Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 716, Sat Nov 4 05:42:35 2000 UTC	revision 840, Fri Jun 15 19:05:19 2001 UTC
#	Line 14	Line 14
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16	Name: Matthias Blume	Name: Matthias Blume
17		Date: 2001/06/15 15:05:00 EDT
18		Tag: blume-20010615-sparc
19		Description:
20
21		The purpose of this update is to provide an implementation of NLFFI
22		on Sparc machines.
23
24		Here are the changes in detail:
25
26		* src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
27		the Sparc implementation of the c-calls API.
28		* The Sparc backend of SML/NJ has been modified to uniformely use %fp
29		for accessing the ML frame.  Thus, we have a real frame pointer and
30		can freely modify %sp without need for an omit-frame-ptr phase.
31		The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
32		this case.
33		* ml-nlffigen has been taught to produce code for different architectures
34		and calling conventions.
35		* In a way similar to what was done in the x86 case, the Sparc
36		backend uses its own specific extension to mltree.  (For example,
37		it needs to be able to generate UNIMP instructions which are part
38		of the calling convention.)
39		* ml-nlffi-lib was reorganized to make it more modular (in particular,
40		to make it easier to plug in new machine- and os-dependent parts).
41
42		There are some other fairly unrelated bug fixes and cleanups as well:
43
44		* I further hacked the .cm files for MLRISC tools (like MDLGen) so
45		that they properly share their libraries with existing SML/NJ libraries.
46		* I fixed a minor cosmetic bug in CM, supressing certain spurious
47		follow-up error messages.
48		* Updates to CM/CMB documentation.
49
50		TODO items:
51
52		* MLRISC should use a different register as its asmTemp on the Sparc.
53		(The current %o2 is a really bad choice because it is part of the
54		calling conventions, so things might interfere in unexpected ways.)
55
56		----------------------------------------------------------------------
57		Name: Matthias Blume
58		Date: 2001/06/07
59		Tag: blume-20010607-calls
60		Description:
61
62		A number of internal changes related to C calls and calling conventions:
63
64		1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
65		number of bytes popped implicitly (by the callee).  In most cases
66		this field is 0 but on x86/win32 it is some non-zero value.  This
67		is information provided for the benefit of the "omit-frameptr" pass.
68		2. The CALL instruction on the x86 carries a similar "pops" field.
69		The instruction selection phase copies its value from the ML-Tree
70		CALL statement.
71		3. On all other architectures, the instruction selection phase checks
72		whether "pops=0" and complains if not.
73		4. The c-calls implementation for x86 now accepts two calling conventions:
74		"ccall" and "stdcall".  When "ccall" is selected, the caller cleans
75		up after the call and pops is set to 0.  For "stdcall", the caller
76		does nothing, leaving the cleanup to the callee; pops is set to
77		the number of bytes that were pushed onto the stack.
78		5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
79		distinguish between "ccall" and "stdcall".
80		6. The UNIMP instruction has been added to the supported Sparc instruction
81		set. (This is needed for implementing the official C calling convention
82		on this architecture.)
83		7. I fixed some of the .cm files under src/MLRISC/Tools to make them
84		work with the latest CM.
85
86		----------------------------------------------------------------------
87		Name: Matthias Blume
88		Date: 2001/06/05 15:10:00 EDT
89		Tag: blume-20010605-cm-index
90		Description:
91
92		0. The "lambdasplit" parameter for class "sml" in CM has been documented.
93
94		1. CM can now generate "index files".  These are human-readable files
95		that list on a per-.cm-file basis each toplevel symbol defined or
96		imported.  The location of the index file for
97		<p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
98		To enable index-file generation, set CM.Control.generate_index to true
99		or export an environment-symbol: export CM_GENERATE_INDEX=true.
100
101		The CM manual has been updated accordingly.
102
103		2. I made some slight modifications to the c-calls API in MLRISC.
104
105		a) There is now a callback to support saving/restoring of
106		dedicated but caller-save registers around the actual call
107		instruction.
108		b) One can optionally specify a comment-annotation for the
109		call instruction.
110
111		3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
112		(For example, the comment annotation shows the C prototype of
113		the function being called.)
114
115		----------------------------------------------------------------------
116		Name: Matthias Blume
117		Date: 2001/06/01 13:30:00 EDT
118		Tag: blume-20010601-nlffi-cleanup
119		Description:
120
121		This is mostly a cleanup of MLFFI stuff:
122
123		- some signature files have been put into a more exposed place
124		- the ugly 'f type parameter is gone (simplifies types tremendously!)
125		- ml-nlffigen changed accordingly
126		- tutorial updated
127
128		Other changes:
129
130		- author's affiliation in CM manual(s) updated
131		- some more recognized keywords added to Allen's sml.sty
132
133		----------------------------------------------------------------------
134		Name: Matthias Blume
135		Date: 2001/05/25 15:30:00 EDT
136		Tag: blume-20010525-iptr
137		Description:
138
139		- put the official 110.33-README (as it appears on the ftp server) under
140		CVS
141		- fixed a small bug related to incomplete pointer types in
142		ml-nlffigen
143		- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
144		(it does not need the 'f type parameter)
145
146		----------------------------------------------------------------------
147		Name: Matthias Blume
148		Date: 2001/05/23 14:30:00 EDT
149		Tag: Release_110_33
150		Description:
151
152		New version number (110.33).  New bootfiles.
153
154		----------------------------------------------------------------------
155		Name: Matthias Blume
156		Date: 2001/05/22 18:06:00 EDT
157		Tag: blume-20010522-targets
158		Description:
159
160		Made install.sh use file config/targets.customized if it exists, falling
161		back to config/targets if it doesn't.  This way one can have a customized
162		version of the targets file without touching the "real thing", thus
163		eliminating the constant fear of accidentally checking something bogus
164		back into the CVS repository...  (File config/targets.customized must
165		not be added to the repository!)
166
167		----------------------------------------------------------------------
168		Name: Matthias Blume
169		Date: 2001/05/22 16:30:00 EDT
170		Tag: blume-20010522-minitut
171		Description:
172
173		1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
174		struct returns.
175		2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
176		incomplete, preliminary documentation for NLFFI.
177
178		----------------------------------------------------------------------
179		Name: Matthias Blume
180		Date: 2001/05/14 11:30:00 EDT
181		Tag: blume-20010514-script
182		Description:
183
184		Some bugs in install script fixed.
185
186		In addition to that I also made a slight change to the NLFFI API:
187		Functors generated by ml-nlffigen now take the dynamic library as a
188		straight functor argument, not as a suspended one.  (The original
189		functor code used to force the suspension right away anyway, so there
190		was nothing gained by this complication of the interface.)
191
192		----------------------------------------------------------------------
193		Name: Matthias Blume
194		Date: 2001/05/11 14:35:00 EDT
195		Tag: blume-20010511-ml-nlffi
196		Description:
197
198		I finally took the plunge and added my new FFI code to the main
199		repository.  For x86-linux it is now ready for prime-time.
200
201		There are two new subdirectories of "src":
202
203		- ml-nlffi-lib:
204		The utility library for programs using the FFI interface.
205		Here is the implementation of $/c.cm and its associated low-level
206		partners $/c-int.cm and $/memory.cm.
207		- ml-nlffigen:
208		A stand-alone program for generating ML glue code from C source
209		code.
210
211		Building ml-nlffigen requires $/ckit-lib.cm.
212
213		The config/install.sh script has been updates to do the Right Thing
214		(hopefully).
215
216		Notice that the source tree for the C-Kit will not be put under "src"
217		but directly under the installation root directory.  (This is the
218		structure that currently exists on the CVS server when you check out
219		module "sml".)  Fortunately, config/install.sh knows about this oddity.
220
221		Bugs: No documentation yet.
222
223		----------------------------------------------------------------------
224		Name: Matthias Blume
225		Date: 2001/05/09 16:35:00 EDT
226		Tag: blume-20010509-cpscontract
227		Description:
228
229		Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
230		wrapper/unwrapper elimination did not decrement usage counts and some
231		dead variables got overlooked by the dead-up logic.)
232
233		----------------------------------------------------------------------
234		Name: Lal George
235		Date: 2001/05/08  17:26:09 EDT
236		Tag: george-20010508-omit-frameptr
237		Description:
238
239		Changes to implement the omit-frame-pointer optimization to support
240		raw C calls. For now, there is only support on the Intel x86, but
241		other architectures will follow as more experience is gained with this.
242
243
244		----------------------------------------------------------------------
245		Name: Matthias Blume
246		Date: 2001/05/07 14:40:00 EDT
247		Tag: blume-20010507-proxies
248		Description:
249
250		I made into "proxy libraries" all libraries that qualify for such a
251		change.  (A qualifying library is a library that has another library or
252		groups as its sole member and repeats that member's export list
253		verbatim.  A proxy library avoids this repetition by omitting its export
254		list, effectively inheriting the list that its (only) member exports.
255		See the CM manual for more explanation.)
256		The main effect is that explicit export lists for these libraries
257		do not have to be kepts in sync, making maintenance a bit easier.
258
259		I also added copyright notices to many .cm-files.
260
261		Last but not least, I made a new set of bootfiles.
262
263		----------------------------------------------------------------------
264		Name: Matthias Blume
265		Date: 2001/05/04 17:00:00 EDT
266		Tag: blume-20010504-cm-lsplit
267		Description:
268
269		0. John merged pending changes to $/smlnj-lib.cm
270
271		1. Allen's previous change accidentally backed out of one of Lal's
272		earlier changes.  I undid this mistake (re-introducing Lal's change).
273
274		2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
275		within the compiler where applicable.  There is some code simplification
276		because of that.
277
278		3. The "split" phase (in FLINT) is now part of the default list of phases.
279		Compiler.Control.LambdaSplitting.* can be used to globally control the
280		lambda-splitting (cross-module-inlining) engine.  In addition to that,
281		it can now also be controlled on a per-source basis: CM has been taught
282		a new tool parameter applicable to ML source files.
283
284		- To turn lambda-splitting off completely:
285		local open Compiler.Control.LambdaSplitting in
286		val _ = set Off
287		end
288		- To make "no lambda-splitting" the global default (but allow per-source
289		overriding); this is the initial setting:
290		local open Compiler.Control.LambdaSplitting in
291		val _ = set (Default NONE)
292		end
293		- To make "lambda-splitting with aggressiveness a" the global default
294		(and allow per-source overriding):
295		local open Compiler.Control.LambdaSplitting in
296		val _ = set (Default (SOME a))
297		end
298
299		- To turn lambda-splitting off for a given ML souce file (say: a.sml)
300		write (in the respective .cm-file):
301		a.sml (lambdasplitting:off)
302		- To turn lambda-splitting for a.sml on with minimal aggressiveness:
303		a.sml (lambdasplitting:on)
304		- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
305		<a> is a decimal non-negative integer):
306		a.sml (lambdasplitting:<a>)
307		- To turn lambda-splitting for a.sml on with maximal aggressiveness:
308		a.sml (lambdasplitting:infinity)
309		- To use the global default for a.sml:
310		a.sml (lambdasplitting:default)
311		or simply
312		a.sml
313
314		----------------------------------------------------------------------
315		Name: Allen Leung
316		Date: 2001/05/04 01:57:00 EDT
317		Tag: leunga-20010504-sync
318		Description:
319
320		MLRISC features.
321
322		1. Fix to CMPXCHG instructions.
323		2. Changed RA interface to allow annotations in callbacks.
324		3. Added a new method to the stream interface to allow annotations updates.
325
326		----------------------------------------------------------------------
327		Name: Matthias Blume
328		Date: 2001/05/01 11:45:00 EDT
329		Tag: blume-20010501-pcedittmp
330		Description:
331
332		Changed install.sh to use the current working directory instead of
333		/usr/tmp for a temporary file (pcedittmp).  The previous choice
334		of /usr/tmp caused trouble with MacOS X because of file premission
335		problems.
336
337		----------------------------------------------------------------------
338		Name: Matthias Blume
339		Date: 2001/04/20 11:10:00 EDT
340		Tag: blume-20010420-inMLflag
341		Description:
342
343		- added vp_limitPtrMask to vproc-state.h
344		(for use by the raw-C-calls mechanism to implement proper interrupt
345		handling)
346		- made the ML compiler aware of various data-structure offsets so it
347		can generate code for accessing the vp_inML flag and vp_limitPtrMask
348		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
349		raw C-calls
350
351		----------------------------------------------------------------------
352		Name: Lal George
353		Date: 2001/04/20 09:15:28 EDT
354		Tag: george-20010420-macosX
355		Description:
356
357		- Changes to port to Mac OS X; Darwin.
358
359		- In the process I found that sqrt was broken on the PPC, because the
360		fsqrt instruction is not implemented.
361
362		----------------------------------------------------------------------
363		Name: Matthias Blume
364		Date: 2001/04/18 12:45:00 EDT
365		Tag: blume-20010418-ccalls
366		Description:
367
368		- fixed two off-by-4 errors in the x86-specific c-calls implementation
369		(this bug prevented structure arguments containing pointers from being
370		passed correctly)
371		- changed the raw-C-call code in mlriscGen.sml in such a way that
372		structure arguments are represented as a pointer to the beginning
373		of the structure (instead of having a series of synthesized arguments,
374		one for each structure member)
375
376		- made makeml script's verbosity level configurable via environment
377		variable (MAKEML_VERBOSITY)
378
379		- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
380		in rawmem-x86.sml; we are now using the real thing
381
382		----------------------------------------------------------------------
383		Name: Matthias Blume
384		Date: 2001/03/22 16:25:00 EST
385		Tag: blume-20010322-bootfiles
386		Description:
387
388		Created a new set of bootfiles (for your automatic installation convenience).
389
390		----------------------------------------------------------------------
391		Name: Matthias Blume
392		Date: 2001/03/22 15:10:00 EST
393		Tag: blume-20010322-rawmem-parcm
394		Description:
395
396		1. All "raw memory access" primitives for the new FFI are implemented now
397		(at least on the x86).
398		2. Some further cleanup of CM's parallel make mechanism.
399
400		----------------------------------------------------------------------
401		Name: Matthias Blume
402		Date: 2001/03/19 17:53:00 EST
403		Tag: blume-20010319-parallel
404		Description:
405
406		Parallel make (using compile servers) now works again.
407
408		To this end, CM.stabilize and CMB.make have been modified to work in
409		two passes when compile servers are attached:
410		1. Compile everything, do not perform stabilization; this pass
411		uses compile servers
412		2. Stabilize everything; this pass does not use compile servers
413		If there are no compile servers, the two passes are combined into one
414		(as before).  Splitting the passes increases the inherent parallelism
415		in the dependency graph because the entire graph including all
416		libraries is available at the same time.  This, in turn, improves
417		server utilization.  The downside is that the master process will
418		have to do some extra work after compilation is done (because for
419		technical reasons it must re-read all the binfiles during stabilization).
420
421		----------------------------------------------------------------------
422		Name: Matthias Blume
423		Date: 2001/03/16 12:22:00 EST
424		Tag: blume-20010316-bootfiles
425		Description:
426
427		Created a new set of bootfiles (for your automatic installation convenience).
428
429		----------------------------------------------------------------------
430		Name: Matthias Blume
431		Date: 2001/03/16 11:00:00 EST
432		Tag: blume-20010316-MLTREE-fixup
433		Description:
434
435		This is a minor fixup for an (untagged) earlier commit by Allen.
436		(A file was missing).
437
438		----------------------------------------------------------------------
439		Name: Allen Leung
440		Date: Mon Mar  5 18:54:57 EST 2001
441		Tag: leunga-20010305-cut-support
442
443		1. New support for alternative control-flow in MLTREE.
444		Currently we support
445
446		FLOW_TO(CALL ...., [k1,...,kn])
447
448		This is needed for 'cuts to' in C-- and try/handle-like constructs
449		in Moby
450
451		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
452
453		2. Register Allocator
454
455		Changes in interface [from Fermin, John]
456
457		3. Alpha 8-bit SLL support [Fermin]
458
459		4. All architectures
460
461		A new module (ClusterExpandCopies) for expanding parallel copies.
462
463		----------------------------------------------------------------------
464		Name: Allen Leung
465		Date: 2001/02/27 23:07:00 EST
466		Tag: leunga-20010227-minor-stuff
467
468		1. Alpha bug fix for CMOVNE
469		2. Handle mltree COND(..,FCMP ...,...)
470		3. Bug fix in simplifier
471
472		----------------------------------------------------------------------
473		Name: Matthias Blume
474		Date: 2001/01/30 17:50:00 EST
475		Tag: blume-20010130-sync
476		Description:
477
478		This is just a minor update to sync my devel branch with the main brach.
479		The only visible change is the addition of some README files.
480
481		----------------------------------------------------------------------
482		Name: Matthias Blume
483		Date: 2001/01/12 23:30:00 JST
484		Tag: blume-20010112-bootfiles
485		Description:
486
487		Made a new set of bootfiles that goes with the current state of the
488		repository.
489
490		----------------------------------------------------------------------
491		Name: Matthias Blume
492		Date: 2001/01/12 21:20:00 JST
493		Tag: blume-20010112-sync
494		Description:
495
496		I am just flushing out some minor changes that had accumulated in
497		my private branch in order to sync with the main tree.  (This is
498		mainly because I had CVS trouble when trying to merge _into_ my
499		private branch.)
500
501		Most people should be completely unaffected by this.
502
503		----------------------------------------------------------------------
504		Name: Allen Leung
505		Date: Thu Jan 11 21:03:00 EST 2001
506		Tag: leunga-20010111-labexp=mltree
507		Description:
508
509		1.  Removed the type LabelExp and replace it by MLTree.
510		2.  Rewritten mltree-simplify with the pattern matcher tool.
511		3.  There were some bugs in alpha code generator which would break
512		64-bit code generation.
513		4.  Redo the tools to generate code with the
514		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
515		are now generated by perl scripts.
516
517		----------------------------------------------------------------------
518		Name: Matthias Blume
519		Date: 2001/01/10 21:55:00 JST
520		Tag: blume-20010110-rcc
521		Description:
522
523		The RCC stuff now seems to work (but only on the x86).
524		This required hacking of the c-calls interface (and -implementation) in
525		MLRISC.
526
527		Normal compiler users should be unaffected.
528
529		----------------------------------------------------------------------
530		Name: Matthias Blume
531		Date: 2001/01/09 01:20:00 JST
532		Tag: blume-20010109-rcc
533		Description:
534
535		This is a fairly big patch, flushing out a large number of pending
536		changes that I made to my development copy over the last couple of days.
537
538		Of practical relevance at this moment is a workaround for a pickling
539		bug that Allen ran into the other day.  The cause of the bug itself is
540		still unknown and it might be hard to fix it properly, but the
541		workaround has some merits of its own (namely somewhat reducing pickling
542		overhead for certain libraries).  Therefore, I think this solution should
543		be satisfactory at this time.
544
545		The rest of the changes (i.e., the vast majority) has to do with my
546		ongoing efforts of providing direct support for C function calls from
547		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
548		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
549		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
550		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
551		handle RCC.
552
553		The last part (the changes to mlriscGen.sml) are still known to be
554		wrong on the x86 and not implemented on all other architectures.  But
555		the infrastructure is in place. I had to change a few functor
556		signatures in the backend to be able to route the CCalls interface
557		from MLRISC there, and I had to specialize the mltree type (on the
558		x86) to include the necessary extensions. (The extensions themselves
559		were already there and redy to go in MLRISC/x86).
560
561		Everything should be very happy as soon as someone helps me with
562		mlriscGen.sml...
563
564		In any case, nothing of this should matter to anyone as long as the
565		new primop is not being used (which is going to be the case unless you
566		find it where I hid it :). The rest of the compiler is completely
567		unaffected.
568
569		----------------------------------------------------------------------
570		Name: Matthias Blume
571		Date: 2001/01/05 00:30:00 JST
572		Tag: blume-20010105-primops
573		Description:
574
575		Added some experimental support for work that I am doing right now.
576		These changes mostly concern added primops, but there is also a new
577		experimental C library in the runtime system (but currently not enabled
578		anywhere except on Linux/X86).
579
580		In the course of adding primops (and playing with them), I discovered that
581		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
582		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
583		To recover, I made type information in INL_PRIM mandatory and changed
584		prim.sml as well as built-in.sml accordingly.  The InLine structure now
585		has complete, correct type information (i.e., no bottom types).
586
587		Since all these changes mean that we need new binfiles, I also bumped the
588		version number to 110.32.1.
589
590		----------------------------------------------------------------------
591		Name: Matthias Blume
592		Date: 2000/12/30 22:10:00 JST
593		Tag: blume-20001230-various
594		Description:
595
596		Added proxy libraries for MLRISC and let MLRISC libraries refer
597		to each other using path anchors.  (See CM manual for explanation.)
598
599		Updated CM documentation.
600
601		Fixed some bugs in CM.
602
603		Implemented "proxy" libraries (= syntactic sugar for CM).
604
605		Added "-quiet" option to makeml and changed runtime system accordingly.
606
607		Added cleanup handler for exportML to reset timers and compiler stats.
608
609		----------------------------------------------------------------------
610		Name: Lal George
611		Date: 2000/12/22 22:22:58 EST 2000
612		Tag: Release_110_32
613		Description:
614
615		Infinite precision used throughout MLRISC.
616		see MLRISC/mltree/machine-int.sig
617
618		----------------------------------------------------------------------
619		Name: Matthias Blume
620		Date: 2000/12/22 23:16:00 JST
621		Tag: blume-20001222-warn
622		Description:
623
624		Corrected wording and formatting of some CM warning message which I
625		broke in my previous patch.
626
627		----------------------------------------------------------------------
628		Name: Matthias Blume
629		Date: 2000/12/22 21:20:00 JST
630		Tag: blume-20001222-anchorenv
631		Description:
632
633		Fixed CM's handling of anchor environments in connection with CMB.make.
634
635		----------------------------------------------------------------------
636		Name: Matthias Blume
637		Date: 2000/12/22 13:15:00 JST
638		Tag: blume-20001222-cleanup
639		Description:
640
641		Removed src/cm/ffi which does not (and did not) belong here.
642
643		----------------------------------------------------------------------
644		Name: Matthias Blume
645		Date: 2000/12/21 23:55:00 JST
646		Tag: blume-20001221-exn
647		Description:
648
649		Probably most important: CM no longer silently swallows all exceptions
650		in the compiler.
651		Plus: some other minor CM changes.  For example, CM now reports some
652		sizes for generated binfiles (code, data, envpickle, lambdapickle).
653
654		----------------------------------------------------------------------
655		Name: Matthias Blume
656		Date: 2000/12/15 00:01:05 JST
657		Tag: blume-20001215-dirtool
658		Description:
659
660		- "dir" tool added.
661		- improvements and cleanup to Tools structure
662		- documentation updates
663
664		----------------------------------------------------------------------
665		Name: Allen Leung
666		Date: Thu Dec 14 03:45:24 EST 2000
667		Description:
668		Tag:  leunga-20001214-int-inf
669		Description:
670
671		In IntInf, added these standard functions, which are missing from our
672		implementation:
673
674		andb : int * int -> int
675		xorb : int * int -> int
676		orb  : int * int -> int
677		notb : int -> int
678		<<   : int * word -> int
679		~>>  : int * word -> int
680
681		Not tested, I hope they are correct.
682
683		----------------------------------------------------------------------
684		Name: Allen Leung
685		Date: Fri Dec  8 19:23:26 EST 2000
686		Description:
687		Tag:  leunga-20001208-nowhere
688		Description:
689
690		Slight improvements to the 'nowhere' tool to handle OR-patterns,
691		to generate better error messages etc.  Plus a brief manual.
692
693		----------------------------------------------------------------------
694		Name: Lal George
695		Date: 2000/12/08 09:54:02 EST 2000
696		Tag: Release_110_31
697		Description:
698
699		- Version 110.31
700		----------------------------------------------------------------------
701		Name: Allen Leung
702		Date: Thu Dec  7 22:01:04 EST 2000
703		Tag:  leunga-20001207-cell-monster-hack
704		Description:
705
706		Major MLRISC internal changes.  Affect all clients.
707		Summary:
708
709		1.  Type CELLS.cell = int is now replaced by a datatype.
710		As a result, the old regmap is now gone.  Almost all interfaces
711		in MLRISC change as a consequence.
712
713		2.  A new brand version of machine description tool (v3.0) that generates
714		modules expecting the new interface.  The old version is removed.
715
716		3.  The RA interface has been further abstracted into two new functors.
717		RISC_RA and X86RA.  These functors have much simpler interfaces.
718		[See also directory MLRISC/demo.]
719
720		4.  Some other new source->source code generation tools are available:
721
722		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
723		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
724		I use this tool to generate the peephole optimizers---with the new
725		cell type changes, peephole rules are becoming difficult to write
726		without conditional pattern matching.
727
728		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
729		cover the entire MLRISC source tree so many things broke.
730
731		6.  CM files have been moved to the subdirectory MLRISC/cm.
732		They are moved because there are a lot of them and they clutter up the
733		root dir.
734
735		7.  More detailed documentation to come...
736
737		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
738		a makeml -rebuild first.  This is because of other other
739		changes that Matthias has made (see below).
740
741
742		----------------------------------------------------------------------
743		Name: Matthias Blume
744		Date: 2000/11/30 23:12:00 JST
745		Tag: blume-20001130-filereorg
746		Description:
747
748		Some manual updates and some file reorganizations in CM.
749
750		----------------------------------------------------------------------
751		Name: Matthias Blume
752		Date: 2000/11/24 17:45:00 JST
753		Tag: blume-20001124-link
754		Description:
755
756		Drastically improved link traversal code for the case that the dynamic
757		value was already loaded at bootstrap time.  As a result, CM and CMB
758		now both load blazingly fast -- even on a very slow machine.  Also,
759		memory consumption has been further reduced by this.
760
761		Warning: The format of the PIDMAP file has changed.  THerefore, to
762		bootstrap you have to do this:
763
764		1. Run CMB.make
765		2. Make a symbolic link for the boot directory:
766		ln -s sml.boot.ARCH-OS xxx
767		3. "Rebuild" the boot directory:
768		./makeml -boot xxx -rebuild sml ; rm xxx
769		4. Boot normally:
770		./makeml
771
772		----------------------------------------------------------------------
773		Name: Matthias Blume
774		Date: 2000/11/21 21:20:00 JST
775		Tag: blume-20001121-tools
776		Description:
777
778		Continued hacking on autoloading problem -- with success this time.
779		Also changed tool-plugin mechanism.  See new CM manual.
780
781		----------------------------------------------------------------------
782		Name: Matthias Blume
783		Date: 2000/11/19 14:30:00 JST
784		Tag:  blume-20001119-autoload
785		Description:
786
787		Some hacking to make autoloading faster.  Success for CMB, no success
788		so far for CM.  There is a reduced structure CM' that autoloads faster.
789		(This is a temporary, non-documented hack to be eliminated again when
790		the general problem is solved.)
791
792		----------------------------------------------------------------------
793		Name: Matthias Blume
794		Date: 2000/11/17 14:10:00 JST
795		Tag: blume-20001117-pickle-lib
796		Description:
797
798		1. Eliminated comp-lib.cm
799		2. Made pickle-lib.cm
800		3. Eliminated all uses of intset.sml (from comp-lib.cm)
801		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
802		equivalent constructs from smlnj-lib.cm (INtHashTable).
803		5. Point 4. also goes for those uses of intmap.* in MLRISC.
804		Duplicated intmap modules thrown out.
805		6. Hunted down all duplicated SCC code and replaced it with
806		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
807		7. Rewrote Feedback module.
808		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
809		should be thrown out and equivalent modules from smlnj-lib.cm
810		should be used (IntRedBlackSet, IntListSet, ...).
811
812		Confirmed that compiler compiles to fixpoint.
813
814		----------------------------------------------------------------------
815		Name: Allen Leung
816		Date: 2000/11/10 18:00:00
817		Tag: leunga-20001110-new-x86-fp
818
819		A new x86 floating point code generator has been added.
820		By default this is turned off.  To turn this on, do:
821
822		CM.autoload "$smlnj/compiler.cm";
823		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
824
825		Changes:
826
827		1.  Changed FTAN to FPTAN so that the assembly output is correct.
828		2.  Changed the extension callback for FTANGENT to generate:
829
830		fptan
831		fstp  %st(0)
832		instead of
833		fptan
834		fstpl ftempmem
835
836		3.  Numerous assembly fixes for x86.
837
838		5.  Cleaned up the machine code output module x86/x86MC.sml and added
839		support for a whole bunch of instructions and addressing modes:
840
841		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
842		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
843		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
844		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
845		fxch %st(n)
846		fld %st(n)
847		fst %st(n)
848		fst mem
849		fstp %st(n)
850		fucom %st(n)
851		fucomp %st(n)
852
853		All these are now generated when the fast fp mode is turned on.
854
855		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
856
857		----------------------------------------------------------------------
858		Name: Matthias Blume
859		Date: 2000/11/09 11:20:00 JST
860		Tag: blume-20001109-scc
861		Description:
862
863		Eliminated some code duplication:
864
865		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
866		(Otherwise the functor is useless.)
867		2. Used GraphSCCFn where SCCUtilFun was used previously.
868		3. Got rid of SCCUtilFun (in comp-lib.cm).
869
870		----------------------------------------------------------------------
871		Name: Lal George
872		Date: 2000/11/06 09:02:21 EST 2000
873		Tag: Release_110_30
874		Description:
875
876		- Version 110.30
877		----------------------------------------------------------------------
878		Name: Matthias Blume
879	Date: 2000/11/04 14:45:00	Date: 2000/11/04 14:45:00
880	Tag: blume-20001104-mlbuild	Tag: blume-20001104-mlbuild
881	Description:	Description:

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.716   changed lines   Added in v.840

---

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