried to
1027		make sure that nothing else still relies on implicit anchors.
1028		(This is the next step on the schedule towards a CM that does not even
1029		have the notion of implicit anchors anymore.)
1030
1031		2. More CM manual updates.
1032
1033		3. I managed to track down and fix the pickling bug I mentioned last
1034		time.  Because of the previously existing workaround, this entails no
1035		immediate practical changes.
1036
1037		----------------------------------------------------------------------
1038		Name: Matthias Blume
1039		Date: 2000/06/06 11:15:00 JST
1040		Tag: blume-20000606-lazierpickle
1041		Description:
1042
1043		!!!! NEW BOOT FILES !!!!
1044
1045		* The main purpose of this update is to make library pickles lazier in
1046		order to reduce the initial space penalty for autoloading a library.
1047		As a result, it is now possible to have $smlnj/compiler.cm
1048		pre-registered.  This should take care of the many complaints or
1049		inquiries about missing structure Compiler.  This required changes to
1050		CM's internal data structures and small tweaks to some algorithms.
1051
1052		As a neat additional effect, it is no longer necessary (for the sake
1053		of lean heap image files) to distinguish between a "minimal" CM and a
1054		"full" CM.  Now, there is only one CM (i.e., the "full" version:
1055		$smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1056		interactive top level. ($smlnj/cm/minimal.cm is gone.)
1057
1058		To make the life of compiler-hackers easier, "makeml" now also
1059		pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1060		words, after you bootstrap a new sml for the first time, you will not
1061		have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1062		around you will still have to do it, though.)
1063
1064		* A second change consists of major updates to the CM manual.  There
1065		are now several appendices with summary information and also a full
1066		specification of the CM description file syntax.
1067
1068		* In directory src/system I added the script "allcross".  This script
1069		invokes sml and cross-compiles the compiler for all supported
1070		architectures.  (Useful when providing a new set of boot files.)
1071
1072		* There seems to be a latent bug in my "lazy pickles" mechanism.  I
1073		added a small tweak to pickle-util.sml to work around this problem,
1074		but it is not a proper fix yet.  I will investigate further.  (The
1075		effect of the bug was an inflation of library pickle size.)
1076
1077		* Version number increased to 110.28.1 (to avoid compatibility problems).
1078
1079		----------------------------------------------------------------------
1080		Name: Allen Leung
1081		Date: 2000/05/25 17:28 EDT
1082		Tag: leunga-20000525-ra
1083		Description:
1084
1085		Fixed a bug in freezing phase of the register allocator.
1086
1087		----------------------------------------------------------------------
1088		Name: Allen Leung
1089		Date: 2000/05/15 22:53 EDT
1090		Tag: leunga-20000515-alpha-x86-ra
1091		Description:
1092
1093		1. Alpha
1094
1095		Slight cleanup.  Removed the instruction SGNXL
1096
1097		2. X86
1098
1099		Added the following instructions to the instruction set:
1100
1101		ROLx, RORx,
1102		BTx, BTSx, BTLx, BTRx,
1103		XCHGx, and variants with the LOCK prefix
1104
1105		3. Register Allocation
1106
1107		The module ra-rewrite-with-renaming has been improved.
1108
1109		These have no effect on SML/NJ.
1110
1111		----------------------------------------------------------------------
1112		Name: Matthias Blume
1113		Date: 2000/05/15 16:20:00 JST
1114		Tag: blume-20000515-lightrebuild
1115		Description:
1116
1117		1. I added an alternative to "-rebuild" to "makeml".  The difference is
1118		that prior to calling CMB.make' the CM-variable "LIGHT" will be
1119		defined.  In effect, the command will not build any cross-compiler
1120		backends and therefore finish more quickly.
1121
1122		The "fixpt" script also takes a "-light" switch to be able to use
1123		this new facility while compiling for a fixpoint.
1124
1125		2. I replaced all mentions of anchored paths in group owner specifications
1126		with simple relative paths (usually starting with "..").
1127		The rationale is that a library's internal workings should not be
1128		compromised by the lack of some anchor.  (An anchor is necessary
1129		for someone who wants to refer to the library by an anchored path,
1130		but it should not be necessary to build the same library in the first
1131		place.)
1132
1133		3. I changed the way CM's tool mechanism determines the shell command
1134		string used for things like ml-yacc etc. so that it does not break
1135		when CM.Control.implicit_anchors is turned off.
1136
1137		----------------------------------------------------------------------
1138		Name: Matthias Blume
1139		Date: 2000/05/12 18:20:00 JST
1140		Tag: blume-20000512-ml-build
1141		Description:
1142
1143		Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1144		getting installed properly (by config/install.sh).
1145
1146		----------------------------------------------------------------------
1147		Name: Matthias Blume
1148		Date: 2000/05/12 17:30:00 JST
1149		Tag: blume-20000512-anchors
1150		Description:
1151
1152		!!! NEW BOOT FILES !!!
1153
1154		This change is in preparation of fading out support for "implicitly
1155		anchored path names".  I went through all sources and used the
1156		explicit (and relatively new) $-notation.  See system/README and the
1157		CM manual for more info on this.
1158
1159		I also modified the anchoring scheme for some things such as "smlnj",
1160		"MLRISC", "cm", etc. to take advantage of the fact that explicit
1161		anchors are more expressive: anchor name and first arc do not have to
1162		coincide.  This entails the following user-visible change:
1163
1164		You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1165		particular, when you fire up sml with a command-line argument, say,
1166		e.g.:
1167
1168		sml '$smlnj/cmb.cm'
1169
1170		At the ML toplevel prompt:
1171
1172		CM.autoload "$smlnj/cmb.cm";
1173
1174		There is also a new controller in CM.Control that can be used to turn
1175		off all remaining support for implicit anchors by saying:
1176
1177		CM.autoload "$smlnj/
1178		#set CM.Control.implicit_anchors false;
1179
1180		This causes CM to reject implicitly anchored paths.  This is (for the
1181		time being) less permissive than the "final" version where there will
1182		be no more such implicit anchors and relative paths will be just that:
1183		relative.
1184
1185		The next step (version after next version?) will be to make the
1186		default for CM.Control.implicit_anchors false.  After the dust has
1187		settled, I can then produce the "final" version of this...
1188
1189		Note: Since bootstrapping is a bit tricky, I provided new boot files.
1190
1191		----------------------------------------------------------------------
1192		Name: Matthias Blume
1193		Date: 2000/05/11 16:30:00 JST
1194		Tag: blume-20000511-sources
1195		Description:
1196
1197		The main change is that I added function CM.sources as a generalized
1198		version of the earlier CM.makedepend.  This entails the following
1199		additional changes:
1200
1201		- CM.makedepend has been dropped.
1202
1203		- CM manual has been updated.
1204
1205		- TOOLS signature and API have been changed.
1206
1207		----------------------------------------------------------------------
1208		Name: Allen Leung
1209		Date: 2000/05/10 21:17 EDT
1210		Tag: leunga-20000510-moby-c--ssa
1211		Description:
1212
1213		Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1214		None of these affect SML/NJ.
1215
1216		1. Register Allocation
1217
1218		a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1219		This module tries to remove local (i.e. basic block level) redundancies
1220		during spilling.
1221
1222		b. A new framework for performing region based register allocation.
1223		Not yet entirely functional.
1224
1225		2. X86
1226
1227		a. DefUse for POP was missing the stack pointer [found by Lal]
1228		b. Reload for CALL was incorrect in X86Spill [found by John]
1229		c. Various fixes in X86Spill so that it can be used correctly for
1230		the new spilling module.
1231
1232		3. SSA/IR
1233
1234		a. New module ir/dj-dataflow.sml implements elimination based
1235		data flow analysis.
1236
1237		4. MLRiscGen
1238
1239		a. Fix for gc type annotation
1240
1241		5. MDGen
1242
1243		Various fixes for machine description -> ml code translation.  For ssa
1244		only.
1245
1246		----------------------------------------------------------------------
1247		Name: Allen Leung
1248		Date: 2000/05/08 22:17 EDT
1249		Tag: leunga-20000508-labexp
1250		Description:
1251
1252		Fermin has found a few assembly problems with constant expressions
1253		generated in LabelExp.  Mostly, the problems involve extra parentheses,
1254		which choke on dumb assemblers.  This is his fix.
1255
1256		----------------------------------------------------------------------
1257		Name: Dave MacQueen
1258		Date: 2000/04/09 14:00 EDT
1259		Tag: dbm-20000502-Version_110_28
1260		Description:
1261
1262		1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1263
1264		2. Updated config/version to 110.28
1265
1266		3. Updated config/srcarchiveurl
1267
1268		3. New boot files!
1269		ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1270
1271		----------------------------------------------------------------------
1272		Name: Matthias Blume
1273		Date: 2000/05/01 19:05:00 JST
1274		Tag: blume-20000501-noweb
1275		Description:
1276
1277		A new noweb tool has been added.  The existing system is entirely
1278		unaffected by this, but some CM users have asked for renewed noweb
1279		support.  Everything is documented in the CM manual.
1280
1281		New (plugin) libraries:
1282
1283		noweb-tool.cm
1284		nw-ext.cm
1285
1286		----------------------------------------------------------------------
1287		Name: Dave MacQueen
1288		Date: 2000/04/30 12:40PM EDT
1289		Tag: dbm-20000430-bug_fixes
1290		Description:
1291
1292		1. Fix for bug 1498
1293		smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1294		smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1295		added toRealArray function
1296		smlnj/src/compiler/MiscUtil/print/ppobj.sml
1297		added check for tag Obj.RealArray to array printing case in ppObj
1298
1299		2. Fix for bug 1510
1300		smlnj/src/compiler/Semant/types/typesutil.sml
1301		fixed definition of dummyargs (used by equalTycon) so that
1302		dummy args are distinct types
1303
1304		----------------------------------------------------------------------
1305		Name: Matthias Blume
1306		Date: 2000/04/30 01:00:00 JST
1307		Tag: blume-20000430-versions
1308		Description:
1309
1310		1. CM version numbering added.  This is an implementation of Lal's
1311		proposal for adding version numbers and version checking to .cm
1312		files.  Lal said that his proposal was just that -- a proposal.
1313		For the time being I went ahead and implemented it so that people
1314		can comment on it.  Everything is completely backward-compatible
1315		(except for the stable library format, i.e., new bootfiles!).
1316
1317		As usual, see the CM manual for details.
1318
1319		2. An alternative syntax for anchored paths has been implemented.
1320		Dave has recently voiced the same concerns that I had when I did
1321		this, so there should be some support.  My take is that eventually
1322		I will let support for the current syntax (where anchors are
1323		"implicit") fade out in favor of the new, explicit syntax.
1324		In order to be backward-compatible, both old and new syntax are
1325		currently supported.
1326
1327		Again, see the CM manual for details.
1328
1329		3. Parallel make is trying to be slightly smarter:  When the master
1330		process finds a "bottleneck", i.e., when there is only one
1331		compilation unit that can be compiled and everybody else is
1332		waiting on it, then it will simply compile it directly instead
1333		of clumsily telling one of the slaves to do it.
1334
1335		4. Support for "unsharing" added.  This is necessary in order to be
1336		able to have two different versions of the same library running
1337		at the same time (e.g., for trying out a new MLRISC while still
1338		having the old MLRISC linked into the current compiler, etc.)
1339		See the CM manual.
1340
1341		5. Simple "makedepend" functionality added for generating Makefile
1342		dependency information.  (This is rather crude at the moment.
1343		Expect some changes here in the future.)
1344
1345		6. ".fun" added as a recognized suffix for ML files. Also documented
1346		explicitly in the manual that the fallback behavior (unknown suffix
1347		-> ML file) is not an official feature!
1348
1349		7. Small changes to the pickler for stable libraries.
1350
1351		8. Several internal changes to CM (for cleanup/improvement).
1352
1353
1354		!!!! NEW BINFILES !!!!
1355
1356		----------------------------------------------------------------------
1357		Name: Matthias Blume
1358		Date: 2000/04/28 17:30:00 JST
1359		Tag: blume-20000428-pathconfig
1360		Description:
1361
1362		1. I changed config/install.sh to remove duplicate entries from the
1363		lib/pathconfig file at the end.  Moreover, the final version of
1364		lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1365		in src/system/installml.
1366
1367		2. The config/install.sh script now consistently uses relative
1368		pathnames in lib/pathconfig whenever the anchor is in the lib
1369		directory.  (So far this was true for the libraries that come
1370		pre-compiled and bundled as part of the bootfiles but not for
1371		libraries that are compiled by the script itself.)
1372
1373		----------------------------------------------------------------------
1374		Name: Matthias Blume
1375		Date: 2000/04/26 13:10:00 JST
1376		Tag: blume-20000426-fun_suffix
1377		Description:
1378
1379		Added ".fun" as a recognized file name suffix (for ML code).
1380
1381		----------------------------------------------------------------------
1382		Name: Allen Leung
1383		Date: 2000/04/25 17:00:00 EST
1384		Tag: leunga-20000425-alpha-ra
1385		Description:
1386
1387		1. Alpha
1388
1389		PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1390		in C--.
1391
1392		2. RA
1393
1394		Added a flag "ra-dump-size" to print out the size of the flowgraph
1395		and the interference graph.
1396
1397		----------------------------------------------------------------------
1398		Name: Dave MacQueen
1399		Date: 2000/04/25/
1400		Tag: dbm-20000425-mlyacc_doc_examples
1401		Description:
1402		Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1403		Updated all three examples in src/ml-yacc/examples to run
1404		under 110.* using CM.make.
1405
1406		----------------------------------------------------------------------
1407		Name: Allen Leung
1408		Date: 2000/04/20 23:04:00 EST
1409		Tag: leunga-20000420-ssa-c---stuff
1410		Description:
1411
1412		This update synchronizes my repository with Yale's.  Most of these
1413		changes, however, do not affect SML/NJ at all (the RA is an exception).
1414
1415		1. Register Allocator
1416
1417		a. An improvement in the interference graph construction:
1418		Given a copy
1419
1420		s <- t
1421
1422		no interference edge between s and t is added for this definition of s.
1423
1424		b. I've added two new spill heuristic modules that Fermin and I developed
1425		(in the new library RA.cm). These are unused in SML/NJ but maybe
1426		useful for others (Moby?)
1427
1428		2. X86
1429
1430		a. Various fixes in the backend provided by Fermin [C--] and Lal.
1431
1432		3. Alpha
1433
1434		a. Added the BSR instruction and code generation that goes with it [C--]
1435		b. Other fixes too numerous to recount provided by Fermin [C--]
1436
1437		4. Regmaps
1438
1439		a. The regmaps are not initialized with the identity physical bindings
1440		at creation time.  This is unneeded.
1441
1442		5. MLRISC Optimizations
1443
1444		a. The DJ-Graph module can now compute the iterated dominance frontiers
1445		intersects with liveness incrementally in linear time! Woohoo!
1446		This is now used in my new SSA construction algorithm.
1447
1448		b. THe branch reorganization module is now smarter about linear chains of
1449		basic blocks.
1450
1451
1452		----------------------------------------------------------------------
1453		Name: Matthias Blume
1454		Date: 2000/04/12 13:52:00 JST
1455		Tag: blume_main_v110p27_1
1456		Description:
1457
1458		Changed install.sh script to handle archive files without version number
1459		and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1460		name of the boot file archive.
1461
1462		----------------------------------------------------------------------
1463		Name: Dave MacQueen
1464		Date: 2000/04/09 14:00 EDT
1465		Tag: dbm-20000410-Version_110_27
1466		Description:
1467
1468		1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1469
1470		2. Updated src/config/version to 110.27
1471
1472		3. New boot files!
1473
1474		----------------------------------------------------------------------
1475		Name: Allen Leung
1476		Date: 2000/04/09 19:09:00 EST
1477		Tag: leunga-20000409-misc
1478		Description:
1479
1480		1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1481
1482		2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1483
1484		----------------------------------------------------------------------
1485		Name: Stefan
1486		Date: 2000/04/07 10:00:00 EDT
1487		Tag: monnier-20000406-branch-handling
1488		Description:
1489
1490		Improved handling of branches (mostly those generated from
1491		polymorphic equality), removed switchoff and changed the
1492		default optimization settings (more cpsopt and less flintopt).
1493
1494		----------------------------------------------------------------------
1495		Name: Allen Leung
1496		Date: 2000/04/06 01:30:00 EST
1497		Tag: leunga-20000406-peephole-x86-SSA-2
1498		Description:
1499
1500		Forgot a few files.
1501
1502		----------------------------------------------------------------------
1503		Name: Allen Leung
1504		Date: 2000/04/06 00:36:00 EST
1505		Tag: leunga-20000406-peephole-x86-SSA
1506		Description:
1507
1508		1.  New Peephole code
1509
1510		2.  Minor improvement to X86 instruction selection
1511
1512		3.  Various fixes to SSA and machine description -> code translator
1513
1514		----------------------------------------------------------------------
1515		Name: Matthias Blume
1516		Date: 2000/04/05 12:30:00 JST
1517		Tag: blume_main_v110p26p2_3
1518		Description:
1519
1520		This update just merges three minor cosmetic updates to CM's sources
1521		to get ready for the 110.27 code freeze on Friday.  No functionality
1522		has changed.
1523
1524		----------------------------------------------------------------------
1525		Name: Allen Leung
1526	Date: 2000/04/04 19:39:00 EST	Date: 2000/04/04 19:39:00 EST
1527	Tag: leunga-20000404-x86-asm	Tag: leunga-20000404-x86-asm
1528	Description:	Description:
#	Line 57	Line 1567
1567
1568	3.  Assembly	3.  Assembly
1569
1570	When generating assemby, resolve the value of client defined constants,	When generating assembly, resolve the value of client defined constants,
1571	instead of generating symbolic values.  This is controlled by the	instead of generating symbolic values.  This is controlled by the
1572	new flag "asm-resolve-constants", which is default to true.	new flag "asm-resolve-constants", which is default to true.
1573
#	Line 80	Line 1590
1590
1591	To this end, I arranged that instead of "structure Core" as "structure	To this end, I arranged that instead of "structure Core" as "structure
1592	_Core" is bound in the pervasive environment.  Core access is done via	_Core" is bound in the pervasive environment.  Core access is done via
1593	_Core (which can never be accidentially rebound because _Core is not a	_Core (which can never be accidentally rebound because _Core is not a
1594	legal surface-syntax symbol).	legal surface-syntax symbol).
1595
1596	The current solution is much cleaner because the core environment is	The current solution is much cleaner because the core environment is
#	Line 90	Line 1600
1600	with dynamic and symbolic parts of the core environment.	with dynamic and symbolic parts of the core environment.
1601
1602	Remaining hackery (to bind the "magic" symbol _Core) is localized in the	Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1603	compilation mananger's bootstrap compiler (actually: in the "init group"	compilation manager's bootstrap compiler (actually: in the "init group"
1604	handling).  See the comments in src/system/smlnj/init/init.cmi for	handling).  See the comments in src/system/smlnj/init/init.cmi for
1605	more details.	more details.
1606
#	Line 205	Line 1715
1715	(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
1716	used).	used).
1717
1718	This was done to accomodate the new "make" and "shell" tools which	This was done to accommodate the new "make" and "shell" tools which
1719	facilitate fairly seemless hookup to portions of code managed using	facilitate fairly seamless hookup to portions of code managed using
1720	Makefiles or Shell scripts.	Makefiles or Shell scripts.
1721
1722	There are no classes "shared" or "private" anymore.  Instead, the	There are no classes "shared" or "private" anymore.  Instead, the
#	Line 218	Line 1728
1728
1729	All existing tools are described in the CM manual.	All existing tools are described in the CM manual.
1730
1731	- Slightly better error handling.  (CM now surpresses many followup	- Slightly better error handling.  (CM now suppresses many followup
1732	error messages that tended to be more annoying than helpful.)	error messages that tended to be more annoying than helpful.)
1733
1734	2. Major changes to the compiler's static environment data structures.	2. Major changes to the compiler's static environment data structures.
#	Line 352	Line 1862
1862
1863	I've changed andl to testl in the floating point test sequence	I've changed andl to testl in the floating point test sequence
1864	whenever appropriate.  The Intel optimization guide states that	whenever appropriate.  The Intel optimization guide states that
1865	testl is perferable to andl.	testl is preferable to andl.
1866
1867	7. RA (x86 only)	7. RA (x86 only)
1868
#	Line 534	Line 2044
2044
2045	1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an	1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2046	additional argument called "template" which is an optional string that	additional argument called "template" which is an optional string that
2047	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
2048	explanation.	explanation.
2049
2050	2. A special-purpose tool can be "regisitered" by simply dropping the	2. A special-purpose tool can be "registered" by simply dropping the
2051	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2052	directory where the .cm file lives that uses this tool.  (The	directory where the .cm file lives that uses this tool.  (The
2053	behavior/misfeature until now was to look for the tool description	behavior/misfeature until now was to look for the tool description
#	Line 581	Line 2091
2091	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
2092	<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,
2093	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
2094	perhaps got accidentially killed).  In this case, fixpt will quickly	perhaps got accidentally killed).  In this case, fixpt will quickly
2095	move through what exists before continuing where it left off earlier,	move through what exists before continuing where it left off earlier,
2096	and, thus, saves a lot of time.	and, thus, saves a lot of time.
2097
#	Line 631	Line 2141
2141	it from that remote directory.	it from that remote directory.
2142	This should simplify installation further:  For machines that have	This should simplify installation further:  For machines that have
2143	access to the internet, just fetch <version>-config.tgz, unpack it,	access to the internet, just fetch <version>-config.tgz, unpack it,
2144	edit config/targets, and go (run config/install.sh).  The scipt will	edit config/targets, and go (run config/install.sh).  The script will
2145	fetch everything else that it might need all by itself.	fetch everything else that it might need all by itself.
2146
2147	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.594   changed lines   Added in v.755

---

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