Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 842, Tue Jun 19 21:24:35 2001 UTC	revision 1015, Tue Jan 15 21:46:00 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		Name: Matthias Blume
16		Date: 2002/01/15
17		Tag: blume-20020115-ffiupdates
18		Description:
19
20		1. Fix for bug resulting in single-precision float values being returned
21		incorrectly from FFI calls.
22
23		2. Small modifications to C FFI API:
24
25		- memory-allocation routines return straight objects (no options)
26		and raise an exception in out-of-memory situations
27		- unsafe extensions to cast between function pointers and pointers
28		from/to ints
29		- added structure C_Debug as an alternative to structure C where
30		pointer-dereferencing (|*| and |*!) always check for null-pointers
31		- added open_lib' to DynLinkage;  open_lib' works like open_lib
32		but also takes a (possibly empty) list of existing library handles
33		that the current library depends on
34
35		----------------------------------------------------------------------
36		Name: Matthias Blume
37		Date: 2002/01/10
38		Tag: blume-20020110-newffigen
39		Description:
40
41		1. Updates to portable graph code.
42
43		2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
44		more scalable now so that even huge interfaces such as the one
45		for GTK compile in finite time and space. :-)
46		See src/ml-nlffigen/README for details on what's new.
47
48		----------------------------------------------------------------------
49		Name: Lal George
50		Date: 2001/01/09 14:31:35 EST 2002
51		Tag: george-20011206-rm-native-copy
52		Description:
53
54		Removed the native COPY and FCOPY instructions
55		from all the architectures and replaced it with the
56		explicit COPY instruction from the previous commit.
57
58		It is now possible to simplify many of the optimizations
59		modules that manipulate copies. This has not been
60		done in this change.
61
62		----------------------------------------------------------------------
63		Name: Lal George
64		Date: 2001/12/06 16:50:13 EST 2001
65		Tag: george-20011206-mlrisc-instruction
66		Description:
67
68		Changed the representation of instructions from being fully abstract
69		to being partially concrete. That is to say:
70
71		from
72		type instruction
73
74		to
75		type instr                              (* machine instruction *)
76
77		datatype instruction =
78		LIVE of {regs: C.cellset, spilled: C.cellset}
79		| KILL of {regs: C.cellset, spilled: C.cellset}
80		| COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
81		| ANNOTATION of {i: instruction, a: Annotations.annotation}
82		| INSTR of instr
83
84		This makes the handling of certain special instructions that appear on
85		all architectures easier and uniform.
86
87		LIVE and KILL say that a list of registers are live or killed at the
88		program point where they appear. No spill code is generated when an
89		element of the 'regs' field is spilled, but the register is moved to
90		the 'spilled' (which is present, more for debugging than anything else).
91
92		LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
93		We used to generate:
94
95		DEFFREG f1
96		f1 := f2 + f3
97		trapb
98
99		but now generate:
100
101		f1 := f2 + f3
102		trapb
103		LIVE {regs=[f1,f2,f3], spilled=[]}
104
105		Furthermore, the DEFFREG (hack) required that all floating point instruction
106		use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
107		defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
108		in a cleaner alpha implementation. (Hopefully, intel will not get rid of
109		this architecture).
110
111		COPYXXX is intended to replace the parallel COPY and FCOPY  available on
112		all the architectures. This will result in further simplification of the
113		register allocator that must be aware of them for coalescing purposes, and
114		will also simplify certain aspects of the machine description that provides
115		callbacks related to parallel copies.
116
117		ANNOTATION should be obvious, and now INSTR represents the honest to God
118		machine instruction set!
119
120		The <arch>/instructions/<arch>Instr.sml files define certain utility
121		functions for making porting easier -- essentially converting upper case
122		to lower case. All machine instructions (of type instr) are in upper case,
123		and the lower case form generates an MLRISC instruction. For example on
124		the alpha we have:
125
126		datatype instr =
127		LDA of {r:cell, b:cell, d:operand}
128		| ...
129
130		val lda : {r:cell, b:cell, d:operand} -> instruction
131		...
132
133		where lda is just (INSTR o LDA), etc.
134
135		----------------------------------------------------------------------
136		Name: Matthias Blume
137		Date: 2001/11/22 21:40:00 EST
138		Tag: Release_110_37
139		Description:
140
141		Release 110.37.  This time for real.
142
143		----------------------------------------------------------------------
144		Name: Matthias Blume
145		Date: 2001/11/21 16:35:00 EST
146		Tag: blume-20011121-foot-in-mouth
147		Description:
148
149		Removed the "Release_110_37" tag because of a serious bug.
150		This will be re-tagged once the bug is fixed.
151
152		----------------------------------------------------------------------
153		Name: Matthias Blume
154		Date: 2001/11/21 16:14:00 EST
155		Tag: blume-20011121-forgottenfile
156		Description:
157
158		Forgot to add a file.  (Just a .tex-file -- part of
159		the CM manual source.)
160
161		----------------------------------------------------------------------
162		Name: Matthias Blume
163		Date: 2001/11/21 16:10:00 EST
164		Tag: blume-20011121-invalid_110_37
165		Description:
166
167		Note: I removed the original tag "Release_110_37" from this commit
168		because we found a serious bug in all non-x86 backends.
169		- Matthias
170
171		1. Modifications to the SML/NJ code generator and to the runtime system
172		so that code object name strings are directly inserted into code
173		objects at code generation time.  The only business the runtime system
174		has with this is now to read the name strings on occasions.
175		(The encoding of the name string has also changed somewhat.)
176
177		2. CM now implements a simple "set calculus" for specifying export lists.
178		In particular, it is now possible to refer to the export lists of
179		other libraries/groups/sources and form unions as well as differences.
180		See the latest CM manual for details.
181
182		3. An separate notion of "proxy" libraries has again be eliminated from
183		CM's model.  (Proxy libraries are now simply a special case of using
184		the export list calculus.)
185
186		4. Some of the existing libraries now take advantage of the new set
187		calculus.
188		(Notice that not all libraries have been converted because some
189		of the existing .cm-files are supposed to be backward compatible
190		with 110.0.x.)
191
192		5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
193		"General.exnMessage"!  The former relies on a certain hook to be
194		initialized, and that often does not happen in the stand-alone case.)
195
196		----------------------------------------------------------------------
197		Name: Lal George
198		Date: 2001/11/21  13:56:18 EST
199		Tag: george-2001121-pseudo-ops
200		Description:
201
202		Implemented a complete redesign of MLRISC pseudo-ops. Now there
203		ought to never be any question of incompatabilities with
204		pseudo-op syntax expected by host assemblers.
205
206		For now, only modules supporting GAS syntax are implemented
207		but more should follow, such as MASM, and vendor assembler
208		syntax, e.g. IBM as, Sun as, etc.
209
210		----------------------------------------------------------------------
211		Name: Matthias Blume
212		Date: 2001/11/14 11:52:00 EST
213		Tag: blume-20011114-srcname
214		Description:
215
216		1. Routed the name of the current source file to mlriscgen where it
217		should be directly emitted into the code object.  (This last part
218		is yet to be done.)
219
220		2. Some cleanup of the pgraph code to make it match the proposal that
221		I put out the other day.  (The proposal notwithstanding, things are
222		still in flux here.)
223
224		----------------------------------------------------------------------
225		Name: Lal George
226		Date: 2001/11/14 09:44:04 EST
227		Tag:
228		Description:
229
230		Fix for a backpatching bug reported by Allen.
231
232		Because the boundary between short and long span-dependent
233		instructions is +/- 128, there are an astounding number of
234		span-dependent instructions whose size is over estimated.
235
236		Allen came up with the idea of letting the size of span
237		dependent instructions be non-monotonic, for a maxIter
238		number of times, after which the size must be monotonically
239		increasing.
240
241		This table shows the number of span-dependent instructions
242		whose size was over-estimated as a function of maxIter, for the
243		file Parse/parse/ml.grm.sml:
244
245		maxIter            # of instructions:
246		10                      687
247		20                      438
248		30                      198
249		40                        0
250
251		In compiling the compiler, there is no significant difference in
252		compilation speed between maxIter=10 and maxIter=40. Actually,
253		my measurements showed that maxIter=40 was a tad faster than
254		maxIter=10! Also 96% of the  files in the compiler reach a fix
255		point within 13 iterations, so fixing maxIter at 40, while high,
256		is okay.
257
258		----------------------------------------------------------------------
259		Name: Matthias Blume
260		Date: 2001/10/31 15:25:00 EST
261		Tag: blume-20011031-pgraph
262		Description:
263
264		CKIT:
265		* Changed the "Function" constructor of type Ast.ctype to carry optional
266		argument identifiers.
267		* Changed the return type of TypeUtil.getFunction accordingly.
268		* Type equality ignores the argument names.
269		* TypeUtil.composite tries to preserve argument names but gives up quickly
270		if there is a mismatch.
271
272		installation script:
273		* attempts to use "curl" if available (unless "wget" is available as well)
274
275		CM:
276		* has an experimental implementation of "portable graphs" which I will
277		soon propose as an implementation-independent library format
278		* there are also new libraries $/pgraph.cm and $/pgraph-util.cm
279
280		NLFFI-LIB:
281		* some cleanup (all cosmetic)
282
283		NLFFIGEN:
284		* temporarily disabled the mechanism that suppresses ML output for
285		C definitions whose identifiers start with an underscore character
286		* generate val bindings for enum constants
287		* user can request that only one style (light or heavy) is being used;
288		default is to use both (command-line arguments: -heavy and -light)
289		* fixed bug in handling of function types involving incomplete pointers
290		* generate ML entry points that take record arguments (i.e., using
291		named arguments) for C functions that have a prototype with named
292		arguments
293		(see changes to CKIT)
294
295		----------------------------------------------------------------------
296		Name: Allen Leung
297		Date: 2001/10/27 20:34:00 EDT
298		Tag: leunga-20011027-x86-fast-fp-call
299		Description:
300
301		Fixed the bug described in blume-20010920-slowfp.
302
303		The fix involves
304		1. generating FCOPYs in FSTP in ia32-svid
305		2. marking a CALL with the appropriate annotation
306
307		----------------------------------------------------------------------
308		Name: Matthias Blume
309		Date: 2001/10/16 11:32:00 EDT
310		Tag: blume-20011016-netbsd
311		Description:
312
313		Underscore patch from Chris Richards (fixing problem with compiling
314		runtime system under recent NetBSD).
315
316		----------------------------------------------------------------------
317		Name: Allen Leung
318		Date: 2001/10/12 17:18:32 EDT 2001
319		Tag: leung-20011012-x86-printflowgraph
320		Description:
321
322		X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
323
324		----------------------------------------------------------------------
325		Name: Lal George
326		Date: 2001/10/11 23:51:34 EDT
327		Tag: george-20011011-too-many-instrs
328		Description:
329
330		The representation of a program point never expected to see more
331		than 65536 instructions in a basic block!
332
333		----------------------------------------------------------------------
334		Name: Lal George
335		Date: 2001/10/09 09:41:37 EDT
336		Tag: george-20011008-mlrisc-labels
337		Description:
338
339		Changed the machine description files to support printing of
340		local and global labels in assembly code, based on host assembler
341		conventions.
342
343		----------------------------------------------------------------------
344		Name: Matthias Blume
345		Date: 2001/09/25 15:25:00 EDT
346		Tag: blume-20010925-exninfo
347		Description:
348
349		I provided a non-hook implementation of exnName (at the toplevel) and
350		made the "dummy" implementation of exnMessage (at the toplevel) more
351		useful: if nothing gets "hooked in", then at least you are going to
352		see the exception name and a message indicating why you don't see more.
353
354		[For the time being, programs that need exnMessage and want to use
355		ml-build should either use General.exnMessage (strongly recommended) or
356		refer to structure General at some other point so that CM sees a
357		static dependency.]
358
359		[Similar remarks go for "print" and "use":  If you want to use their
360		functionality in stand-alone programs generated by ml-build, then use
361		TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
362
363		----------------------------------------------------------------------
364		Name: Matthias Blume
365		Date: 2001/09/20 17:28:00 EDT
366		Tag: blume-20010920-slowfp
367		Description:
368
369		Allen says that x86-fast-fp is not safe yet, so I turned it off again...
370
371		----------------------------------------------------------------------
372		Name: Matthias Blume
373		Date: 2001/09/20 17:20:00 EDT
374		Tag: blume-20010920-canonicalpaths
375		Description:
376
377		0. Updated the BOOT file (something that I forgot to do earlier).
378
379		1. Small internal change to CM so that it avoids "/../" in filenames
380		as much as possible (but only where it is safe).
381
382		2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
383		that arguments that contain delimiters are passed through correctly.
384		This change also means that all "special" arguments of the form
385		@SMLxxx... must come first.
386
387		3. Changed install script to put relative anchor names for tool commands
388		into pathconfig.
389
390		----------------------------------------------------------------------
391		Name: Matthias Blume
392		>>>>>>> 1.169
393		Date: 2001/09/18 15:35:00 EDT
394		Tag: blume-20010918-readme11036
395		Description:
396
397		Added README files.
398
399		----------------------------------------------------------------------
400		Name: Matthias Blume
401		Date: 2001/09/18 11:45:00 EDT
402		Tag: Release_110_36 (retag)
403		Description:
404
405		Fixed mistake in config/preloads. Retagged as 110.36.
406
407		----------------------------------------------------------------------
408		Name: Matthias Blume
409		Date: 2001/09/18 09:40:00 EDT
410		Tag: Release_110_36_orig (tag changed)
411		Description:
412
413		New version (110.36).  New bootfiles.
414
415		----------------------------------------------------------------------
416		Name: Matthias Blume
417		Date: 2001/09/14 16:15:00 EDT
418		Tag: blume-20010914-x86fastfp
419		Description:
420
421		John committed some changes that Allen made, in particular a (hopefully)
422		correctly working version of the x86-fp module.
423
424		I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
425		flag to "true".  Everything seems to compile to a fixpoint ok, and
426		"mandelbrot" speeds up by about 15%.
427
428		----------------------------------------------------------------------
429		Name: Matthias Blume
430		Date: 2001/09/13 11:20:00 EDT
431		Tag: blume-20010913-minimal
432		Description:
433
434		1. Stefan Monnier's patch to fix a miscompilation problem that
435		was brought to light by John Reppy's work on Moby.
436
437		2. Implemented a minimal "structure Compiler" that contains just
438		"version" and "architecture".  The minimal version will be
439		available when the full version is not.  This is for backward-
440		compatibility with code that wants to test Compiler.version.
441
442		----------------------------------------------------------------------
443		Name: Matthias Blume
444		Date: 2001/08/28 14:03:00 EDT
445		Tag: blume-20010828-ml-lex
446		Description:
447
448		Fix for bug 1581, received from Neophytos Michael.
449
450		----------------------------------------------------------------------
451		Name: Matthias Blume
452		Date: 2001/08/27 11:20:00 EDT
453		Tag: blume-20010827-readme11035
454		Description:
455
456		Fleshed out the README file for 110.35.
457
458		----------------------------------------------------------------------
459		Name: Matthias Blume
460		Date: 2001/08/24 17:10:00 EDT
461		Tag: Release_110_35
462		Description:
463
464		New version number (110.35).  New bootfiles.
465
466		----------------------------------------------------------------------
467		Name: Lal George
468		Date: 2001/08/24 13:47:18 EDT 2001
469		Tag: george-20010824-MLRISC-graphs
470		Description:
471
472		removed clusters from MLRISC completely and replaced with graphs.
473
474		----------------------------------------------------------------------
475		Name: Matthias Blume
476		Date: 2001/08/23 17:50:00 EDT
477		Tag: blume-20010823-toplevel
478		Description:
479
480		- some reorganization of the code that implements various kinds of
481		environments in the compiler (static, dynamic, symbolic, combined)
482		- re-implemented the EnvRef module so that evalStream works properly
483		(if the stream contains references to "use", "CM.make", etc.)
484		- cleaned up evalloop.sml and interact.sml (but they need more cleaning)
485
486		----------------------------------------------------------------------
487		Name: Matthias Blume
488		Date: 2001/08/20 15:50 EDT
489		Tag: blume20010820-slipup
490		Description:
491
492		I forgot to commit a few files.  Here they are...
493
494		----------------------------------------------------------------------
495		Name: Matthias Blume
496		Date: 2001/08/20 15:35:00 EDT
497		Tag: blume-20010820-debugprof
498		Description:
499
500		!!!! NEW BOOTFILES !!!!
501
502		This is another round of reorganizing the compiler sources.  This
503		time the main goal was to factor out all the "instrumentation"
504		passes (for profiling and backtracing) into their own library.
505		The difficulty was to do it in such a way that it does not depend
506		on elaborate.cm but only on elabdata.cm.
507
508		Therefore there have been further changes to both elaborate.cm and
509		elabdata.cm -- more "generic" things have been moved from the former
510		to the latter.  As a result, I was forced to split the assignment
511		of numbers indicating "primtyc"s into two portions: SML-generic and
512		SML/NJ-specific.  Since it would have been awkward to maintain,
513		I bit the bullet and actually _changed_ the mapping between these
514		numbers and primtycs.  The bottom line of this is that you need
515		a new set of bin- and bootfiles.
516
517		I have built new bootfiles for all architectures, so doing a fresh
518		checkout and config/install.sh should be all you need.
519
520		The newly created library's name is
521
522		$smlnj/viscomp/debugprof.cm
523
524		and its sources live under
525
526		src/compiler/DebugProf
527
528		----------------------------------------------------------------------
529		Name: Matthias Blume
530		Date: 2001/08/15 17:15:00 EDT
531		Tag: blume-20010815-compreorg
532		Description:
533
534		This is a first cut at reorganizing the CM libraries that make up the
535		core of the compiler.  The idea is to separate out pieces that could
536		be used independently by tools, e.g., the parser, the typechecker, etc.
537
538		The current status is a step in this direction, but it is not quite
539		satisfactory yet.  Expect more changes in the future.
540
541		Here is the current (new) organization...
542
543		What used to be $smlnj/viscomp/core.cm is now divided into
544		six CM libraries:
545
546		$smlnj/viscomp/basics.cm
547		/parser.cm
548		/elabdata.cm
549		/elaborate.cm
550		/execute.cm
551		/core.cm
552
553		The CM files for these libraries live under src/system/smlnj/viscomp.
554		All these libraries are proxy libraries that contain precisely
555		one CM library component.  Here are the locations of the components
556		(all within the src/compiler tree):
557
558		Basics/basics.cm
559		Parse/parser.cm
560		ElabData/elabdata.cm
561		Elaborator/elaborate.cm
562		Execution/execute.cm
563		core.cm
564
565		[This organization is the same that has been used already
566		for a while for the architecture-specific parts of the visible
567		compiler and for the old version of core.cm.]
568
569		As you will notice, many source files have been moved from their
570		respective original locations to a new home in one of the above
571		subtrees.
572
573		The division of labor between the new libraries is the following:
574
575		basics.cm:
576		- Simple, basic definitions that pertain to many (or all) of
577		the other libraries.
578		parser.cm:
579		- The SML parser, producing output of type Ast.dec.
580		- The type family for Ast is also defined and exported here.
581		elabdata.cm:
582		- The datatypes that describe input and output of the elaborator.
583		This includes types, absyn, and static environments.
584		elaborator.cm:
585		- The SML/NJ type checker and elaborator.
586		This maps an Ast.dec (with a given static environment) to
587		an Absyn.dec (with a new static environment).
588		- This libraries implements certain modules that used to be
589		structures as functors (to remove dependencies on FLINT).
590		execute.cm:
591		- Everything having to do with executing binary code objects.
592		- Dynamic environments.
593		core.cm:
594		- SML/NJ-specific instantiations of the elaborator and MLRISC.
595		- Top-level modules.
596		- FLINT (this should eventually become its own library)
597
598		Notes:
599
600		I am not 100% happy with the way I separated the elaborator (and its
601		data structures) from FLINT.  Two instances of the same problem:
602
603		1. Data structures contain certain fields that carry FLINT-specific
604		information.  I hacked around this using exn and the property list
605		module from smlnj-lib.  But the fact that there are middle-end
606		specific fields around at all is a bit annoying.
607
608		2. The elaborator calculates certain FLINT-related information.  I tried
609		to make this as abstract as I could using functorization, but, again,
610		the fact that the elaborator has to perform calculations on behalf
611		of the middle-end at all is not nice.
612
613		3. Having to used exn and property lists is unfortunate because it
614		weakens type checking.  The other alternative (parameterizing
615		nearly *everything*) is not appealing, though.
616
617		I removed the "rebinding =" warning hack because due to the new organization
618		it was awkward to maintain it.  As a result, the compiler now issues some of
619		these warnings when compiling init.cmi during bootstrap compilation. On
620		the plus side, you also get a warning when you do, for example:
621		val op = = Int32.+
622		which was not the case up to now.
623
624		I placed "assign" and "deref" into the _Core structure so that the
625		code that deals with the "lazy" keyword can find them there.  This
626		removes the need for having access to the primitive environment
627		during elaboration.
628
629		----------------------------------------------------------------------
630		Name: Matthias Blume
631		Date: 2001/08/13
632		Tag: blume-20010813-closures
633		Description:
634
635		This fix was sent to us by Zhong Shao.  It is supposed to improve the
636		performance of certain loops by avoiding needless closure allocation.
637
638		----------------------------------------------------------------------
639		Name: Lal George
640		Date: 2001/07/31 10:03:23 EDT 2001
641		Tag: george-20010731-x86-fmalloc
642		Description: Fixed bug in x86 calls
643
644		There was a bug where call instructions would mysteriously
645		vanish. The call instruction had to be one that returned
646		a floating point value.
647
648		----------------------------------------------------------------------
649		Name: Lal George
650		Date: 2001/07/19 16:36:29 EDT 2001
651		Tag: george-20010719-simple-cells
652		Description:
653
654		I have dramatically simplified the interface for CELLS in MLRISC.
655
656		In summary, the cells interface is broken up into three parts:
657
658		1. CellsBasis : CELLS_BASIS
659
660		CellsBasis is a top level structure and common for all
661		architectures.  it contains the definitions of basic datatypes
662		and utility  functions over these types.
663
664		2. functor Cells() : CELLS
665
666		Cells generates an interface for CELLS that incorporates the
667		specific resources on the target architecture, such as the
668		presence of special register classes, their number and size,
669		and various useful substructures.
670
671		3. <ARCH>CELLS
672
673		e.g. SparcCells: SPARCCELLS
674
675		<ARCH>CELLS usually contains additional bindings for special
676		registers  on the architecture, such as:
677
678		val r0 : cell           (* register zero *)
679		val y : cell            (* Y register *)
680		val psr : cell          (* processor status register *)
681		...
682
683		The structure returned by applying the Cells functor is opened
684		in this interface.
685
686		The main implication of all this is that the datatypes for cells is
687		split between CellsBasis and CELLS -- a fairly simple change for user
688		code.
689
690		In the old scheme the CELLS interface had a definitional binding of
691		the form:
692
693		signature CELLS = sig
694
695		structure CellsBasis = CellsBasis
696
697		...
698
699		end
700
701		With all the sharing constraints that goes on in MLRISC, this old
702		design  quickly leads to errors such as:
703
704		"structure definition spec inside of sharing ... "
705
706
707		and appears to require an unacceptable amount of sharing and where
708		constraint hackery.
709
710		I think this error message (the interaction of definitional specs and
711		sharing) requires more explanation on our web page.
712
713		----------------------------------------------------------------------
714		Name: Matthias Blume
715		Date: 2001/07/19 15:00:00 EDT
716		Tag: blume-20010719-libreorg
717		Description:
718
719		This update puts together a fairly extensive but straightforward change
720		to the way the libraries that implement the interactive system are
721		organized:
722
723		The biggest change is the elimination of structure Compiler.  As a
724		replacement for this structure, there is now a CM library
725		(known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
726		that exports all the substructures of the original structure Compiler
727		directly.  So instead of saying Compiler.Foo.bar one now simply
728		says Foo.bar.  (The CM libraries actually export a collection of
729		structures that is richer than the collection of substructures of
730		structure Compiler.)
731
732		To make the transition smooth, there is a separate library called
733		$smlnj/compiler/compiler.cm which puts together and exports the
734		original structure Compiler (or at least something very close to it).
735
736		There are five members of the original structure Compiler
737		that are not exported directly but which instead became members
738		of a new structure Backend (described by signature BACKEND).  These are:
739		structure Profile (: PROFILE), structure Compile (: COMPILE), structure
740		Interact (: INTERACT), structure Machine (: MACHINE), and val
741		architecture (: string).
742
743		Structure Compiler.Version has become structure CompilerVersion.
744
745		Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
746		by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
747		or x86, respectively.
748		Each of these exports the same frontend structures that
749		$smlnj/compiler.cm exports.  But they do not have a structure Backend
750		and instead export some structure <Arch>Backend where <Arch> is Alpha32,
751		Hppa, PPC, Sparc, or X86, respectively.
752
753		Library $smlnj/compiler/all.cm exports the union of the exports of
754		$smlnj/compiler/<arch>.cm
755
756		There are no structures <Arch>Compiler anymore, use
757		$smlnj/compiler/<arch>.cm instead.
758
759		Library host-compiler-0.cm is gone.  Instead, the internal library
760		that instantiates CM is now called cm0.cm.  Selection of the host
761		compiler (backend) is no longer done here but. (Responsibility for it
762		now lies with $smlnj/compiler/current.cm.  This seems to be more
763		logical.)
764
765		Many individual files have been moved or renamed.  Some files have
766		been split into multiple files, and some "dead" files have been deleted.
767
768		Aside from these changes to library organization, there are also changes
769		to the way the code itself is organized:
770
771		Structure Binfile has been re-implemented in such a way that it no
772		longer needs any knowledge of the compiler.  It exclusively deals
773		with the details of binfile layout.  It no longer invokes the
774		compiler (for the purpose of creating new prospective binfile
775		content), and it no longer has any knowledge of how to interpret
776		pickles.
777
778		Structure Compile (: COMPILE) has been stripped down to the bare
779		essentials of compilation.  It no longer deals with linking/execution.
780		The interface has been cleaned up considerably.
781
782		Utility routines for dealing with linking and execution have been
783		moved into their own substructures.
784
785		(The ultimate goal of these changes is to provide a light-weight
786		binfile loader/linker (at least for, e.g., stable libraries) that
787		does not require CM or the compiler to be present.)
788
789		CM documentation has been updated to reflect the changes to library
790		organization.
791
792		----------------------------------------------------------------------
793		Name: Matthias Blume
794		Date: 2001/07/10 17:30:00 EDT
795		Tag: Release_110_34
796		Description:
797
798		Minor tweak to 110.34 (re-tagged):
799
800		- README.html file added to CVS repository
801		- runtime compiles properly under FreeBSD 3.X and 4.X
802
803		----------------------------------------------------------------------
804		Name: Matthias Blume
805		Date: 2001/07/10 17:30:00 EDT
806		Tag: Release_110_34
807		Description:
808
809		New version number (110.34). New bootfiles.
810
811		----------------------------------------------------------------------
812		Name: Matthias Blume
813		Date: 2001/07/09 16:00:00 EDT
814		Tag: blume-20010709-more-varargs
815		Description:
816
817		I changed the handling of varargs in ml-nlffigen again:
818		The ellipsis ... will now simply be ignored (with an accompanying warning).
819
820		The immediate effect is that you can actually call a varargs function
821		from ML -- but you can't actually supply any arguments beyond the ones
822		specified explicitly.  (For example, you can call printf with its format
823		string, but you cannot pass additional arguments.)
824
825		This behavior is only marginally more useful than the one before, but
826		it has the advantage that a function or, more importantly, a function
827		type never gets dropped on the floor, thus avoiding follow-up problems with
828		other types that refer to the offending one.
829
830		----------------------------------------------------------------------
831		Name: Matthias Blume
832		Date: 2001/07/09 11:25:00 EDT
833		Tag: blume-20010709-varargs
834		Description:
835
836		1. ckit-lib.cm now exports structure Error
837		2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
838		with a warning accompanied by a source location.  Moreover, it
839		merely skips the offending function or type and proceeds with the
840		rest of its work.u  As a result, one can safely feed C code containing
841		"..." to ml-nlffigen.
842		3. There are some internal improvements to CM, providing slightly
843		more general string substitutions in the tools subsystem.
844
845		----------------------------------------------------------------------
846		Name: Matthias Blume
847		Date: 2001/06/27 15:10:00 EDT
848		Tag: blume-20010627-concur
849		Description:
850
851		Fixed a small bug in CM's handling of parallel compilation.
852		(You could observe the bug by Control-C-interrupting an ordinary
853		CMB.make or CM.stabilize and then attaching some compile servers.
854		The result was that all of a sudden the previously interrupted
855		compilation would continue on its own.  This was because of
856		an over-optimization: CM did not bother to clean out certain queues
857		when no servers were attached "anyway", resulting in the contents
858		of these queues to grab control when new servers did get attached.)
859
860		There is also another minor update to the CM manual.
861
862		----------------------------------------------------------------------
863		Name: Matthias Blume
864		Date: 2001/06/26 16:15:00 EDT
865		Tag: blume-20010626-cmdoc
866		Description:
867
868		Minor typo fixed in CM manual (syntax diagram for libraries).
869
870		----------------------------------------------------------------------
871		Name: Matthias Blume
872		Date: 2001/06/25 22:55:00 EDT
873		Tag: blume-20010625-x86pc
874		Description:
875
876		Fixed a nasty bug in the X86 assembly code that caused signal
877		handlers to fail (crash) randomly.
878
879		----------------------------------------------------------------------
880		Name: Matthias Blume
881		Date: 2001/06/25 12:05:00 EDT
882		Tag: blume-20010625-nlffigen
883		Description:
884
885		This update fixes a number of minor bugs in ml-nlffigen as reported by
886		Nick Carter <nbc@andrew.cmu.edu>.
887
888		1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
889		2. Default names for generated files are now derived from the name of
890		the C file *without its directory*.  In particular, this causes generated
891		files to be placed locally even if the C file is in some system directory.
892		3. Default names for generated signatures and structures are also derived
893		from the C file name without its directory.  This avoids silly things
894		like "structure GL/GL".
895		(Other silly names are still possible because ml-nlffigen does not do
896		a thorough check of whether generated names are legal ML identifiers.
897		When in doubt, use command line arguments to force particular names.)
898
899		----------------------------------------------------------------------
900		Name: Matthias Blume
901		Date: 2001/06/21 12:25:00 EDT
902		Tag: blume-20010621-eXene
903		Description:
904
905		eXene now compiles and (sort of) works again.
906
907		The library name (for version > 110.33) is $/eXene.cm.
908
909		I also added an new example in src/eXene/examples/nbody.  See the
910		README file there for details.
911
912		----------------------------------------------------------------------
913		Name: Matthias Blume
914		Date: 2001/06/20 16:40:00 EDT
915		Tag: blume-20010620-cml
916		Description:
917
918		CML now compiles and works again.
919
920		Libraries (for version > 110.33):
921
922		$cml/cml.cm            Main CML library.
923		$cml/basis.cm          CML's version of $/basis.cm.
924		$cml/cml-internal.cm   Internal helper library.
925		$cml/core-cml.cm       Internal helper library.
926		$cml-lib/trace-cml.cm  Tracing facility.
927		$cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
928
929		The installer (config/install.sh) has been taught how to properly
930		install this stuff.
931
932		----------------------------------------------------------------------
933		Name: Matthias Blume
934		Date: 2001/06/19 17:55:00 EDT
935		Tag: blume-20010619-instantiate
936		Description:
937
938		This un-breaks the fix for bug 1432.
939		(The bug was originally fixed in 110.9 but I broke it again some
940		time after that.)
941
942	----------------------------------------------------------------------	----------------------------------------------------------------------
943	Name: Matthias Blume	Name: Matthias Blume
#	Line 2439	Line 3366
3366	elaborator). There were a lot of changes during my "linkpath" trials	elaborator). There were a lot of changes during my "linkpath" trials
3367	that could have been reverted to their original state but weren't.	that could have been reverted to their original state but weren't.
3368	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
3369	than what was strictly necessary...  (I _did_ resist the tempation	than what was strictly necessary...  (I _did_ resist the temptation
3370	of doing any "global reformatting" to avoid an untimely death at	of doing any "global reformatting" to avoid an untimely death at
3371	Dave's hands. :)	Dave's hands. :)
3372

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.842   changed lines   Added in v.1015

---

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