Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 759, Fri Dec 22 12:29:26 2000 UTC	revision 987, Wed Nov 21 21:14:44 2001 UTC
#	Line 11	Line 11
11	Date:	Date:
12	Tag: <post-commit CVS tag>	Tag: <post-commit CVS tag>
13	Description:	Description:
14
15		----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2001/11/21 16:14:00 EST
18		Tag: blume-20011121-forgottenfile
19		Description:
20
21		Forgot to add a file.  (Just a .tex-file -- part of
22		the CM manual source.)
23
24		----------------------------------------------------------------------
25		Name: Matthias Blume
26		Date: 2001/11/21 16:10:00 EST
27		Tag: Release_110_37
28		Description:
29
30		1. Modifications to the SML/NJ code generator and to the runtime system
31		so that code object name strings are directly inserted into code
32		objects at code generation time.  The only business the runtime system
33		has with this is now to read the name strings on occasions.
34		(The encoding of the name string has also changed somewhat.)
35
36		2. CM now implements a simple "set calculus" for specifying export lists.
37		In particular, it is now possible to refer to the export lists of
38		other libraries/groups/sources and form unions as well as differences.
39		See the latest CM manual for details.
40
41		3. An separate notion of "proxy" libraries has again be eliminated from
42		CM's model.  (Proxy libraries are now simply a special case of using
43		the export list calculus.)
44
45		4. Some of the existing libraries now take advantage of the new set
46		calculus.
47		(Notice that not all libraries have been converted because some
48		of the existing .cm-files are supposed to be backward compatible
49		with 110.0.x.)
50
51		5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
52		"General.exnMessage"!  The former relies on a certain hook to be
53		initialized, and that often does not happen in the stand-alone case.)
54
55		----------------------------------------------------------------------
56		Name: Lal George
57		Date: 2001/11/21  13:56:18 EST
58		Tag: george-2001121-pseudo-ops
59		Description:
60
61		Implemented a complete redesign of MLRISC pseudo-ops. Now there
62		ought to never be any question of incompatabilities with
63		pseudo-op syntax expected by host assemblers.
64
65		For now, only modules supporting GAS syntax are implemented
66		but more should follow, such as MASM, and vendor assembler
67		syntax, e.g. IBM as, Sun as, etc.
68
69		----------------------------------------------------------------------
70		Name: Matthias Blume
71		Date: 2001/11/14 11:52:00 EST
72		Tag: blume-20011114-srcname
73		Description:
74
75		1. Routed the name of the current source file to mlriscgen where it
76		should be directly emitted into the code object.  (This last part
77		is yet to be done.)
78
79		2. Some cleanup of the pgraph code to make it match the proposal that
80		I put out the other day.  (The proposal notwithstanding, things are
81		still in flux here.)
82
83		----------------------------------------------------------------------
84		Name: Lal George
85		Date: 2001/11/14 09:44:04 EST
86		Tag:
87		Description:
88
89		Fix for a backpatching bug reported by Allen.
90
91		Because the boundary between short and long span-dependent
92		instructions is +/- 128, there are an astounding number of
93		span-dependent instructions whose size is over estimated.
94
95		Allen came up with the idea of letting the size of span
96		dependent instructions be non-monotonic, for a maxIter
97		number of times, after which the size must be monotonically
98		increasing.
99
100		This table shows the number of span-dependent instructions
101		whose size was over-estimated as a function of maxIter, for the
102		file Parse/parse/ml.grm.sml:
103
104		maxIter            # of instructions:
105		10                      687
106		20                      438
107		30                      198
108		40                        0
109
110		In compiling the compiler, there is no significant difference in
111		compilation speed between maxIter=10 and maxIter=40. Actually,
112		my measurements showed that maxIter=40 was a tad faster than
113		maxIter=10! Also 96% of the  files in the compiler reach a fix
114		point within 13 iterations, so fixing maxIter at 40, while high,
115		is okay.
116
117		----------------------------------------------------------------------
118		Name: Matthias Blume
119		Date: 2001/10/31 15:25:00 EST
120		Tag: blume-20011031-pgraph
121		Description:
122
123		CKIT:
124		* Changed the "Function" constructor of type Ast.ctype to carry optional
125		argument identifiers.
126		* Changed the return type of TypeUtil.getFunction accordingly.
127		* Type equality ignores the argument names.
128		* TypeUtil.composite tries to preserve argument names but gives up quickly
129		if there is a mismatch.
130
131		installation script:
132		* attempts to use "curl" if available (unless "wget" is available as well)
133
134		CM:
135		* has an experimental implementation of "portable graphs" which I will
136		soon propose as an implementation-independent library format
137		* there are also new libraries $/pgraph.cm and $/pgraph-util.cm
138
139		NLFFI-LIB:
140		* some cleanup (all cosmetic)
141
142		NLFFIGEN:
143		* temporarily disabled the mechanism that suppresses ML output for
144		C definitions whose identifiers start with an underscore character
145		* generate val bindings for enum constants
146		* user can request that only one style (light or heavy) is being used;
147		default is to use both (command-line arguments: -heavy and -light)
148		* fixed bug in handling of function types involving incomplete pointers
149		* generate ML entry points that take record arguments (i.e., using
150		named arguments) for C functions that have a prototype with named
151		arguments
152		(see changes to CKIT)
153
154		----------------------------------------------------------------------
155		Name: Allen Leung
156		Date: 2001/10/27 20:34:00 EDT
157		Tag: leunga-20011027-x86-fast-fp-call
158		Description:
159
160		Fixed the bug described in blume-20010920-slowfp.
161
162		The fix involves
163		1. generating FCOPYs in FSTP in ia32-svid
164		2. marking a CALL with the appropriate annotation
165
166		----------------------------------------------------------------------
167		Name: Matthias Blume
168		Date: 2001/10/16 11:32:00 EDT
169		Tag: blume-20011016-netbsd
170		Description:
171
172		Underscore patch from Chris Richards (fixing problem with compiling
173		runtime system under recent NetBSD).
174
175		----------------------------------------------------------------------
176		Name: Allen Leung
177		Date: 2001/10/12 17:18:32 EDT 2001
178		Tag: leung-20011012-x86-printflowgraph
179		Description:
180
181		X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
182
183		----------------------------------------------------------------------
184		Name: Lal George
185		Date: 2001/10/11 23:51:34 EDT
186		Tag: george-20011011-too-many-instrs
187		Description:
188
189		The representation of a program point never expected to see more
190		than 65536 instructions in a basic block!
191
192		----------------------------------------------------------------------
193		Name: Lal George
194		Date: 2001/10/09 09:41:37 EDT
195		Tag: george-20011008-mlrisc-labels
196		Description:
197
198		Changed the machine description files to support printing of
199		local and global labels in assembly code, based on host assembler
200		conventions.
201
202		----------------------------------------------------------------------
203		Name: Matthias Blume
204		Date: 2001/09/25 15:25:00 EDT
205		Tag: blume-20010925-exninfo
206		Description:
207
208		I provided a non-hook implementation of exnName (at the toplevel) and
209		made the "dummy" implementation of exnMessage (at the toplevel) more
210		useful: if nothing gets "hooked in", then at least you are going to
211		see the exception name and a message indicating why you don't see more.
212
213		[For the time being, programs that need exnMessage and want to use
214		ml-build should either use General.exnMessage (strongly recommended) or
215		refer to structure General at some other point so that CM sees a
216		static dependency.]
217
218		[Similar remarks go for "print" and "use":  If you want to use their
219		functionality in stand-alone programs generated by ml-build, then use
220		TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
221
222		----------------------------------------------------------------------
223		Name: Matthias Blume
224		Date: 2001/09/20 17:28:00 EDT
225		Tag: blume-20010920-slowfp
226		Description:
227
228		Allen says that x86-fast-fp is not safe yet, so I turned it off again...
229
230		----------------------------------------------------------------------
231		Name: Matthias Blume
232		Date: 2001/09/20 17:20:00 EDT
233		Tag: blume-20010920-canonicalpaths
234		Description:
235
236		0. Updated the BOOT file (something that I forgot to do earlier).
237
238		1. Small internal change to CM so that it avoids "/../" in filenames
239		as much as possible (but only where it is safe).
240
241		2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
242		that arguments that contain delimiters are passed through correctly.
243		This change also means that all "special" arguments of the form
244		@SMLxxx... must come first.
245
246		3. Changed install script to put relative anchor names for tool commands
247		into pathconfig.
248
249		----------------------------------------------------------------------
250		Name: Matthias Blume
251		>>>>>>> 1.169
252		Date: 2001/09/18 15:35:00 EDT
253		Tag: blume-20010918-readme11036
254		Description:
255
256		Added README files.
257
258		----------------------------------------------------------------------
259		Name: Matthias Blume
260		Date: 2001/09/18 11:45:00 EDT
261		Tag: Release_110_36 (retag)
262		Description:
263
264		Fixed mistake in config/preloads. Retagged as 110.36.
265
266		----------------------------------------------------------------------
267		Name: Matthias Blume
268		Date: 2001/09/18 09:40:00 EDT
269		Tag: Release_110_36_orig (tag changed)
270		Description:
271
272		New version (110.36).  New bootfiles.
273
274		----------------------------------------------------------------------
275		Name: Matthias Blume
276		Date: 2001/09/14 16:15:00 EDT
277		Tag: blume-20010914-x86fastfp
278		Description:
279
280		John committed some changes that Allen made, in particular a (hopefully)
281		correctly working version of the x86-fp module.
282
283		I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
284		flag to "true".  Everything seems to compile to a fixpoint ok, and
285		"mandelbrot" speeds up by about 15%.
286
287		----------------------------------------------------------------------
288		Name: Matthias Blume
289		Date: 2001/09/13 11:20:00 EDT
290		Tag: blume-20010913-minimal
291		Description:
292
293		1. Stefan Monnier's patch to fix a miscompilation problem that
294		was brought to light by John Reppy's work on Moby.
295
296		2. Implemented a minimal "structure Compiler" that contains just
297		"version" and "architecture".  The minimal version will be
298		available when the full version is not.  This is for backward-
299		compatibility with code that wants to test Compiler.version.
300
301		----------------------------------------------------------------------
302		Name: Matthias Blume
303		Date: 2001/08/28 14:03:00 EDT
304		Tag: blume-20010828-ml-lex
305		Description:
306
307		Fix for bug 1581, received from Neophytos Michael.
308
309		----------------------------------------------------------------------
310		Name: Matthias Blume
311		Date: 2001/08/27 11:20:00 EDT
312		Tag: blume-20010827-readme11035
313		Description:
314
315		Fleshed out the README file for 110.35.
316
317		----------------------------------------------------------------------
318		Name: Matthias Blume
319		Date: 2001/08/24 17:10:00 EDT
320		Tag: Release_110_35
321		Description:
322
323		New version number (110.35).  New bootfiles.
324
325		----------------------------------------------------------------------
326		Name: Lal George
327		Date: 2001/08/24 13:47:18 EDT 2001
328		Tag: george-20010824-MLRISC-graphs
329		Description:
330
331		removed clusters from MLRISC completely and replaced with graphs.
332
333		----------------------------------------------------------------------
334		Name: Matthias Blume
335		Date: 2001/08/23 17:50:00 EDT
336		Tag: blume-20010823-toplevel
337		Description:
338
339		- some reorganization of the code that implements various kinds of
340		environments in the compiler (static, dynamic, symbolic, combined)
341		- re-implemented the EnvRef module so that evalStream works properly
342		(if the stream contains references to "use", "CM.make", etc.)
343		- cleaned up evalloop.sml and interact.sml (but they need more cleaning)
344
345		----------------------------------------------------------------------
346		Name: Matthias Blume
347		Date: 2001/08/20 15:50 EDT
348		Tag: blume20010820-slipup
349		Description:
350
351		I forgot to commit a few files.  Here they are...
352
353		----------------------------------------------------------------------
354		Name: Matthias Blume
355		Date: 2001/08/20 15:35:00 EDT
356		Tag: blume-20010820-debugprof
357		Description:
358
359		!!!! NEW BOOTFILES !!!!
360
361		This is another round of reorganizing the compiler sources.  This
362		time the main goal was to factor out all the "instrumentation"
363		passes (for profiling and backtracing) into their own library.
364		The difficulty was to do it in such a way that it does not depend
365		on elaborate.cm but only on elabdata.cm.
366
367		Therefore there have been further changes to both elaborate.cm and
368		elabdata.cm -- more "generic" things have been moved from the former
369		to the latter.  As a result, I was forced to split the assignment
370		of numbers indicating "primtyc"s into two portions: SML-generic and
371		SML/NJ-specific.  Since it would have been awkward to maintain,
372		I bit the bullet and actually _changed_ the mapping between these
373		numbers and primtycs.  The bottom line of this is that you need
374		a new set of bin- and bootfiles.
375
376		I have built new bootfiles for all architectures, so doing a fresh
377		checkout and config/install.sh should be all you need.
378
379		The newly created library's name is
380
381		$smlnj/viscomp/debugprof.cm
382
383		and its sources live under
384
385		src/compiler/DebugProf
386
387		----------------------------------------------------------------------
388		Name: Matthias Blume
389		Date: 2001/08/15 17:15:00 EDT
390		Tag: blume-20010815-compreorg
391		Description:
392
393		This is a first cut at reorganizing the CM libraries that make up the
394		core of the compiler.  The idea is to separate out pieces that could
395		be used independently by tools, e.g., the parser, the typechecker, etc.
396
397		The current status is a step in this direction, but it is not quite
398		satisfactory yet.  Expect more changes in the future.
399
400		Here is the current (new) organization...
401
402		What used to be $smlnj/viscomp/core.cm is now divided into
403		six CM libraries:
404
405		$smlnj/viscomp/basics.cm
406		/parser.cm
407		/elabdata.cm
408		/elaborate.cm
409		/execute.cm
410		/core.cm
411
412		The CM files for these libraries live under src/system/smlnj/viscomp.
413		All these libraries are proxy libraries that contain precisely
414		one CM library component.  Here are the locations of the components
415		(all within the src/compiler tree):
416
417		Basics/basics.cm
418		Parse/parser.cm
419		ElabData/elabdata.cm
420		Elaborator/elaborate.cm
421		Execution/execute.cm
422		core.cm
423
424		[This organization is the same that has been used already
425		for a while for the architecture-specific parts of the visible
426		compiler and for the old version of core.cm.]
427
428		As you will notice, many source files have been moved from their
429		respective original locations to a new home in one of the above
430		subtrees.
431
432		The division of labor between the new libraries is the following:
433
434		basics.cm:
435		- Simple, basic definitions that pertain to many (or all) of
436		the other libraries.
437		parser.cm:
438		- The SML parser, producing output of type Ast.dec.
439		- The type family for Ast is also defined and exported here.
440		elabdata.cm:
441		- The datatypes that describe input and output of the elaborator.
442		This includes types, absyn, and static environments.
443		elaborator.cm:
444		- The SML/NJ type checker and elaborator.
445		This maps an Ast.dec (with a given static environment) to
446		an Absyn.dec (with a new static environment).
447		- This libraries implements certain modules that used to be
448		structures as functors (to remove dependencies on FLINT).
449		execute.cm:
450		- Everything having to do with executing binary code objects.
451		- Dynamic environments.
452		core.cm:
453		- SML/NJ-specific instantiations of the elaborator and MLRISC.
454		- Top-level modules.
455		- FLINT (this should eventually become its own library)
456
457		Notes:
458
459		I am not 100% happy with the way I separated the elaborator (and its
460		data structures) from FLINT.  Two instances of the same problem:
461
462		1. Data structures contain certain fields that carry FLINT-specific
463		information.  I hacked around this using exn and the property list
464		module from smlnj-lib.  But the fact that there are middle-end
465		specific fields around at all is a bit annoying.
466
467		2. The elaborator calculates certain FLINT-related information.  I tried
468		to make this as abstract as I could using functorization, but, again,
469		the fact that the elaborator has to perform calculations on behalf
470		of the middle-end at all is not nice.
471
472		3. Having to used exn and property lists is unfortunate because it
473		weakens type checking.  The other alternative (parameterizing
474		nearly *everything*) is not appealing, though.
475
476		I removed the "rebinding =" warning hack because due to the new organization
477		it was awkward to maintain it.  As a result, the compiler now issues some of
478		these warnings when compiling init.cmi during bootstrap compilation. On
479		the plus side, you also get a warning when you do, for example:
480		val op = = Int32.+
481		which was not the case up to now.
482
483		I placed "assign" and "deref" into the _Core structure so that the
484		code that deals with the "lazy" keyword can find them there.  This
485		removes the need for having access to the primitive environment
486		during elaboration.
487
488		----------------------------------------------------------------------
489		Name: Matthias Blume
490		Date: 2001/08/13
491		Tag: blume-20010813-closures
492		Description:
493
494		This fix was sent to us by Zhong Shao.  It is supposed to improve the
495		performance of certain loops by avoiding needless closure allocation.
496
497		----------------------------------------------------------------------
498		Name: Lal George
499		Date: 2001/07/31 10:03:23 EDT 2001
500		Tag: george-20010731-x86-fmalloc
501		Description: Fixed bug in x86 calls
502
503		There was a bug where call instructions would mysteriously
504		vanish. The call instruction had to be one that returned
505		a floating point value.
506
507		----------------------------------------------------------------------
508		Name: Lal George
509		Date: 2001/07/19 16:36:29 EDT 2001
510		Tag: george-20010719-simple-cells
511		Description:
512
513		I have dramatically simplified the interface for CELLS in MLRISC.
514
515		In summary, the cells interface is broken up into three parts:
516
517		1. CellsBasis : CELLS_BASIS
518
519		CellsBasis is a top level structure and common for all
520		architectures.  it contains the definitions of basic datatypes
521		and utility  functions over these types.
522
523		2. functor Cells() : CELLS
524
525		Cells generates an interface for CELLS that incorporates the
526		specific resources on the target architecture, such as the
527		presence of special register classes, their number and size,
528		and various useful substructures.
529
530		3. <ARCH>CELLS
531
532		e.g. SparcCells: SPARCCELLS
533
534		<ARCH>CELLS usually contains additional bindings for special
535		registers  on the architecture, such as:
536
537		val r0 : cell           (* register zero *)
538		val y : cell            (* Y register *)
539		val psr : cell          (* processor status register *)
540		...
541
542		The structure returned by applying the Cells functor is opened
543		in this interface.
544
545		The main implication of all this is that the datatypes for cells is
546		split between CellsBasis and CELLS -- a fairly simple change for user
547		code.
548
549		In the old scheme the CELLS interface had a definitional binding of
550		the form:
551
552		signature CELLS = sig
553
554		structure CellsBasis = CellsBasis
555
556		...
557
558		end
559
560		With all the sharing constraints that goes on in MLRISC, this old
561		design  quickly leads to errors such as:
562
563		"structure definition spec inside of sharing ... "
564
565
566		and appears to require an unacceptable amount of sharing and where
567		constraint hackery.
568
569		I think this error message (the interaction of definitional specs and
570		sharing) requires more explanation on our web page.
571
572		----------------------------------------------------------------------
573		Name: Matthias Blume
574		Date: 2001/07/19 15:00:00 EDT
575		Tag: blume-20010719-libreorg
576		Description:
577
578		This update puts together a fairly extensive but straightforward change
579		to the way the libraries that implement the interactive system are
580		organized:
581
582		The biggest change is the elimination of structure Compiler.  As a
583		replacement for this structure, there is now a CM library
584		(known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
585		that exports all the substructures of the original structure Compiler
586		directly.  So instead of saying Compiler.Foo.bar one now simply
587		says Foo.bar.  (The CM libraries actually export a collection of
588		structures that is richer than the collection of substructures of
589		structure Compiler.)
590
591		To make the transition smooth, there is a separate library called
592		$smlnj/compiler/compiler.cm which puts together and exports the
593		original structure Compiler (or at least something very close to it).
594
595		There are five members of the original structure Compiler
596		that are not exported directly but which instead became members
597		of a new structure Backend (described by signature BACKEND).  These are:
598		structure Profile (: PROFILE), structure Compile (: COMPILE), structure
599		Interact (: INTERACT), structure Machine (: MACHINE), and val
600		architecture (: string).
601
602		Structure Compiler.Version has become structure CompilerVersion.
603
604		Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
605		by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
606		or x86, respectively.
607		Each of these exports the same frontend structures that
608		$smlnj/compiler.cm exports.  But they do not have a structure Backend
609		and instead export some structure <Arch>Backend where <Arch> is Alpha32,
610		Hppa, PPC, Sparc, or X86, respectively.
611
612		Library $smlnj/compiler/all.cm exports the union of the exports of
613		$smlnj/compiler/<arch>.cm
614
615		There are no structures <Arch>Compiler anymore, use
616		$smlnj/compiler/<arch>.cm instead.
617
618		Library host-compiler-0.cm is gone.  Instead, the internal library
619		that instantiates CM is now called cm0.cm.  Selection of the host
620		compiler (backend) is no longer done here but. (Responsibility for it
621		now lies with $smlnj/compiler/current.cm.  This seems to be more
622		logical.)
623
624		Many individual files have been moved or renamed.  Some files have
625		been split into multiple files, and some "dead" files have been deleted.
626
627		Aside from these changes to library organization, there are also changes
628		to the way the code itself is organized:
629
630		Structure Binfile has been re-implemented in such a way that it no
631		longer needs any knowledge of the compiler.  It exclusively deals
632		with the details of binfile layout.  It no longer invokes the
633		compiler (for the purpose of creating new prospective binfile
634		content), and it no longer has any knowledge of how to interpret
635		pickles.
636
637		Structure Compile (: COMPILE) has been stripped down to the bare
638		essentials of compilation.  It no longer deals with linking/execution.
639		The interface has been cleaned up considerably.
640
641		Utility routines for dealing with linking and execution have been
642		moved into their own substructures.
643
644		(The ultimate goal of these changes is to provide a light-weight
645		binfile loader/linker (at least for, e.g., stable libraries) that
646		does not require CM or the compiler to be present.)
647
648		CM documentation has been updated to reflect the changes to library
649		organization.
650
651		----------------------------------------------------------------------
652		Name: Matthias Blume
653		Date: 2001/07/10 17:30:00 EDT
654		Tag: Release_110_34
655		Description:
656
657		Minor tweak to 110.34 (re-tagged):
658
659		- README.html file added to CVS repository
660		- runtime compiles properly under FreeBSD 3.X and 4.X
661
662		----------------------------------------------------------------------
663		Name: Matthias Blume
664		Date: 2001/07/10 17:30:00 EDT
665		Tag: Release_110_34
666		Description:
667
668		New version number (110.34). New bootfiles.
669
670		----------------------------------------------------------------------
671		Name: Matthias Blume
672		Date: 2001/07/09 16:00:00 EDT
673		Tag: blume-20010709-more-varargs
674		Description:
675
676		I changed the handling of varargs in ml-nlffigen again:
677		The ellipsis ... will now simply be ignored (with an accompanying warning).
678
679		The immediate effect is that you can actually call a varargs function
680		from ML -- but you can't actually supply any arguments beyond the ones
681		specified explicitly.  (For example, you can call printf with its format
682		string, but you cannot pass additional arguments.)
683
684		This behavior is only marginally more useful than the one before, but
685		it has the advantage that a function or, more importantly, a function
686		type never gets dropped on the floor, thus avoiding follow-up problems with
687		other types that refer to the offending one.
688
689		----------------------------------------------------------------------
690		Name: Matthias Blume
691		Date: 2001/07/09 11:25:00 EDT
692		Tag: blume-20010709-varargs
693		Description:
694
695		1. ckit-lib.cm now exports structure Error
696		2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
697		with a warning accompanied by a source location.  Moreover, it
698		merely skips the offending function or type and proceeds with the
699		rest of its work.u  As a result, one can safely feed C code containing
700		"..." to ml-nlffigen.
701		3. There are some internal improvements to CM, providing slightly
702		more general string substitutions in the tools subsystem.
703
704		----------------------------------------------------------------------
705		Name: Matthias Blume
706		Date: 2001/06/27 15:10:00 EDT
707		Tag: blume-20010627-concur
708		Description:
709
710		Fixed a small bug in CM's handling of parallel compilation.
711		(You could observe the bug by Control-C-interrupting an ordinary
712		CMB.make or CM.stabilize and then attaching some compile servers.
713		The result was that all of a sudden the previously interrupted
714		compilation would continue on its own.  This was because of
715		an over-optimization: CM did not bother to clean out certain queues
716		when no servers were attached "anyway", resulting in the contents
717		of these queues to grab control when new servers did get attached.)
718
719		There is also another minor update to the CM manual.
720
721		----------------------------------------------------------------------
722		Name: Matthias Blume
723		Date: 2001/06/26 16:15:00 EDT
724		Tag: blume-20010626-cmdoc
725		Description:
726
727		Minor typo fixed in CM manual (syntax diagram for libraries).
728
729		----------------------------------------------------------------------
730		Name: Matthias Blume
731		Date: 2001/06/25 22:55:00 EDT
732		Tag: blume-20010625-x86pc
733		Description:
734
735		Fixed a nasty bug in the X86 assembly code that caused signal
736		handlers to fail (crash) randomly.
737
738		----------------------------------------------------------------------
739		Name: Matthias Blume
740		Date: 2001/06/25 12:05:00 EDT
741		Tag: blume-20010625-nlffigen
742		Description:
743
744		This update fixes a number of minor bugs in ml-nlffigen as reported by
745		Nick Carter <nbc@andrew.cmu.edu>.
746
747		1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
748		2. Default names for generated files are now derived from the name of
749		the C file *without its directory*.  In particular, this causes generated
750		files to be placed locally even if the C file is in some system directory.
751		3. Default names for generated signatures and structures are also derived
752		from the C file name without its directory.  This avoids silly things
753		like "structure GL/GL".
754		(Other silly names are still possible because ml-nlffigen does not do
755		a thorough check of whether generated names are legal ML identifiers.
756		When in doubt, use command line arguments to force particular names.)
757
758		----------------------------------------------------------------------
759		Name: Matthias Blume
760		Date: 2001/06/21 12:25:00 EDT
761		Tag: blume-20010621-eXene
762		Description:
763
764		eXene now compiles and (sort of) works again.
765
766		The library name (for version > 110.33) is $/eXene.cm.
767
768		I also added an new example in src/eXene/examples/nbody.  See the
769		README file there for details.
770
771		----------------------------------------------------------------------
772		Name: Matthias Blume
773		Date: 2001/06/20 16:40:00 EDT
774		Tag: blume-20010620-cml
775		Description:
776
777		CML now compiles and works again.
778
779		Libraries (for version > 110.33):
780
781		$cml/cml.cm            Main CML library.
782		$cml/basis.cm          CML's version of $/basis.cm.
783		$cml/cml-internal.cm   Internal helper library.
784		$cml/core-cml.cm       Internal helper library.
785		$cml-lib/trace-cml.cm  Tracing facility.
786		$cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
787
788		The installer (config/install.sh) has been taught how to properly
789		install this stuff.
790
791		----------------------------------------------------------------------
792		Name: Matthias Blume
793		Date: 2001/06/19 17:55:00 EDT
794		Tag: blume-20010619-instantiate
795		Description:
796
797		This un-breaks the fix for bug 1432.
798		(The bug was originally fixed in 110.9 but I broke it again some
799		time after that.)
800
801		----------------------------------------------------------------------
802		Name: Matthias Blume
803		Date: 2001/06/19 17:25:00 EDT
804		Tag: blume-20010619-signals
805		Description:
806
807		This should (hopefully) fix the long-standing signal handling bug.
808		(The runtime system was constructing a continuation record with an
809		incorrect descriptor which would cause the GC to drop data on the floor...)
810
811		----------------------------------------------------------------------
812		Name: Matthias Blume
813		Date: 2001/06/15 15:05:00 EDT
814		Tag: blume-20010615-moresparc
815		Description:
816
817		Here is a short late-hour update related to Sparc c-calls:
818
819		-- made handling of double-word arguments a bit smarter
820
821		-- instruction selection phase tries to collapse certain clumsily
822		constructed ML-Trees; typical example:
823
824		ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
825
826		This currently has no further impact on SML/NJ since mlriscGen does
827		not seem to generate such patterns in the first place, and c-calls
828		(which did generate them in the beginning) has meanwhile been fixed
829		so as to avoid them as well.
830
831		----------------------------------------------------------------------
832		Name: Matthias Blume
833		Date: 2001/06/15 15:05:00 EDT
834		Tag: blume-20010615-sparc
835		Description:
836
837		The purpose of this update is to provide an implementation of NLFFI
838		on Sparc machines.
839
840		Here are the changes in detail:
841
842		* src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
843		the Sparc implementation of the c-calls API.
844		* The Sparc backend of SML/NJ has been modified to uniformely use %fp
845		for accessing the ML frame.  Thus, we have a real frame pointer and
846		can freely modify %sp without need for an omit-frame-ptr phase.
847		The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
848		this case.
849		* ml-nlffigen has been taught to produce code for different architectures
850		and calling conventions.
851		* In a way similar to what was done in the x86 case, the Sparc
852		backend uses its own specific extension to mltree.  (For example,
853		it needs to be able to generate UNIMP instructions which are part
854		of the calling convention.)
855		* ml-nlffi-lib was reorganized to make it more modular (in particular,
856		to make it easier to plug in new machine- and os-dependent parts).
857
858		There are some other fairly unrelated bug fixes and cleanups as well:
859
860		* I further hacked the .cm files for MLRISC tools (like MDLGen) so
861		that they properly share their libraries with existing SML/NJ libraries.
862		* I fixed a minor cosmetic bug in CM, supressing certain spurious
863		follow-up error messages.
864		* Updates to CM/CMB documentation.
865
866		TODO items:
867
868		* MLRISC should use a different register as its asmTemp on the Sparc.
869		(The current %o2 is a really bad choice because it is part of the
870		calling conventions, so things might interfere in unexpected ways.)
871
872		----------------------------------------------------------------------
873		Name: Matthias Blume
874		Date: 2001/06/07
875		Tag: blume-20010607-calls
876		Description:
877
878		A number of internal changes related to C calls and calling conventions:
879
880		1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
881		number of bytes popped implicitly (by the callee).  In most cases
882		this field is 0 but on x86/win32 it is some non-zero value.  This
883		is information provided for the benefit of the "omit-frameptr" pass.
884		2. The CALL instruction on the x86 carries a similar "pops" field.
885		The instruction selection phase copies its value from the ML-Tree
886		CALL statement.
887		3. On all other architectures, the instruction selection phase checks
888		whether "pops=0" and complains if not.
889		4. The c-calls implementation for x86 now accepts two calling conventions:
890		"ccall" and "stdcall".  When "ccall" is selected, the caller cleans
891		up after the call and pops is set to 0.  For "stdcall", the caller
892		does nothing, leaving the cleanup to the callee; pops is set to
893		the number of bytes that were pushed onto the stack.
894		5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
895		distinguish between "ccall" and "stdcall".
896		6. The UNIMP instruction has been added to the supported Sparc instruction
897		set. (This is needed for implementing the official C calling convention
898		on this architecture.)
899		7. I fixed some of the .cm files under src/MLRISC/Tools to make them
900		work with the latest CM.
901
902		----------------------------------------------------------------------
903		Name: Matthias Blume
904		Date: 2001/06/05 15:10:00 EDT
905		Tag: blume-20010605-cm-index
906		Description:
907
908		0. The "lambdasplit" parameter for class "sml" in CM has been documented.
909
910		1. CM can now generate "index files".  These are human-readable files
911		that list on a per-.cm-file basis each toplevel symbol defined or
912		imported.  The location of the index file for
913		<p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
914		To enable index-file generation, set CM.Control.generate_index to true
915		or export an environment-symbol: export CM_GENERATE_INDEX=true.
916
917		The CM manual has been updated accordingly.
918
919		2. I made some slight modifications to the c-calls API in MLRISC.
920
921		a) There is now a callback to support saving/restoring of
922		dedicated but caller-save registers around the actual call
923		instruction.
924		b) One can optionally specify a comment-annotation for the
925		call instruction.
926
927		3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
928		(For example, the comment annotation shows the C prototype of
929		the function being called.)
930
931		----------------------------------------------------------------------
932		Name: Matthias Blume
933		Date: 2001/06/01 13:30:00 EDT
934		Tag: blume-20010601-nlffi-cleanup
935		Description:
936
937		This is mostly a cleanup of MLFFI stuff:
938
939		- some signature files have been put into a more exposed place
940		- the ugly 'f type parameter is gone (simplifies types tremendously!)
941		- ml-nlffigen changed accordingly
942		- tutorial updated
943
944		Other changes:
945
946		- author's affiliation in CM manual(s) updated
947		- some more recognized keywords added to Allen's sml.sty
948
949		----------------------------------------------------------------------
950		Name: Matthias Blume
951		Date: 2001/05/25 15:30:00 EDT
952		Tag: blume-20010525-iptr
953		Description:
954
955		- put the official 110.33-README (as it appears on the ftp server) under
956		CVS
957		- fixed a small bug related to incomplete pointer types in
958		ml-nlffigen
959		- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
960		(it does not need the 'f type parameter)
961
962		----------------------------------------------------------------------
963		Name: Matthias Blume
964		Date: 2001/05/23 14:30:00 EDT
965		Tag: Release_110_33
966		Description:
967
968		New version number (110.33).  New bootfiles.
969
970		----------------------------------------------------------------------
971		Name: Matthias Blume
972		Date: 2001/05/22 18:06:00 EDT
973		Tag: blume-20010522-targets
974		Description:
975
976		Made install.sh use file config/targets.customized if it exists, falling
977		back to config/targets if it doesn't.  This way one can have a customized
978		version of the targets file without touching the "real thing", thus
979		eliminating the constant fear of accidentally checking something bogus
980		back into the CVS repository...  (File config/targets.customized must
981		not be added to the repository!)
982
983		----------------------------------------------------------------------
984		Name: Matthias Blume
985		Date: 2001/05/22 16:30:00 EDT
986		Tag: blume-20010522-minitut
987		Description:
988
989		1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
990		struct returns.
991		2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
992		incomplete, preliminary documentation for NLFFI.
993
994		----------------------------------------------------------------------
995		Name: Matthias Blume
996		Date: 2001/05/14 11:30:00 EDT
997		Tag: blume-20010514-script
998		Description:
999
1000		Some bugs in install script fixed.
1001
1002		In addition to that I also made a slight change to the NLFFI API:
1003		Functors generated by ml-nlffigen now take the dynamic library as a
1004		straight functor argument, not as a suspended one.  (The original
1005		functor code used to force the suspension right away anyway, so there
1006		was nothing gained by this complication of the interface.)
1007
1008		----------------------------------------------------------------------
1009		Name: Matthias Blume
1010		Date: 2001/05/11 14:35:00 EDT
1011		Tag: blume-20010511-ml-nlffi
1012		Description:
1013
1014		I finally took the plunge and added my new FFI code to the main
1015		repository.  For x86-linux it is now ready for prime-time.
1016
1017		There are two new subdirectories of "src":
1018
1019		- ml-nlffi-lib:
1020		The utility library for programs using the FFI interface.
1021		Here is the implementation of $/c.cm and its associated low-level
1022		partners $/c-int.cm and $/memory.cm.
1023		- ml-nlffigen:
1024		A stand-alone program for generating ML glue code from C source
1025		code.
1026
1027		Building ml-nlffigen requires $/ckit-lib.cm.
1028
1029		The config/install.sh script has been updates to do the Right Thing
1030		(hopefully).
1031
1032		Notice that the source tree for the C-Kit will not be put under "src"
1033		but directly under the installation root directory.  (This is the
1034		structure that currently exists on the CVS server when you check out
1035		module "sml".)  Fortunately, config/install.sh knows about this oddity.
1036
1037		Bugs: No documentation yet.
1038
1039		----------------------------------------------------------------------
1040		Name: Matthias Blume
1041		Date: 2001/05/09 16:35:00 EDT
1042		Tag: blume-20010509-cpscontract
1043		Description:
1044
1045		Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
1046		wrapper/unwrapper elimination did not decrement usage counts and some
1047		dead variables got overlooked by the dead-up logic.)
1048
1049		----------------------------------------------------------------------
1050		Name: Lal George
1051		Date: 2001/05/08  17:26:09 EDT
1052		Tag: george-20010508-omit-frameptr
1053		Description:
1054
1055		Changes to implement the omit-frame-pointer optimization to support
1056		raw C calls. For now, there is only support on the Intel x86, but
1057		other architectures will follow as more experience is gained with this.
1058
1059
1060		----------------------------------------------------------------------
1061		Name: Matthias Blume
1062		Date: 2001/05/07 14:40:00 EDT
1063		Tag: blume-20010507-proxies
1064		Description:
1065
1066		I made into "proxy libraries" all libraries that qualify for such a
1067		change.  (A qualifying library is a library that has another library or
1068		groups as its sole member and repeats that member's export list
1069		verbatim.  A proxy library avoids this repetition by omitting its export
1070		list, effectively inheriting the list that its (only) member exports.
1071		See the CM manual for more explanation.)
1072		The main effect is that explicit export lists for these libraries
1073		do not have to be kepts in sync, making maintenance a bit easier.
1074
1075		I also added copyright notices to many .cm-files.
1076
1077		Last but not least, I made a new set of bootfiles.
1078
1079		----------------------------------------------------------------------
1080		Name: Matthias Blume
1081		Date: 2001/05/04 17:00:00 EDT
1082		Tag: blume-20010504-cm-lsplit
1083		Description:
1084
1085		0. John merged pending changes to $/smlnj-lib.cm
1086
1087		1. Allen's previous change accidentally backed out of one of Lal's
1088		earlier changes.  I undid this mistake (re-introducing Lal's change).
1089
1090		2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1091		within the compiler where applicable.  There is some code simplification
1092		because of that.
1093
1094		3. The "split" phase (in FLINT) is now part of the default list of phases.
1095		Compiler.Control.LambdaSplitting.* can be used to globally control the
1096		lambda-splitting (cross-module-inlining) engine.  In addition to that,
1097		it can now also be controlled on a per-source basis: CM has been taught
1098		a new tool parameter applicable to ML source files.
1099
1100		- To turn lambda-splitting off completely:
1101		local open Compiler.Control.LambdaSplitting in
1102		val _ = set Off
1103		end
1104		- To make "no lambda-splitting" the global default (but allow per-source
1105		overriding); this is the initial setting:
1106		local open Compiler.Control.LambdaSplitting in
1107		val _ = set (Default NONE)
1108		end
1109		- To make "lambda-splitting with aggressiveness a" the global default
1110		(and allow per-source overriding):
1111		local open Compiler.Control.LambdaSplitting in
1112		val _ = set (Default (SOME a))
1113		end
1114
1115		- To turn lambda-splitting off for a given ML souce file (say: a.sml)
1116		write (in the respective .cm-file):
1117		a.sml (lambdasplitting:off)
1118		- To turn lambda-splitting for a.sml on with minimal aggressiveness:
1119		a.sml (lambdasplitting:on)
1120		- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1121		<a> is a decimal non-negative integer):
1122		a.sml (lambdasplitting:<a>)
1123		- To turn lambda-splitting for a.sml on with maximal aggressiveness:
1124		a.sml (lambdasplitting:infinity)
1125		- To use the global default for a.sml:
1126		a.sml (lambdasplitting:default)
1127		or simply
1128		a.sml
1129
1130		----------------------------------------------------------------------
1131		Name: Allen Leung
1132		Date: 2001/05/04 01:57:00 EDT
1133		Tag: leunga-20010504-sync
1134		Description:
1135
1136		MLRISC features.
1137
1138		1. Fix to CMPXCHG instructions.
1139		2. Changed RA interface to allow annotations in callbacks.
1140		3. Added a new method to the stream interface to allow annotations updates.
1141
1142		----------------------------------------------------------------------
1143		Name: Matthias Blume
1144		Date: 2001/05/01 11:45:00 EDT
1145		Tag: blume-20010501-pcedittmp
1146		Description:
1147
1148		Changed install.sh to use the current working directory instead of
1149		/usr/tmp for a temporary file (pcedittmp).  The previous choice
1150		of /usr/tmp caused trouble with MacOS X because of file premission
1151		problems.
1152
1153		----------------------------------------------------------------------
1154		Name: Matthias Blume
1155		Date: 2001/04/20 11:10:00 EDT
1156		Tag: blume-20010420-inMLflag
1157		Description:
1158
1159		- added vp_limitPtrMask to vproc-state.h
1160		(for use by the raw-C-calls mechanism to implement proper interrupt
1161		handling)
1162		- made the ML compiler aware of various data-structure offsets so it
1163		can generate code for accessing the vp_inML flag and vp_limitPtrMask
1164		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
1165		raw C-calls
1166
1167		----------------------------------------------------------------------
1168		Name: Lal George
1169		Date: 2001/04/20 09:15:28 EDT
1170		Tag: george-20010420-macosX
1171		Description:
1172
1173		- Changes to port to Mac OS X; Darwin.
1174
1175		- In the process I found that sqrt was broken on the PPC, because the
1176		fsqrt instruction is not implemented.
1177
1178		----------------------------------------------------------------------
1179		Name: Matthias Blume
1180		Date: 2001/04/18 12:45:00 EDT
1181		Tag: blume-20010418-ccalls
1182		Description:
1183
1184		- fixed two off-by-4 errors in the x86-specific c-calls implementation
1185		(this bug prevented structure arguments containing pointers from being
1186		passed correctly)
1187		- changed the raw-C-call code in mlriscGen.sml in such a way that
1188		structure arguments are represented as a pointer to the beginning
1189		of the structure (instead of having a series of synthesized arguments,
1190		one for each structure member)
1191
1192		- made makeml script's verbosity level configurable via environment
1193		variable (MAKEML_VERBOSITY)
1194
1195		- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1196		in rawmem-x86.sml; we are now using the real thing
1197
1198		----------------------------------------------------------------------
1199		Name: Matthias Blume
1200		Date: 2001/03/22 16:25:00 EST
1201		Tag: blume-20010322-bootfiles
1202		Description:
1203
1204		Created a new set of bootfiles (for your automatic installation convenience).
1205
1206		----------------------------------------------------------------------
1207		Name: Matthias Blume
1208		Date: 2001/03/22 15:10:00 EST
1209		Tag: blume-20010322-rawmem-parcm
1210		Description:
1211
1212		1. All "raw memory access" primitives for the new FFI are implemented now
1213		(at least on the x86).
1214		2. Some further cleanup of CM's parallel make mechanism.
1215
1216		----------------------------------------------------------------------
1217		Name: Matthias Blume
1218		Date: 2001/03/19 17:53:00 EST
1219		Tag: blume-20010319-parallel
1220		Description:
1221
1222		Parallel make (using compile servers) now works again.
1223
1224		To this end, CM.stabilize and CMB.make have been modified to work in
1225		two passes when compile servers are attached:
1226		1. Compile everything, do not perform stabilization; this pass
1227		uses compile servers
1228		2. Stabilize everything; this pass does not use compile servers
1229		If there are no compile servers, the two passes are combined into one
1230		(as before).  Splitting the passes increases the inherent parallelism
1231		in the dependency graph because the entire graph including all
1232		libraries is available at the same time.  This, in turn, improves
1233		server utilization.  The downside is that the master process will
1234		have to do some extra work after compilation is done (because for
1235		technical reasons it must re-read all the binfiles during stabilization).
1236
1237		----------------------------------------------------------------------
1238		Name: Matthias Blume
1239		Date: 2001/03/16 12:22:00 EST
1240		Tag: blume-20010316-bootfiles
1241		Description:
1242
1243		Created a new set of bootfiles (for your automatic installation convenience).
1244
1245		----------------------------------------------------------------------
1246		Name: Matthias Blume
1247		Date: 2001/03/16 11:00:00 EST
1248		Tag: blume-20010316-MLTREE-fixup
1249		Description:
1250
1251		This is a minor fixup for an (untagged) earlier commit by Allen.
1252		(A file was missing).
1253
1254		----------------------------------------------------------------------
1255		Name: Allen Leung
1256		Date: Mon Mar  5 18:54:57 EST 2001
1257		Tag: leunga-20010305-cut-support
1258
1259		1. New support for alternative control-flow in MLTREE.
1260		Currently we support
1261
1262		FLOW_TO(CALL ...., [k1,...,kn])
1263
1264		This is needed for 'cuts to' in C-- and try/handle-like constructs
1265		in Moby
1266
1267		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1268
1269		2. Register Allocator
1270
1271		Changes in interface [from Fermin, John]
1272
1273		3. Alpha 8-bit SLL support [Fermin]
1274
1275		4. All architectures
1276
1277		A new module (ClusterExpandCopies) for expanding parallel copies.
1278
1279		----------------------------------------------------------------------
1280		Name: Allen Leung
1281		Date: 2001/02/27 23:07:00 EST
1282		Tag: leunga-20010227-minor-stuff
1283
1284		1. Alpha bug fix for CMOVNE
1285		2. Handle mltree COND(..,FCMP ...,...)
1286		3. Bug fix in simplifier
1287
1288		----------------------------------------------------------------------
1289		Name: Matthias Blume
1290		Date: 2001/01/30 17:50:00 EST
1291		Tag: blume-20010130-sync
1292		Description:
1293
1294		This is just a minor update to sync my devel branch with the main brach.
1295		The only visible change is the addition of some README files.
1296
1297		----------------------------------------------------------------------
1298		Name: Matthias Blume
1299		Date: 2001/01/12 23:30:00 JST
1300		Tag: blume-20010112-bootfiles
1301		Description:
1302
1303		Made a new set of bootfiles that goes with the current state of the
1304		repository.
1305
1306		----------------------------------------------------------------------
1307		Name: Matthias Blume
1308		Date: 2001/01/12 21:20:00 JST
1309		Tag: blume-20010112-sync
1310		Description:
1311
1312		I am just flushing out some minor changes that had accumulated in
1313		my private branch in order to sync with the main tree.  (This is
1314		mainly because I had CVS trouble when trying to merge _into_ my
1315		private branch.)
1316
1317		Most people should be completely unaffected by this.
1318
1319		----------------------------------------------------------------------
1320		Name: Allen Leung
1321		Date: Thu Jan 11 21:03:00 EST 2001
1322		Tag: leunga-20010111-labexp=mltree
1323		Description:
1324
1325		1.  Removed the type LabelExp and replace it by MLTree.
1326		2.  Rewritten mltree-simplify with the pattern matcher tool.
1327		3.  There were some bugs in alpha code generator which would break
1328		64-bit code generation.
1329		4.  Redo the tools to generate code with the
1330		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1331		are now generated by perl scripts.
1332
1333		----------------------------------------------------------------------
1334		Name: Matthias Blume
1335		Date: 2001/01/10 21:55:00 JST
1336		Tag: blume-20010110-rcc
1337		Description:
1338
1339		The RCC stuff now seems to work (but only on the x86).
1340		This required hacking of the c-calls interface (and -implementation) in
1341		MLRISC.
1342
1343		Normal compiler users should be unaffected.
1344
1345		----------------------------------------------------------------------
1346		Name: Matthias Blume
1347		Date: 2001/01/09 01:20:00 JST
1348		Tag: blume-20010109-rcc
1349		Description:
1350
1351		This is a fairly big patch, flushing out a large number of pending
1352		changes that I made to my development copy over the last couple of days.
1353
1354		Of practical relevance at this moment is a workaround for a pickling
1355		bug that Allen ran into the other day.  The cause of the bug itself is
1356		still unknown and it might be hard to fix it properly, but the
1357		workaround has some merits of its own (namely somewhat reducing pickling
1358		overhead for certain libraries).  Therefore, I think this solution should
1359		be satisfactory at this time.
1360
1361		The rest of the changes (i.e., the vast majority) has to do with my
1362		ongoing efforts of providing direct support for C function calls from
1363		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
1364		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1365		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1366		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1367		handle RCC.
1368
1369		The last part (the changes to mlriscGen.sml) are still known to be
1370		wrong on the x86 and not implemented on all other architectures.  But
1371		the infrastructure is in place. I had to change a few functor
1372		signatures in the backend to be able to route the CCalls interface
1373		from MLRISC there, and I had to specialize the mltree type (on the
1374		x86) to include the necessary extensions. (The extensions themselves
1375		were already there and redy to go in MLRISC/x86).
1376
1377		Everything should be very happy as soon as someone helps me with
1378		mlriscGen.sml...
1379
1380		In any case, nothing of this should matter to anyone as long as the
1381		new primop is not being used (which is going to be the case unless you
1382		find it where I hid it :). The rest of the compiler is completely
1383		unaffected.
1384
1385		----------------------------------------------------------------------
1386		Name: Matthias Blume
1387		Date: 2001/01/05 00:30:00 JST
1388		Tag: blume-20010105-primops
1389		Description:
1390
1391		Added some experimental support for work that I am doing right now.
1392		These changes mostly concern added primops, but there is also a new
1393		experimental C library in the runtime system (but currently not enabled
1394		anywhere except on Linux/X86).
1395
1396		In the course of adding primops (and playing with them), I discovered that
1397		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1398		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
1399		To recover, I made type information in INL_PRIM mandatory and changed
1400		prim.sml as well as built-in.sml accordingly.  The InLine structure now
1401		has complete, correct type information (i.e., no bottom types).
1402
1403		Since all these changes mean that we need new binfiles, I also bumped the
1404		version number to 110.32.1.
1405
1406		----------------------------------------------------------------------
1407		Name: Matthias Blume
1408		Date: 2000/12/30 22:10:00 JST
1409		Tag: blume-20001230-various
1410		Description:
1411
1412		Added proxy libraries for MLRISC and let MLRISC libraries refer
1413		to each other using path anchors.  (See CM manual for explanation.)
1414
1415		Updated CM documentation.
1416
1417		Fixed some bugs in CM.
1418
1419		Implemented "proxy" libraries (= syntactic sugar for CM).
1420
1421		Added "-quiet" option to makeml and changed runtime system accordingly.
1422
1423		Added cleanup handler for exportML to reset timers and compiler stats.
1424
1425		----------------------------------------------------------------------
1426		Name: Lal George
1427		Date: 2000/12/22 22:22:58 EST 2000
1428		Tag: Release_110_32
1429		Description:
1430
1431		Infinite precision used throughout MLRISC.
1432		see MLRISC/mltree/machine-int.sig
1433
1434		----------------------------------------------------------------------
1435		Name: Matthias Blume
1436		Date: 2000/12/22 23:16:00 JST
1437		Tag: blume-20001222-warn
1438		Description:
1439
1440		Corrected wording and formatting of some CM warning message which I
1441		broke in my previous patch.
1442
1443	----------------------------------------------------------------------	----------------------------------------------------------------------
1444	Name: Matthias Blume	Name: Matthias Blume
1445	Date: 2000/12/22 21:20:00 JST	Date: 2000/12/22 21:20:00 JST
1446	Tag: blume-20001222-anchorenv	Tag: blume-20001222-anchorenv
1447	Description:	Description:
1448
1449	Fixed a CM problem related to anchor environments and their interaction	Fixed CM's handling of anchor environments in connection with CMB.make.
	with CMB.make.
1450
1451	----------------------------------------------------------------------	----------------------------------------------------------------------
1452	Name: Matthias Blume	Name: Matthias Blume
#	Line 1797	Line 3225
3225	elaborator). There were a lot of changes during my "linkpath" trials	elaborator). There were a lot of changes during my "linkpath" trials
3226	that could have been reverted to their original state but weren't.	that could have been reverted to their original state but weren't.
3227	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
3228	than what was strictly necessary...  (I _did_ resist the tempation	than what was strictly necessary...  (I _did_ resist the temptation
3229	of doing any "global reformatting" to avoid an untimely death at	of doing any "global reformatting" to avoid an untimely death at
3230	Dave's hands. :)	Dave's hands. :)
3231

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.759   changed lines   Added in v.987

---

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