Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 695, Mon Aug 7 23:57:38 2000 UTC	revision 1183, Fri Mar 29 19:09:48 2002 UTC
#	Line 8	Line 8
8	The form of an entry should be:	The form of an entry should be:
9
10	Name:	Name:
11	Date:	Date: yyyy/mm/dd
12	Tag: <post-commit CVS tag>	Tag: <post-commit CVS tag>
13	Description:	Description:
14
15		----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2002/03/29 14:10:00 EST
18		Tag: blume-20020329-inlprims
19		Description:
20
21		NEW BOOTFILES!!!    Version number bumped to 110.39.3.
22
23		Primops have changed. This means that the bin/boot-file formats have
24		changed as well.
25
26		To make sure that there is no confusion, I made a new version.
27
28
29		CHANGES:
30
31		* removed REMT from mltree (remainder should never overflow).
32
33		* added primops to deal with divisions of all flavors to the frontend
34
35		* handled these primops all the way through so they map to their respective
36		MLRISC support
37
38		* used these primops in the implementation of Int, Int32, Word, Word32
39
40		* removed INLDIV, INLMOD, and INLREM as they are no longer necessary
41
42		* parameterized INLMIN, INLMAX, and INLABS by a numkind
43
44		* translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
45		floating point
46
47		* used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
48		and Real (but Real.abs maps to a separate floating-point-only primop)
49
50
51		TODO items:
52
53		* Hacked Alpha32 instruction selection, disabling the selection of REMx
54		instructions because the machine instruction encoder cannot handle
55		them.  (Hppa, PPC, and Sparc instruction selection did not handle
56		REM in the first place, and REM is supported by the x86 machine coder.)
57
58		* Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
59		selection phase.  (The two can be streamlined because the hardware
60		delivers both quotient and remainder at the same time anyway.)
61
62		* Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
63		(Currently the behavior is inconsistent both across architectures and
64		wrt. the draft Basis spec.)
65
66		* Word8 should eventually be handled natively, too.
67
68		* There seems to be one serious bug in mltree-gen.sml.  It appears, though,
69		as if there currently is no execution path that could trigger it in
70		SML/NJ.  (The assumptions underlying functions arith and promotable do not
71		hold for things like multiplication and division.)
72
73		----------------------------------------------------------------------
74		Name: Matthias Blume
75		Date: 2002/03/27 16:27:00 EST
76		Tag: blume-20020327-mlrisc-divisions
77		Description:
78
79		Added support for all four division operations (ML's div, mod, quot,
80		and rem) to MLRISC.  In the course of doing so, I also rationalized
81		the naming (no more annoying switch-around of DIV and QUOT), by
82		parameterizing the operation by div_rounding_mode (which can be either
83		DIV_TO_ZERO or DIV_TO_NEGINF).
84
85		The generic MLTreeGen functor takes care of compiling all four
86		operations down to only round-to-zero div.
87
88		Missing pieces:
89
90		* Doing something smarter than relying on MLTreeGen on architectures
91		like, e.g., the x86 where hardware division delivers both quotient and
92		remainder at the same time.  With this, the implementation of the
93		round-to-neginf operations could be further streamlined.
94
95		* Remove inlining support for div/mod/rem from the frontend and replace it
96		with primops that get carried through to the backend.  Do this for all
97		int and word types.
98
99		----------------------------------------------------------------------
100		Name: Matthias Blume
101		Date: 2002/03/25 17:25:00 EST
102		Tag: blume-20020325-divmod
103		Description:
104
105		I improved (hopefully without breaking them) the implementation of Int.div,
106		Int.mod, and Int.rem.   For this, the code in translate.sml now takes
107		advantage of the following observations:
108
109		Let  q = x quot y      r = x rem y
110		d = x div  y      m = x mod y
111
112		where "quot" is the round-to-zero version of integer division that
113		hardware usually provides.  Then we have:
114
115		r = x - q * y        where neither the * nor the - will overflow
116		d = if q >= 0 orelse x = q * y then q else q - 1
117		where neither the * nor the - will overflow
118		m = if q >= 0 orelse r = 0 then r else r + y
119		where the + will not overflow
120
121		This results in substantial simplification of the generated code.
122		The following table shows the number of CFG nodes and edges generated
123		for
124		fun f (x, y) = x OPER y
125		(* with OPER \in div, mod, quot, rem *)
126
127
128		OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
129		--------------------------------------------------------
130		div |         24 |         39 |         12 |         16
131		mod |         41 |         71 |         12 |         16
132		quot |          8 |         10 |          8 |         10
133		rem |         10 |         14 |          8 |         10
134
135
136		----------------------------------------------------------------------
137		Name: Matthias Blume
138		Date: 2002/03/25 22:06:00 EST
139		Tag: blume-20020325-cprotobug
140		Description:
141
142		Fixed a bug in cproto (c prototype decoder).
143
144		----------------------------------------------------------------------
145		Name: Matthias Blume
146		Date: 2002/03/25 16:00:00 EST
147		Tag: blume-20020325-raw-primops
148		Description:
149
150		I did some cleanup to Allen's new primop code and
151		replaced yesterday's bootfiles with new ones.
152		(But they are stored in the same place.)
153
154		----------------------------------------------------------------------
155		Name: Matthias Blume
156		Date: 2002/03/24 22:40:00 EST
157		Tag: blume-20020324-bootfiles
158		Description:
159
160		Made the bootfiles that Allen asked for.
161
162		----------------------------------------------------------------------
163		Name: Allen Leung
164		Date: 2002/03/23 15:50:00 EST
165		Tag: leunga-20020323-flint-cps-rcc-primops
166		Description:
167
168		1. Changes to FLINT primops:
169
170		(* make a call to a C-function;
171		* The primop carries C function prototype information and specifies
172		* which of its (ML-) arguments are floating point. C prototype
173		* information is for use by the backend, ML information is for
174		* use by the CPS converter. *)
175		| RAW_CCALL of { c_proto: CTypes.c_proto,
176		ml_args: ccall_type list,
177		ml_res_opt: ccall_type option,
178		reentrant : bool
179		} option
180		(* Allocate uninitialized storage on the heap.
181		* The record is meant to hold short-lived C objects, i.e., they
182		* are not ML pointers.  With the tag, the representation is
183		* the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
184		*)
185		| RAW_RECORD of {tag:bool,sz:int}
186		and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
187
188		2.  These CPS primops are now overloaded:
189
190		rawload of {kind:numkind}
191		rawstore of {kind:numkind}
192
193		The one argument form is:
194
195		rawload {kind} address
196
197		The two argument form is:
198
199		rawload {kind} [ml object, byte-offset]
200
201		3. RAW_CCALL/RCC now takes two extra arguments:
202
203		a. The first is whether the C call is reentrant, i.e., whether
204		ML state should be saved and restored.
205		b. The second argument is a string argument specifying the name of
206		library and the C function.
207
208		These things are currently not handled in the code generator, yet.
209
210		4. In CProto,
211
212		An encoding type of "bool" means "ml object" and is mapped into
213		C prototype of PTR.  Note that "bool" is different than "string",
214		even though "string" is also mapped into PTR, because "bool"
215		is assigned an CPS type of BOGt, while "string" is assigned INT32t.
216
217		5. Pickler/unpicker
218
219		Changed to handle RAW_RECORD and newest RAW_CCALL
220
221		6. MLRiscGen,
222
223		1. Changed to handle the new rawload/rawstore/rawrecord operators.
224		2. Code for handling C Calls has been moved to a new module CPSCCalls,
225		in the file CodeGen/cpscompile/cps-c-calls.sml
226
227		7. Added the conditional move operator
228
229		condmove of branch
230
231		to cps.  Generation of this is still buggy so it is currently
232		disabled.
233
234		----------------------------------------------------------------------
235		Name: Lal George
236		Date: 2002/03/22 14:18:25 EST
237		Tag: george-20020322-cps-branch-prob
238		Description:
239
240		Implemented the Ball-Larus branch prediction-heuristics, and
241		incorporated graphical viewers for control flow graphs.
242
243		Ball-Larus Heuristics:
244		---------------------
245		See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
246
247		By design it uses the Dempster-Shafer theory for combining
248		probabilities.  For example, in the function:
249
250		fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
251
252		the ball-larus heuristics predicts that the n=0 is unlikely
253		(OH-heuristic), and the 'then' branch is unlikely because of the
254		RH-heuristic -- giving the 'then' branch an even lower combined
255		probability using the Dempster-Shafer theory.
256
257		Finally, John Reppy's loop analysis in MLRISC, further lowers the
258		probability of the 'then' branch because of the loop in the else
259		branch.
260
261
262		Graphical Viewing:
263		------------------
264		I merely plugged in Allen's graphical viewers into the compiler. The
265		additional code is not much. At the top level, saying:
266
267		Control.MLRISC.getFlag "cfg-graphical-view" := true;
268
269		will display the graphical view of the control flow graph just before
270		back-patching.  daVinci must be in your path for this to work. If
271		daVinci is not available, then the default viewer can be changed
272		using:
273
274		Control.MLRISC.getString "viewer"
275
276		which can be set to "dot" or "vcg" for the corresponding viewers. Of
277		course, these viewers must be in your path.
278
279		The above will display the compilation unit at the level of clusters,
280		many of which are small, boring, and un-interesting. Also setting:
281
282		Control.MLRISC.getInt "cfg-graphical-view_size"
283
284		will display clusters that are larger than the value set by the above.
285
286
287		----------------------------------------------------------------------
288		Name: Matthias Blume
289		Date: 2002/03/21 22:20:00 EST
290		Tag: blume-20020321-kmp-bugfix
291		Description:
292
293		Changed the interface to the KMP routine in PreString and fixed
294		a minor bug in one place where it was used.
295
296		----------------------------------------------------------------------
297		Name: Allen Leung
298		Date: 2002/03/21 20:30:00 EST
299		Tag: leunga-20020321-cfg
300		Description:
301
302		Fixed a potential problem in cfg edge splitting.
303
304		----------------------------------------------------------------------
305		Name: Allen Leung
306		Date: 2002/03/21 17:15:00 EST
307		Tag: leunga-20020321-x86-fp-cfg
308		Description:
309
310		1. Recoded the buggy parts of x86-fp.
311
312		a. All the block reordering code has been removed.
313		We now depend on the block placement phases to do this work.
314
315		b. Critical edge splitting code has been simplified and moved into the
316		CFG modules, as where they belong.
317
318		Both of these were quite buggy and complex.  The code is now much, much
319		simpler.
320
321		2. X86 backend.
322
323		a. Added instructions for 64-bit support.  Instruction selection for
324		64-bit has not been committed, however, since that
325		requires changes to MLTREE which haven't been approved by
326		Lal and John.
327
328		b. Added support for FUCOMI and FUCOMIP when generating code for
329		PentiumPro and above.  We only generate these instructions in
330		the fast-fp mode.
331
332		c. Added cases for JP and JNP in X86FreqProps.
333
334		3. CFG
335
336		CFG now has a bunch of methods for edge splitting and merging.
337
338		4. Machine description.
339
340		John's simplification of MLTREE_BASIS.fcond broke a few machine
341		description things:
342
343		rtl-build.{sig,sml} and hppa.mdl fixed.
344
345		NOTE: the machine description stuff in the repository is still broken.
346		Again, I can't put my fixes in because that involves
347		changes to MLTREE.
348
349		----------------------------------------------------------------------
350		Name: Matthias Blume
351		Date: 2002/03/20 15:55:00 EST
352		Tag: blume-20020320-kmp
353		Description:
354
355		Implemented Knuth-Morris-Pratt string matching in PreString and used
356		it for String.isSubstring, Substring.isSubstring, and
357		Substring.position.
358
359		(Might need some stress-testing.  Simple examples worked fine.)
360
361		----------------------------------------------------------------------
362		Name: Matthias Blume
363		Date: 2002/03/19 16:37:00 EST
364		Tag: blume-20020319-witnesses
365		Description:
366
367		Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
368
369		This implements a generic mechanism for changing constness qualifiers
370		anywhere within big C types without resorting to outright "casts".
371		(So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
372		modify the constness at the outermost level.)
373		The implementation of "convert" is based on the idea of "witness"
374		values -- values that are not used by the operation but whose types
375		"testify" to their applicability.  On the implementation side, "convert"
376		is simply a projection (returning its second curried argument).  With
377		cross-module inlining, it should not result in any machine code being
378		generated.
379
380		----------------------------------------------------------------------
381		Name: Matthias Blume
382		Date: 2002/03/15 16:40:00 EST
383		Tag: blume-20020315-basis
384		Description:
385
386		Provided (preliminary?) implementations for
387
388		{String,Substring}.{concatWith,isSuffix,isSubstring}
389
390		and
391
392		Substring.full
393
394		Those are in the Basis spec but they were missing in SML/NJ.
395
396		----------------------------------------------------------------------
397		Name: Matthias Blume
398		Date: 2002/03/14 21:30:00 EST
399		Tag: blume-20020314-controls
400		Description:
401
402		Controls:
403		---------
404
405		1. Factored out the recently-added Controls : CONTROLS stuff and put
406		it into its own library $/controls-lib.cm.  The source tree for
407		this is under src/smlnj-lib/Controls.
408
409		2. Changed the names of types and functions in this interface, so they
410		make a bit more "sense":
411
412		module -> registry
413		'a registry -> 'a group
414
415		3. The interface now deals in ref cells only.  The getter/setter interface
416		is (mostly) gone.
417
418		4. Added a function that lets one register an already-existing ref cell.
419
420		5. Made the corresponding modifications to the rest of the code so that
421		everything compiles again.
422
423		6. Changed the implementation of Controls.MLRISC back to something closer
424		to the original.  In particular, this module (and therefore MLRISC)
425		does not depend on Controls.  There now is some link-time code in
426		int-sys.sml that registers the MLRISC controls with the Controls
427		module.
428
429		CM:
430		---
431
432		* One can now specify the lambda-split aggressiveness in init.cmi.
433
434		----------------------------------------------------------------------
435		Name: Allen Leung
436		Date: 2002/03/13 17:30:00 EST
437		Tag: leunga-20020313-x86-fp-unary
438		Description:
439
440		Bug fix for:
441
442		> leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
443		> Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
444		> - fun f(x,(y,z)) = Real.~ y;
445		> [autoloading]
446		> [autoloading done]
447		>       fchsl   (%eax), 184(%esp)
448		> Error: MLRisc bug: X86MCEmitter.emitInstr
449		>
450		> uncaught exception Error
451		>   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
452
453		The problem was that the code generator did not generate any fp registers
454		in this case, and the ra didn't know that it needed to run the X86FP phase to
455		translate the pseudo fp instruction.   This only happened with unary fp
456		operators in certain situations.
457
458		----------------------------------------------------------------------
459		Name: Matthias Blume
460		Date: 2002/03/13 14:00:00 EST
461		Tag: blume-20020313-overload-etc
462		Description:
463
464		1. Added _overload as a synonym for overload for backward compatibility.
465		(Control.overloadKW must be true for either version to be accepted.)
466
467		2. Fixed bug in install script that caused more things to be installed
468		than what was requested in config/targets.
469
470		3. Made CM aware of the (_)overload construct so that autoloading
471		works.
472
473		----------------------------------------------------------------------
474		Name: Matthias Blume
475		Date: 2002/03/12 22:03:00 EST
476		Tag: blume-20020312-url
477		Description:
478
479		Forgot to update BOOT and srcarchiveurl.
480
481		----------------------------------------------------------------------
482		Name: Matthias Blume
483		Date: 2002/03/12 17:30:00 EST
484		Tag: blume-20020312-version110392
485		Description:
486
487		Yet another version number bump (because of small changes to the
488		binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
489
490		Changes:
491
492		The new pid generation scheme described a few weeks ago was overly
493		complicated.  I implemented a new mechanism that is simpler and
494		provides a bit more "stability":  Once CM has seen a compilation
495		unit, it keeps its identity constant (as long as you do not delete
496		those crucial CM/GUID/* files).  This means that when you change
497		an interface, compile, then go back to the old interface, and
498		compile again, you arrive at the original pid.
499
500		There now also is a mechanism that instructs CM to use the plain
501		environment hash as a module's pid (effectively making its GUID
502		the empty string).  For this, "noguid" must be specified as an
503		option to the .sml file in question within its .cm file.
504		This is most useful for code that is being generated by tools such
505		as ml-nlffigen (because during development programmers tend to
506		erase the tool's entire output directory tree including CM's cached
507		GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
508		revert to the old, broken behavior of SML/NJ, but in specific cases
509		where there is no danger of interface confusion, its use is ok
510		(I think).
511
512		ml-nlffigen by default generates "noguid" annotations.  They can be
513		turned off by specifying -guid in its command line.
514
515		----------------------------------------------------------------------
516		Name: Lal George
517		Date: 2002/03/12 12 14:42:36 EST
518		Tag: george-20020312-frequency-computation
519		Description:
520
521		Integrated jump chaining and static block frequency into the
522		compiler. More details and numbers later.
523
524		----------------------------------------------------------------------
525		Name: Lal George
526		Date: 2002/03/11 11 22:38:53 EST
527		Tag: george-20020311-jump-chain-elim
528		Description:
529
530		Tested the jump chain elimination on all architectures (except the
531		hppa).  This is on by default right now and is profitable for the
532		alpha and x86, however, it may not be profitable for the sparc and ppc
533		when compiling the compiler.
534
535		The gc test will typically jump to a label at the end of the cluster,
536		where there is another jump to an external cluster containing the actual
537		code to invoke gc. This is to allow factoring of common gc invocation
538		sequences. That is to say, we generate:
539
540		f:
541		testgc
542		ja   L1      % jump if above to L1
543
544		L1:
545		jmp L2
546
547
548		After jump chain elimination the 'ja L1' instructions is converted to
549		'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
550		up being implemented in their long form (if L2 is far away) using:
551
552		jbe     L3      % jump if below or equal to L3
553		jmp     L2
554		L3:
555		...
556
557
558		For large compilation units L2  may be far away.
559
560
561		----------------------------------------------------------------------
562		Name: Matthias Blume
563		Date: 2002/03/11 13:30:00 EST
564		Tag: blume-20020311-mltreeeval
565		Description:
566
567		A functor parameter was missing.
568
569		----------------------------------------------------------------------
570		Name: Allen Leung
571		Date: 2002/03/11 10:30:00 EST
572		Tag: leunga-20020311-runtime-string0
573		Description:
574
575		The representation of the empty string now points to a
576		legal null terminated C string instead of unit.  It is now possible
577		to convert an ML string into C string with InlineT.CharVector.getData.
578		This compiles into one single machine instruction.
579
580		----------------------------------------------------------------------
581		Name: Allen Leung
582		Date: 2002/03/10 23:55:00 EST
583		Tag: leunga-20020310-x86-call
584		Description:
585
586		Added machine generation for CALL instruction (relative displacement mode)
587
588		----------------------------------------------------------------------
589		Name: Matthias Blume
590		Date: 2002/03/08 16:05:00
591		Tag: blume-20020308-entrypoints
592		Description:
593
594		Version number bumped to 110.39.1.  NEW BOOTFILES!
595
596		Entrypoints: non-zero offset into a code object where execution should begin.
597
598		- Added the notion of an entrypoint to CodeObj.
599		- Added reading/writing of entrypoint info to Binfile.
600		- Made runtime system bootloader aware of entrypoints.
601		- Use the address of the label of the first function given to mlriscGen
602		as the entrypoint.  This address is currently always 0, but it will
603		not be 0 once we turn on block placement.
604		- Removed the linkage cluster code (which was The Other Way(tm) of dealing
605		with entry points) from mlriscGen.
606
607		----------------------------------------------------------------------
608		Name: Allen Leung
609		Date: 2002/03/07 20:45:00 EST
610		Tag: leunga-20020307-x86-cmov
611		Description:
612
613		Bug fixes for CMOVcc on x86.
614
615		1. Added machine code generation for CMOVcc
616		2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
617		3. CMOVcc cannot have an immediate operand as argument.
618
619		----------------------------------------------------------------------
620		Name: Matthias Blume
621		Date: 2002/03/07 16:15:00 EST
622		Tag: blume-20020307-controls
623		Description:
624
625		This is a very large but mostly boring patch which makes (almost)
626		every tuneable compiler knob (i.e., pretty much everything under
627		Control.* plus a few other things) configurable via both the command
628		line and environment variables in the style CM did its configuration
629		until now.
630
631		Try starting sml with '-h' (or, if you are brave, '-H')
632
633		To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
634		implements the underlying generic mechanism.
635
636		The interface to some of the existing such facilities has changed somewhat.
637		For example, the MLRiscControl module now provides mkFoo instead of getFoo.
638		(The getFoo interface is still there for backward-compatibility, but its
639		use is deprecated.)
640
641		The ml-build script passes -Cxxx=yyy command-line arguments through so
642		that one can now twiddle the compiler settings when using this "batch"
643		compiler.
644
645		TODO items:
646
647		We should go through and throw out all controls that are no longer
648		connected to anything.  Moreover, we should go through and provide
649		meaningful (and correct!) documentation strings for those controls
650		that still are connected.
651
652		Currently, multiple calls to Controls.new are accepted (only the first
653		has any effect).  Eventually we should make sure that every control
654		is being made (via Controls.new) exactly once.  Future access can then
655		be done using Controls.acc.
656
657		Finally, it would probably be a good idea to use the getter-setter
658		interface to controls rather than ref cells.  For the time being, both
659		styles are provided by the Controls module, but getter-setter pairs are
660		better if thread-safety is of any concern because they can be wrapped.
661
662		*****************************************
663
664		One bug fix: The function blockPlacement in three of the MLRISC
665		backpatch files used to be hard-wired to one of two possibilities at
666		link time (according to the value of the placementFlag).  But (I
667		think) it should rather sense the flag every time.
668
669		*****************************************
670
671		Other assorted changes (by other people who did not supply a HISTORY entry):
672
673		1. the cross-module inliner now works much better (Monnier)
674		2. representation of weights, frequencies, and probabilities in MLRISC
675		changed in preparation of using those for weighted block placement
676		(Reppy, George)
677
678		----------------------------------------------------------------------
679		Name: Lal George
680		Date: 2002/03/07 14:44:24 EST 2002
681		Tag: george-20020307-weighted-block-placement
682
683		Tested the weighted block placement optimization on all architectures
684		(except the hppa) using AMPL to generate the block and edge frequencies.
685		Changes were required in the machine properties to correctly
686		categorize trap instructions. There is an MLRISC flag
687		"weighted-block-placement" that can be used to enable weighted block
688		placement, but this will be ineffective without block/edge
689		frequencies (coming soon).
690
691
692		----------------------------------------------------------------------
693		Name: Lal George
694		Date: 2002/03/05 17:24:48 EST
695		Tag: george-20020305-linkage-cluster
696
697		In order to support the block placement optimization, a new cluster
698		is generated as the very first cluster (called the linkage cluster).
699		It contains a single jump to the 'real' entry point for the compilation
700		unit. Block placement has no effect on the linkage cluster itself, but
701		all the other clusters  have full freedom in the manner in which they
702		reorder blocks or functions.
703
704		On the x86 the typical linkage code that is generated is:
705		----------------------
706		.align 2
707		L0:
708		addl    $L1-L0, 72(%esp)
709		jmp     L1
710
711
712		.align  2
713		L1:
714		----------------------
715
716		72(%esp) is the memory location for the stdlink register. This
717		must contain the address of the CPS function being called. In the
718		above example, it contains the address of  L0; before
719		calling L1 (the real entry point for the compilation unit), it
720		must contain the address for L1, and hence
721
722		addl $L1-L0, 72(%esp)
723
724		I have tested this on all architectures except the hppa.The increase
725		in code size is of course negligible
726
727		----------------------------------------------------------------------
728		Name: Allen Leung
729		Date: 2002/03/03 13:20:00 EST
730		Tag: leunga-20020303-mlrisc-tools
731
732		Added #[ ... ] expressions to mlrisc tools
733
734		----------------------------------------------------------------------
735		Name: Matthias Blume
736		Date: 2002/02/27 12:29:00 EST
737		Tag: blume-20020227-cdebug
738		Description:
739
740		- made types in structure C and C_Debug to be equal
741		- got rid of code duplication (c-int.sml vs. c-int-debug.sml)
742		- there no longer is a C_Int_Debug (C_Debug is directly derived from C)
743
744		----------------------------------------------------------------------
745		Name: Matthias Blume
746		Date: 2002/02/26 12:00:00 EST
747		Tag: blume-20020226-ffi
748		Description:
749
750		1. Fixed a minor bug in CM's "noweb" tool:
751		If numbering is turned off, then truly don't number (i.e., do not
752		supply the -L option to noweb).  The previous behavior was to supply
753		-L'' -- which caused noweb to use the "default" line numbering scheme.
754		Thanks to Chris Richards for pointing this out (and supplying the fix).
755
756		2. Once again, I reworked some aspects of the FFI:
757
758		A. The incomplete/complete type business:
759
760		- Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
761		gone!
762		- ML types representing an incomplete type are now *equal* to
763		ML types representing their corresponding complete types (just like
764		in C).  This is still safe because ml-nlffigen will not generate
765		RTTI for incomplete types, nor will it generate functions that
766		require access to such RTTI.   But when ML code generated from both
767		incomplete and complete versions of the C type meet, the ML types
768		are trivially interoperable.
769
770		NOTE:  These changes restore the full generality of the translation
771		(which was previously lost when I eliminated functorization)!
772
773		B. Enum types:
774
775		- Structure C now has a type constructor "enum" that is similar to
776		how the "su" constructor works.  However, "enum" is not a phantom
777		type because each "T enum" has values (and is isomorphic to
778		MLRep.Signed.int).
779		- There are generic access operations for enum objects (using
780		MLRep.Signed.int).
781		- ml-nlffigen will generate a structure E_foo for each "enum foo".
782		* The structure contains the definition of type "mlrep" (the ML-side
783		representation type of the enum).  Normally, mlrep is the same
784		as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
785		then mlrep will be defined as a datatype -- thus facilitating
786		pattern matching on mlrep values.
787		("-ec" will be suppressed if there are duplicate values in an
788		enumeration.)
789		* Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
790		will be generated for each C enum constant xxx.
791		* Conversion functions m2i and i2m convert between mlrep and
792		MLRep.Signed.int.  (Without "-ec", these functions are identities.)
793		* Coversion functions c and ml convert between mlrep and "tag enum".
794		* Access functions (get/set) fetch and store mlrep values.
795		- By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
796		enumerations are merged into one single enumeration represented by
797		structure E_'.
798
799		----------------------------------------------------------------------
800		Name: Allen Leung
801		Date: 2002/02/25 04:45:00 EST
802		Tag: leunga-20020225-cps-spill
803
804		This is a new implementation of the CPS spill phase.
805		The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
806		In case of problems, replace it with the old file spill.sml
807
808		The current compiler runs into some serious performance problems when
809		constructing a large record.  This can happen when we try to compile a
810		structure with many items.  Even a very simple structure like the following
811		makes the compiler slow down.
812
813		structure Foo = struct
814		val x_1 = 0w1 : Word32.int
815		val x_2 = 0w2 : Word32.int
816		val x_3 = 0w3 : Word32.int
817		...
818		val x_N = 0wN : Word32.int
819		end
820
821		The following table shows the compile time, from N=1000 to N=4000,
822		with the old compiler:
823
824		N
825		1000   CPS 100 spill                           0.04u  0.00s  0.00g
826		MLRISC ra                               0.06u  0.00s  0.05g
827		(spills = 0 reloads = 0)
828		TOTAL                                   0.63u  0.07s  0.21g
829
830		1100   CPS 100 spill                           8.25u  0.32s  0.64g
831		MLRISC ra                               5.68u  0.59s  3.93g
832		(spills = 0 reloads = 0)
833		TOTAL                                   14.71u  0.99s  4.81g
834
835		1500   CPS 100 spill                           58.55u  2.34s  1.74g
836		MLRISC ra                               5.54u  0.65s  3.91g
837		(spills = 543 reloads = 1082)
838		TOTAL                                   65.40u  3.13s  6.00g
839
840		2000   CPS 100 spill                           126.69u  4.84s  3.08g
841		MLRISC ra                               0.80u  0.10s  0.55g
842		(spills = 42 reloads = 84)
843		TOTAL                                   129.42u  5.10s  4.13g
844
845		3000   CPS 100 spill                           675.59u  19.03s  11.64g
846		MLRISC ra                               2.69u  0.27s  1.38g
847		(spills = 62 reloads = 124)
848		TOTAL                                   682.48u  19.61s  13.99g
849
850		4000   CPS 100 spill                           2362.82u  56.28s  43.60g
851		MLRISC ra                               4.96u  0.27s  2.72g
852		(spills = 85 reloads = 170)
853		TOTAL                                   2375.26u  57.21s  48.00g
854
855		As you can see the old cps spill module suffers from some serious
856		performance problem.  But since I cannot decipher the old code fully,
857		instead of patching the problems up, I'm reimplementing it
858		with a different algorithm.  The new code is more modular,
859		smaller when compiled, and substantially faster
860		(O(n log n) time and O(n) space).  Timing of the new spill module:
861
862		4000  CPS 100 spill                           0.02u  0.00s  0.00g
863		MLRISC ra                               0.25u  0.02s  0.15g
864		(spills=1 reloads=3)
865		TOTAL                                   7.74u  0.34s  1.62g
866
867		Implementation details:
868
869		As far as I can tell, the purpose of the CPS spill module is to make sure the
870		number of live variables at any program point (the bandwidth)
871		does not exceed a certain limit, which is determined by the
872		size of the spill area.
873
874		When the bandwidth is too large, we decrease the register pressure by
875		packing live variables into spill records.  How we achieve this is
876		completely different than what we did in the old code.
877
878		First, there is something about the MLRiscGen code generator
879		that we should be aware of:
880
881		o MLRiscGen performs code motion!
882
883		In particular, it will move floating point computations and
884		address computations involving only the heap pointer to
885		their use sites (if there is only a single use).
886		What this means is that if we have a CPS record construction
887		statement
888
889		RECORD(k,vl,w,e)
890
891		we should never count the new record address w as live if w
892		has only one use (which is often the case).
893
894		We should do something similar to floating point, but the transformation
895		there is much more complex, so I won't deal with that.
896
897		Secondly, there are now two new cps primops at our disposal:
898
899		1. rawrecord of record_kind option
900		This pure operator allocates some uninitialized storage from the heap.
901		There are two forms:
902
903		rawrecord NONE [INT n]  allocates a tagless record of length n
904		rawrecord (SOME rk) [INT n] allocates a tagged record of length n
905		and initializes the tag.
906
907		2. rawupdate of cty
908		rawupdate cty (v,i,x)
909		Assigns to x to the ith component of record v.
910		The storelist is not updated.
911
912		We use these new primops for both spilling and increment record construction.
913
914		1. Spilling.
915
916		This is implemented with a linear scan algorithm (but generalized
917		to trees).  The algorithm will create a single spill record at the
918		beginning of the cps function and use rawupdate to spill to it,
919		and SELECT or SELp to reload from it.  So both spills and reloads
920		are fine-grain operations.  In contrast, in the old algorithm
921		"spills" have to be bundled together in records.
922
923		Ideally, we should sink the spill record construction to where
924		it is needed.  We can even split the spill record into multiple ones
925		at the places where they are needed.  But CPS is not a good
926		representation for global code motion, so I'll keep it simple and
927		am not attempting this.
928
929		2. Incremental record construction (aka record splitting).
930
931		Long records with many component values which are simulatenously live
932		(recall that single use record addresses are not considered to
933		be live) are constructed with rawrecord and rawupdate.
934		We allocate space on the heap with rawrecord first, then gradually
935		fill it in with rawupdate.  This is the technique suggested to me
936		by Matthias.
937
938		Some restrictions on when this is applicable:
939		1. It is not a VECTOR record.  The code generator currently does not handle
940		this case. VECTOR record uses double indirection like arrays.
941		2. All the record component values are defined in the same "basic block"
942		as the record constructor.  This is to prevent speculative
943		record construction.
944
945		----------------------------------------------------------------------
946		Name: Allen Leung
947		Date: 2002/02/22 01:02:00 EST
948		Tag: leunga-20020222-mlrisc-tools
949
950		Minor bug fixes in the parser and rewriter
951
952		----------------------------------------------------------------------
953		Name: Allen Leung
954		Date: 2002/02/21 20:20:00 EST
955		Tag: leunga-20020221-peephole
956
957		Regenerated the peephole files.  Some contained typos in the specification
958		and some didn't compile because of pretty printing bugs in the old version
959		of 'nowhere'.
960
961		----------------------------------------------------------------------
962		Name: Allen Leung
963		Date: 2002/02/19 20:20:00 EST
964		Tag: leunga-20020219-mlrisc-tools
965		Description:
966
967		Minor bug fixes to the mlrisc-tools library:
968
969		1.  Fixed up parsing colon suffixed keywords
970		2.  Added the ability to shut the error messages up
971		3.  Reimplemented the pretty printer and fixed up/improved
972		the pretty printing of handle and -> types.
973		4.  Fixed up generation of literal symbols in the nowhere tool.
974		5.  Added some SML keywords to to sml.sty
975
976		----------------------------------------------------------------------
977		Name: Matthias Blume
978		Date: 2002/02/19 16:20:00 EST
979		Tag: blume-20020219-cmffi
980		Description:
981
982		A wild mix of changes, some minor, some major:
983
984		* All C FFI-related libraries are now anchored under $c:
985		$/c.cm      --> $c/c.cm
986		$/c-int.cm  --> $c/internals/c-int.cm
987		$/memory.cm --> $c/memory/memory.cm
988
989		* "make" tool (in CM) now treats its argument pathname slightly
990		differently:
991		1. If the native expansion is an absolute name, then before invoking
992		the "make" command on it, CM will apply OS.Path.mkRelative
993		(with relativeTo = OS.FileSys.getDir()) to it.
994		2. The argument will be passed through to subsequent phases of CM
995		processing without "going native".  In particular, if the argument
996		was an anchored path, then "make" will not lose track of that anchor.
997
998		* Compiler backends now "know" their respective C calling conventions
999		instead of having to be told about it by ml-nlffigen.  This relieves
1000		ml-nlffigen from one of its burdens.
1001
1002		* The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
1003
1004		* Export C_DEBUG and C_Debug from $c/c.cm.
1005
1006		* C type encoding in ml-nlffi-lib has been improved to model the conceptual
1007		subtyping relationship between incomplete pointers and their complete
1008		counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
1009		with the convention of instantiating 'o with ('t, 'c) obj whenever
1010		the pointer target type is complete.  In the incomplete case, 'o
1011		will be instantiated with some "'c iobj" -- a type obtained by
1012		using one of the functors PointerToIncompleteType or PointerToCompleteType.
1013
1014		Operations that work on both incomplete and complete pointer types are
1015		typed as taking an 'o ptr while operations that require the target to
1016		be known are typed as taking some ('t, 'c) obj ptr.
1017
1018		voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
1019		where void is an eqtype without any values.  This makes it possible
1020		to work on voidptr values using functions meant to operate on light
1021		incomplete pointers.
1022
1023		* As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
1024		been vastly simplified.
1025
1026		----------------------------------------------------------------------
1027		Name: Matthias Blume
1028		Date: 2002/02/19 10:48:00 EST
1029		Tag: blume-20020219-pqfix
1030		Description:
1031
1032		Applied Chris Okasaki's bug fix for priority queues.
1033
1034		----------------------------------------------------------------------
1035		Name: Matthias Blume
1036		Date: 2002/02/15 17:05:00
1037		Tag: Release_110_39
1038		Description:
1039
1040		Last-minute retagging is becoming a tradition... :-(
1041
1042		This is the working release 110.39.
1043
1044		----------------------------------------------------------------------
1045		Name: Matthias Blume
1046		Date: 2002/02/15 16:00:00 EST
1047		Tag: Release_110_39-orig
1048		Description:
1049
1050		Working release 110.39.  New bootfiles.
1051
1052		(Update: There was a small bug in the installer so it wouldn't work
1053		with all shells.  So I retagged. -Matthias)
1054
1055		----------------------------------------------------------------------
1056		Name: Matthias Blume
1057		Date: 2002/02/15 14:17:00 EST
1058		Tag: blume-20020215-showbindings
1059		Description:
1060
1061		Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
1062		the latter can be useful for exploring what bindings are available at
1063		the interactive prompt.  (The first function returns only the list
1064		of symbols that are really bound, the second prints those but also the
1065		ones that CM's autoloading mechanism knows about.)
1066
1067		----------------------------------------------------------------------
1068		Name: Matthias Blume
1069		Date: 2002/02/15 12:08:00 EST
1070		Tag: blume-20020215-iptrs
1071		Description:
1072
1073		Two improvements to ml-nlffigen:
1074
1075		1. Write files only if they do not exist or if their current contents
1076		do not coincide with what's being written.  (That is, avoid messing
1077		with the time stamps unless absolutely necessary.)
1078
1079		2. Implement a "repository" mechanism for generated files related
1080		to "incomplete pointer types".   See the README file for details.
1081
1082		----------------------------------------------------------------------
1083		Name: Matthias Blume
1084		Date: 2002/02/14 11:50:00 EST
1085		Tag: blume-20020214-quote
1086		Description:
1087
1088		Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
1089		because of the new and improved tag generation scheme.  (Thanks to Allen
1090		Leung for pointing it out.)
1091
1092		----------------------------------------------------------------------
1093		Name: Lal George
1094		Date: 2002/02/14 09:55:27 EST 2002
1095		Tag: george-20020214-isabelle-bug
1096		Description:
1097
1098		Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
1099		of Isabelle on the x86.
1100
1101		From Allen:
1102		-----------
1103		I've found the problem:
1104
1105		in ra-core.sml, I use the counter "blocked" to keep track of the
1106		true number of elements in the freeze queue.  When the counter goes
1107		to zero, I skip examining the queue.  But I've messed up the
1108		bookkeeping in combine():
1109
1110		else ();
1111		case !ucol of
1112		PSEUDO => (if !cntv > 0 then
1113		(if !cntu > 0 then blocked := !blocked - 1 else ();
1114		^^^^^^^^^^^^^^^^^^^^^^^
1115		moveu := mergeMoveList(!movev, !moveu)
1116		)
1117		else ();
1118
1119		combine() is called to coalesce two nodes u and v.
1120		I think I was thinking that if the move counts of u and v are both
1121		greater than zero then after they are coalesced then one node is
1122		removed from the freeze queue.  Apparently I was thinking that
1123		both u and v are of low degree, but that's clearly not necessarily true.
1124
1125
1126		02/12/2002:
1127		Here's the patch.  HOL now compiles.
1128
1129		I don't know how this impact on performance (compile
1130		time or runtime).  This bug caused the RA (especially on the x86)
1131		to go thru the potential spill phase when there are still nodes on the
1132		freeze queue.
1133
1134
1135
1136
1137		----------------------------------------------------------------------
1138		Name: Matthias Blume
1139		Date: 2002/02/13 22:40:00 EST
1140		Tag: blume-20020213-fptr-rtti
1141		Description:
1142
1143		Fixed a bug in ml-nlffigen that was introduced with one of the previous
1144		updates.
1145
1146		----------------------------------------------------------------------
1147		Name: Matthias Blume
1148		Date: 2002/02/13 16:41:00 EST
1149		Tag: blume-20020213-cmlpq
1150		Description:
1151
1152		Added new priority queue export symbols (which have just been added to
1153		smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
1154		not compile and the installer would choke.)
1155
1156		----------------------------------------------------------------------
1157		Name: Matthias Blume
1158		Date: 2002/02/13 16:15:00 EST
1159		Tag: blume-20020213-various
1160		Description:
1161
1162		1. More tweaks to ml-nlffigen:
1163
1164		- better internal datastructures (resulting in slight speedup)
1165		- "-match" option requires exact match
1166		- "localized" gensym counters (untagged structs/unions nested within
1167		other structs/unions or within typedefs get a fresh counter; their
1168		tag will be prefixed by a concatenation of their parents' tags)
1169		- bug fixes (related to calculation of transitive closure of types
1170		to be included in the output)
1171
1172		2. Minor Basis updates:
1173
1174		- added implementations for List.collate and Option.app
1175
1176		----------------------------------------------------------------------
1177		Name: Matthias Blume
1178		Date: 2002/02/11 15:55:00 EST
1179		Tag: blume-20020211-gensym
1180		Description:
1181
1182		Added a "-gensym" option to command line of ml-nlffigen.  This can be
1183		used to specify a "stem" -- a string that is inserted in all "gensym'd"
1184		names (ML structure names that correspond to unnamed C structs, unions,
1185		and enums), so that separate runs of ml-nlffigen do not clash.
1186
1187		----------------------------------------------------------------------
1188		Name: Matthias Blume
1189		Date: 2002/02/11 12:05:00 EST
1190		Tag: blume-20020211-gensml
1191		Description:
1192
1193		A quick fix for a problem with GenSML (in the pgraph-util library):
1194		Make generation of toplevel "local" optional.  (Strictly speaking,
1195		signature definitions within "local" are not legal SML.)
1196
1197		Other than that: updates to INSTALL and cm/TODO.
1198
1199		----------------------------------------------------------------------
1200		Name: Matthias Blume
1201		Date: 2002/02/08 15:00:00 EST
1202		Tag: blume-20020208-uniquepid
1203		Description:
1204
1205		0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
1206
1207		1. The installer (config/install.sh) has gotten smarter:
1208
1209		- Configuration options are a bit easier to specify now
1210		(in config/targets).
1211		- Bug in recognizing .tar.bz2 files fixed.
1212		- Installer automatically resolves dependencies between
1213		configuration options (e.g., if you ask for eXene, you will
1214		also get cml -- regardless whether you asked for it or not).
1215		- Installer can run in "quieter mode" by setting the environment
1216		variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
1217		"completely silent", though.
1218		- Build HashCons library as part of smlnj-lib.
1219
1220		2. A new scheme for assigning persistent identifiers to compilation
1221		units (and, by extension, to types etc.) has been put into place.
1222		This fixes a long-standing bug where types and even dynamic values
1223		can get internally confused, thereby compromising type safety
1224		(abstraction) and dynamic correctness.  See
1225
1226		http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
1227
1228		for an example of how things could go wrong until now.
1229
1230		The downside of the new scheme is that pids are not quite as
1231		persistent as they used to be: CM will generate a fresh pid
1232		for every compilation unit that it thinks it sees for the first
1233		time.  That means that if you compile starting from a clean, fresh
1234		source tree at two different times, you end up with different
1235		binaries.
1236
1237		Cutoff recompilation, however, has not been compromised because
1238		CM keeps pid information in special caches between runs.
1239
1240		----------------------------------------------------------------------
1241		Name: Lal George
1242		Date: 2002/02/07 15:34:13 EST 2002
1243		Tag: <none>
1244		Description:
1245
1246		Compilers that generate assembly code may produce  global labels
1247		whose value is resolved at link time. The various peephole optimization
1248		modules did not take this in account.
1249
1250		TODO. The Labels.addrOf function should really return an option
1251		type so that clients are forced to deal with this issue, rather
1252		than an exception being raised.
1253
1254		----------------------------------------------------------------------
1255		Name: Lal George
1256		Date: 2002/02/06 13:55:02 EST
1257		Tag: george-20020206-ra-breakup
1258		Description:
1259
1260		1. A bug fix from Allen.
1261
1262		A typo causes extra fstp %st(0)'s to be generated at compensation
1263		edges, which might cause stack underflow traps at runtime.  This
1264		occurs in fft where there are extraneous fstps right before the 'into'
1265		trap instruction (in this case they are harmless since none of the
1266		integers overflow.)
1267
1268		2. Pulled out various utility modules that were embedded in the modules
1269		of the register allocator. I need these modules for other purposes, but
1270		they are not complete enough to put into a library (just yet).
1271		----------------------------------------------------------------------
1272		Name: Matthias Blume
1273		Date: 2002/01/31 16:05:00 EST
1274		Tag: blume-20020131-sparc-ccalls
1275		Description:
1276
1277		1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
1278		of extra stack space by mistake.  Fixed.
1279
1280		2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
1281
1282		----------------------------------------------------------------------
1283		Name: Allen Leung
1284		Date: 2002/01/30
1285		Tag: leunga-20020130-nowhere-bug-fix
1286		Description:
1287
1288		MLRISC bug fixes:
1289		1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
1290		2. MachineInt.fromString was negating its value.
1291
1292		----------------------------------------------------------------------
1293		Name: Matthias Blume
1294		Date: 2002/01/29
1295		Tag: blume-20020129-INSTALL
1296		Description:
1297
1298		- Added somewhat detailed installation instructions (file INSTALL).
1299		- Fixed curl-detection bug in config/install.sh.
1300		- It is now possible to select the URL getter using the URLGETTER
1301		environment variable:
1302
1303		not set / "unknown"      --> automatic detection (script tries wget,
1304		curl, and lynx)
1305		"wget" / "curl" / "lynx" --> use the specified program (script "knows"
1306		how to properly invoke them)
1307		other                    --> use $URLGETTER directly, it must take
1308		precisely two command-line arguments
1309		(source URL and destination file name)
1310
1311		----------------------------------------------------------------------
1312		Name: Matthias Blume
1313		Date: 2002/01/28
1314		Tag: blume-20020128-sparc-ccalls
1315		Description:
1316
1317		- Fixed problem with calculation of "used" registers in sparc-c-calls.
1318		- Make use of the allocParam argument in sparc-c-calls.
1319
1320		----------------------------------------------------------------------
1321		Name: Matthias Blume
1322		Date: 2002/01/28
1323		Tag: blume-20020128-allocParam
1324		Description:
1325
1326		John Reppy:  Changes c-calls API to accept client-callback for
1327		allocating extra stack space.
1328		me: Corresponding changes to mlriscGen (using a dummy argument that
1329		does not change the current behavior).
1330
1331		----------------------------------------------------------------------
1332		Name: Matthias Blume
1333		Date: 2002/01/28 12:00:00
1334		Tag: Release_110_38
1335		Description:
1336
1337		This time for real!!!
1338
1339		----------------------------------------------------------------------
1340		Name: Matthias Blume
1341		Date: 2002/01/28 10:56:00 EST
1342		Tag: blume-20020128-retraction
1343		Description:
1344
1345		0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
1346		with blume-Release_110_38-retracted.)
1347
1348		1. Fixed a problem with incorrect rounding modes in real64.sml.
1349		(Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
1350
1351		2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
1352		and enums fixed.  The naming of corresponding ML identifiers should
1353		now be consistent again.
1354
1355		----------------------------------------------------------------------
1356		Name: Allen Leung
1357		Date: 2002/01/27
1358		Tag: leunga-20020127-nowhere
1359		Description:
1360
1361		Added a target called nowhere in the configuration scripts.
1362		Enabling this will build the MLRISC 'nowhere' tool (for translating
1363		programs with where-clauses into legal SML code) during installation.
1364
1365		----------------------------------------------------------------------
1366		Name: Matthias Blume
1367		Date: 2002/01/25 21:27:00 EST
1368		Tag: blume-Release_110_38-retracted
1369		Description:
1370
1371		Call it a (working) release!  Version is 110.38. Bootfiles are ready.
1372
1373		README will be added later.
1374
1375		!!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
1376		(Release_110_38) removed.  Reason: Last-minute bug fixes.
1377
1378		----------------------------------------------------------------------
1379		Name: Matthias Blume
1380		Date: 2002/01/25
1381		Tag: blume-20020125-ffi
1382		Description:
1383
1384		A large number of tweaks and improvements to ml-nlffi-lib and
1385		ml-nlffigen:
1386
1387		- ML represenation types have been streamlined
1388		- getter and setter functions work with concrete values, not abstract
1389		ones where possible
1390		- ml-nlffigen command line more flexible (see README file there)
1391		- some bugs have been fixed (hopefully)
1392
1393		----------------------------------------------------------------------
1394		Name: Lal George
1395		Date: 2002/01/24
1396		Tag: george-20020124-risc-ra-interface
1397		Description:
1398
1399		There is a dramatic simplification in the interface to the
1400		register allocator for RISC architectures as a result of making
1401		parallel copy instructions explicit.
1402
1403		----------------------------------------------------------------------
1404		Name: Matthias Blume
1405		Date: 2002/01/22
1406		Tag: blume-20020122-x86-ccalls
1407		Description:
1408
1409		Bug fix for c-calls on x86 (having to do with how char- and
1410		short-arguments are being handled).
1411
1412		----------------------------------------------------------------------
1413		Name: Matthias Blume
1414		Date: 2002/01/21
1415		Tag: blume-20020121-ff
1416		Description:
1417
1418		Another day of fiddling with the FFI...
1419
1420		1. Bug fix/workaround:  CKIT does not complain about negative array
1421		dimensions, so ml-nlffigen has to guard itself against this possibility.
1422		(Otherwise a negative dimension would send it into an infinite loop.)
1423
1424		2. Some of the abstract types (light objects, light pointers, most "base"
1425		types) in structure C are now eqtypes.
1426
1427		3. Added constructors and test functions for NULL function pointers.
1428
1429		----------------------------------------------------------------------
1430		Name: Matthias Blume
1431		Date: 2002/01/18
1432		Tag: blume-20020118-ready-for-new-release
1433		Description:
1434
1435		Made config/srcarchiveurl point to a new place.  (Will provide boot
1436		files shortly.)
1437
1438		Maybe we christen this to be 110.38?
1439
1440		----------------------------------------------------------------------
1441		Name: Matthias Blume
1442		Date: 2002/01/18
1443		Tag: blume-20020118-more-ffifiddle
1444		Description:
1445
1446		Today's FFI fiddling:
1447
1448		- Provided a structure CGetSet with "convenient" versions of C.Get.* and
1449		C.Set.* that use concrete (MLRep.*) arguments and results instead
1450		of abstract ones.
1451
1452		- Provided word-style bit operations etc. for "int" representation
1453		types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
1454		and Long.
1455
1456		----------------------------------------------------------------------
1457		Name: Matthias Blume
1458		Date: 2002/01/18
1459		Tag: blume-20020118-use-x86-fp
1460		Description:
1461
1462		Now that x86-fast-fp seems to be working, I turned it back on again
1463		by default.  (Seems to work fine now, even with the FFI.)
1464
1465		Other than that, I added some documentation about the FFI to
1466		src/ml-nlffigen/README and updated the FFI test examples in
1467		src/ml-nlffi-lib/Tests/*.
1468
1469		----------------------------------------------------------------------
1470		Name: Allen Leung
1471		Date: 2002/01/17
1472		Tag: leunga-20020117-x86-fast-fp-call
1473		Description:
1474
1475		1. Fixed a problem with handling return fp values when x86's fast fp
1476		mode is turned on.
1477
1478		2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
1479		of %f32.
1480
1481		3. Added a constructor INT32lit to the ast of MLRISC tools.
1482
1483		----------------------------------------------------------------------
1484		Name: Matthias Blume
1485		Date: 2002/01/16
1486		Tag: blume-20020116-ffifiddle
1487		Description:
1488
1489		More fiddling with the FFI interface:
1490
1491		- Make constness 'c instead of rw wherever possible.  This eliminates
1492		the need for certain explicit coercions.  (However, due to ML's
1493		value polymorphism, there will still be many cases where explicit
1494		coercions are necessary.  Phantom types are not the whole answer
1495		to modeling a subtyping relationship in ML.)
1496
1497		- ro/rw coersions for pointers added.  (Avoids the detour through */&.)
1498
1499		- "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
1500		clumsy workaround for varargs problem.)
1501
1502		----------------------------------------------------------------------
1503		Name: Lal George
1504		Date: 2002/01/15
1505		Tag: <none>
1506		Description:
1507
1508		1. Since COPY instructions are no longer native to the architecture,
1509		a generic functor can be used to implement the expandCopies function.
1510
1511		2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
1512		TEXT segment.
1513
1514		----------------------------------------------------------------------
1515		Name: Matthias Blume
1516		Date: 2002/01/15
1517		Tag: blume-20020115-ffiupdates
1518		Description:
1519
1520		1. Fix for bug resulting in single-precision float values being returned
1521		incorrectly from FFI calls.
1522
1523		2. Small modifications to C FFI API:
1524
1525		- memory-allocation routines return straight objects (no options)
1526		and raise an exception in out-of-memory situations
1527		- unsafe extensions to cast between function pointers and pointers
1528		from/to ints
1529		- added structure C_Debug as an alternative to structure C where
1530		pointer-dereferencing (|*| and |*!) always check for null-pointers
1531		- added open_lib' to DynLinkage;  open_lib' works like open_lib
1532		but also takes a (possibly empty) list of existing library handles
1533		that the current library depends on
1534
1535		----------------------------------------------------------------------
1536		Name: Matthias Blume
1537		Date: 2002/01/10
1538		Tag: blume-20020110-newffigen
1539		Description:
1540
1541		1. Updates to portable graph code.
1542
1543		2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
1544		more scalable now so that even huge interfaces such as the one
1545		for GTK compile in finite time and space. :-)
1546		See src/ml-nlffigen/README for details on what's new.
1547
1548		----------------------------------------------------------------------
1549		Name: Lal George
1550		Date: 2001/01/09 14:31:35 EST 2002
1551		Tag: george-20011206-rm-native-copy
1552		Description:
1553
1554		Removed the native COPY and FCOPY instructions
1555		from all the architectures and replaced it with the
1556		explicit COPY instruction from the previous commit.
1557
1558		It is now possible to simplify many of the optimizations
1559		modules that manipulate copies. This has not been
1560		done in this change.
1561
1562		----------------------------------------------------------------------
1563		Name: Lal George
1564		Date: 2001/12/06 16:50:13 EST 2001
1565		Tag: george-20011206-mlrisc-instruction
1566		Description:
1567
1568		Changed the representation of instructions from being fully abstract
1569		to being partially concrete. That is to say:
1570
1571		from
1572		type instruction
1573
1574		to
1575		type instr                              (* machine instruction *)
1576
1577		datatype instruction =
1578		LIVE of {regs: C.cellset, spilled: C.cellset}
1579		| KILL of {regs: C.cellset, spilled: C.cellset}
1580		| COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
1581		| ANNOTATION of {i: instruction, a: Annotations.annotation}
1582		| INSTR of instr
1583
1584		This makes the handling of certain special instructions that appear on
1585		all architectures easier and uniform.
1586
1587		LIVE and KILL say that a list of registers are live or killed at the
1588		program point where they appear. No spill code is generated when an
1589		element of the 'regs' field is spilled, but the register is moved to
1590		the 'spilled' (which is present, more for debugging than anything else).
1591
1592		LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
1593		We used to generate:
1594
1595		DEFFREG f1
1596		f1 := f2 + f3
1597		trapb
1598
1599		but now generate:
1600
1601		f1 := f2 + f3
1602		trapb
1603		LIVE {regs=[f1,f2,f3], spilled=[]}
1604
1605		Furthermore, the DEFFREG (hack) required that all floating point instruction
1606		use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
1607		defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
1608		in a cleaner alpha implementation. (Hopefully, intel will not get rid of
1609		this architecture).
1610
1611		COPYXXX is intended to replace the parallel COPY and FCOPY  available on
1612		all the architectures. This will result in further simplification of the
1613		register allocator that must be aware of them for coalescing purposes, and
1614		will also simplify certain aspects of the machine description that provides
1615		callbacks related to parallel copies.
1616
1617		ANNOTATION should be obvious, and now INSTR represents the honest to God
1618		machine instruction set!
1619
1620		The <arch>/instructions/<arch>Instr.sml files define certain utility
1621		functions for making porting easier -- essentially converting upper case
1622		to lower case. All machine instructions (of type instr) are in upper case,
1623		and the lower case form generates an MLRISC instruction. For example on
1624		the alpha we have:
1625
1626		datatype instr =
1627		LDA of {r:cell, b:cell, d:operand}
1628		| ...
1629
1630		val lda : {r:cell, b:cell, d:operand} -> instruction
1631		...
1632
1633		where lda is just (INSTR o LDA), etc.
1634
1635		----------------------------------------------------------------------
1636		Name: Matthias Blume
1637		Date: 2001/11/22 21:40:00 EST
1638		Tag: Release_110_37
1639		Description:
1640
1641		Release 110.37.  This time for real.
1642
1643		----------------------------------------------------------------------
1644		Name: Matthias Blume
1645		Date: 2001/11/21 16:35:00 EST
1646		Tag: blume-20011121-foot-in-mouth
1647		Description:
1648
1649		Removed the "Release_110_37" tag because of a serious bug.
1650		This will be re-tagged once the bug is fixed.
1651
1652		----------------------------------------------------------------------
1653		Name: Matthias Blume
1654		Date: 2001/11/21 16:14:00 EST
1655		Tag: blume-20011121-forgottenfile
1656		Description:
1657
1658		Forgot to add a file.  (Just a .tex-file -- part of
1659		the CM manual source.)
1660
1661		----------------------------------------------------------------------
1662		Name: Matthias Blume
1663		Date: 2001/11/21 16:10:00 EST
1664		Tag: blume-20011121-invalid_110_37
1665		Description:
1666
1667		Note: I removed the original tag "Release_110_37" from this commit
1668		because we found a serious bug in all non-x86 backends.
1669		- Matthias
1670
1671		1. Modifications to the SML/NJ code generator and to the runtime system
1672		so that code object name strings are directly inserted into code
1673		objects at code generation time.  The only business the runtime system
1674		has with this is now to read the name strings on occasions.
1675		(The encoding of the name string has also changed somewhat.)
1676
1677		2. CM now implements a simple "set calculus" for specifying export lists.
1678		In particular, it is now possible to refer to the export lists of
1679		other libraries/groups/sources and form unions as well as differences.
1680		See the latest CM manual for details.
1681
1682		3. An separate notion of "proxy" libraries has again be eliminated from
1683		CM's model.  (Proxy libraries are now simply a special case of using
1684		the export list calculus.)
1685
1686		4. Some of the existing libraries now take advantage of the new set
1687		calculus.
1688		(Notice that not all libraries have been converted because some
1689		of the existing .cm-files are supposed to be backward compatible
1690		with 110.0.x.)
1691
1692		5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
1693		"General.exnMessage"!  The former relies on a certain hook to be
1694		initialized, and that often does not happen in the stand-alone case.)
1695
1696		----------------------------------------------------------------------
1697		Name: Lal George
1698		Date: 2001/11/21  13:56:18 EST
1699		Tag: george-2001121-pseudo-ops
1700		Description:
1701
1702		Implemented a complete redesign of MLRISC pseudo-ops. Now there
1703		ought to never be any question of incompatabilities with
1704		pseudo-op syntax expected by host assemblers.
1705
1706		For now, only modules supporting GAS syntax are implemented
1707		but more should follow, such as MASM, and vendor assembler
1708		syntax, e.g. IBM as, Sun as, etc.
1709
1710		----------------------------------------------------------------------
1711		Name: Matthias Blume
1712		Date: 2001/11/14 11:52:00 EST
1713		Tag: blume-20011114-srcname
1714		Description:
1715
1716		1. Routed the name of the current source file to mlriscgen where it
1717		should be directly emitted into the code object.  (This last part
1718		is yet to be done.)
1719
1720		2. Some cleanup of the pgraph code to make it match the proposal that
1721		I put out the other day.  (The proposal notwithstanding, things are
1722		still in flux here.)
1723
1724		----------------------------------------------------------------------
1725		Name: Lal George
1726		Date: 2001/11/14 09:44:04 EST
1727		Tag:
1728		Description:
1729
1730		Fix for a backpatching bug reported by Allen.
1731
1732		Because the boundary between short and long span-dependent
1733		instructions is +/- 128, there are an astounding number of
1734		span-dependent instructions whose size is over estimated.
1735
1736		Allen came up with the idea of letting the size of span
1737		dependent instructions be non-monotonic, for a maxIter
1738		number of times, after which the size must be monotonically
1739		increasing.
1740
1741		This table shows the number of span-dependent instructions
1742		whose size was over-estimated as a function of maxIter, for the
1743		file Parse/parse/ml.grm.sml:
1744
1745		maxIter            # of instructions:
1746		10                      687
1747		20                      438
1748		30                      198
1749		40                        0
1750
1751		In compiling the compiler, there is no significant difference in
1752		compilation speed between maxIter=10 and maxIter=40. Actually,
1753		my measurements showed that maxIter=40 was a tad faster than
1754		maxIter=10! Also 96% of the  files in the compiler reach a fix
1755		point within 13 iterations, so fixing maxIter at 40, while high,
1756		is okay.
1757
1758		----------------------------------------------------------------------
1759		Name: Matthias Blume
1760		Date: 2001/10/31 15:25:00 EST
1761		Tag: blume-20011031-pgraph
1762		Description:
1763
1764		CKIT:
1765		* Changed the "Function" constructor of type Ast.ctype to carry optional
1766		argument identifiers.
1767		* Changed the return type of TypeUtil.getFunction accordingly.
1768		* Type equality ignores the argument names.
1769		* TypeUtil.composite tries to preserve argument names but gives up quickly
1770		if there is a mismatch.
1771
1772		installation script:
1773		* attempts to use "curl" if available (unless "wget" is available as well)
1774
1775		CM:
1776		* has an experimental implementation of "portable graphs" which I will
1777		soon propose as an implementation-independent library format
1778		* there are also new libraries $/pgraph.cm and $/pgraph-util.cm
1779
1780		NLFFI-LIB:
1781		* some cleanup (all cosmetic)
1782
1783		NLFFIGEN:
1784		* temporarily disabled the mechanism that suppresses ML output for
1785		C definitions whose identifiers start with an underscore character
1786		* generate val bindings for enum constants
1787		* user can request that only one style (light or heavy) is being used;
1788		default is to use both (command-line arguments: -heavy and -light)
1789		* fixed bug in handling of function types involving incomplete pointers
1790		* generate ML entry points that take record arguments (i.e., using
1791		named arguments) for C functions that have a prototype with named
1792		arguments
1793		(see changes to CKIT)
1794
1795		----------------------------------------------------------------------
1796		Name: Allen Leung
1797		Date: 2001/10/27 20:34:00 EDT
1798		Tag: leunga-20011027-x86-fast-fp-call
1799		Description:
1800
1801		Fixed the bug described in blume-20010920-slowfp.
1802
1803		The fix involves
1804		1. generating FCOPYs in FSTP in ia32-svid
1805		2. marking a CALL with the appropriate annotation
1806
1807		----------------------------------------------------------------------
1808		Name: Matthias Blume
1809		Date: 2001/10/16 11:32:00 EDT
1810		Tag: blume-20011016-netbsd
1811		Description:
1812
1813		Underscore patch from Chris Richards (fixing problem with compiling
1814		runtime system under recent NetBSD).
1815
1816		----------------------------------------------------------------------
1817		Name: Allen Leung
1818		Date: 2001/10/12 17:18:32 EDT 2001
1819		Tag: leung-20011012-x86-printflowgraph
1820		Description:
1821
1822		X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
1823
1824		----------------------------------------------------------------------
1825		Name: Lal George
1826		Date: 2001/10/11 23:51:34 EDT
1827		Tag: george-20011011-too-many-instrs
1828		Description:
1829
1830		The representation of a program point never expected to see more
1831		than 65536 instructions in a basic block!
1832
1833		----------------------------------------------------------------------
1834		Name: Lal George
1835		Date: 2001/10/09 09:41:37 EDT
1836		Tag: george-20011008-mlrisc-labels
1837		Description:
1838
1839		Changed the machine description files to support printing of
1840		local and global labels in assembly code, based on host assembler
1841		conventions.
1842
1843		----------------------------------------------------------------------
1844		Name: Matthias Blume
1845		Date: 2001/09/25 15:25:00 EDT
1846		Tag: blume-20010925-exninfo
1847		Description:
1848
1849		I provided a non-hook implementation of exnName (at the toplevel) and
1850		made the "dummy" implementation of exnMessage (at the toplevel) more
1851		useful: if nothing gets "hooked in", then at least you are going to
1852		see the exception name and a message indicating why you don't see more.
1853
1854		[For the time being, programs that need exnMessage and want to use
1855		ml-build should either use General.exnMessage (strongly recommended) or
1856		refer to structure General at some other point so that CM sees a
1857		static dependency.]
1858
1859		[Similar remarks go for "print" and "use":  If you want to use their
1860		functionality in stand-alone programs generated by ml-build, then use
1861		TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
1862
1863		----------------------------------------------------------------------
1864		Name: Matthias Blume
1865		Date: 2001/09/20 17:28:00 EDT
1866		Tag: blume-20010920-slowfp
1867		Description:
1868
1869		Allen says that x86-fast-fp is not safe yet, so I turned it off again...
1870
1871		----------------------------------------------------------------------
1872		Name: Matthias Blume
1873		Date: 2001/09/20 17:20:00 EDT
1874		Tag: blume-20010920-canonicalpaths
1875		Description:
1876
1877		0. Updated the BOOT file (something that I forgot to do earlier).
1878
1879		1. Small internal change to CM so that it avoids "/../" in filenames
1880		as much as possible (but only where it is safe).
1881
1882		2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
1883		that arguments that contain delimiters are passed through correctly.
1884		This change also means that all "special" arguments of the form
1885		@SMLxxx... must come first.
1886
1887		3. Changed install script to put relative anchor names for tool commands
1888		into pathconfig.
1889
1890		----------------------------------------------------------------------
1891		Name: Matthias Blume
1892		Date: 2001/09/18 15:35:00 EDT
1893		Tag: blume-20010918-readme11036
1894		Description:
1895
1896		Added README files.
1897
1898		----------------------------------------------------------------------
1899		Name: Matthias Blume
1900		Date: 2001/09/18 11:45:00 EDT
1901		Tag: Release_110_36 (retag)
1902		Description:
1903
1904		Fixed mistake in config/preloads. Retagged as 110.36.
1905
1906		----------------------------------------------------------------------
1907		Name: Matthias Blume
1908		Date: 2001/09/18 09:40:00 EDT
1909		Tag: Release_110_36_orig (tag changed)
1910		Description:
1911
1912		New version (110.36).  New bootfiles.
1913
1914		----------------------------------------------------------------------
1915		Name: Matthias Blume
1916		Date: 2001/09/14 16:15:00 EDT
1917		Tag: blume-20010914-x86fastfp
1918		Description:
1919
1920		John committed some changes that Allen made, in particular a (hopefully)
1921		correctly working version of the x86-fp module.
1922
1923		I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
1924		flag to "true".  Everything seems to compile to a fixpoint ok, and
1925		"mandelbrot" speeds up by about 15%.
1926
1927		----------------------------------------------------------------------
1928		Name: Matthias Blume
1929		Date: 2001/09/13 11:20:00 EDT
1930		Tag: blume-20010913-minimal
1931		Description:
1932
1933		1. Stefan Monnier's patch to fix a miscompilation problem that
1934		was brought to light by John Reppy's work on Moby.
1935
1936		2. Implemented a minimal "structure Compiler" that contains just
1937		"version" and "architecture".  The minimal version will be
1938		available when the full version is not.  This is for backward-
1939		compatibility with code that wants to test Compiler.version.
1940
1941		----------------------------------------------------------------------
1942		Name: Matthias Blume
1943		Date: 2001/08/28 14:03:00 EDT
1944		Tag: blume-20010828-ml-lex
1945		Description:
1946
1947		Fix for bug 1581, received from Neophytos Michael.
1948
1949		----------------------------------------------------------------------
1950		Name: Matthias Blume
1951		Date: 2001/08/27 11:20:00 EDT
1952		Tag: blume-20010827-readme11035
1953		Description:
1954
1955		Fleshed out the README file for 110.35.
1956
1957		----------------------------------------------------------------------
1958		Name: Matthias Blume
1959		Date: 2001/08/24 17:10:00 EDT
1960		Tag: Release_110_35
1961		Description:
1962
1963		New version number (110.35).  New bootfiles.
1964
1965		----------------------------------------------------------------------
1966		Name: Lal George
1967		Date: 2001/08/24 13:47:18 EDT 2001
1968		Tag: george-20010824-MLRISC-graphs
1969		Description:
1970
1971		removed clusters from MLRISC completely and replaced with graphs.
1972
1973		----------------------------------------------------------------------
1974		Name: Matthias Blume
1975		Date: 2001/08/23 17:50:00 EDT
1976		Tag: blume-20010823-toplevel
1977		Description:
1978
1979		- some reorganization of the code that implements various kinds of
1980		environments in the compiler (static, dynamic, symbolic, combined)
1981		- re-implemented the EnvRef module so that evalStream works properly
1982		(if the stream contains references to "use", "CM.make", etc.)
1983		- cleaned up evalloop.sml and interact.sml (but they need more cleaning)
1984
1985		----------------------------------------------------------------------
1986		Name: Matthias Blume
1987		Date: 2001/08/20 15:50 EDT
1988		Tag: blume20010820-slipup
1989		Description:
1990
1991		I forgot to commit a few files.  Here they are...
1992
1993		----------------------------------------------------------------------
1994		Name: Matthias Blume
1995		Date: 2001/08/20 15:35:00 EDT
1996		Tag: blume-20010820-debugprof
1997		Description:
1998
1999		!!!! NEW BOOTFILES !!!!
2000
2001		This is another round of reorganizing the compiler sources.  This
2002		time the main goal was to factor out all the "instrumentation"
2003		passes (for profiling and backtracing) into their own library.
2004		The difficulty was to do it in such a way that it does not depend
2005		on elaborate.cm but only on elabdata.cm.
2006
2007		Therefore there have been further changes to both elaborate.cm and
2008		elabdata.cm -- more "generic" things have been moved from the former
2009		to the latter.  As a result, I was forced to split the assignment
2010		of numbers indicating "primtyc"s into two portions: SML-generic and
2011		SML/NJ-specific.  Since it would have been awkward to maintain,
2012		I bit the bullet and actually _changed_ the mapping between these
2013		numbers and primtycs.  The bottom line of this is that you need
2014		a new set of bin- and bootfiles.
2015
2016		I have built new bootfiles for all architectures, so doing a fresh
2017		checkout and config/install.sh should be all you need.
2018
2019		The newly created library's name is
2020
2021		$smlnj/viscomp/debugprof.cm
2022
2023		and its sources live under
2024
2025		src/compiler/DebugProf
2026
2027		----------------------------------------------------------------------
2028		Name: Matthias Blume
2029		Date: 2001/08/15 17:15:00 EDT
2030		Tag: blume-20010815-compreorg
2031		Description:
2032
2033		This is a first cut at reorganizing the CM libraries that make up the
2034		core of the compiler.  The idea is to separate out pieces that could
2035		be used independently by tools, e.g., the parser, the typechecker, etc.
2036
2037		The current status is a step in this direction, but it is not quite
2038		satisfactory yet.  Expect more changes in the future.
2039
2040		Here is the current (new) organization...
2041
2042		What used to be $smlnj/viscomp/core.cm is now divided into
2043		six CM libraries:
2044
2045		$smlnj/viscomp/basics.cm
2046		/parser.cm
2047		/elabdata.cm
2048		/elaborate.cm
2049		/execute.cm
2050		/core.cm
2051
2052		The CM files for these libraries live under src/system/smlnj/viscomp.
2053		All these libraries are proxy libraries that contain precisely
2054		one CM library component.  Here are the locations of the components
2055		(all within the src/compiler tree):
2056
2057		Basics/basics.cm
2058		Parse/parser.cm
2059		ElabData/elabdata.cm
2060		Elaborator/elaborate.cm
2061		Execution/execute.cm
2062		core.cm
2063
2064		[This organization is the same that has been used already
2065		for a while for the architecture-specific parts of the visible
2066		compiler and for the old version of core.cm.]
2067
2068		As you will notice, many source files have been moved from their
2069		respective original locations to a new home in one of the above
2070		subtrees.
2071
2072		The division of labor between the new libraries is the following:
2073
2074		basics.cm:
2075		- Simple, basic definitions that pertain to many (or all) of
2076		the other libraries.
2077		parser.cm:
2078		- The SML parser, producing output of type Ast.dec.
2079		- The type family for Ast is also defined and exported here.
2080		elabdata.cm:
2081		- The datatypes that describe input and output of the elaborator.
2082		This includes types, absyn, and static environments.
2083		elaborator.cm:
2084		- The SML/NJ type checker and elaborator.
2085		This maps an Ast.dec (with a given static environment) to
2086		an Absyn.dec (with a new static environment).
2087		- This libraries implements certain modules that used to be
2088		structures as functors (to remove dependencies on FLINT).
2089		execute.cm:
2090		- Everything having to do with executing binary code objects.
2091		- Dynamic environments.
2092		core.cm:
2093		- SML/NJ-specific instantiations of the elaborator and MLRISC.
2094		- Top-level modules.
2095		- FLINT (this should eventually become its own library)
2096
2097		Notes:
2098
2099		I am not 100% happy with the way I separated the elaborator (and its
2100		data structures) from FLINT.  Two instances of the same problem:
2101
2102		1. Data structures contain certain fields that carry FLINT-specific
2103		information.  I hacked around this using exn and the property list
2104		module from smlnj-lib.  But the fact that there are middle-end
2105		specific fields around at all is a bit annoying.
2106
2107		2. The elaborator calculates certain FLINT-related information.  I tried
2108		to make this as abstract as I could using functorization, but, again,
2109		the fact that the elaborator has to perform calculations on behalf
2110		of the middle-end at all is not nice.
2111
2112		3. Having to used exn and property lists is unfortunate because it
2113		weakens type checking.  The other alternative (parameterizing
2114		nearly *everything*) is not appealing, though.
2115
2116		I removed the "rebinding =" warning hack because due to the new organization
2117		it was awkward to maintain it.  As a result, the compiler now issues some of
2118		these warnings when compiling init.cmi during bootstrap compilation. On
2119		the plus side, you also get a warning when you do, for example:
2120		val op = = Int32.+
2121		which was not the case up to now.
2122
2123		I placed "assign" and "deref" into the _Core structure so that the
2124		code that deals with the "lazy" keyword can find them there.  This
2125		removes the need for having access to the primitive environment
2126		during elaboration.
2127
2128		----------------------------------------------------------------------
2129		Name: Matthias Blume
2130		Date: 2001/08/13
2131		Tag: blume-20010813-closures
2132		Description:
2133
2134		This fix was sent to us by Zhong Shao.  It is supposed to improve the
2135		performance of certain loops by avoiding needless closure allocation.
2136
2137		----------------------------------------------------------------------
2138		Name: Lal George
2139		Date: 2001/07/31 10:03:23 EDT 2001
2140		Tag: george-20010731-x86-fmalloc
2141		Description: Fixed bug in x86 calls
2142
2143		There was a bug where call instructions would mysteriously
2144		vanish. The call instruction had to be one that returned
2145		a floating point value.
2146
2147		----------------------------------------------------------------------
2148		Name: Lal George
2149		Date: 2001/07/19 16:36:29 EDT 2001
2150		Tag: george-20010719-simple-cells
2151		Description:
2152
2153		I have dramatically simplified the interface for CELLS in MLRISC.
2154
2155		In summary, the cells interface is broken up into three parts:
2156
2157		1. CellsBasis : CELLS_BASIS
2158
2159		CellsBasis is a top level structure and common for all
2160		architectures.  it contains the definitions of basic datatypes
2161		and utility  functions over these types.
2162
2163		2. functor Cells() : CELLS
2164
2165		Cells generates an interface for CELLS that incorporates the
2166		specific resources on the target architecture, such as the
2167		presence of special register classes, their number and size,
2168		and various useful substructures.
2169
2170		3. <ARCH>CELLS
2171
2172		e.g. SparcCells: SPARCCELLS
2173
2174		<ARCH>CELLS usually contains additional bindings for special
2175		registers  on the architecture, such as:
2176
2177		val r0 : cell           (* register zero *)
2178		val y : cell            (* Y register *)
2179		val psr : cell          (* processor status register *)
2180		...
2181
2182		The structure returned by applying the Cells functor is opened
2183		in this interface.
2184
2185		The main implication of all this is that the datatypes for cells is
2186		split between CellsBasis and CELLS -- a fairly simple change for user
2187		code.
2188
2189		In the old scheme the CELLS interface had a definitional binding of
2190		the form:
2191
2192		signature CELLS = sig
2193
2194		structure CellsBasis = CellsBasis
2195
2196		...
2197
2198		end
2199
2200		With all the sharing constraints that goes on in MLRISC, this old
2201		design  quickly leads to errors such as:
2202
2203		"structure definition spec inside of sharing ... "
2204
2205
2206		and appears to require an unacceptable amount of sharing and where
2207		constraint hackery.
2208
2209		I think this error message (the interaction of definitional specs and
2210		sharing) requires more explanation on our web page.
2211
2212		----------------------------------------------------------------------
2213		Name: Matthias Blume
2214		Date: 2001/07/19 15:00:00 EDT
2215		Tag: blume-20010719-libreorg
2216		Description:
2217
2218		This update puts together a fairly extensive but straightforward change
2219		to the way the libraries that implement the interactive system are
2220		organized:
2221
2222		The biggest change is the elimination of structure Compiler.  As a
2223		replacement for this structure, there is now a CM library
2224		(known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
2225		that exports all the substructures of the original structure Compiler
2226		directly.  So instead of saying Compiler.Foo.bar one now simply
2227		says Foo.bar.  (The CM libraries actually export a collection of
2228		structures that is richer than the collection of substructures of
2229		structure Compiler.)
2230
2231		To make the transition smooth, there is a separate library called
2232		$smlnj/compiler/compiler.cm which puts together and exports the
2233		original structure Compiler (or at least something very close to it).
2234
2235		There are five members of the original structure Compiler
2236		that are not exported directly but which instead became members
2237		of a new structure Backend (described by signature BACKEND).  These are:
2238		structure Profile (: PROFILE), structure Compile (: COMPILE), structure
2239		Interact (: INTERACT), structure Machine (: MACHINE), and val
2240		architecture (: string).
2241
2242		Structure Compiler.Version has become structure CompilerVersion.
2243
2244		Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
2245		by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
2246		or x86, respectively.
2247		Each of these exports the same frontend structures that
2248		$smlnj/compiler.cm exports.  But they do not have a structure Backend
2249		and instead export some structure <Arch>Backend where <Arch> is Alpha32,
2250		Hppa, PPC, Sparc, or X86, respectively.
2251
2252		Library $smlnj/compiler/all.cm exports the union of the exports of
2253		$smlnj/compiler/<arch>.cm
2254
2255		There are no structures <Arch>Compiler anymore, use
2256		$smlnj/compiler/<arch>.cm instead.
2257
2258		Library host-compiler-0.cm is gone.  Instead, the internal library
2259		that instantiates CM is now called cm0.cm.  Selection of the host
2260		compiler (backend) is no longer done here but. (Responsibility for it
2261		now lies with $smlnj/compiler/current.cm.  This seems to be more
2262		logical.)
2263
2264		Many individual files have been moved or renamed.  Some files have
2265		been split into multiple files, and some "dead" files have been deleted.
2266
2267		Aside from these changes to library organization, there are also changes
2268		to the way the code itself is organized:
2269
2270		Structure Binfile has been re-implemented in such a way that it no
2271		longer needs any knowledge of the compiler.  It exclusively deals
2272		with the details of binfile layout.  It no longer invokes the
2273		compiler (for the purpose of creating new prospective binfile
2274		content), and it no longer has any knowledge of how to interpret
2275		pickles.
2276
2277		Structure Compile (: COMPILE) has been stripped down to the bare
2278		essentials of compilation.  It no longer deals with linking/execution.
2279		The interface has been cleaned up considerably.
2280
2281		Utility routines for dealing with linking and execution have been
2282		moved into their own substructures.
2283
2284		(The ultimate goal of these changes is to provide a light-weight
2285		binfile loader/linker (at least for, e.g., stable libraries) that
2286		does not require CM or the compiler to be present.)
2287
2288		CM documentation has been updated to reflect the changes to library
2289		organization.
2290
2291		----------------------------------------------------------------------
2292		Name: Matthias Blume
2293		Date: 2001/07/10 17:30:00 EDT
2294		Tag: Release_110_34
2295		Description:
2296
2297		Minor tweak to 110.34 (re-tagged):
2298
2299		- README.html file added to CVS repository
2300		- runtime compiles properly under FreeBSD 3.X and 4.X
2301
2302		----------------------------------------------------------------------
2303		Name: Matthias Blume
2304		Date: 2001/07/10 17:30:00 EDT
2305		Tag: Release_110_34
2306		Description:
2307
2308		New version number (110.34). New bootfiles.
2309
2310		----------------------------------------------------------------------
2311		Name: Matthias Blume
2312		Date: 2001/07/09 16:00:00 EDT
2313		Tag: blume-20010709-more-varargs
2314		Description:
2315
2316		I changed the handling of varargs in ml-nlffigen again:
2317		The ellipsis ... will now simply be ignored (with an accompanying warning).
2318
2319		The immediate effect is that you can actually call a varargs function
2320		from ML -- but you can't actually supply any arguments beyond the ones
2321		specified explicitly.  (For example, you can call printf with its format
2322		string, but you cannot pass additional arguments.)
2323
2324		This behavior is only marginally more useful than the one before, but
2325		it has the advantage that a function or, more importantly, a function
2326		type never gets dropped on the floor, thus avoiding follow-up problems with
2327		other types that refer to the offending one.
2328
2329		----------------------------------------------------------------------
2330		Name: Matthias Blume
2331		Date: 2001/07/09 11:25:00 EDT
2332		Tag: blume-20010709-varargs
2333		Description:
2334
2335		1. ckit-lib.cm now exports structure Error
2336		2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
2337		with a warning accompanied by a source location.  Moreover, it
2338		merely skips the offending function or type and proceeds with the
2339		rest of its work.u  As a result, one can safely feed C code containing
2340		"..." to ml-nlffigen.
2341		3. There are some internal improvements to CM, providing slightly
2342		more general string substitutions in the tools subsystem.
2343
2344		----------------------------------------------------------------------
2345		Name: Matthias Blume
2346		Date: 2001/06/27 15:10:00 EDT
2347		Tag: blume-20010627-concur
2348		Description:
2349
2350		Fixed a small bug in CM's handling of parallel compilation.
2351		(You could observe the bug by Control-C-interrupting an ordinary
2352		CMB.make or CM.stabilize and then attaching some compile servers.
2353		The result was that all of a sudden the previously interrupted
2354		compilation would continue on its own.  This was because of
2355		an over-optimization: CM did not bother to clean out certain queues
2356		when no servers were attached "anyway", resulting in the contents
2357		of these queues to grab control when new servers did get attached.)
2358
2359		There is also another minor update to the CM manual.
2360
2361		----------------------------------------------------------------------
2362		Name: Matthias Blume
2363		Date: 2001/06/26 16:15:00 EDT
2364		Tag: blume-20010626-cmdoc
2365		Description:
2366
2367		Minor typo fixed in CM manual (syntax diagram for libraries).
2368
2369		----------------------------------------------------------------------
2370		Name: Matthias Blume
2371		Date: 2001/06/25 22:55:00 EDT
2372		Tag: blume-20010625-x86pc
2373		Description:
2374
2375		Fixed a nasty bug in the X86 assembly code that caused signal
2376		handlers to fail (crash) randomly.
2377
2378		----------------------------------------------------------------------
2379		Name: Matthias Blume
2380		Date: 2001/06/25 12:05:00 EDT
2381		Tag: blume-20010625-nlffigen
2382		Description:
2383
2384		This update fixes a number of minor bugs in ml-nlffigen as reported by
2385		Nick Carter <nbc@andrew.cmu.edu>.
2386
2387		1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
2388		2. Default names for generated files are now derived from the name of
2389		the C file *without its directory*.  In particular, this causes generated
2390		files to be placed locally even if the C file is in some system directory.
2391		3. Default names for generated signatures and structures are also derived
2392		from the C file name without its directory.  This avoids silly things
2393		like "structure GL/GL".
2394		(Other silly names are still possible because ml-nlffigen does not do
2395		a thorough check of whether generated names are legal ML identifiers.
2396		When in doubt, use command line arguments to force particular names.)
2397
2398		----------------------------------------------------------------------
2399		Name: Matthias Blume
2400		Date: 2001/06/21 12:25:00 EDT
2401		Tag: blume-20010621-eXene
2402		Description:
2403
2404		eXene now compiles and (sort of) works again.
2405
2406		The library name (for version > 110.33) is $/eXene.cm.
2407
2408		I also added an new example in src/eXene/examples/nbody.  See the
2409		README file there for details.
2410
2411		----------------------------------------------------------------------
2412		Name: Matthias Blume
2413		Date: 2001/06/20 16:40:00 EDT
2414		Tag: blume-20010620-cml
2415		Description:
2416
2417		CML now compiles and works again.
2418
2419		Libraries (for version > 110.33):
2420
2421		$cml/cml.cm            Main CML library.
2422		$cml/basis.cm          CML's version of $/basis.cm.
2423		$cml/cml-internal.cm   Internal helper library.
2424		$cml/core-cml.cm       Internal helper library.
2425		$cml-lib/trace-cml.cm  Tracing facility.
2426		$cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
2427
2428		The installer (config/install.sh) has been taught how to properly
2429		install this stuff.
2430
2431		----------------------------------------------------------------------
2432		Name: Matthias Blume
2433		Date: 2001/06/19 17:55:00 EDT
2434		Tag: blume-20010619-instantiate
2435		Description:
2436
2437		This un-breaks the fix for bug 1432.
2438		(The bug was originally fixed in 110.9 but I broke it again some
2439		time after that.)
2440
2441		----------------------------------------------------------------------
2442		Name: Matthias Blume
2443		Date: 2001/06/19 17:25:00 EDT
2444		Tag: blume-20010619-signals
2445		Description:
2446
2447		This should (hopefully) fix the long-standing signal handling bug.
2448		(The runtime system was constructing a continuation record with an
2449		incorrect descriptor which would cause the GC to drop data on the floor...)
2450
2451		----------------------------------------------------------------------
2452		Name: Matthias Blume
2453		Date: 2001/06/15 15:05:00 EDT
2454		Tag: blume-20010615-moresparc
2455		Description:
2456
2457		Here is a short late-hour update related to Sparc c-calls:
2458
2459		-- made handling of double-word arguments a bit smarter
2460
2461		-- instruction selection phase tries to collapse certain clumsily
2462		constructed ML-Trees; typical example:
2463
2464		ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
2465
2466		This currently has no further impact on SML/NJ since mlriscGen does
2467		not seem to generate such patterns in the first place, and c-calls
2468		(which did generate them in the beginning) has meanwhile been fixed
2469		so as to avoid them as well.
2470
2471		----------------------------------------------------------------------
2472		Name: Matthias Blume
2473		Date: 2001/06/15 15:05:00 EDT
2474		Tag: blume-20010615-sparc
2475		Description:
2476
2477		The purpose of this update is to provide an implementation of NLFFI
2478		on Sparc machines.
2479
2480		Here are the changes in detail:
2481
2482		* src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
2483		the Sparc implementation of the c-calls API.
2484		* The Sparc backend of SML/NJ has been modified to uniformely use %fp
2485		for accessing the ML frame.  Thus, we have a real frame pointer and
2486		can freely modify %sp without need for an omit-frame-ptr phase.
2487		The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
2488		this case.
2489		* ml-nlffigen has been taught to produce code for different architectures
2490		and calling conventions.
2491		* In a way similar to what was done in the x86 case, the Sparc
2492		backend uses its own specific extension to mltree.  (For example,
2493		it needs to be able to generate UNIMP instructions which are part
2494		of the calling convention.)
2495		* ml-nlffi-lib was reorganized to make it more modular (in particular,
2496		to make it easier to plug in new machine- and os-dependent parts).
2497
2498		There are some other fairly unrelated bug fixes and cleanups as well:
2499
2500		* I further hacked the .cm files for MLRISC tools (like MDLGen) so
2501		that they properly share their libraries with existing SML/NJ libraries.
2502		* I fixed a minor cosmetic bug in CM, supressing certain spurious
2503		follow-up error messages.
2504		* Updates to CM/CMB documentation.
2505
2506		TODO items:
2507
2508		* MLRISC should use a different register as its asmTemp on the Sparc.
2509		(The current %o2 is a really bad choice because it is part of the
2510		calling conventions, so things might interfere in unexpected ways.)
2511
2512		----------------------------------------------------------------------
2513		Name: Matthias Blume
2514		Date: 2001/06/07
2515		Tag: blume-20010607-calls
2516		Description:
2517
2518		A number of internal changes related to C calls and calling conventions:
2519
2520		1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
2521		number of bytes popped implicitly (by the callee).  In most cases
2522		this field is 0 but on x86/win32 it is some non-zero value.  This
2523		is information provided for the benefit of the "omit-frameptr" pass.
2524		2. The CALL instruction on the x86 carries a similar "pops" field.
2525		The instruction selection phase copies its value from the ML-Tree
2526		CALL statement.
2527		3. On all other architectures, the instruction selection phase checks
2528		whether "pops=0" and complains if not.
2529		4. The c-calls implementation for x86 now accepts two calling conventions:
2530		"ccall" and "stdcall".  When "ccall" is selected, the caller cleans
2531		up after the call and pops is set to 0.  For "stdcall", the caller
2532		does nothing, leaving the cleanup to the callee; pops is set to
2533		the number of bytes that were pushed onto the stack.
2534		5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
2535		distinguish between "ccall" and "stdcall".
2536		6. The UNIMP instruction has been added to the supported Sparc instruction
2537		set. (This is needed for implementing the official C calling convention
2538		on this architecture.)
2539		7. I fixed some of the .cm files under src/MLRISC/Tools to make them
2540		work with the latest CM.
2541
2542		----------------------------------------------------------------------
2543		Name: Matthias Blume
2544		Date: 2001/06/05 15:10:00 EDT
2545		Tag: blume-20010605-cm-index
2546		Description:
2547
2548		0. The "lambdasplit" parameter for class "sml" in CM has been documented.
2549
2550		1. CM can now generate "index files".  These are human-readable files
2551		that list on a per-.cm-file basis each toplevel symbol defined or
2552		imported.  The location of the index file for
2553		<p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
2554		To enable index-file generation, set CM.Control.generate_index to true
2555		or export an environment-symbol: export CM_GENERATE_INDEX=true.
2556
2557		The CM manual has been updated accordingly.
2558
2559		2. I made some slight modifications to the c-calls API in MLRISC.
2560
2561		a) There is now a callback to support saving/restoring of
2562		dedicated but caller-save registers around the actual call
2563		instruction.
2564		b) One can optionally specify a comment-annotation for the
2565		call instruction.
2566
2567		3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
2568		(For example, the comment annotation shows the C prototype of
2569		the function being called.)
2570
2571		----------------------------------------------------------------------
2572		Name: Matthias Blume
2573		Date: 2001/06/01 13:30:00 EDT
2574		Tag: blume-20010601-nlffi-cleanup
2575		Description:
2576
2577		This is mostly a cleanup of MLFFI stuff:
2578
2579		- some signature files have been put into a more exposed place
2580		- the ugly 'f type parameter is gone (simplifies types tremendously!)
2581		- ml-nlffigen changed accordingly
2582		- tutorial updated
2583
2584		Other changes:
2585
2586		- author's affiliation in CM manual(s) updated
2587		- some more recognized keywords added to Allen's sml.sty
2588
2589		----------------------------------------------------------------------
2590		Name: Matthias Blume
2591		Date: 2001/05/25 15:30:00 EDT
2592		Tag: blume-20010525-iptr
2593		Description:
2594
2595		- put the official 110.33-README (as it appears on the ftp server) under
2596		CVS
2597		- fixed a small bug related to incomplete pointer types in
2598		ml-nlffigen
2599		- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
2600		(it does not need the 'f type parameter)
2601
2602		----------------------------------------------------------------------
2603		Name: Matthias Blume
2604		Date: 2001/05/23 14:30:00 EDT
2605		Tag: Release_110_33
2606		Description:
2607
2608		New version number (110.33).  New bootfiles.
2609
2610		----------------------------------------------------------------------
2611		Name: Matthias Blume
2612		Date: 2001/05/22 18:06:00 EDT
2613		Tag: blume-20010522-targets
2614		Description:
2615
2616		Made install.sh use file config/targets.customized if it exists, falling
2617		back to config/targets if it doesn't.  This way one can have a customized
2618		version of the targets file without touching the "real thing", thus
2619		eliminating the constant fear of accidentally checking something bogus
2620		back into the CVS repository...  (File config/targets.customized must
2621		not be added to the repository!)
2622
2623		----------------------------------------------------------------------
2624		Name: Matthias Blume
2625		Date: 2001/05/22 16:30:00 EDT
2626		Tag: blume-20010522-minitut
2627		Description:
2628
2629		1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
2630		struct returns.
2631		2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
2632		incomplete, preliminary documentation for NLFFI.
2633
2634		----------------------------------------------------------------------
2635		Name: Matthias Blume
2636		Date: 2001/05/14 11:30:00 EDT
2637		Tag: blume-20010514-script
2638		Description:
2639
2640		Some bugs in install script fixed.
2641
2642		In addition to that I also made a slight change to the NLFFI API:
2643		Functors generated by ml-nlffigen now take the dynamic library as a
2644		straight functor argument, not as a suspended one.  (The original
2645		functor code used to force the suspension right away anyway, so there
2646		was nothing gained by this complication of the interface.)
2647
2648		----------------------------------------------------------------------
2649		Name: Matthias Blume
2650		Date: 2001/05/11 14:35:00 EDT
2651		Tag: blume-20010511-ml-nlffi
2652		Description:
2653
2654		I finally took the plunge and added my new FFI code to the main
2655		repository.  For x86-linux it is now ready for prime-time.
2656
2657		There are two new subdirectories of "src":
2658
2659		- ml-nlffi-lib:
2660		The utility library for programs using the FFI interface.
2661		Here is the implementation of $/c.cm and its associated low-level
2662		partners $/c-int.cm and $/memory.cm.
2663		- ml-nlffigen:
2664		A stand-alone program for generating ML glue code from C source
2665		code.
2666
2667		Building ml-nlffigen requires $/ckit-lib.cm.
2668
2669		The config/install.sh script has been updates to do the Right Thing
2670		(hopefully).
2671
2672		Notice that the source tree for the C-Kit will not be put under "src"
2673		but directly under the installation root directory.  (This is the
2674		structure that currently exists on the CVS server when you check out
2675		module "sml".)  Fortunately, config/install.sh knows about this oddity.
2676
2677		Bugs: No documentation yet.
2678
2679		----------------------------------------------------------------------
2680		Name: Matthias Blume
2681		Date: 2001/05/09 16:35:00 EDT
2682		Tag: blume-20010509-cpscontract
2683		Description:
2684
2685		Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
2686		wrapper/unwrapper elimination did not decrement usage counts and some
2687		dead variables got overlooked by the dead-up logic.)
2688
2689		----------------------------------------------------------------------
2690		Name: Lal George
2691		Date: 2001/05/08  17:26:09 EDT
2692		Tag: george-20010508-omit-frameptr
2693		Description:
2694
2695		Changes to implement the omit-frame-pointer optimization to support
2696		raw C calls. For now, there is only support on the Intel x86, but
2697		other architectures will follow as more experience is gained with this.
2698
2699
2700		----------------------------------------------------------------------
2701		Name: Matthias Blume
2702		Date: 2001/05/07 14:40:00 EDT
2703		Tag: blume-20010507-proxies
2704		Description:
2705
2706		I made into "proxy libraries" all libraries that qualify for such a
2707		change.  (A qualifying library is a library that has another library or
2708		groups as its sole member and repeats that member's export list
2709		verbatim.  A proxy library avoids this repetition by omitting its export
2710		list, effectively inheriting the list that its (only) member exports.
2711		See the CM manual for more explanation.)
2712		The main effect is that explicit export lists for these libraries
2713		do not have to be kepts in sync, making maintenance a bit easier.
2714
2715		I also added copyright notices to many .cm-files.
2716
2717		Last but not least, I made a new set of bootfiles.
2718
2719		----------------------------------------------------------------------
2720		Name: Matthias Blume
2721		Date: 2001/05/04 17:00:00 EDT
2722		Tag: blume-20010504-cm-lsplit
2723		Description:
2724
2725		0. John merged pending changes to $/smlnj-lib.cm
2726
2727		1. Allen's previous change accidentally backed out of one of Lal's
2728		earlier changes.  I undid this mistake (re-introducing Lal's change).
2729
2730		2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
2731		within the compiler where applicable.  There is some code simplification
2732		because of that.
2733
2734		3. The "split" phase (in FLINT) is now part of the default list of phases.
2735		Compiler.Control.LambdaSplitting.* can be used to globally control the
2736		lambda-splitting (cross-module-inlining) engine.  In addition to that,
2737		it can now also be controlled on a per-source basis: CM has been taught
2738		a new tool parameter applicable to ML source files.
2739
2740		- To turn lambda-splitting off completely:
2741		local open Compiler.Control.LambdaSplitting in
2742		val _ = set Off
2743		end
2744		- To make "no lambda-splitting" the global default (but allow per-source
2745		overriding); this is the initial setting:
2746		local open Compiler.Control.LambdaSplitting in
2747		val _ = set (Default NONE)
2748		end
2749		- To make "lambda-splitting with aggressiveness a" the global default
2750		(and allow per-source overriding):
2751		local open Compiler.Control.LambdaSplitting in
2752		val _ = set (Default (SOME a))
2753		end
2754
2755		- To turn lambda-splitting off for a given ML souce file (say: a.sml)
2756		write (in the respective .cm-file):
2757		a.sml (lambdasplitting:off)
2758		- To turn lambda-splitting for a.sml on with minimal aggressiveness:
2759		a.sml (lambdasplitting:on)
2760		- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
2761		<a> is a decimal non-negative integer):
2762		a.sml (lambdasplitting:<a>)
2763		- To turn lambda-splitting for a.sml on with maximal aggressiveness:
2764		a.sml (lambdasplitting:infinity)
2765		- To use the global default for a.sml:
2766		a.sml (lambdasplitting:default)
2767		or simply
2768		a.sml
2769
2770		----------------------------------------------------------------------
2771		Name: Allen Leung
2772		Date: 2001/05/04 01:57:00 EDT
2773		Tag: leunga-20010504-sync
2774		Description:
2775
2776		MLRISC features.
2777
2778		1. Fix to CMPXCHG instructions.
2779		2. Changed RA interface to allow annotations in callbacks.
2780		3. Added a new method to the stream interface to allow annotations updates.
2781
2782		----------------------------------------------------------------------
2783		Name: Matthias Blume
2784		Date: 2001/05/01 11:45:00 EDT
2785		Tag: blume-20010501-pcedittmp
2786		Description:
2787
2788		Changed install.sh to use the current working directory instead of
2789		/usr/tmp for a temporary file (pcedittmp).  The previous choice
2790		of /usr/tmp caused trouble with MacOS X because of file premission
2791		problems.
2792
2793		----------------------------------------------------------------------
2794		Name: Matthias Blume
2795		Date: 2001/04/20 11:10:00 EDT
2796		Tag: blume-20010420-inMLflag
2797		Description:
2798
2799		- added vp_limitPtrMask to vproc-state.h
2800		(for use by the raw-C-calls mechanism to implement proper interrupt
2801		handling)
2802		- made the ML compiler aware of various data-structure offsets so it
2803		can generate code for accessing the vp_inML flag and vp_limitPtrMask
2804		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
2805		raw C-calls
2806
2807		----------------------------------------------------------------------
2808		Name: Lal George
2809		Date: 2001/04/20 09:15:28 EDT
2810		Tag: george-20010420-macosX
2811		Description:
2812
2813		- Changes to port to Mac OS X; Darwin.
2814
2815		- In the process I found that sqrt was broken on the PPC, because the
2816		fsqrt instruction is not implemented.
2817
2818		----------------------------------------------------------------------
2819		Name: Matthias Blume
2820		Date: 2001/04/18 12:45:00 EDT
2821		Tag: blume-20010418-ccalls
2822		Description:
2823
2824		- fixed two off-by-4 errors in the x86-specific c-calls implementation
2825		(this bug prevented structure arguments containing pointers from being
2826		passed correctly)
2827		- changed the raw-C-call code in mlriscGen.sml in such a way that
2828		structure arguments are represented as a pointer to the beginning
2829		of the structure (instead of having a series of synthesized arguments,
2830		one for each structure member)
2831
2832		- made makeml script's verbosity level configurable via environment
2833		variable (MAKEML_VERBOSITY)
2834
2835		- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
2836		in rawmem-x86.sml; we are now using the real thing
2837
2838		----------------------------------------------------------------------
2839		Name: Matthias Blume
2840		Date: 2001/03/22 16:25:00 EST
2841		Tag: blume-20010322-bootfiles
2842		Description:
2843
2844		Created a new set of bootfiles (for your automatic installation convenience).
2845
2846		----------------------------------------------------------------------
2847		Name: Matthias Blume
2848		Date: 2001/03/22 15:10:00 EST
2849		Tag: blume-20010322-rawmem-parcm
2850		Description:
2851
2852		1. All "raw memory access" primitives for the new FFI are implemented now
2853		(at least on the x86).
2854		2. Some further cleanup of CM's parallel make mechanism.
2855
2856		----------------------------------------------------------------------
2857		Name: Matthias Blume
2858		Date: 2001/03/19 17:53:00 EST
2859		Tag: blume-20010319-parallel
2860		Description:
2861
2862		Parallel make (using compile servers) now works again.
2863
2864		To this end, CM.stabilize and CMB.make have been modified to work in
2865		two passes when compile servers are attached:
2866		1. Compile everything, do not perform stabilization; this pass
2867		uses compile servers
2868		2. Stabilize everything; this pass does not use compile servers
2869		If there are no compile servers, the two passes are combined into one
2870		(as before).  Splitting the passes increases the inherent parallelism
2871		in the dependency graph because the entire graph including all
2872		libraries is available at the same time.  This, in turn, improves
2873		server utilization.  The downside is that the master process will
2874		have to do some extra work after compilation is done (because for
2875		technical reasons it must re-read all the binfiles during stabilization).
2876
2877		----------------------------------------------------------------------
2878		Name: Matthias Blume
2879		Date: 2001/03/16 12:22:00 EST
2880		Tag: blume-20010316-bootfiles
2881		Description:
2882
2883		Created a new set of bootfiles (for your automatic installation convenience).
2884
2885		----------------------------------------------------------------------
2886		Name: Matthias Blume
2887		Date: 2001/03/16 11:00:00 EST
2888		Tag: blume-20010316-MLTREE-fixup
2889		Description:
2890
2891		This is a minor fixup for an (untagged) earlier commit by Allen.
2892		(A file was missing).
2893
2894		----------------------------------------------------------------------
2895		Name: Allen Leung
2896		Date: Mon Mar  5 18:54:57 EST 2001
2897		Tag: leunga-20010305-cut-support
2898
2899		1. New support for alternative control-flow in MLTREE.
2900		Currently we support
2901
2902		FLOW_TO(CALL ...., [k1,...,kn])
2903
2904		This is needed for 'cuts to' in C-- and try/handle-like constructs
2905		in Moby
2906
2907		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
2908
2909		2. Register Allocator
2910
2911		Changes in interface [from Fermin, John]
2912
2913		3. Alpha 8-bit SLL support [Fermin]
2914
2915		4. All architectures
2916
2917		A new module (ClusterExpandCopies) for expanding parallel copies.
2918
2919		----------------------------------------------------------------------
2920		Name: Allen Leung
2921		Date: 2001/02/27 23:07:00 EST
2922		Tag: leunga-20010227-minor-stuff
2923
2924		1. Alpha bug fix for CMOVNE
2925		2. Handle mltree COND(..,FCMP ...,...)
2926		3. Bug fix in simplifier
2927
2928		----------------------------------------------------------------------
2929		Name: Matthias Blume
2930		Date: 2001/01/30 17:50:00 EST
2931		Tag: blume-20010130-sync
2932		Description:
2933
2934		This is just a minor update to sync my devel branch with the main brach.
2935		The only visible change is the addition of some README files.
2936
2937		----------------------------------------------------------------------
2938		Name: Matthias Blume
2939		Date: 2001/01/12 23:30:00 JST
2940		Tag: blume-20010112-bootfiles
2941		Description:
2942
2943		Made a new set of bootfiles that goes with the current state of the
2944		repository.
2945
2946		----------------------------------------------------------------------
2947		Name: Matthias Blume
2948		Date: 2001/01/12 21:20:00 JST
2949		Tag: blume-20010112-sync
2950		Description:
2951
2952		I am just flushing out some minor changes that had accumulated in
2953		my private branch in order to sync with the main tree.  (This is
2954		mainly because I had CVS trouble when trying to merge _into_ my
2955		private branch.)
2956
2957		Most people should be completely unaffected by this.
2958
2959		----------------------------------------------------------------------
2960		Name: Allen Leung
2961		Date: Thu Jan 11 21:03:00 EST 2001
2962		Tag: leunga-20010111-labexp=mltree
2963		Description:
2964
2965		1.  Removed the type LabelExp and replace it by MLTree.
2966		2.  Rewritten mltree-simplify with the pattern matcher tool.
2967		3.  There were some bugs in alpha code generator which would break
2968		64-bit code generation.
2969		4.  Redo the tools to generate code with the
2970		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
2971		are now generated by perl scripts.
2972
2973		----------------------------------------------------------------------
2974		Name: Matthias Blume
2975		Date: 2001/01/10 21:55:00 JST
2976		Tag: blume-20010110-rcc
2977		Description:
2978
2979		The RCC stuff now seems to work (but only on the x86).
2980		This required hacking of the c-calls interface (and -implementation) in
2981		MLRISC.
2982
2983		Normal compiler users should be unaffected.
2984
2985		----------------------------------------------------------------------
2986		Name: Matthias Blume
2987		Date: 2001/01/09 01:20:00 JST
2988		Tag: blume-20010109-rcc
2989		Description:
2990
2991		This is a fairly big patch, flushing out a large number of pending
2992		changes that I made to my development copy over the last couple of days.
2993
2994		Of practical relevance at this moment is a workaround for a pickling
2995		bug that Allen ran into the other day.  The cause of the bug itself is
2996		still unknown and it might be hard to fix it properly, but the
2997		workaround has some merits of its own (namely somewhat reducing pickling
2998		overhead for certain libraries).  Therefore, I think this solution should
2999		be satisfactory at this time.
3000
3001		The rest of the changes (i.e., the vast majority) has to do with my
3002		ongoing efforts of providing direct support for C function calls from
3003		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
3004		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
3005		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
3006		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
3007		handle RCC.
3008
3009		The last part (the changes to mlriscGen.sml) are still known to be
3010		wrong on the x86 and not implemented on all other architectures.  But
3011		the infrastructure is in place. I had to change a few functor
3012		signatures in the backend to be able to route the CCalls interface
3013		from MLRISC there, and I had to specialize the mltree type (on the
3014		x86) to include the necessary extensions. (The extensions themselves
3015		were already there and redy to go in MLRISC/x86).
3016
3017		Everything should be very happy as soon as someone helps me with
3018		mlriscGen.sml...
3019
3020		In any case, nothing of this should matter to anyone as long as the
3021		new primop is not being used (which is going to be the case unless you
3022		find it where I hid it :). The rest of the compiler is completely
3023		unaffected.
3024
3025		----------------------------------------------------------------------
3026		Name: Matthias Blume
3027		Date: 2001/01/05 00:30:00 JST
3028		Tag: blume-20010105-primops
3029		Description:
3030
3031		Added some experimental support for work that I am doing right now.
3032		These changes mostly concern added primops, but there is also a new
3033		experimental C library in the runtime system (but currently not enabled
3034		anywhere except on Linux/X86).
3035
3036		In the course of adding primops (and playing with them), I discovered that
3037		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
3038		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
3039		To recover, I made type information in INL_PRIM mandatory and changed
3040		prim.sml as well as built-in.sml accordingly.  The InLine structure now
3041		has complete, correct type information (i.e., no bottom types).
3042
3043		Since all these changes mean that we need new binfiles, I also bumped the
3044		version number to 110.32.1.
3045
3046		----------------------------------------------------------------------
3047		Name: Matthias Blume
3048		Date: 2000/12/30 22:10:00 JST
3049		Tag: blume-20001230-various
3050		Description:
3051
3052		Added proxy libraries for MLRISC and let MLRISC libraries refer
3053		to each other using path anchors.  (See CM manual for explanation.)
3054
3055		Updated CM documentation.
3056
3057		Fixed some bugs in CM.
3058
3059		Implemented "proxy" libraries (= syntactic sugar for CM).
3060
3061		Added "-quiet" option to makeml and changed runtime system accordingly.
3062
3063		Added cleanup handler for exportML to reset timers and compiler stats.
3064
3065		----------------------------------------------------------------------
3066		Name: Lal George
3067		Date: 2000/12/22 22:22:58 EST 2000
3068		Tag: Release_110_32
3069		Description:
3070
3071		Infinite precision used throughout MLRISC.
3072		see MLRISC/mltree/machine-int.sig
3073
3074		----------------------------------------------------------------------
3075		Name: Matthias Blume
3076		Date: 2000/12/22 23:16:00 JST
3077		Tag: blume-20001222-warn
3078		Description:
3079
3080		Corrected wording and formatting of some CM warning message which I
3081		broke in my previous patch.
3082
3083		----------------------------------------------------------------------
3084		Name: Matthias Blume
3085		Date: 2000/12/22 21:20:00 JST
3086		Tag: blume-20001222-anchorenv
3087		Description:
3088
3089		Fixed CM's handling of anchor environments in connection with CMB.make.
3090
3091		----------------------------------------------------------------------
3092		Name: Matthias Blume
3093		Date: 2000/12/22 13:15:00 JST
3094		Tag: blume-20001222-cleanup
3095		Description:
3096
3097		Removed src/cm/ffi which does not (and did not) belong here.
3098
3099		----------------------------------------------------------------------
3100		Name: Matthias Blume
3101		Date: 2000/12/21 23:55:00 JST
3102		Tag: blume-20001221-exn
3103		Description:
3104
3105		Probably most important: CM no longer silently swallows all exceptions
3106		in the compiler.
3107		Plus: some other minor CM changes.  For example, CM now reports some
3108		sizes for generated binfiles (code, data, envpickle, lambdapickle).
3109
3110		----------------------------------------------------------------------
3111		Name: Matthias Blume
3112		Date: 2000/12/15 00:01:05 JST
3113		Tag: blume-20001215-dirtool
3114		Description:
3115
3116		- "dir" tool added.
3117		- improvements and cleanup to Tools structure
3118		- documentation updates
3119
3120		----------------------------------------------------------------------
3121		Name: Allen Leung
3122		Date: Thu Dec 14 03:45:24 EST 2000
3123		Description:
3124		Tag:  leunga-20001214-int-inf
3125		Description:
3126
3127		In IntInf, added these standard functions, which are missing from our
3128		implementation:
3129
3130		andb : int * int -> int
3131		xorb : int * int -> int
3132		orb  : int * int -> int
3133		notb : int -> int
3134		<<   : int * word -> int
3135		~>>  : int * word -> int
3136
3137		Not tested, I hope they are correct.
3138
3139		----------------------------------------------------------------------
3140		Name: Allen Leung
3141		Date: Fri Dec  8 19:23:26 EST 2000
3142		Description:
3143		Tag:  leunga-20001208-nowhere
3144		Description:
3145
3146		Slight improvements to the 'nowhere' tool to handle OR-patterns,
3147		to generate better error messages etc.  Plus a brief manual.
3148
3149		----------------------------------------------------------------------
3150		Name: Lal George
3151		Date: 2000/12/08 09:54:02 EST 2000
3152		Tag: Release_110_31
3153		Description:
3154
3155		- Version 110.31
3156		----------------------------------------------------------------------
3157		Name: Allen Leung
3158		Date: Thu Dec  7 22:01:04 EST 2000
3159		Tag:  leunga-20001207-cell-monster-hack
3160		Description:
3161
3162		Major MLRISC internal changes.  Affect all clients.
3163		Summary:
3164
3165		1.  Type CELLS.cell = int is now replaced by a datatype.
3166		As a result, the old regmap is now gone.  Almost all interfaces
3167		in MLRISC change as a consequence.
3168
3169		2.  A new brand version of machine description tool (v3.0) that generates
3170		modules expecting the new interface.  The old version is removed.
3171
3172		3.  The RA interface has been further abstracted into two new functors.
3173		RISC_RA and X86RA.  These functors have much simpler interfaces.
3174		[See also directory MLRISC/demo.]
3175
3176		4.  Some other new source->source code generation tools are available:
3177
3178		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
3179		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
3180		I use this tool to generate the peephole optimizers---with the new
3181		cell type changes, peephole rules are becoming difficult to write
3182		without conditional pattern matching.
3183
3184		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
3185		cover the entire MLRISC source tree so many things broke.
3186
3187		6.  CM files have been moved to the subdirectory MLRISC/cm.
3188		They are moved because there are a lot of them and they clutter up the
3189		root dir.
3190
3191		7.  More detailed documentation to come...
3192
3193		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
3194		a makeml -rebuild first.  This is because of other other
3195		changes that Matthias has made (see below).
3196
3197
3198		----------------------------------------------------------------------
3199		Name: Matthias Blume
3200		Date: 2000/11/30 23:12:00 JST
3201		Tag: blume-20001130-filereorg
3202		Description:
3203
3204		Some manual updates and some file reorganizations in CM.
3205
3206		----------------------------------------------------------------------
3207		Name: Matthias Blume
3208		Date: 2000/11/24 17:45:00 JST
3209		Tag: blume-20001124-link
3210		Description:
3211
3212		Drastically improved link traversal code for the case that the dynamic
3213		value was already loaded at bootstrap time.  As a result, CM and CMB
3214		now both load blazingly fast -- even on a very slow machine.  Also,
3215		memory consumption has been further reduced by this.
3216
3217		Warning: The format of the PIDMAP file has changed.  THerefore, to
3218		bootstrap you have to do this:
3219
3220		1. Run CMB.make
3221		2. Make a symbolic link for the boot directory:
3222		ln -s sml.boot.ARCH-OS xxx
3223		3. "Rebuild" the boot directory:
3224		./makeml -boot xxx -rebuild sml ; rm xxx
3225		4. Boot normally:
3226		./makeml
3227
3228		----------------------------------------------------------------------
3229		Name: Matthias Blume
3230		Date: 2000/11/21 21:20:00 JST
3231		Tag: blume-20001121-tools
3232		Description:
3233
3234		Continued hacking on autoloading problem -- with success this time.
3235		Also changed tool-plugin mechanism.  See new CM manual.
3236
3237		----------------------------------------------------------------------
3238		Name: Matthias Blume
3239		Date: 2000/11/19 14:30:00 JST
3240		Tag:  blume-20001119-autoload
3241		Description:
3242
3243		Some hacking to make autoloading faster.  Success for CMB, no success
3244		so far for CM.  There is a reduced structure CM' that autoloads faster.
3245		(This is a temporary, non-documented hack to be eliminated again when
3246		the general problem is solved.)
3247
3248		----------------------------------------------------------------------
3249		Name: Matthias Blume
3250		Date: 2000/11/17 14:10:00 JST
3251		Tag: blume-20001117-pickle-lib
3252		Description:
3253
3254		1. Eliminated comp-lib.cm
3255		2. Made pickle-lib.cm
3256		3. Eliminated all uses of intset.sml (from comp-lib.cm)
3257		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
3258		equivalent constructs from smlnj-lib.cm (INtHashTable).
3259		5. Point 4. also goes for those uses of intmap.* in MLRISC.
3260		Duplicated intmap modules thrown out.
3261		6. Hunted down all duplicated SCC code and replaced it with
3262		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
3263		7. Rewrote Feedback module.
3264		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
3265		should be thrown out and equivalent modules from smlnj-lib.cm
3266		should be used (IntRedBlackSet, IntListSet, ...).
3267
3268		Confirmed that compiler compiles to fixpoint.
3269
3270		----------------------------------------------------------------------
3271		Name: Allen Leung
3272		Date: 2000/11/10 18:00:00
3273		Tag: leunga-20001110-new-x86-fp
3274
3275		A new x86 floating point code generator has been added.
3276		By default this is turned off.  To turn this on, do:
3277
3278		CM.autoload "$smlnj/compiler.cm";
3279		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
3280
3281		Changes:
3282
3283		1.  Changed FTAN to FPTAN so that the assembly output is correct.
3284		2.  Changed the extension callback for FTANGENT to generate:
3285
3286		fptan
3287		fstp  %st(0)
3288		instead of
3289		fptan
3290		fstpl ftempmem
3291
3292		3.  Numerous assembly fixes for x86.
3293
3294		5.  Cleaned up the machine code output module x86/x86MC.sml and added
3295		support for a whole bunch of instructions and addressing modes:
3296
3297		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
3298		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
3299		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
3300		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
3301		fxch %st(n)
3302		fld %st(n)
3303		fst %st(n)
3304		fst mem
3305		fstp %st(n)
3306		fucom %st(n)
3307		fucomp %st(n)
3308
3309		All these are now generated when the fast fp mode is turned on.
3310
3311		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
3312
3313		----------------------------------------------------------------------
3314		Name: Matthias Blume
3315		Date: 2000/11/09 11:20:00 JST
3316		Tag: blume-20001109-scc
3317		Description:
3318
3319		Eliminated some code duplication:
3320
3321		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
3322		(Otherwise the functor is useless.)
3323		2. Used GraphSCCFn where SCCUtilFun was used previously.
3324		3. Got rid of SCCUtilFun (in comp-lib.cm).
3325
3326		----------------------------------------------------------------------
3327		Name: Lal George
3328		Date: 2000/11/06 09:02:21 EST 2000
3329		Tag: Release_110_30
3330		Description:
3331
3332		- Version 110.30
3333		----------------------------------------------------------------------
3334		Name: Matthias Blume
3335		Date: 2000/11/04 14:45:00
3336		Tag: blume-20001104-mlbuild
3337		Description:
3338
3339		- Made ml-build faster on startup.
3340		- Documentation fixes.
3341
3342		----------------------------------------------------------------------
3343		Name: Matthias Blume
3344		Date: 2000/11/02 17:00:00 JST
3345		Tag: blume-20001102-condcomp
3346		Description:
3347
3348		- Small tweaks to pickler -- new BOOTFILES!
3349		- Version bumped to 110.29.2.
3350		- Added conditional compilation facility to init.cmi (see comment there).
3351		----------------------------------------------------------------------
3352		Name: Allen Leung
3353		Date: 2000/10/23 19:31:00
3354		Tag: leunga-20001023-demo-ra
3355
3356		1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
3357		2. Test programs for the graph library updated
3358		3. Some new MLRISC demo programs added
3359
3360		----------------------------------------------------------------------
3361		Name: Matthias Blume
3362		Date: 2000/08/31 22:15:00 JST
3363		Tag: blume-20001017-errmsg
3364		Description:
3365
3366		More error message grief: Where there used to be no messages, there
3367		now were some that had bogus error regions.  Fixed.
3368
3369		----------------------------------------------------------------------
3370		Name: Matthias Blume
3371		Date: 2000/08/31 17:30:00 JST
3372		Tag: blume-20001017-v110p29p1
3373		Description:
3374
3375		I made a version 110.29.1 with new bootfiles.
3376
3377		Changes:  Modified pickler/unpickler for faster and leaner unpickling.
3378		CM documentation changes and a small bugfix in CM's error reporting.
3379
3380		----------------------------------------------------------------------
3381		Name: Lal George
3382		Date: 2000/09/27 14:42:35 EDT
3383		Tag: george-20000927-nodestatus
3384		Description:
3385
3386		Changed the type of the nodestatus, so that:
3387
3388		SPILLED(~1)             is now SPILLED
3389		SPILLED(m) where m>=0   is now MEMREG(m)
3390		SPILLED(s) where s<~1   is now SPILL_LOC(~s)
3391
3392		----------------------------------------------------------------------
3393		Name: Matthias Blume
3394		Date: 2000/09/07 14:45:00 JST
3395		Tag: blume-20000907-cmerrmsg
3396		Description:
3397
3398		Small tweak to CM to avoid getting ML syntax error messages twice.
3399
3400		----------------------------------------------------------------------
3401		Name: Matthias Blume
3402		Date: 2000/08/31 18:00:00 JST
3403		Tag: blume-20000831-cvsbootfiles
3404		Description:
3405
3406		New URL for boot files (because the 110.29 files on the BL server do
3407		now work correctly with my updated install scripts for yacc and lex).
3408
3409		----------------------------------------------------------------------
3410		Name: Matthias Blume
3411		Date: 2000/08/08 12:33:00 JST
3412		Tag: blume-20000808-manual
3413		Description:
3414
3415		Tiny update to CM manual.
3416
3417	----------------------------------------------------------------------	----------------------------------------------------------------------
3418	Name: Allen Leung	Name: Allen Leung
3419	Date: 2000/08/7 19:31:00	Date: 2000/08/7 19:31:00
#	Line 1462	Line 4865
4865	elaborator). There were a lot of changes during my "linkpath" trials	elaborator). There were a lot of changes during my "linkpath" trials
4866	that could have been reverted to their original state but weren't.	that could have been reverted to their original state but weren't.
4867	Please, don't be too harsh on me for messing with this code a bit more	Please, don't be too harsh on me for messing with this code a bit more
4868	than what was strictly necessary...  (I _did_ resist the tempation	than what was strictly necessary...  (I _did_ resist the temptation
4869	of doing any "global reformatting" to avoid an untimely death at	of doing any "global reformatting" to avoid an untimely death at
4870	Dave's hands. :)	Dave's hands. :)
4871

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.695   changed lines   Added in v.1183

---

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