Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 814, Tue May 1 15:44:45 2001 UTC	revision 1044, Mon Jan 28 21:36:08 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	Name: Matthias Blume
17		Date: 2002/01/28
18		Tag: blume-20020128-allocParam
19		Description:
20
21		John Reppy:  Changes c-calls API to accept client-callback for
22		allocating extra stack space.
23		me: Corresponding changes to mlriscGen (using a dummy argument that
24		does not change the current behavior).
25
26		----------------------------------------------------------------------
27		Name: Matthias Blume
28		Date: 2002/01/28 12:00:00
29		Tag: Release_110_38
30		Description:
31
32		This time for real!!!
33
34		----------------------------------------------------------------------
35		Name: Matthias Blume
36		Date: 2002/01/28 10:56:00 EST
37		Tag: blume-20020128-retraction
38		Description:
39
40		0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
41		with blume-Release_110_38-retracted.)
42
43		1. Fixed a problem with incorrect rounding modes in real64.sml.
44		(Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
45
46		2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
47		and enums fixed.  The naming of corresponding ML identifiers should
48		now be consistent again.
49
50		----------------------------------------------------------------------
51		Name: Allen Leung
52		Date: 2002/01/27
53		Tag: leunga-20020127-nowhere
54		Description:
55
56		Added a target called nowhere in the configuration scripts.
57		Enabling this will build the MLRISC 'nowhere' tool (for translating
58		programs with where-clauses into legal SML code) during installation.
59
60		----------------------------------------------------------------------
61		Name: Matthias Blume
62		Date: 2002/01/25 21:27:00 EST
63		Tag: blume-Release_110_38-retracted
64		Description:
65
66		Call it a (working) release!  Version is 110.38. Bootfiles are ready.
67
68		README will be added later.
69
70		!!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
71		(Release_110_38) removed.  Reason: Last-minute bug fixes.
72
73		----------------------------------------------------------------------
74		Name: Matthias Blume
75		Date: 2002/01/25
76		Tag: blume-20020125-ffi
77		Description:
78
79		A large number of tweaks and improvements to ml-nlffi-lib and
80		ml-nlffigen:
81
82		- ML represenation types have been streamlined
83		- getter and setter functions work with concrete values, not abstract
84		ones where possible
85		- ml-nlffigen command line more flexible (see README file there)
86		- some bugs have been fixed (hopefully)
87
88		----------------------------------------------------------------------
89		Name: Lal George
90		Date: 2002/01/24
91		Tag: george-20020124-risc-ra-interface
92		Description:
93
94		There is a dramatic simplification in the interface to the
95		register allocator for RISC architectures as a result of making
96		parallel copy instructions explicit.
97
98		----------------------------------------------------------------------
99		Name: Matthias Blume
100		Date: 2002/01/22
101		Tag: blume-20020122-x86-ccalls
102		Description:
103
104		Bug fix for c-calls on x86 (having to do with how char- and
105		short-arguments are being handled).
106
107		----------------------------------------------------------------------
108		Name: Matthias Blume
109		Date: 2002/01/21
110		Tag: blume-20020121-ff
111		Description:
112
113		Another day of fiddling with the FFI...
114
115		1. Bug fix/workaround:  CKIT does not complain about negative array
116		dimensions, so ml-nlffigen has to guard itself against this possibility.
117		(Otherwise a negative dimension would send it into an infinite loop.)
118
119		2. Some of the abstract types (light objects, light pointers, most "base"
120		types) in structure C are now eqtypes.
121
122		3. Added constructors and test functions for NULL function pointers.
123
124		----------------------------------------------------------------------
125		Name: Matthias Blume
126		Date: 2002/01/18
127		Tag: blume-20020118-ready-for-new-release
128		Description:
129
130		Made config/srcarchiveurl point to a new place.  (Will provide boot
131		files shortly.)
132
133		Maybe we christen this to be 110.38?
134
135		----------------------------------------------------------------------
136		Name: Matthias Blume
137		Date: 2002/01/18
138		Tag: blume-20020118-more-ffifiddle
139		Description:
140
141		Today's FFI fiddling:
142
143		- Provided a structure CGetSet with "convenient" versions of C.Get.* and
144		C.Set.* that use concrete (MLRep.*) arguments and results instead
145		of abstract ones.
146
147		- Provided word-style bit operations etc. for "int" representation
148		types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
149		and Long.
150
151		----------------------------------------------------------------------
152		Name: Matthias Blume
153		Date: 2002/01/18
154		Tag: blume-20020118-use-x86-fp
155		Description:
156
157		Now that x86-fast-fp seems to be working, I turned it back on again
158		by default.  (Seems to work fine now, even with the FFI.)
159
160		Other than that, I added some documentation about the FFI to
161		src/ml-nlffigen/README and updated the FFI test examples in
162		src/ml-nlffi-lib/Tests/*.
163
164		----------------------------------------------------------------------
165		Name: Allen Leung
166		Date: 2002/01/17
167		Tag: leunga-20020117-x86-fast-fp-call
168		Description:
169
170		1. Fixed a problem with handling return fp values when x86's fast fp
171		mode is turned on.
172
173		2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
174		of %f32.
175
176		3. Added a constructor INT32lit to the ast of MLRISC tools.
177
178		----------------------------------------------------------------------
179		Name: Matthias Blume
180		Date: 2002/01/16
181		Tag: blume-20020116-ffifiddle
182		Description:
183
184		More fiddling with the FFI interface:
185
186		- Make constness 'c instead of rw wherever possible.  This eliminates
187		the need for certain explicit coercions.  (However, due to ML's
188		value polymorphism, there will still be many cases where explicit
189		coercions are necessary.  Phantom types are not the whole answer
190		to modeling a subtyping relationship in ML.)
191
192		- ro/rw coersions for pointers added.  (Avoids the detour through */&.)
193
194		- "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
195		clumsy workaround for varargs problem.)
196
197		----------------------------------------------------------------------
198		Name: Lal George
199		Date: 2002/01/15
200		Tag: <none>
201		Description:
202
203		1. Since COPY instructions are no longer native to the architecture,
204		a generic functor can be used to implement the expandCopies function.
205
206		2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
207		TEXT segment.
208
209		----------------------------------------------------------------------
210		Name: Matthias Blume
211		Date: 2002/01/15
212		Tag: blume-20020115-ffiupdates
213		Description:
214
215		1. Fix for bug resulting in single-precision float values being returned
216		incorrectly from FFI calls.
217
218		2. Small modifications to C FFI API:
219
220		- memory-allocation routines return straight objects (no options)
221		and raise an exception in out-of-memory situations
222		- unsafe extensions to cast between function pointers and pointers
223		from/to ints
224		- added structure C_Debug as an alternative to structure C where
225		pointer-dereferencing (|*| and |*!) always check for null-pointers
226		- added open_lib' to DynLinkage;  open_lib' works like open_lib
227		but also takes a (possibly empty) list of existing library handles
228		that the current library depends on
229
230		----------------------------------------------------------------------
231		Name: Matthias Blume
232		Date: 2002/01/10
233		Tag: blume-20020110-newffigen
234		Description:
235
236		1. Updates to portable graph code.
237
238		2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
239		more scalable now so that even huge interfaces such as the one
240		for GTK compile in finite time and space. :-)
241		See src/ml-nlffigen/README for details on what's new.
242
243		----------------------------------------------------------------------
244		Name: Lal George
245		Date: 2001/01/09 14:31:35 EST 2002
246		Tag: george-20011206-rm-native-copy
247		Description:
248
249		Removed the native COPY and FCOPY instructions
250		from all the architectures and replaced it with the
251		explicit COPY instruction from the previous commit.
252
253		It is now possible to simplify many of the optimizations
254		modules that manipulate copies. This has not been
255		done in this change.
256
257		----------------------------------------------------------------------
258		Name: Lal George
259		Date: 2001/12/06 16:50:13 EST 2001
260		Tag: george-20011206-mlrisc-instruction
261		Description:
262
263		Changed the representation of instructions from being fully abstract
264		to being partially concrete. That is to say:
265
266		from
267		type instruction
268
269		to
270		type instr                              (* machine instruction *)
271
272		datatype instruction =
273		LIVE of {regs: C.cellset, spilled: C.cellset}
274		| KILL of {regs: C.cellset, spilled: C.cellset}
275		| COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
276		| ANNOTATION of {i: instruction, a: Annotations.annotation}
277		| INSTR of instr
278
279		This makes the handling of certain special instructions that appear on
280		all architectures easier and uniform.
281
282		LIVE and KILL say that a list of registers are live or killed at the
283		program point where they appear. No spill code is generated when an
284		element of the 'regs' field is spilled, but the register is moved to
285		the 'spilled' (which is present, more for debugging than anything else).
286
287		LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
288		We used to generate:
289
290		DEFFREG f1
291		f1 := f2 + f3
292		trapb
293
294		but now generate:
295
296		f1 := f2 + f3
297		trapb
298		LIVE {regs=[f1,f2,f3], spilled=[]}
299
300		Furthermore, the DEFFREG (hack) required that all floating point instruction
301		use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
302		defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
303		in a cleaner alpha implementation. (Hopefully, intel will not get rid of
304		this architecture).
305
306		COPYXXX is intended to replace the parallel COPY and FCOPY  available on
307		all the architectures. This will result in further simplification of the
308		register allocator that must be aware of them for coalescing purposes, and
309		will also simplify certain aspects of the machine description that provides
310		callbacks related to parallel copies.
311
312		ANNOTATION should be obvious, and now INSTR represents the honest to God
313		machine instruction set!
314
315		The <arch>/instructions/<arch>Instr.sml files define certain utility
316		functions for making porting easier -- essentially converting upper case
317		to lower case. All machine instructions (of type instr) are in upper case,
318		and the lower case form generates an MLRISC instruction. For example on
319		the alpha we have:
320
321		datatype instr =
322		LDA of {r:cell, b:cell, d:operand}
323		| ...
324
325		val lda : {r:cell, b:cell, d:operand} -> instruction
326		...
327
328		where lda is just (INSTR o LDA), etc.
329
330		----------------------------------------------------------------------
331		Name: Matthias Blume
332		Date: 2001/11/22 21:40:00 EST
333		Tag: Release_110_37
334		Description:
335
336		Release 110.37.  This time for real.
337
338		----------------------------------------------------------------------
339		Name: Matthias Blume
340		Date: 2001/11/21 16:35:00 EST
341		Tag: blume-20011121-foot-in-mouth
342		Description:
343
344		Removed the "Release_110_37" tag because of a serious bug.
345		This will be re-tagged once the bug is fixed.
346
347		----------------------------------------------------------------------
348		Name: Matthias Blume
349		Date: 2001/11/21 16:14:00 EST
350		Tag: blume-20011121-forgottenfile
351		Description:
352
353		Forgot to add a file.  (Just a .tex-file -- part of
354		the CM manual source.)
355
356		----------------------------------------------------------------------
357		Name: Matthias Blume
358		Date: 2001/11/21 16:10:00 EST
359		Tag: blume-20011121-invalid_110_37
360		Description:
361
362		Note: I removed the original tag "Release_110_37" from this commit
363		because we found a serious bug in all non-x86 backends.
364		- Matthias
365
366		1. Modifications to the SML/NJ code generator and to the runtime system
367		so that code object name strings are directly inserted into code
368		objects at code generation time.  The only business the runtime system
369		has with this is now to read the name strings on occasions.
370		(The encoding of the name string has also changed somewhat.)
371
372		2. CM now implements a simple "set calculus" for specifying export lists.
373		In particular, it is now possible to refer to the export lists of
374		other libraries/groups/sources and form unions as well as differences.
375		See the latest CM manual for details.
376
377		3. An separate notion of "proxy" libraries has again be eliminated from
378		CM's model.  (Proxy libraries are now simply a special case of using
379		the export list calculus.)
380
381		4. Some of the existing libraries now take advantage of the new set
382		calculus.
383		(Notice that not all libraries have been converted because some
384		of the existing .cm-files are supposed to be backward compatible
385		with 110.0.x.)
386
387		5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
388		"General.exnMessage"!  The former relies on a certain hook to be
389		initialized, and that often does not happen in the stand-alone case.)
390
391		----------------------------------------------------------------------
392		Name: Lal George
393		Date: 2001/11/21  13:56:18 EST
394		Tag: george-2001121-pseudo-ops
395		Description:
396
397		Implemented a complete redesign of MLRISC pseudo-ops. Now there
398		ought to never be any question of incompatabilities with
399		pseudo-op syntax expected by host assemblers.
400
401		For now, only modules supporting GAS syntax are implemented
402		but more should follow, such as MASM, and vendor assembler
403		syntax, e.g. IBM as, Sun as, etc.
404
405		----------------------------------------------------------------------
406		Name: Matthias Blume
407		Date: 2001/11/14 11:52:00 EST
408		Tag: blume-20011114-srcname
409		Description:
410
411		1. Routed the name of the current source file to mlriscgen where it
412		should be directly emitted into the code object.  (This last part
413		is yet to be done.)
414
415		2. Some cleanup of the pgraph code to make it match the proposal that
416		I put out the other day.  (The proposal notwithstanding, things are
417		still in flux here.)
418
419		----------------------------------------------------------------------
420		Name: Lal George
421		Date: 2001/11/14 09:44:04 EST
422		Tag:
423		Description:
424
425		Fix for a backpatching bug reported by Allen.
426
427		Because the boundary between short and long span-dependent
428		instructions is +/- 128, there are an astounding number of
429		span-dependent instructions whose size is over estimated.
430
431		Allen came up with the idea of letting the size of span
432		dependent instructions be non-monotonic, for a maxIter
433		number of times, after which the size must be monotonically
434		increasing.
435
436		This table shows the number of span-dependent instructions
437		whose size was over-estimated as a function of maxIter, for the
438		file Parse/parse/ml.grm.sml:
439
440		maxIter            # of instructions:
441		10                      687
442		20                      438
443		30                      198
444		40                        0
445
446		In compiling the compiler, there is no significant difference in
447		compilation speed between maxIter=10 and maxIter=40. Actually,
448		my measurements showed that maxIter=40 was a tad faster than
449		maxIter=10! Also 96% of the  files in the compiler reach a fix
450		point within 13 iterations, so fixing maxIter at 40, while high,
451		is okay.
452
453		----------------------------------------------------------------------
454		Name: Matthias Blume
455		Date: 2001/10/31 15:25:00 EST
456		Tag: blume-20011031-pgraph
457		Description:
458
459		CKIT:
460		* Changed the "Function" constructor of type Ast.ctype to carry optional
461		argument identifiers.
462		* Changed the return type of TypeUtil.getFunction accordingly.
463		* Type equality ignores the argument names.
464		* TypeUtil.composite tries to preserve argument names but gives up quickly
465		if there is a mismatch.
466
467		installation script:
468		* attempts to use "curl" if available (unless "wget" is available as well)
469
470		CM:
471		* has an experimental implementation of "portable graphs" which I will
472		soon propose as an implementation-independent library format
473		* there are also new libraries $/pgraph.cm and $/pgraph-util.cm
474
475		NLFFI-LIB:
476		* some cleanup (all cosmetic)
477
478		NLFFIGEN:
479		* temporarily disabled the mechanism that suppresses ML output for
480		C definitions whose identifiers start with an underscore character
481		* generate val bindings for enum constants
482		* user can request that only one style (light or heavy) is being used;
483		default is to use both (command-line arguments: -heavy and -light)
484		* fixed bug in handling of function types involving incomplete pointers
485		* generate ML entry points that take record arguments (i.e., using
486		named arguments) for C functions that have a prototype with named
487		arguments
488		(see changes to CKIT)
489
490		----------------------------------------------------------------------
491		Name: Allen Leung
492		Date: 2001/10/27 20:34:00 EDT
493		Tag: leunga-20011027-x86-fast-fp-call
494		Description:
495
496		Fixed the bug described in blume-20010920-slowfp.
497
498		The fix involves
499		1. generating FCOPYs in FSTP in ia32-svid
500		2. marking a CALL with the appropriate annotation
501
502		----------------------------------------------------------------------
503		Name: Matthias Blume
504		Date: 2001/10/16 11:32:00 EDT
505		Tag: blume-20011016-netbsd
506		Description:
507
508		Underscore patch from Chris Richards (fixing problem with compiling
509		runtime system under recent NetBSD).
510
511		----------------------------------------------------------------------
512		Name: Allen Leung
513		Date: 2001/10/12 17:18:32 EDT 2001
514		Tag: leung-20011012-x86-printflowgraph
515		Description:
516
517		X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
518
519		----------------------------------------------------------------------
520		Name: Lal George
521		Date: 2001/10/11 23:51:34 EDT
522		Tag: george-20011011-too-many-instrs
523		Description:
524
525		The representation of a program point never expected to see more
526		than 65536 instructions in a basic block!
527
528		----------------------------------------------------------------------
529		Name: Lal George
530		Date: 2001/10/09 09:41:37 EDT
531		Tag: george-20011008-mlrisc-labels
532		Description:
533
534		Changed the machine description files to support printing of
535		local and global labels in assembly code, based on host assembler
536		conventions.
537
538		----------------------------------------------------------------------
539		Name: Matthias Blume
540		Date: 2001/09/25 15:25:00 EDT
541		Tag: blume-20010925-exninfo
542		Description:
543
544		I provided a non-hook implementation of exnName (at the toplevel) and
545		made the "dummy" implementation of exnMessage (at the toplevel) more
546		useful: if nothing gets "hooked in", then at least you are going to
547		see the exception name and a message indicating why you don't see more.
548
549		[For the time being, programs that need exnMessage and want to use
550		ml-build should either use General.exnMessage (strongly recommended) or
551		refer to structure General at some other point so that CM sees a
552		static dependency.]
553
554		[Similar remarks go for "print" and "use":  If you want to use their
555		functionality in stand-alone programs generated by ml-build, then use
556		TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
557
558		----------------------------------------------------------------------
559		Name: Matthias Blume
560		Date: 2001/09/20 17:28:00 EDT
561		Tag: blume-20010920-slowfp
562		Description:
563
564		Allen says that x86-fast-fp is not safe yet, so I turned it off again...
565
566		----------------------------------------------------------------------
567		Name: Matthias Blume
568		Date: 2001/09/20 17:20:00 EDT
569		Tag: blume-20010920-canonicalpaths
570		Description:
571
572		0. Updated the BOOT file (something that I forgot to do earlier).
573
574		1. Small internal change to CM so that it avoids "/../" in filenames
575		as much as possible (but only where it is safe).
576
577		2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
578		that arguments that contain delimiters are passed through correctly.
579		This change also means that all "special" arguments of the form
580		@SMLxxx... must come first.
581
582		3. Changed install script to put relative anchor names for tool commands
583		into pathconfig.
584
585		----------------------------------------------------------------------
586		Name: Matthias Blume
587		>>>>>>> 1.169
588		Date: 2001/09/18 15:35:00 EDT
589		Tag: blume-20010918-readme11036
590		Description:
591
592		Added README files.
593
594		----------------------------------------------------------------------
595		Name: Matthias Blume
596		Date: 2001/09/18 11:45:00 EDT
597		Tag: Release_110_36 (retag)
598		Description:
599
600		Fixed mistake in config/preloads. Retagged as 110.36.
601
602		----------------------------------------------------------------------
603		Name: Matthias Blume
604		Date: 2001/09/18 09:40:00 EDT
605		Tag: Release_110_36_orig (tag changed)
606		Description:
607
608		New version (110.36).  New bootfiles.
609
610		----------------------------------------------------------------------
611		Name: Matthias Blume
612		Date: 2001/09/14 16:15:00 EDT
613		Tag: blume-20010914-x86fastfp
614		Description:
615
616		John committed some changes that Allen made, in particular a (hopefully)
617		correctly working version of the x86-fp module.
618
619		I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
620		flag to "true".  Everything seems to compile to a fixpoint ok, and
621		"mandelbrot" speeds up by about 15%.
622
623		----------------------------------------------------------------------
624		Name: Matthias Blume
625		Date: 2001/09/13 11:20:00 EDT
626		Tag: blume-20010913-minimal
627		Description:
628
629		1. Stefan Monnier's patch to fix a miscompilation problem that
630		was brought to light by John Reppy's work on Moby.
631
632		2. Implemented a minimal "structure Compiler" that contains just
633		"version" and "architecture".  The minimal version will be
634		available when the full version is not.  This is for backward-
635		compatibility with code that wants to test Compiler.version.
636
637		----------------------------------------------------------------------
638		Name: Matthias Blume
639		Date: 2001/08/28 14:03:00 EDT
640		Tag: blume-20010828-ml-lex
641		Description:
642
643		Fix for bug 1581, received from Neophytos Michael.
644
645		----------------------------------------------------------------------
646		Name: Matthias Blume
647		Date: 2001/08/27 11:20:00 EDT
648		Tag: blume-20010827-readme11035
649		Description:
650
651		Fleshed out the README file for 110.35.
652
653		----------------------------------------------------------------------
654		Name: Matthias Blume
655		Date: 2001/08/24 17:10:00 EDT
656		Tag: Release_110_35
657		Description:
658
659		New version number (110.35).  New bootfiles.
660
661		----------------------------------------------------------------------
662		Name: Lal George
663		Date: 2001/08/24 13:47:18 EDT 2001
664		Tag: george-20010824-MLRISC-graphs
665		Description:
666
667		removed clusters from MLRISC completely and replaced with graphs.
668
669		----------------------------------------------------------------------
670		Name: Matthias Blume
671		Date: 2001/08/23 17:50:00 EDT
672		Tag: blume-20010823-toplevel
673		Description:
674
675		- some reorganization of the code that implements various kinds of
676		environments in the compiler (static, dynamic, symbolic, combined)
677		- re-implemented the EnvRef module so that evalStream works properly
678		(if the stream contains references to "use", "CM.make", etc.)
679		- cleaned up evalloop.sml and interact.sml (but they need more cleaning)
680
681		----------------------------------------------------------------------
682		Name: Matthias Blume
683		Date: 2001/08/20 15:50 EDT
684		Tag: blume20010820-slipup
685		Description:
686
687		I forgot to commit a few files.  Here they are...
688
689		----------------------------------------------------------------------
690		Name: Matthias Blume
691		Date: 2001/08/20 15:35:00 EDT
692		Tag: blume-20010820-debugprof
693		Description:
694
695		!!!! NEW BOOTFILES !!!!
696
697		This is another round of reorganizing the compiler sources.  This
698		time the main goal was to factor out all the "instrumentation"
699		passes (for profiling and backtracing) into their own library.
700		The difficulty was to do it in such a way that it does not depend
701		on elaborate.cm but only on elabdata.cm.
702
703		Therefore there have been further changes to both elaborate.cm and
704		elabdata.cm -- more "generic" things have been moved from the former
705		to the latter.  As a result, I was forced to split the assignment
706		of numbers indicating "primtyc"s into two portions: SML-generic and
707		SML/NJ-specific.  Since it would have been awkward to maintain,
708		I bit the bullet and actually _changed_ the mapping between these
709		numbers and primtycs.  The bottom line of this is that you need
710		a new set of bin- and bootfiles.
711
712		I have built new bootfiles for all architectures, so doing a fresh
713		checkout and config/install.sh should be all you need.
714
715		The newly created library's name is
716
717		$smlnj/viscomp/debugprof.cm
718
719		and its sources live under
720
721		src/compiler/DebugProf
722
723		----------------------------------------------------------------------
724		Name: Matthias Blume
725		Date: 2001/08/15 17:15:00 EDT
726		Tag: blume-20010815-compreorg
727		Description:
728
729		This is a first cut at reorganizing the CM libraries that make up the
730		core of the compiler.  The idea is to separate out pieces that could
731		be used independently by tools, e.g., the parser, the typechecker, etc.
732
733		The current status is a step in this direction, but it is not quite
734		satisfactory yet.  Expect more changes in the future.
735
736		Here is the current (new) organization...
737
738		What used to be $smlnj/viscomp/core.cm is now divided into
739		six CM libraries:
740
741		$smlnj/viscomp/basics.cm
742		/parser.cm
743		/elabdata.cm
744		/elaborate.cm
745		/execute.cm
746		/core.cm
747
748		The CM files for these libraries live under src/system/smlnj/viscomp.
749		All these libraries are proxy libraries that contain precisely
750		one CM library component.  Here are the locations of the components
751		(all within the src/compiler tree):
752
753		Basics/basics.cm
754		Parse/parser.cm
755		ElabData/elabdata.cm
756		Elaborator/elaborate.cm
757		Execution/execute.cm
758		core.cm
759
760		[This organization is the same that has been used already
761		for a while for the architecture-specific parts of the visible
762		compiler and for the old version of core.cm.]
763
764		As you will notice, many source files have been moved from their
765		respective original locations to a new home in one of the above
766		subtrees.
767
768		The division of labor between the new libraries is the following:
769
770		basics.cm:
771		- Simple, basic definitions that pertain to many (or all) of
772		the other libraries.
773		parser.cm:
774		- The SML parser, producing output of type Ast.dec.
775		- The type family for Ast is also defined and exported here.
776		elabdata.cm:
777		- The datatypes that describe input and output of the elaborator.
778		This includes types, absyn, and static environments.
779		elaborator.cm:
780		- The SML/NJ type checker and elaborator.
781		This maps an Ast.dec (with a given static environment) to
782		an Absyn.dec (with a new static environment).
783		- This libraries implements certain modules that used to be
784		structures as functors (to remove dependencies on FLINT).
785		execute.cm:
786		- Everything having to do with executing binary code objects.
787		- Dynamic environments.
788		core.cm:
789		- SML/NJ-specific instantiations of the elaborator and MLRISC.
790		- Top-level modules.
791		- FLINT (this should eventually become its own library)
792
793		Notes:
794
795		I am not 100% happy with the way I separated the elaborator (and its
796		data structures) from FLINT.  Two instances of the same problem:
797
798		1. Data structures contain certain fields that carry FLINT-specific
799		information.  I hacked around this using exn and the property list
800		module from smlnj-lib.  But the fact that there are middle-end
801		specific fields around at all is a bit annoying.
802
803		2. The elaborator calculates certain FLINT-related information.  I tried
804		to make this as abstract as I could using functorization, but, again,
805		the fact that the elaborator has to perform calculations on behalf
806		of the middle-end at all is not nice.
807
808		3. Having to used exn and property lists is unfortunate because it
809		weakens type checking.  The other alternative (parameterizing
810		nearly *everything*) is not appealing, though.
811
812		I removed the "rebinding =" warning hack because due to the new organization
813		it was awkward to maintain it.  As a result, the compiler now issues some of
814		these warnings when compiling init.cmi during bootstrap compilation. On
815		the plus side, you also get a warning when you do, for example:
816		val op = = Int32.+
817		which was not the case up to now.
818
819		I placed "assign" and "deref" into the _Core structure so that the
820		code that deals with the "lazy" keyword can find them there.  This
821		removes the need for having access to the primitive environment
822		during elaboration.
823
824		----------------------------------------------------------------------
825		Name: Matthias Blume
826		Date: 2001/08/13
827		Tag: blume-20010813-closures
828		Description:
829
830		This fix was sent to us by Zhong Shao.  It is supposed to improve the
831		performance of certain loops by avoiding needless closure allocation.
832
833		----------------------------------------------------------------------
834		Name: Lal George
835		Date: 2001/07/31 10:03:23 EDT 2001
836		Tag: george-20010731-x86-fmalloc
837		Description: Fixed bug in x86 calls
838
839		There was a bug where call instructions would mysteriously
840		vanish. The call instruction had to be one that returned
841		a floating point value.
842
843		----------------------------------------------------------------------
844		Name: Lal George
845		Date: 2001/07/19 16:36:29 EDT 2001
846		Tag: george-20010719-simple-cells
847		Description:
848
849		I have dramatically simplified the interface for CELLS in MLRISC.
850
851		In summary, the cells interface is broken up into three parts:
852
853		1. CellsBasis : CELLS_BASIS
854
855		CellsBasis is a top level structure and common for all
856		architectures.  it contains the definitions of basic datatypes
857		and utility  functions over these types.
858
859		2. functor Cells() : CELLS
860
861		Cells generates an interface for CELLS that incorporates the
862		specific resources on the target architecture, such as the
863		presence of special register classes, their number and size,
864		and various useful substructures.
865
866		3. <ARCH>CELLS
867
868		e.g. SparcCells: SPARCCELLS
869
870		<ARCH>CELLS usually contains additional bindings for special
871		registers  on the architecture, such as:
872
873		val r0 : cell           (* register zero *)
874		val y : cell            (* Y register *)
875		val psr : cell          (* processor status register *)
876		...
877
878		The structure returned by applying the Cells functor is opened
879		in this interface.
880
881		The main implication of all this is that the datatypes for cells is
882		split between CellsBasis and CELLS -- a fairly simple change for user
883		code.
884
885		In the old scheme the CELLS interface had a definitional binding of
886		the form:
887
888		signature CELLS = sig
889
890		structure CellsBasis = CellsBasis
891
892		...
893
894		end
895
896		With all the sharing constraints that goes on in MLRISC, this old
897		design  quickly leads to errors such as:
898
899		"structure definition spec inside of sharing ... "
900
901
902		and appears to require an unacceptable amount of sharing and where
903		constraint hackery.
904
905		I think this error message (the interaction of definitional specs and
906		sharing) requires more explanation on our web page.
907
908		----------------------------------------------------------------------
909		Name: Matthias Blume
910		Date: 2001/07/19 15:00:00 EDT
911		Tag: blume-20010719-libreorg
912		Description:
913
914		This update puts together a fairly extensive but straightforward change
915		to the way the libraries that implement the interactive system are
916		organized:
917
918		The biggest change is the elimination of structure Compiler.  As a
919		replacement for this structure, there is now a CM library
920		(known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
921		that exports all the substructures of the original structure Compiler
922		directly.  So instead of saying Compiler.Foo.bar one now simply
923		says Foo.bar.  (The CM libraries actually export a collection of
924		structures that is richer than the collection of substructures of
925		structure Compiler.)
926
927		To make the transition smooth, there is a separate library called
928		$smlnj/compiler/compiler.cm which puts together and exports the
929		original structure Compiler (or at least something very close to it).
930
931		There are five members of the original structure Compiler
932		that are not exported directly but which instead became members
933		of a new structure Backend (described by signature BACKEND).  These are:
934		structure Profile (: PROFILE), structure Compile (: COMPILE), structure
935		Interact (: INTERACT), structure Machine (: MACHINE), and val
936		architecture (: string).
937
938		Structure Compiler.Version has become structure CompilerVersion.
939
940		Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
941		by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
942		or x86, respectively.
943		Each of these exports the same frontend structures that
944		$smlnj/compiler.cm exports.  But they do not have a structure Backend
945		and instead export some structure <Arch>Backend where <Arch> is Alpha32,
946		Hppa, PPC, Sparc, or X86, respectively.
947
948		Library $smlnj/compiler/all.cm exports the union of the exports of
949		$smlnj/compiler/<arch>.cm
950
951		There are no structures <Arch>Compiler anymore, use
952		$smlnj/compiler/<arch>.cm instead.
953
954		Library host-compiler-0.cm is gone.  Instead, the internal library
955		that instantiates CM is now called cm0.cm.  Selection of the host
956		compiler (backend) is no longer done here but. (Responsibility for it
957		now lies with $smlnj/compiler/current.cm.  This seems to be more
958		logical.)
959
960		Many individual files have been moved or renamed.  Some files have
961		been split into multiple files, and some "dead" files have been deleted.
962
963		Aside from these changes to library organization, there are also changes
964		to the way the code itself is organized:
965
966		Structure Binfile has been re-implemented in such a way that it no
967		longer needs any knowledge of the compiler.  It exclusively deals
968		with the details of binfile layout.  It no longer invokes the
969		compiler (for the purpose of creating new prospective binfile
970		content), and it no longer has any knowledge of how to interpret
971		pickles.
972
973		Structure Compile (: COMPILE) has been stripped down to the bare
974		essentials of compilation.  It no longer deals with linking/execution.
975		The interface has been cleaned up considerably.
976
977		Utility routines for dealing with linking and execution have been
978		moved into their own substructures.
979
980		(The ultimate goal of these changes is to provide a light-weight
981		binfile loader/linker (at least for, e.g., stable libraries) that
982		does not require CM or the compiler to be present.)
983
984		CM documentation has been updated to reflect the changes to library
985		organization.
986
987		----------------------------------------------------------------------
988		Name: Matthias Blume
989		Date: 2001/07/10 17:30:00 EDT
990		Tag: Release_110_34
991		Description:
992
993		Minor tweak to 110.34 (re-tagged):
994
995		- README.html file added to CVS repository
996		- runtime compiles properly under FreeBSD 3.X and 4.X
997
998		----------------------------------------------------------------------
999		Name: Matthias Blume
1000		Date: 2001/07/10 17:30:00 EDT
1001		Tag: Release_110_34
1002		Description:
1003
1004		New version number (110.34). New bootfiles.
1005
1006		----------------------------------------------------------------------
1007		Name: Matthias Blume
1008		Date: 2001/07/09 16:00:00 EDT
1009		Tag: blume-20010709-more-varargs
1010		Description:
1011
1012		I changed the handling of varargs in ml-nlffigen again:
1013		The ellipsis ... will now simply be ignored (with an accompanying warning).
1014
1015		The immediate effect is that you can actually call a varargs function
1016		from ML -- but you can't actually supply any arguments beyond the ones
1017		specified explicitly.  (For example, you can call printf with its format
1018		string, but you cannot pass additional arguments.)
1019
1020		This behavior is only marginally more useful than the one before, but
1021		it has the advantage that a function or, more importantly, a function
1022		type never gets dropped on the floor, thus avoiding follow-up problems with
1023		other types that refer to the offending one.
1024
1025		----------------------------------------------------------------------
1026		Name: Matthias Blume
1027		Date: 2001/07/09 11:25:00 EDT
1028		Tag: blume-20010709-varargs
1029		Description:
1030
1031		1. ckit-lib.cm now exports structure Error
1032		2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1033		with a warning accompanied by a source location.  Moreover, it
1034		merely skips the offending function or type and proceeds with the
1035		rest of its work.u  As a result, one can safely feed C code containing
1036		"..." to ml-nlffigen.
1037		3. There are some internal improvements to CM, providing slightly
1038		more general string substitutions in the tools subsystem.
1039
1040		----------------------------------------------------------------------
1041		Name: Matthias Blume
1042		Date: 2001/06/27 15:10:00 EDT
1043		Tag: blume-20010627-concur
1044		Description:
1045
1046		Fixed a small bug in CM's handling of parallel compilation.
1047		(You could observe the bug by Control-C-interrupting an ordinary
1048		CMB.make or CM.stabilize and then attaching some compile servers.
1049		The result was that all of a sudden the previously interrupted
1050		compilation would continue on its own.  This was because of
1051		an over-optimization: CM did not bother to clean out certain queues
1052		when no servers were attached "anyway", resulting in the contents
1053		of these queues to grab control when new servers did get attached.)
1054
1055		There is also another minor update to the CM manual.
1056
1057		----------------------------------------------------------------------
1058		Name: Matthias Blume
1059		Date: 2001/06/26 16:15:00 EDT
1060		Tag: blume-20010626-cmdoc
1061		Description:
1062
1063		Minor typo fixed in CM manual (syntax diagram for libraries).
1064
1065		----------------------------------------------------------------------
1066		Name: Matthias Blume
1067		Date: 2001/06/25 22:55:00 EDT
1068		Tag: blume-20010625-x86pc
1069		Description:
1070
1071		Fixed a nasty bug in the X86 assembly code that caused signal
1072		handlers to fail (crash) randomly.
1073
1074		----------------------------------------------------------------------
1075		Name: Matthias Blume
1076		Date: 2001/06/25 12:05:00 EDT
1077		Tag: blume-20010625-nlffigen
1078		Description:
1079
1080		This update fixes a number of minor bugs in ml-nlffigen as reported by
1081		Nick Carter <nbc@andrew.cmu.edu>.
1082
1083		1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1084		2. Default names for generated files are now derived from the name of
1085		the C file *without its directory*.  In particular, this causes generated
1086		files to be placed locally even if the C file is in some system directory.
1087		3. Default names for generated signatures and structures are also derived
1088		from the C file name without its directory.  This avoids silly things
1089		like "structure GL/GL".
1090		(Other silly names are still possible because ml-nlffigen does not do
1091		a thorough check of whether generated names are legal ML identifiers.
1092		When in doubt, use command line arguments to force particular names.)
1093
1094		----------------------------------------------------------------------
1095		Name: Matthias Blume
1096		Date: 2001/06/21 12:25:00 EDT
1097		Tag: blume-20010621-eXene
1098		Description:
1099
1100		eXene now compiles and (sort of) works again.
1101
1102		The library name (for version > 110.33) is $/eXene.cm.
1103
1104		I also added an new example in src/eXene/examples/nbody.  See the
1105		README file there for details.
1106
1107		----------------------------------------------------------------------
1108		Name: Matthias Blume
1109		Date: 2001/06/20 16:40:00 EDT
1110		Tag: blume-20010620-cml
1111		Description:
1112
1113		CML now compiles and works again.
1114
1115		Libraries (for version > 110.33):
1116
1117		$cml/cml.cm            Main CML library.
1118		$cml/basis.cm          CML's version of $/basis.cm.
1119		$cml/cml-internal.cm   Internal helper library.
1120		$cml/core-cml.cm       Internal helper library.
1121		$cml-lib/trace-cml.cm  Tracing facility.
1122		$cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1123
1124		The installer (config/install.sh) has been taught how to properly
1125		install this stuff.
1126
1127		----------------------------------------------------------------------
1128		Name: Matthias Blume
1129		Date: 2001/06/19 17:55:00 EDT
1130		Tag: blume-20010619-instantiate
1131		Description:
1132
1133		This un-breaks the fix for bug 1432.
1134		(The bug was originally fixed in 110.9 but I broke it again some
1135		time after that.)
1136
1137		----------------------------------------------------------------------
1138		Name: Matthias Blume
1139		Date: 2001/06/19 17:25:00 EDT
1140		Tag: blume-20010619-signals
1141		Description:
1142
1143		This should (hopefully) fix the long-standing signal handling bug.
1144		(The runtime system was constructing a continuation record with an
1145		incorrect descriptor which would cause the GC to drop data on the floor...)
1146
1147		----------------------------------------------------------------------
1148		Name: Matthias Blume
1149		Date: 2001/06/15 15:05:00 EDT
1150		Tag: blume-20010615-moresparc
1151		Description:
1152
1153		Here is a short late-hour update related to Sparc c-calls:
1154
1155		-- made handling of double-word arguments a bit smarter
1156
1157		-- instruction selection phase tries to collapse certain clumsily
1158		constructed ML-Trees; typical example:
1159
1160		ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1161
1162		This currently has no further impact on SML/NJ since mlriscGen does
1163		not seem to generate such patterns in the first place, and c-calls
1164		(which did generate them in the beginning) has meanwhile been fixed
1165		so as to avoid them as well.
1166
1167		----------------------------------------------------------------------
1168		Name: Matthias Blume
1169		Date: 2001/06/15 15:05:00 EDT
1170		Tag: blume-20010615-sparc
1171		Description:
1172
1173		The purpose of this update is to provide an implementation of NLFFI
1174		on Sparc machines.
1175
1176		Here are the changes in detail:
1177
1178		* src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1179		the Sparc implementation of the c-calls API.
1180		* The Sparc backend of SML/NJ has been modified to uniformely use %fp
1181		for accessing the ML frame.  Thus, we have a real frame pointer and
1182		can freely modify %sp without need for an omit-frame-ptr phase.
1183		The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1184		this case.
1185		* ml-nlffigen has been taught to produce code for different architectures
1186		and calling conventions.
1187		* In a way similar to what was done in the x86 case, the Sparc
1188		backend uses its own specific extension to mltree.  (For example,
1189		it needs to be able to generate UNIMP instructions which are part
1190		of the calling convention.)
1191		* ml-nlffi-lib was reorganized to make it more modular (in particular,
1192		to make it easier to plug in new machine- and os-dependent parts).
1193
1194		There are some other fairly unrelated bug fixes and cleanups as well:
1195
1196		* I further hacked the .cm files for MLRISC tools (like MDLGen) so
1197		that they properly share their libraries with existing SML/NJ libraries.
1198		* I fixed a minor cosmetic bug in CM, supressing certain spurious
1199		follow-up error messages.
1200		* Updates to CM/CMB documentation.
1201
1202		TODO items:
1203
1204		* MLRISC should use a different register as its asmTemp on the Sparc.
1205		(The current %o2 is a really bad choice because it is part of the
1206		calling conventions, so things might interfere in unexpected ways.)
1207
1208		----------------------------------------------------------------------
1209		Name: Matthias Blume
1210		Date: 2001/06/07
1211		Tag: blume-20010607-calls
1212		Description:
1213
1214		A number of internal changes related to C calls and calling conventions:
1215
1216		1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1217		number of bytes popped implicitly (by the callee).  In most cases
1218		this field is 0 but on x86/win32 it is some non-zero value.  This
1219		is information provided for the benefit of the "omit-frameptr" pass.
1220		2. The CALL instruction on the x86 carries a similar "pops" field.
1221		The instruction selection phase copies its value from the ML-Tree
1222		CALL statement.
1223		3. On all other architectures, the instruction selection phase checks
1224		whether "pops=0" and complains if not.
1225		4. The c-calls implementation for x86 now accepts two calling conventions:
1226		"ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1227		up after the call and pops is set to 0.  For "stdcall", the caller
1228		does nothing, leaving the cleanup to the callee; pops is set to
1229		the number of bytes that were pushed onto the stack.
1230		5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1231		distinguish between "ccall" and "stdcall".
1232		6. The UNIMP instruction has been added to the supported Sparc instruction
1233		set. (This is needed for implementing the official C calling convention
1234		on this architecture.)
1235		7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1236		work with the latest CM.
1237
1238		----------------------------------------------------------------------
1239		Name: Matthias Blume
1240		Date: 2001/06/05 15:10:00 EDT
1241		Tag: blume-20010605-cm-index
1242		Description:
1243
1244		0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1245
1246		1. CM can now generate "index files".  These are human-readable files
1247		that list on a per-.cm-file basis each toplevel symbol defined or
1248		imported.  The location of the index file for
1249		<p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1250		To enable index-file generation, set CM.Control.generate_index to true
1251		or export an environment-symbol: export CM_GENERATE_INDEX=true.
1252
1253		The CM manual has been updated accordingly.
1254
1255		2. I made some slight modifications to the c-calls API in MLRISC.
1256
1257		a) There is now a callback to support saving/restoring of
1258		dedicated but caller-save registers around the actual call
1259		instruction.
1260		b) One can optionally specify a comment-annotation for the
1261		call instruction.
1262
1263		3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1264		(For example, the comment annotation shows the C prototype of
1265		the function being called.)
1266
1267		----------------------------------------------------------------------
1268		Name: Matthias Blume
1269		Date: 2001/06/01 13:30:00 EDT
1270		Tag: blume-20010601-nlffi-cleanup
1271		Description:
1272
1273		This is mostly a cleanup of MLFFI stuff:
1274
1275		- some signature files have been put into a more exposed place
1276		- the ugly 'f type parameter is gone (simplifies types tremendously!)
1277		- ml-nlffigen changed accordingly
1278		- tutorial updated
1279
1280		Other changes:
1281
1282		- author's affiliation in CM manual(s) updated
1283		- some more recognized keywords added to Allen's sml.sty
1284
1285		----------------------------------------------------------------------
1286		Name: Matthias Blume
1287		Date: 2001/05/25 15:30:00 EDT
1288		Tag: blume-20010525-iptr
1289		Description:
1290
1291		- put the official 110.33-README (as it appears on the ftp server) under
1292		CVS
1293		- fixed a small bug related to incomplete pointer types in
1294		ml-nlffigen
1295		- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1296		(it does not need the 'f type parameter)
1297
1298		----------------------------------------------------------------------
1299		Name: Matthias Blume
1300		Date: 2001/05/23 14:30:00 EDT
1301		Tag: Release_110_33
1302		Description:
1303
1304		New version number (110.33).  New bootfiles.
1305
1306		----------------------------------------------------------------------
1307		Name: Matthias Blume
1308		Date: 2001/05/22 18:06:00 EDT
1309		Tag: blume-20010522-targets
1310		Description:
1311
1312		Made install.sh use file config/targets.customized if it exists, falling
1313		back to config/targets if it doesn't.  This way one can have a customized
1314		version of the targets file without touching the "real thing", thus
1315		eliminating the constant fear of accidentally checking something bogus
1316		back into the CVS repository...  (File config/targets.customized must
1317		not be added to the repository!)
1318
1319		----------------------------------------------------------------------
1320		Name: Matthias Blume
1321		Date: 2001/05/22 16:30:00 EDT
1322		Tag: blume-20010522-minitut
1323		Description:
1324
1325		1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1326		struct returns.
1327		2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
1328		incomplete, preliminary documentation for NLFFI.
1329
1330		----------------------------------------------------------------------
1331		Name: Matthias Blume
1332		Date: 2001/05/14 11:30:00 EDT
1333		Tag: blume-20010514-script
1334		Description:
1335
1336		Some bugs in install script fixed.
1337
1338		In addition to that I also made a slight change to the NLFFI API:
1339		Functors generated by ml-nlffigen now take the dynamic library as a
1340		straight functor argument, not as a suspended one.  (The original
1341		functor code used to force the suspension right away anyway, so there
1342		was nothing gained by this complication of the interface.)
1343
1344		----------------------------------------------------------------------
1345		Name: Matthias Blume
1346		Date: 2001/05/11 14:35:00 EDT
1347		Tag: blume-20010511-ml-nlffi
1348		Description:
1349
1350		I finally took the plunge and added my new FFI code to the main
1351		repository.  For x86-linux it is now ready for prime-time.
1352
1353		There are two new subdirectories of "src":
1354
1355		- ml-nlffi-lib:
1356		The utility library for programs using the FFI interface.
1357		Here is the implementation of $/c.cm and its associated low-level
1358		partners $/c-int.cm and $/memory.cm.
1359		- ml-nlffigen:
1360		A stand-alone program for generating ML glue code from C source
1361		code.
1362
1363		Building ml-nlffigen requires $/ckit-lib.cm.
1364
1365		The config/install.sh script has been updates to do the Right Thing
1366		(hopefully).
1367
1368		Notice that the source tree for the C-Kit will not be put under "src"
1369		but directly under the installation root directory.  (This is the
1370		structure that currently exists on the CVS server when you check out
1371		module "sml".)  Fortunately, config/install.sh knows about this oddity.
1372
1373		Bugs: No documentation yet.
1374
1375		----------------------------------------------------------------------
1376		Name: Matthias Blume
1377		Date: 2001/05/09 16:35:00 EDT
1378		Tag: blume-20010509-cpscontract
1379		Description:
1380
1381		Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
1382		wrapper/unwrapper elimination did not decrement usage counts and some
1383		dead variables got overlooked by the dead-up logic.)
1384
1385		----------------------------------------------------------------------
1386		Name: Lal George
1387		Date: 2001/05/08  17:26:09 EDT
1388		Tag: george-20010508-omit-frameptr
1389		Description:
1390
1391		Changes to implement the omit-frame-pointer optimization to support
1392		raw C calls. For now, there is only support on the Intel x86, but
1393		other architectures will follow as more experience is gained with this.
1394
1395
1396		----------------------------------------------------------------------
1397		Name: Matthias Blume
1398		Date: 2001/05/07 14:40:00 EDT
1399		Tag: blume-20010507-proxies
1400		Description:
1401
1402		I made into "proxy libraries" all libraries that qualify for such a
1403		change.  (A qualifying library is a library that has another library or
1404		groups as its sole member and repeats that member's export list
1405		verbatim.  A proxy library avoids this repetition by omitting its export
1406		list, effectively inheriting the list that its (only) member exports.
1407		See the CM manual for more explanation.)
1408		The main effect is that explicit export lists for these libraries
1409		do not have to be kepts in sync, making maintenance a bit easier.
1410
1411		I also added copyright notices to many .cm-files.
1412
1413		Last but not least, I made a new set of bootfiles.
1414
1415		----------------------------------------------------------------------
1416		Name: Matthias Blume
1417		Date: 2001/05/04 17:00:00 EDT
1418		Tag: blume-20010504-cm-lsplit
1419		Description:
1420
1421		0. John merged pending changes to $/smlnj-lib.cm
1422
1423		1. Allen's previous change accidentally backed out of one of Lal's
1424		earlier changes.  I undid this mistake (re-introducing Lal's change).
1425
1426		2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1427		within the compiler where applicable.  There is some code simplification
1428		because of that.
1429
1430		3. The "split" phase (in FLINT) is now part of the default list of phases.
1431		Compiler.Control.LambdaSplitting.* can be used to globally control the
1432		lambda-splitting (cross-module-inlining) engine.  In addition to that,
1433		it can now also be controlled on a per-source basis: CM has been taught
1434		a new tool parameter applicable to ML source files.
1435
1436		- To turn lambda-splitting off completely:
1437		local open Compiler.Control.LambdaSplitting in
1438		val _ = set Off
1439		end
1440		- To make "no lambda-splitting" the global default (but allow per-source
1441		overriding); this is the initial setting:
1442		local open Compiler.Control.LambdaSplitting in
1443		val _ = set (Default NONE)
1444		end
1445		- To make "lambda-splitting with aggressiveness a" the global default
1446		(and allow per-source overriding):
1447		local open Compiler.Control.LambdaSplitting in
1448		val _ = set (Default (SOME a))
1449		end
1450
1451		- To turn lambda-splitting off for a given ML souce file (say: a.sml)
1452		write (in the respective .cm-file):
1453		a.sml (lambdasplitting:off)
1454		- To turn lambda-splitting for a.sml on with minimal aggressiveness:
1455		a.sml (lambdasplitting:on)
1456		- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1457		<a> is a decimal non-negative integer):
1458		a.sml (lambdasplitting:<a>)
1459		- To turn lambda-splitting for a.sml on with maximal aggressiveness:
1460		a.sml (lambdasplitting:infinity)
1461		- To use the global default for a.sml:
1462		a.sml (lambdasplitting:default)
1463		or simply
1464		a.sml
1465
1466		----------------------------------------------------------------------
1467		Name: Allen Leung
1468		Date: 2001/05/04 01:57:00 EDT
1469		Tag: leunga-20010504-sync
1470		Description:
1471
1472		MLRISC features.
1473
1474		1. Fix to CMPXCHG instructions.
1475		2. Changed RA interface to allow annotations in callbacks.
1476		3. Added a new method to the stream interface to allow annotations updates.
1477
1478		----------------------------------------------------------------------
1479		Name: Matthias Blume
1480	Date: 2001/05/01 11:45:00 EDT	Date: 2001/05/01 11:45:00 EDT
1481	Tag: blume-20010501-pcedittmp	Tag: blume-20010501-pcedittmp
1482	Description:	Description:
#	Line 2098	Line 3561
3561	elaborator). There were a lot of changes during my "linkpath" trials	elaborator). There were a lot of changes during my "linkpath" trials
3562	that could have been reverted to their original state but weren't.	that could have been reverted to their original state but weren't.
3563	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
3564	than what was strictly necessary...  (I _did_ resist the tempation	than what was strictly necessary...  (I _did_ resist the temptation
3565	of doing any "global reformatting" to avoid an untimely death at	of doing any "global reformatting" to avoid an untimely death at
3566	Dave's hands. :)	Dave's hands. :)
3567

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.814   changed lines   Added in v.1044

---

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