Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 829, Mon May 14 15:30:56 2001 UTC	revision 959, Fri Oct 12 21:21:41 2001 UTC
#	Line 13	Line 13
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: Allen Leung
17		Date: 2001/10/12 17:18:32 EDT 2001
18		Tag: leung-20011012-x86-printflowgraph
19		Description:
20
21		X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
22
23		----------------------------------------------------------------------
24		Name: Lal George
25		Date: 2001/10/11 23:51:34 EDT
26		Tag: george-20011011-too-many-instrs
27		Description:
28
29		The representation of a program point never expected to see more
30		than 65536 instructions in a basic block!
31
32		----------------------------------------------------------------------
33		Name: Lal George
34		Date: 2001/10/09 09:41:37 EDT
35		Tag: george-20011008-mlrisc-labels
36		Description:
37
38		Changed the machine description files to support printing of
39		local and global labels in assembly code, based on host assembler
40		conventions.
41
42		----------------------------------------------------------------------
43		Name: Matthias Blume
44		Date: 2001/09/25 15:25:00 EDT
45		Tag: blume-20010925-exninfo
46		Description:
47
48		I provided a non-hook implementation of exnName (at the toplevel) and
49		made the "dummy" implementation of exnMessage (at the toplevel) more
50		useful: if nothing gets "hooked in", then at least you are going to
51		see the exception name and a message indicating why you don't see more.
52
53		[For the time being, programs that need exnMessage and want to use
54		ml-build should either use General.exnMessage (strongly recommended) or
55		refer to structure General at some other point so that CM sees a
56		static dependency.]
57
58		[Similar remarks go for "print" and "use":  If you want to use their
59		functionality in stand-alone programs generated by ml-build, then use
60		TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
61
62		----------------------------------------------------------------------
63		Name: Matthias Blume
64		Date: 2001/09/20 17:28:00 EDT
65		Tag: blume-20010920-slowfp
66		Description:
67
68		Allen says that x86-fast-fp is not safe yet, so I turned it off again...
69
70		----------------------------------------------------------------------
71		Name: Matthias Blume
72		Date: 2001/09/20 17:20:00 EDT
73		Tag: blume-20010920-canonicalpaths
74		Description:
75
76		0. Updated the BOOT file (something that I forgot to do earlier).
77
78		1. Small internal change to CM so that it avoids "/../" in filenames
79		as much as possible (but only where it is safe).
80
81		2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
82		that arguments that contain delimiters are passed through correctly.
83		This change also means that all "special" arguments of the form
84		@SMLxxx... must come first.
85
86		3. Changed install script to put relative anchor names for tool commands
87		into pathconfig.
88
89		----------------------------------------------------------------------
90		Name: Matthias Blume
91		>>>>>>> 1.169
92		Date: 2001/09/18 15:35:00 EDT
93		Tag: blume-20010918-readme11036
94		Description:
95
96		Added README files.
97
98		----------------------------------------------------------------------
99		Name: Matthias Blume
100		Date: 2001/09/18 11:45:00 EDT
101		Tag: Release_110_36 (retag)
102		Description:
103
104		Fixed mistake in config/preloads. Retagged as 110.36.
105
106		----------------------------------------------------------------------
107		Name: Matthias Blume
108		Date: 2001/09/18 09:40:00 EDT
109		Tag: Release_110_36_orig (tag changed)
110		Description:
111
112		New version (110.36).  New bootfiles.
113
114		----------------------------------------------------------------------
115		Name: Matthias Blume
116		Date: 2001/09/14 16:15:00 EDT
117		Tag: blume-20010914-x86fastfp
118		Description:
119
120		John committed some changes that Allen made, in particular a (hopefully)
121		correctly working version of the x86-fp module.
122
123		I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
124		flag to "true".  Everything seems to compile to a fixpoint ok, and
125		"mandelbrot" speeds up by about 15%.
126
127		----------------------------------------------------------------------
128		Name: Matthias Blume
129		Date: 2001/09/13 11:20:00 EDT
130		Tag: blume-20010913-minimal
131		Description:
132
133		1. Stefan Monnier's patch to fix a miscompilation problem that
134		was brought to light by John Reppy's work on Moby.
135
136		2. Implemented a minimal "structure Compiler" that contains just
137		"version" and "architecture".  The minimal version will be
138		available when the full version is not.  This is for backward-
139		compatibility with code that wants to test Compiler.version.
140
141		----------------------------------------------------------------------
142		Name: Matthias Blume
143		Date: 2001/08/28 14:03:00 EDT
144		Tag: blume-20010828-ml-lex
145		Description:
146
147		Fix for bug 1581, received from Neophytos Michael.
148
149		----------------------------------------------------------------------
150		Name: Matthias Blume
151		Date: 2001/08/27 11:20:00 EDT
152		Tag: blume-20010827-readme11035
153		Description:
154
155		Fleshed out the README file for 110.35.
156
157		----------------------------------------------------------------------
158		Name: Matthias Blume
159		Date: 2001/08/24 17:10:00 EDT
160		Tag: Release_110_35
161		Description:
162
163		New version number (110.35).  New bootfiles.
164
165		----------------------------------------------------------------------
166		Name: Lal George
167		Date: 2001/08/24 13:47:18 EDT 2001
168		Tag: george-20010824-MLRISC-graphs
169		Description:
170
171		removed clusters from MLRISC completely and replaced with graphs.
172
173		----------------------------------------------------------------------
174		Name: Matthias Blume
175		Date: 2001/08/23 17:50:00 EDT
176		Tag: blume-20010823-toplevel
177		Description:
178
179		- some reorganization of the code that implements various kinds of
180		environments in the compiler (static, dynamic, symbolic, combined)
181		- re-implemented the EnvRef module so that evalStream works properly
182		(if the stream contains references to "use", "CM.make", etc.)
183		- cleaned up evalloop.sml and interact.sml (but they need more cleaning)
184
185		----------------------------------------------------------------------
186		Name: Matthias Blume
187		Date: 2001/08/20 15:50 EDT
188		Tag: blume20010820-slipup
189		Description:
190
191		I forgot to commit a few files.  Here they are...
192
193		----------------------------------------------------------------------
194		Name: Matthias Blume
195		Date: 2001/08/20 15:35:00 EDT
196		Tag: blume-20010820-debugprof
197		Description:
198
199		!!!! NEW BOOTFILES !!!!
200
201		This is another round of reorganizing the compiler sources.  This
202		time the main goal was to factor out all the "instrumentation"
203		passes (for profiling and backtracing) into their own library.
204		The difficulty was to do it in such a way that it does not depend
205		on elaborate.cm but only on elabdata.cm.
206
207		Therefore there have been further changes to both elaborate.cm and
208		elabdata.cm -- more "generic" things have been moved from the former
209		to the latter.  As a result, I was forced to split the assignment
210		of numbers indicating "primtyc"s into two portions: SML-generic and
211		SML/NJ-specific.  Since it would have been awkward to maintain,
212		I bit the bullet and actually _changed_ the mapping between these
213		numbers and primtycs.  The bottom line of this is that you need
214		a new set of bin- and bootfiles.
215
216		I have built new bootfiles for all architectures, so doing a fresh
217		checkout and config/install.sh should be all you need.
218
219		The newly created library's name is
220
221		$smlnj/viscomp/debugprof.cm
222
223		and its sources live under
224
225		src/compiler/DebugProf
226
227		----------------------------------------------------------------------
228		Name: Matthias Blume
229		Date: 2001/08/15 17:15:00 EDT
230		Tag: blume-20010815-compreorg
231		Description:
232
233		This is a first cut at reorganizing the CM libraries that make up the
234		core of the compiler.  The idea is to separate out pieces that could
235		be used independently by tools, e.g., the parser, the typechecker, etc.
236
237		The current status is a step in this direction, but it is not quite
238		satisfactory yet.  Expect more changes in the future.
239
240		Here is the current (new) organization...
241
242		What used to be $smlnj/viscomp/core.cm is now divided into
243		six CM libraries:
244
245		$smlnj/viscomp/basics.cm
246		/parser.cm
247		/elabdata.cm
248		/elaborate.cm
249		/execute.cm
250		/core.cm
251
252		The CM files for these libraries live under src/system/smlnj/viscomp.
253		All these libraries are proxy libraries that contain precisely
254		one CM library component.  Here are the locations of the components
255		(all within the src/compiler tree):
256
257		Basics/basics.cm
258		Parse/parser.cm
259		ElabData/elabdata.cm
260		Elaborator/elaborate.cm
261		Execution/execute.cm
262		core.cm
263
264		[This organization is the same that has been used already
265		for a while for the architecture-specific parts of the visible
266		compiler and for the old version of core.cm.]
267
268		As you will notice, many source files have been moved from their
269		respective original locations to a new home in one of the above
270		subtrees.
271
272		The division of labor between the new libraries is the following:
273
274		basics.cm:
275		- Simple, basic definitions that pertain to many (or all) of
276		the other libraries.
277		parser.cm:
278		- The SML parser, producing output of type Ast.dec.
279		- The type family for Ast is also defined and exported here.
280		elabdata.cm:
281		- The datatypes that describe input and output of the elaborator.
282		This includes types, absyn, and static environments.
283		elaborator.cm:
284		- The SML/NJ type checker and elaborator.
285		This maps an Ast.dec (with a given static environment) to
286		an Absyn.dec (with a new static environment).
287		- This libraries implements certain modules that used to be
288		structures as functors (to remove dependencies on FLINT).
289		execute.cm:
290		- Everything having to do with executing binary code objects.
291		- Dynamic environments.
292		core.cm:
293		- SML/NJ-specific instantiations of the elaborator and MLRISC.
294		- Top-level modules.
295		- FLINT (this should eventually become its own library)
296
297		Notes:
298
299		I am not 100% happy with the way I separated the elaborator (and its
300		data structures) from FLINT.  Two instances of the same problem:
301
302		1. Data structures contain certain fields that carry FLINT-specific
303		information.  I hacked around this using exn and the property list
304		module from smlnj-lib.  But the fact that there are middle-end
305		specific fields around at all is a bit annoying.
306
307		2. The elaborator calculates certain FLINT-related information.  I tried
308		to make this as abstract as I could using functorization, but, again,
309		the fact that the elaborator has to perform calculations on behalf
310		of the middle-end at all is not nice.
311
312		3. Having to used exn and property lists is unfortunate because it
313		weakens type checking.  The other alternative (parameterizing
314		nearly *everything*) is not appealing, though.
315
316		I removed the "rebinding =" warning hack because due to the new organization
317		it was awkward to maintain it.  As a result, the compiler now issues some of
318		these warnings when compiling init.cmi during bootstrap compilation. On
319		the plus side, you also get a warning when you do, for example:
320		val op = = Int32.+
321		which was not the case up to now.
322
323		I placed "assign" and "deref" into the _Core structure so that the
324		code that deals with the "lazy" keyword can find them there.  This
325		removes the need for having access to the primitive environment
326		during elaboration.
327
328		----------------------------------------------------------------------
329		Name: Matthias Blume
330		Date: 2001/08/13
331		Tag: blume-20010813-closures
332		Description:
333
334		This fix was sent to us by Zhong Shao.  It is supposed to improve the
335		performance of certain loops by avoiding needless closure allocation.
336
337		----------------------------------------------------------------------
338		Name: Lal George
339		Date: 2001/07/31 10:03:23 EDT 2001
340		Tag: george-20010731-x86-fmalloc
341		Description: Fixed bug in x86 calls
342
343		There was a bug where call instructions would mysteriously
344		vanish. The call instruction had to be one that returned
345		a floating point value.
346
347		----------------------------------------------------------------------
348		Name: Lal George
349		Date: 2001/07/19 16:36:29 EDT 2001
350		Tag: george-20010719-simple-cells
351		Description:
352
353		I have dramatically simplified the interface for CELLS in MLRISC.
354
355		In summary, the cells interface is broken up into three parts:
356
357		1. CellsBasis : CELLS_BASIS
358
359		CellsBasis is a top level structure and common for all
360		architectures.  it contains the definitions of basic datatypes
361		and utility  functions over these types.
362
363		2. functor Cells() : CELLS
364
365		Cells generates an interface for CELLS that incorporates the
366		specific resources on the target architecture, such as the
367		presence of special register classes, their number and size,
368		and various useful substructures.
369
370		3. <ARCH>CELLS
371
372		e.g. SparcCells: SPARCCELLS
373
374		<ARCH>CELLS usually contains additional bindings for special
375		registers  on the architecture, such as:
376
377		val r0 : cell           (* register zero *)
378		val y : cell            (* Y register *)
379		val psr : cell          (* processor status register *)
380		...
381
382		The structure returned by applying the Cells functor is opened
383		in this interface.
384
385		The main implication of all this is that the datatypes for cells is
386		split between CellsBasis and CELLS -- a fairly simple change for user
387		code.
388
389		In the old scheme the CELLS interface had a definitional binding of
390		the form:
391
392		signature CELLS = sig
393
394		structure CellsBasis = CellsBasis
395
396		...
397
398		end
399
400		With all the sharing constraints that goes on in MLRISC, this old
401		design  quickly leads to errors such as:
402
403		"structure definition spec inside of sharing ... "
404
405
406		and appears to require an unacceptable amount of sharing and where
407		constraint hackery.
408
409		I think this error message (the interaction of definitional specs and
410		sharing) requires more explanation on our web page.
411
412		----------------------------------------------------------------------
413		Name: Matthias Blume
414		Date: 2001/07/19 15:00:00 EDT
415		Tag: blume-20010719-libreorg
416		Description:
417
418		This update puts together a fairly extensive but straightforward change
419		to the way the libraries that implement the interactive system are
420		organized:
421
422		The biggest change is the elimination of structure Compiler.  As a
423		replacement for this structure, there is now a CM library
424		(known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
425		that exports all the substructures of the original structure Compiler
426		directly.  So instead of saying Compiler.Foo.bar one now simply
427		says Foo.bar.  (The CM libraries actually export a collection of
428		structures that is richer than the collection of substructures of
429		structure Compiler.)
430
431		To make the transition smooth, there is a separate library called
432		$smlnj/compiler/compiler.cm which puts together and exports the
433		original structure Compiler (or at least something very close to it).
434
435		There are five members of the original structure Compiler
436		that are not exported directly but which instead became members
437		of a new structure Backend (described by signature BACKEND).  These are:
438		structure Profile (: PROFILE), structure Compile (: COMPILE), structure
439		Interact (: INTERACT), structure Machine (: MACHINE), and val
440		architecture (: string).
441
442		Structure Compiler.Version has become structure CompilerVersion.
443
444		Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
445		by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
446		or x86, respectively.
447		Each of these exports the same frontend structures that
448		$smlnj/compiler.cm exports.  But they do not have a structure Backend
449		and instead export some structure <Arch>Backend where <Arch> is Alpha32,
450		Hppa, PPC, Sparc, or X86, respectively.
451
452		Library $smlnj/compiler/all.cm exports the union of the exports of
453		$smlnj/compiler/<arch>.cm
454
455		There are no structures <Arch>Compiler anymore, use
456		$smlnj/compiler/<arch>.cm instead.
457
458		Library host-compiler-0.cm is gone.  Instead, the internal library
459		that instantiates CM is now called cm0.cm.  Selection of the host
460		compiler (backend) is no longer done here but. (Responsibility for it
461		now lies with $smlnj/compiler/current.cm.  This seems to be more
462		logical.)
463
464		Many individual files have been moved or renamed.  Some files have
465		been split into multiple files, and some "dead" files have been deleted.
466
467		Aside from these changes to library organization, there are also changes
468		to the way the code itself is organized:
469
470		Structure Binfile has been re-implemented in such a way that it no
471		longer needs any knowledge of the compiler.  It exclusively deals
472		with the details of binfile layout.  It no longer invokes the
473		compiler (for the purpose of creating new prospective binfile
474		content), and it no longer has any knowledge of how to interpret
475		pickles.
476
477		Structure Compile (: COMPILE) has been stripped down to the bare
478		essentials of compilation.  It no longer deals with linking/execution.
479		The interface has been cleaned up considerably.
480
481		Utility routines for dealing with linking and execution have been
482		moved into their own substructures.
483
484		(The ultimate goal of these changes is to provide a light-weight
485		binfile loader/linker (at least for, e.g., stable libraries) that
486		does not require CM or the compiler to be present.)
487
488		CM documentation has been updated to reflect the changes to library
489		organization.
490
491		----------------------------------------------------------------------
492		Name: Matthias Blume
493		Date: 2001/07/10 17:30:00 EDT
494		Tag: Release_110_34
495		Description:
496
497		Minor tweak to 110.34 (re-tagged):
498
499		- README.html file added to CVS repository
500		- runtime compiles properly under FreeBSD 3.X and 4.X
501
502		----------------------------------------------------------------------
503		Name: Matthias Blume
504		Date: 2001/07/10 17:30:00 EDT
505		Tag: Release_110_34
506		Description:
507
508		New version number (110.34). New bootfiles.
509
510		----------------------------------------------------------------------
511		Name: Matthias Blume
512		Date: 2001/07/09 16:00:00 EDT
513		Tag: blume-20010709-more-varargs
514		Description:
515
516		I changed the handling of varargs in ml-nlffigen again:
517		The ellipsis ... will now simply be ignored (with an accompanying warning).
518
519		The immediate effect is that you can actually call a varargs function
520		from ML -- but you can't actually supply any arguments beyond the ones
521		specified explicitly.  (For example, you can call printf with its format
522		string, but you cannot pass additional arguments.)
523
524		This behavior is only marginally more useful than the one before, but
525		it has the advantage that a function or, more importantly, a function
526		type never gets dropped on the floor, thus avoiding follow-up problems with
527		other types that refer to the offending one.
528
529		----------------------------------------------------------------------
530		Name: Matthias Blume
531		Date: 2001/07/09 11:25:00 EDT
532		Tag: blume-20010709-varargs
533		Description:
534
535		1. ckit-lib.cm now exports structure Error
536		2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
537		with a warning accompanied by a source location.  Moreover, it
538		merely skips the offending function or type and proceeds with the
539		rest of its work.u  As a result, one can safely feed C code containing
540		"..." to ml-nlffigen.
541		3. There are some internal improvements to CM, providing slightly
542		more general string substitutions in the tools subsystem.
543
544		----------------------------------------------------------------------
545		Name: Matthias Blume
546		Date: 2001/06/27 15:10:00 EDT
547		Tag: blume-20010627-concur
548		Description:
549
550		Fixed a small bug in CM's handling of parallel compilation.
551		(You could observe the bug by Control-C-interrupting an ordinary
552		CMB.make or CM.stabilize and then attaching some compile servers.
553		The result was that all of a sudden the previously interrupted
554		compilation would continue on its own.  This was because of
555		an over-optimization: CM did not bother to clean out certain queues
556		when no servers were attached "anyway", resulting in the contents
557		of these queues to grab control when new servers did get attached.)
558
559		There is also another minor update to the CM manual.
560
561		----------------------------------------------------------------------
562		Name: Matthias Blume
563		Date: 2001/06/26 16:15:00 EDT
564		Tag: blume-20010626-cmdoc
565		Description:
566
567		Minor typo fixed in CM manual (syntax diagram for libraries).
568
569		----------------------------------------------------------------------
570		Name: Matthias Blume
571		Date: 2001/06/25 22:55:00 EDT
572		Tag: blume-20010625-x86pc
573		Description:
574
575		Fixed a nasty bug in the X86 assembly code that caused signal
576		handlers to fail (crash) randomly.
577
578		----------------------------------------------------------------------
579		Name: Matthias Blume
580		Date: 2001/06/25 12:05:00 EDT
581		Tag: blume-20010625-nlffigen
582		Description:
583
584		This update fixes a number of minor bugs in ml-nlffigen as reported by
585		Nick Carter <nbc@andrew.cmu.edu>.
586
587		1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
588		2. Default names for generated files are now derived from the name of
589		the C file *without its directory*.  In particular, this causes generated
590		files to be placed locally even if the C file is in some system directory.
591		3. Default names for generated signatures and structures are also derived
592		from the C file name without its directory.  This avoids silly things
593		like "structure GL/GL".
594		(Other silly names are still possible because ml-nlffigen does not do
595		a thorough check of whether generated names are legal ML identifiers.
596		When in doubt, use command line arguments to force particular names.)
597
598		----------------------------------------------------------------------
599		Name: Matthias Blume
600		Date: 2001/06/21 12:25:00 EDT
601		Tag: blume-20010621-eXene
602		Description:
603
604		eXene now compiles and (sort of) works again.
605
606		The library name (for version > 110.33) is $/eXene.cm.
607
608		I also added an new example in src/eXene/examples/nbody.  See the
609		README file there for details.
610
611		----------------------------------------------------------------------
612		Name: Matthias Blume
613		Date: 2001/06/20 16:40:00 EDT
614		Tag: blume-20010620-cml
615		Description:
616
617		CML now compiles and works again.
618
619		Libraries (for version > 110.33):
620
621		$cml/cml.cm            Main CML library.
622		$cml/basis.cm          CML's version of $/basis.cm.
623		$cml/cml-internal.cm   Internal helper library.
624		$cml/core-cml.cm       Internal helper library.
625		$cml-lib/trace-cml.cm  Tracing facility.
626		$cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
627
628		The installer (config/install.sh) has been taught how to properly
629		install this stuff.
630
631		----------------------------------------------------------------------
632		Name: Matthias Blume
633		Date: 2001/06/19 17:55:00 EDT
634		Tag: blume-20010619-instantiate
635		Description:
636
637		This un-breaks the fix for bug 1432.
638		(The bug was originally fixed in 110.9 but I broke it again some
639		time after that.)
640
641		----------------------------------------------------------------------
642		Name: Matthias Blume
643		Date: 2001/06/19 17:25:00 EDT
644		Tag: blume-20010619-signals
645		Description:
646
647		This should (hopefully) fix the long-standing signal handling bug.
648		(The runtime system was constructing a continuation record with an
649		incorrect descriptor which would cause the GC to drop data on the floor...)
650
651		----------------------------------------------------------------------
652		Name: Matthias Blume
653		Date: 2001/06/15 15:05:00 EDT
654		Tag: blume-20010615-moresparc
655		Description:
656
657		Here is a short late-hour update related to Sparc c-calls:
658
659		-- made handling of double-word arguments a bit smarter
660
661		-- instruction selection phase tries to collapse certain clumsily
662		constructed ML-Trees; typical example:
663
664		ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
665
666		This currently has no further impact on SML/NJ since mlriscGen does
667		not seem to generate such patterns in the first place, and c-calls
668		(which did generate them in the beginning) has meanwhile been fixed
669		so as to avoid them as well.
670
671		----------------------------------------------------------------------
672		Name: Matthias Blume
673		Date: 2001/06/15 15:05:00 EDT
674		Tag: blume-20010615-sparc
675		Description:
676
677		The purpose of this update is to provide an implementation of NLFFI
678		on Sparc machines.
679
680		Here are the changes in detail:
681
682		* src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
683		the Sparc implementation of the c-calls API.
684		* The Sparc backend of SML/NJ has been modified to uniformely use %fp
685		for accessing the ML frame.  Thus, we have a real frame pointer and
686		can freely modify %sp without need for an omit-frame-ptr phase.
687		The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
688		this case.
689		* ml-nlffigen has been taught to produce code for different architectures
690		and calling conventions.
691		* In a way similar to what was done in the x86 case, the Sparc
692		backend uses its own specific extension to mltree.  (For example,
693		it needs to be able to generate UNIMP instructions which are part
694		of the calling convention.)
695		* ml-nlffi-lib was reorganized to make it more modular (in particular,
696		to make it easier to plug in new machine- and os-dependent parts).
697
698		There are some other fairly unrelated bug fixes and cleanups as well:
699
700		* I further hacked the .cm files for MLRISC tools (like MDLGen) so
701		that they properly share their libraries with existing SML/NJ libraries.
702		* I fixed a minor cosmetic bug in CM, supressing certain spurious
703		follow-up error messages.
704		* Updates to CM/CMB documentation.
705
706		TODO items:
707
708		* MLRISC should use a different register as its asmTemp on the Sparc.
709		(The current %o2 is a really bad choice because it is part of the
710		calling conventions, so things might interfere in unexpected ways.)
711
712		----------------------------------------------------------------------
713		Name: Matthias Blume
714		Date: 2001/06/07
715		Tag: blume-20010607-calls
716		Description:
717
718		A number of internal changes related to C calls and calling conventions:
719
720		1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
721		number of bytes popped implicitly (by the callee).  In most cases
722		this field is 0 but on x86/win32 it is some non-zero value.  This
723		is information provided for the benefit of the "omit-frameptr" pass.
724		2. The CALL instruction on the x86 carries a similar "pops" field.
725		The instruction selection phase copies its value from the ML-Tree
726		CALL statement.
727		3. On all other architectures, the instruction selection phase checks
728		whether "pops=0" and complains if not.
729		4. The c-calls implementation for x86 now accepts two calling conventions:
730		"ccall" and "stdcall".  When "ccall" is selected, the caller cleans
731		up after the call and pops is set to 0.  For "stdcall", the caller
732		does nothing, leaving the cleanup to the callee; pops is set to
733		the number of bytes that were pushed onto the stack.
734		5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
735		distinguish between "ccall" and "stdcall".
736		6. The UNIMP instruction has been added to the supported Sparc instruction
737		set. (This is needed for implementing the official C calling convention
738		on this architecture.)
739		7. I fixed some of the .cm files under src/MLRISC/Tools to make them
740		work with the latest CM.
741
742		----------------------------------------------------------------------
743		Name: Matthias Blume
744		Date: 2001/06/05 15:10:00 EDT
745		Tag: blume-20010605-cm-index
746		Description:
747
748		0. The "lambdasplit" parameter for class "sml" in CM has been documented.
749
750		1. CM can now generate "index files".  These are human-readable files
751		that list on a per-.cm-file basis each toplevel symbol defined or
752		imported.  The location of the index file for
753		<p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
754		To enable index-file generation, set CM.Control.generate_index to true
755		or export an environment-symbol: export CM_GENERATE_INDEX=true.
756
757		The CM manual has been updated accordingly.
758
759		2. I made some slight modifications to the c-calls API in MLRISC.
760
761		a) There is now a callback to support saving/restoring of
762		dedicated but caller-save registers around the actual call
763		instruction.
764		b) One can optionally specify a comment-annotation for the
765		call instruction.
766
767		3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
768		(For example, the comment annotation shows the C prototype of
769		the function being called.)
770
771		----------------------------------------------------------------------
772		Name: Matthias Blume
773		Date: 2001/06/01 13:30:00 EDT
774		Tag: blume-20010601-nlffi-cleanup
775		Description:
776
777		This is mostly a cleanup of MLFFI stuff:
778
779		- some signature files have been put into a more exposed place
780		- the ugly 'f type parameter is gone (simplifies types tremendously!)
781		- ml-nlffigen changed accordingly
782		- tutorial updated
783
784		Other changes:
785
786		- author's affiliation in CM manual(s) updated
787		- some more recognized keywords added to Allen's sml.sty
788
789		----------------------------------------------------------------------
790		Name: Matthias Blume
791		Date: 2001/05/25 15:30:00 EDT
792		Tag: blume-20010525-iptr
793		Description:
794
795		- put the official 110.33-README (as it appears on the ftp server) under
796		CVS
797		- fixed a small bug related to incomplete pointer types in
798		ml-nlffigen
799		- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
800		(it does not need the 'f type parameter)
801
802		----------------------------------------------------------------------
803		Name: Matthias Blume
804		Date: 2001/05/23 14:30:00 EDT
805		Tag: Release_110_33
806		Description:
807
808		New version number (110.33).  New bootfiles.
809
810		----------------------------------------------------------------------
811		Name: Matthias Blume
812		Date: 2001/05/22 18:06:00 EDT
813		Tag: blume-20010522-targets
814		Description:
815
816		Made install.sh use file config/targets.customized if it exists, falling
817		back to config/targets if it doesn't.  This way one can have a customized
818		version of the targets file without touching the "real thing", thus
819		eliminating the constant fear of accidentally checking something bogus
820		back into the CVS repository...  (File config/targets.customized must
821		not be added to the repository!)
822
823		----------------------------------------------------------------------
824		Name: Matthias Blume
825		Date: 2001/05/22 16:30:00 EDT
826		Tag: blume-20010522-minitut
827		Description:
828
829		1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
830		struct returns.
831		2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
832		incomplete, preliminary documentation for NLFFI.
833
834		----------------------------------------------------------------------
835	Name: Matthias Blume	Name: Matthias Blume
836	Date: 2001/05/14 11:30:00 EDT	Date: 2001/05/14 11:30:00 EDT
837	Tag: blume-20010514-script	Tag: blume-20010514-script
#	Line 2246	Line 3065
3065	elaborator). There were a lot of changes during my "linkpath" trials	elaborator). There were a lot of changes during my "linkpath" trials
3066	that could have been reverted to their original state but weren't.	that could have been reverted to their original state but weren't.
3067	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
3068	than what was strictly necessary...  (I _did_ resist the tempation	than what was strictly necessary...  (I _did_ resist the temptation
3069	of doing any "global reformatting" to avoid an untimely death at	of doing any "global reformatting" to avoid an untimely death at
3070	Dave's hands. :)	Dave's hands. :)
3071

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.829   changed lines   Added in v.959

---

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