Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 713, Mon Oct 23 23:38:38 2000 UTC	revision 842, Tue Jun 19 21:24:35 2001 UTC
#	Line 13	Line 13
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2001/06/19 17:25:00 EDT
18		Tag: blume-20010619-signals
19		Description:
20
21		This should (hopefully) fix the long-standing signal handling bug.
22		(The runtime system was constructing a continuation record with an
23		incorrect descriptor which would cause the GC to drop data on the floor...)
24
25		----------------------------------------------------------------------
26		Name: Matthias Blume
27		Date: 2001/06/15 15:05:00 EDT
28		Tag: blume-20010615-moresparc
29		Description:
30
31		Here is a short late-hour update related to Sparc c-calls:
32
33		-- made handling of double-word arguments a bit smarter
34
35		-- instruction selection phase tries to collapse certain clumsily
36		constructed ML-Trees; typical example:
37
38		ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
39
40		This currently has no further impact on SML/NJ since mlriscGen does
41		not seem to generate such patterns in the first place, and c-calls
42		(which did generate them in the beginning) has meanwhile been fixed
43		so as to avoid them as well.
44
45		----------------------------------------------------------------------
46		Name: Matthias Blume
47		Date: 2001/06/15 15:05:00 EDT
48		Tag: blume-20010615-sparc
49		Description:
50
51		The purpose of this update is to provide an implementation of NLFFI
52		on Sparc machines.
53
54		Here are the changes in detail:
55
56		* src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
57		the Sparc implementation of the c-calls API.
58		* The Sparc backend of SML/NJ has been modified to uniformely use %fp
59		for accessing the ML frame.  Thus, we have a real frame pointer and
60		can freely modify %sp without need for an omit-frame-ptr phase.
61		The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
62		this case.
63		* ml-nlffigen has been taught to produce code for different architectures
64		and calling conventions.
65		* In a way similar to what was done in the x86 case, the Sparc
66		backend uses its own specific extension to mltree.  (For example,
67		it needs to be able to generate UNIMP instructions which are part
68		of the calling convention.)
69		* ml-nlffi-lib was reorganized to make it more modular (in particular,
70		to make it easier to plug in new machine- and os-dependent parts).
71
72		There are some other fairly unrelated bug fixes and cleanups as well:
73
74		* I further hacked the .cm files for MLRISC tools (like MDLGen) so
75		that they properly share their libraries with existing SML/NJ libraries.
76		* I fixed a minor cosmetic bug in CM, supressing certain spurious
77		follow-up error messages.
78		* Updates to CM/CMB documentation.
79
80		TODO items:
81
82		* MLRISC should use a different register as its asmTemp on the Sparc.
83		(The current %o2 is a really bad choice because it is part of the
84		calling conventions, so things might interfere in unexpected ways.)
85
86		----------------------------------------------------------------------
87		Name: Matthias Blume
88		Date: 2001/06/07
89		Tag: blume-20010607-calls
90		Description:
91
92		A number of internal changes related to C calls and calling conventions:
93
94		1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
95		number of bytes popped implicitly (by the callee).  In most cases
96		this field is 0 but on x86/win32 it is some non-zero value.  This
97		is information provided for the benefit of the "omit-frameptr" pass.
98		2. The CALL instruction on the x86 carries a similar "pops" field.
99		The instruction selection phase copies its value from the ML-Tree
100		CALL statement.
101		3. On all other architectures, the instruction selection phase checks
102		whether "pops=0" and complains if not.
103		4. The c-calls implementation for x86 now accepts two calling conventions:
104		"ccall" and "stdcall".  When "ccall" is selected, the caller cleans
105		up after the call and pops is set to 0.  For "stdcall", the caller
106		does nothing, leaving the cleanup to the callee; pops is set to
107		the number of bytes that were pushed onto the stack.
108		5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
109		distinguish between "ccall" and "stdcall".
110		6. The UNIMP instruction has been added to the supported Sparc instruction
111		set. (This is needed for implementing the official C calling convention
112		on this architecture.)
113		7. I fixed some of the .cm files under src/MLRISC/Tools to make them
114		work with the latest CM.
115
116		----------------------------------------------------------------------
117		Name: Matthias Blume
118		Date: 2001/06/05 15:10:00 EDT
119		Tag: blume-20010605-cm-index
120		Description:
121
122		0. The "lambdasplit" parameter for class "sml" in CM has been documented.
123
124		1. CM can now generate "index files".  These are human-readable files
125		that list on a per-.cm-file basis each toplevel symbol defined or
126		imported.  The location of the index file for
127		<p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
128		To enable index-file generation, set CM.Control.generate_index to true
129		or export an environment-symbol: export CM_GENERATE_INDEX=true.
130
131		The CM manual has been updated accordingly.
132
133		2. I made some slight modifications to the c-calls API in MLRISC.
134
135		a) There is now a callback to support saving/restoring of
136		dedicated but caller-save registers around the actual call
137		instruction.
138		b) One can optionally specify a comment-annotation for the
139		call instruction.
140
141		3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
142		(For example, the comment annotation shows the C prototype of
143		the function being called.)
144
145		----------------------------------------------------------------------
146		Name: Matthias Blume
147		Date: 2001/06/01 13:30:00 EDT
148		Tag: blume-20010601-nlffi-cleanup
149		Description:
150
151		This is mostly a cleanup of MLFFI stuff:
152
153		- some signature files have been put into a more exposed place
154		- the ugly 'f type parameter is gone (simplifies types tremendously!)
155		- ml-nlffigen changed accordingly
156		- tutorial updated
157
158		Other changes:
159
160		- author's affiliation in CM manual(s) updated
161		- some more recognized keywords added to Allen's sml.sty
162
163		----------------------------------------------------------------------
164		Name: Matthias Blume
165		Date: 2001/05/25 15:30:00 EDT
166		Tag: blume-20010525-iptr
167		Description:
168
169		- put the official 110.33-README (as it appears on the ftp server) under
170		CVS
171		- fixed a small bug related to incomplete pointer types in
172		ml-nlffigen
173		- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
174		(it does not need the 'f type parameter)
175
176		----------------------------------------------------------------------
177		Name: Matthias Blume
178		Date: 2001/05/23 14:30:00 EDT
179		Tag: Release_110_33
180		Description:
181
182		New version number (110.33).  New bootfiles.
183
184		----------------------------------------------------------------------
185		Name: Matthias Blume
186		Date: 2001/05/22 18:06:00 EDT
187		Tag: blume-20010522-targets
188		Description:
189
190		Made install.sh use file config/targets.customized if it exists, falling
191		back to config/targets if it doesn't.  This way one can have a customized
192		version of the targets file without touching the "real thing", thus
193		eliminating the constant fear of accidentally checking something bogus
194		back into the CVS repository...  (File config/targets.customized must
195		not be added to the repository!)
196
197		----------------------------------------------------------------------
198		Name: Matthias Blume
199		Date: 2001/05/22 16:30:00 EDT
200		Tag: blume-20010522-minitut
201		Description:
202
203		1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
204		struct returns.
205		2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
206		incomplete, preliminary documentation for NLFFI.
207
208		----------------------------------------------------------------------
209		Name: Matthias Blume
210		Date: 2001/05/14 11:30:00 EDT
211		Tag: blume-20010514-script
212		Description:
213
214		Some bugs in install script fixed.
215
216		In addition to that I also made a slight change to the NLFFI API:
217		Functors generated by ml-nlffigen now take the dynamic library as a
218		straight functor argument, not as a suspended one.  (The original
219		functor code used to force the suspension right away anyway, so there
220		was nothing gained by this complication of the interface.)
221
222		----------------------------------------------------------------------
223		Name: Matthias Blume
224		Date: 2001/05/11 14:35:00 EDT
225		Tag: blume-20010511-ml-nlffi
226		Description:
227
228		I finally took the plunge and added my new FFI code to the main
229		repository.  For x86-linux it is now ready for prime-time.
230
231		There are two new subdirectories of "src":
232
233		- ml-nlffi-lib:
234		The utility library for programs using the FFI interface.
235		Here is the implementation of $/c.cm and its associated low-level
236		partners $/c-int.cm and $/memory.cm.
237		- ml-nlffigen:
238		A stand-alone program for generating ML glue code from C source
239		code.
240
241		Building ml-nlffigen requires $/ckit-lib.cm.
242
243		The config/install.sh script has been updates to do the Right Thing
244		(hopefully).
245
246		Notice that the source tree for the C-Kit will not be put under "src"
247		but directly under the installation root directory.  (This is the
248		structure that currently exists on the CVS server when you check out
249		module "sml".)  Fortunately, config/install.sh knows about this oddity.
250
251		Bugs: No documentation yet.
252
253		----------------------------------------------------------------------
254		Name: Matthias Blume
255		Date: 2001/05/09 16:35:00 EDT
256		Tag: blume-20010509-cpscontract
257		Description:
258
259		Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
260		wrapper/unwrapper elimination did not decrement usage counts and some
261		dead variables got overlooked by the dead-up logic.)
262
263		----------------------------------------------------------------------
264		Name: Lal George
265		Date: 2001/05/08  17:26:09 EDT
266		Tag: george-20010508-omit-frameptr
267		Description:
268
269		Changes to implement the omit-frame-pointer optimization to support
270		raw C calls. For now, there is only support on the Intel x86, but
271		other architectures will follow as more experience is gained with this.
272
273
274		----------------------------------------------------------------------
275		Name: Matthias Blume
276		Date: 2001/05/07 14:40:00 EDT
277		Tag: blume-20010507-proxies
278		Description:
279
280		I made into "proxy libraries" all libraries that qualify for such a
281		change.  (A qualifying library is a library that has another library or
282		groups as its sole member and repeats that member's export list
283		verbatim.  A proxy library avoids this repetition by omitting its export
284		list, effectively inheriting the list that its (only) member exports.
285		See the CM manual for more explanation.)
286		The main effect is that explicit export lists for these libraries
287		do not have to be kepts in sync, making maintenance a bit easier.
288
289		I also added copyright notices to many .cm-files.
290
291		Last but not least, I made a new set of bootfiles.
292
293		----------------------------------------------------------------------
294		Name: Matthias Blume
295		Date: 2001/05/04 17:00:00 EDT
296		Tag: blume-20010504-cm-lsplit
297		Description:
298
299		0. John merged pending changes to $/smlnj-lib.cm
300
301		1. Allen's previous change accidentally backed out of one of Lal's
302		earlier changes.  I undid this mistake (re-introducing Lal's change).
303
304		2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
305		within the compiler where applicable.  There is some code simplification
306		because of that.
307
308		3. The "split" phase (in FLINT) is now part of the default list of phases.
309		Compiler.Control.LambdaSplitting.* can be used to globally control the
310		lambda-splitting (cross-module-inlining) engine.  In addition to that,
311		it can now also be controlled on a per-source basis: CM has been taught
312		a new tool parameter applicable to ML source files.
313
314		- To turn lambda-splitting off completely:
315		local open Compiler.Control.LambdaSplitting in
316		val _ = set Off
317		end
318		- To make "no lambda-splitting" the global default (but allow per-source
319		overriding); this is the initial setting:
320		local open Compiler.Control.LambdaSplitting in
321		val _ = set (Default NONE)
322		end
323		- To make "lambda-splitting with aggressiveness a" the global default
324		(and allow per-source overriding):
325		local open Compiler.Control.LambdaSplitting in
326		val _ = set (Default (SOME a))
327		end
328
329		- To turn lambda-splitting off for a given ML souce file (say: a.sml)
330		write (in the respective .cm-file):
331		a.sml (lambdasplitting:off)
332		- To turn lambda-splitting for a.sml on with minimal aggressiveness:
333		a.sml (lambdasplitting:on)
334		- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
335		<a> is a decimal non-negative integer):
336		a.sml (lambdasplitting:<a>)
337		- To turn lambda-splitting for a.sml on with maximal aggressiveness:
338		a.sml (lambdasplitting:infinity)
339		- To use the global default for a.sml:
340		a.sml (lambdasplitting:default)
341		or simply
342		a.sml
343
344		----------------------------------------------------------------------
345		Name: Allen Leung
346		Date: 2001/05/04 01:57:00 EDT
347		Tag: leunga-20010504-sync
348		Description:
349
350		MLRISC features.
351
352		1. Fix to CMPXCHG instructions.
353		2. Changed RA interface to allow annotations in callbacks.
354		3. Added a new method to the stream interface to allow annotations updates.
355
356		----------------------------------------------------------------------
357		Name: Matthias Blume
358		Date: 2001/05/01 11:45:00 EDT
359		Tag: blume-20010501-pcedittmp
360		Description:
361
362		Changed install.sh to use the current working directory instead of
363		/usr/tmp for a temporary file (pcedittmp).  The previous choice
364		of /usr/tmp caused trouble with MacOS X because of file premission
365		problems.
366
367		----------------------------------------------------------------------
368		Name: Matthias Blume
369		Date: 2001/04/20 11:10:00 EDT
370		Tag: blume-20010420-inMLflag
371		Description:
372
373		- added vp_limitPtrMask to vproc-state.h
374		(for use by the raw-C-calls mechanism to implement proper interrupt
375		handling)
376		- made the ML compiler aware of various data-structure offsets so it
377		can generate code for accessing the vp_inML flag and vp_limitPtrMask
378		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
379		raw C-calls
380
381		----------------------------------------------------------------------
382		Name: Lal George
383		Date: 2001/04/20 09:15:28 EDT
384		Tag: george-20010420-macosX
385		Description:
386
387		- Changes to port to Mac OS X; Darwin.
388
389		- In the process I found that sqrt was broken on the PPC, because the
390		fsqrt instruction is not implemented.
391
392		----------------------------------------------------------------------
393		Name: Matthias Blume
394		Date: 2001/04/18 12:45:00 EDT
395		Tag: blume-20010418-ccalls
396		Description:
397
398		- fixed two off-by-4 errors in the x86-specific c-calls implementation
399		(this bug prevented structure arguments containing pointers from being
400		passed correctly)
401		- changed the raw-C-call code in mlriscGen.sml in such a way that
402		structure arguments are represented as a pointer to the beginning
403		of the structure (instead of having a series of synthesized arguments,
404		one for each structure member)
405
406		- made makeml script's verbosity level configurable via environment
407		variable (MAKEML_VERBOSITY)
408
409		- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
410		in rawmem-x86.sml; we are now using the real thing
411
412		----------------------------------------------------------------------
413		Name: Matthias Blume
414		Date: 2001/03/22 16:25:00 EST
415		Tag: blume-20010322-bootfiles
416		Description:
417
418		Created a new set of bootfiles (for your automatic installation convenience).
419
420		----------------------------------------------------------------------
421		Name: Matthias Blume
422		Date: 2001/03/22 15:10:00 EST
423		Tag: blume-20010322-rawmem-parcm
424		Description:
425
426		1. All "raw memory access" primitives for the new FFI are implemented now
427		(at least on the x86).
428		2. Some further cleanup of CM's parallel make mechanism.
429
430		----------------------------------------------------------------------
431		Name: Matthias Blume
432		Date: 2001/03/19 17:53:00 EST
433		Tag: blume-20010319-parallel
434		Description:
435
436		Parallel make (using compile servers) now works again.
437
438		To this end, CM.stabilize and CMB.make have been modified to work in
439		two passes when compile servers are attached:
440		1. Compile everything, do not perform stabilization; this pass
441		uses compile servers
442		2. Stabilize everything; this pass does not use compile servers
443		If there are no compile servers, the two passes are combined into one
444		(as before).  Splitting the passes increases the inherent parallelism
445		in the dependency graph because the entire graph including all
446		libraries is available at the same time.  This, in turn, improves
447		server utilization.  The downside is that the master process will
448		have to do some extra work after compilation is done (because for
449		technical reasons it must re-read all the binfiles during stabilization).
450
451		----------------------------------------------------------------------
452		Name: Matthias Blume
453		Date: 2001/03/16 12:22:00 EST
454		Tag: blume-20010316-bootfiles
455		Description:
456
457		Created a new set of bootfiles (for your automatic installation convenience).
458
459		----------------------------------------------------------------------
460		Name: Matthias Blume
461		Date: 2001/03/16 11:00:00 EST
462		Tag: blume-20010316-MLTREE-fixup
463		Description:
464
465		This is a minor fixup for an (untagged) earlier commit by Allen.
466		(A file was missing).
467
468		----------------------------------------------------------------------
469		Name: Allen Leung
470		Date: Mon Mar  5 18:54:57 EST 2001
471		Tag: leunga-20010305-cut-support
472
473		1. New support for alternative control-flow in MLTREE.
474		Currently we support
475
476		FLOW_TO(CALL ...., [k1,...,kn])
477
478		This is needed for 'cuts to' in C-- and try/handle-like constructs
479		in Moby
480
481		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
482
483		2. Register Allocator
484
485		Changes in interface [from Fermin, John]
486
487		3. Alpha 8-bit SLL support [Fermin]
488
489		4. All architectures
490
491		A new module (ClusterExpandCopies) for expanding parallel copies.
492
493		----------------------------------------------------------------------
494		Name: Allen Leung
495		Date: 2001/02/27 23:07:00 EST
496		Tag: leunga-20010227-minor-stuff
497
498		1. Alpha bug fix for CMOVNE
499		2. Handle mltree COND(..,FCMP ...,...)
500		3. Bug fix in simplifier
501
502		----------------------------------------------------------------------
503		Name: Matthias Blume
504		Date: 2001/01/30 17:50:00 EST
505		Tag: blume-20010130-sync
506		Description:
507
508		This is just a minor update to sync my devel branch with the main brach.
509		The only visible change is the addition of some README files.
510
511		----------------------------------------------------------------------
512		Name: Matthias Blume
513		Date: 2001/01/12 23:30:00 JST
514		Tag: blume-20010112-bootfiles
515		Description:
516
517		Made a new set of bootfiles that goes with the current state of the
518		repository.
519
520		----------------------------------------------------------------------
521		Name: Matthias Blume
522		Date: 2001/01/12 21:20:00 JST
523		Tag: blume-20010112-sync
524		Description:
525
526		I am just flushing out some minor changes that had accumulated in
527		my private branch in order to sync with the main tree.  (This is
528		mainly because I had CVS trouble when trying to merge _into_ my
529		private branch.)
530
531		Most people should be completely unaffected by this.
532
533		----------------------------------------------------------------------
534		Name: Allen Leung
535		Date: Thu Jan 11 21:03:00 EST 2001
536		Tag: leunga-20010111-labexp=mltree
537		Description:
538
539		1.  Removed the type LabelExp and replace it by MLTree.
540		2.  Rewritten mltree-simplify with the pattern matcher tool.
541		3.  There were some bugs in alpha code generator which would break
542		64-bit code generation.
543		4.  Redo the tools to generate code with the
544		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
545		are now generated by perl scripts.
546
547		----------------------------------------------------------------------
548		Name: Matthias Blume
549		Date: 2001/01/10 21:55:00 JST
550		Tag: blume-20010110-rcc
551		Description:
552
553		The RCC stuff now seems to work (but only on the x86).
554		This required hacking of the c-calls interface (and -implementation) in
555		MLRISC.
556
557		Normal compiler users should be unaffected.
558
559		----------------------------------------------------------------------
560		Name: Matthias Blume
561		Date: 2001/01/09 01:20:00 JST
562		Tag: blume-20010109-rcc
563		Description:
564
565		This is a fairly big patch, flushing out a large number of pending
566		changes that I made to my development copy over the last couple of days.
567
568		Of practical relevance at this moment is a workaround for a pickling
569		bug that Allen ran into the other day.  The cause of the bug itself is
570		still unknown and it might be hard to fix it properly, but the
571		workaround has some merits of its own (namely somewhat reducing pickling
572		overhead for certain libraries).  Therefore, I think this solution should
573		be satisfactory at this time.
574
575		The rest of the changes (i.e., the vast majority) has to do with my
576		ongoing efforts of providing direct support for C function calls from
577		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
578		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
579		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
580		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
581		handle RCC.
582
583		The last part (the changes to mlriscGen.sml) are still known to be
584		wrong on the x86 and not implemented on all other architectures.  But
585		the infrastructure is in place. I had to change a few functor
586		signatures in the backend to be able to route the CCalls interface
587		from MLRISC there, and I had to specialize the mltree type (on the
588		x86) to include the necessary extensions. (The extensions themselves
589		were already there and redy to go in MLRISC/x86).
590
591		Everything should be very happy as soon as someone helps me with
592		mlriscGen.sml...
593
594		In any case, nothing of this should matter to anyone as long as the
595		new primop is not being used (which is going to be the case unless you
596		find it where I hid it :). The rest of the compiler is completely
597		unaffected.
598
599		----------------------------------------------------------------------
600		Name: Matthias Blume
601		Date: 2001/01/05 00:30:00 JST
602		Tag: blume-20010105-primops
603		Description:
604
605		Added some experimental support for work that I am doing right now.
606		These changes mostly concern added primops, but there is also a new
607		experimental C library in the runtime system (but currently not enabled
608		anywhere except on Linux/X86).
609
610		In the course of adding primops (and playing with them), I discovered that
611		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
612		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
613		To recover, I made type information in INL_PRIM mandatory and changed
614		prim.sml as well as built-in.sml accordingly.  The InLine structure now
615		has complete, correct type information (i.e., no bottom types).
616
617		Since all these changes mean that we need new binfiles, I also bumped the
618		version number to 110.32.1.
619
620		----------------------------------------------------------------------
621		Name: Matthias Blume
622		Date: 2000/12/30 22:10:00 JST
623		Tag: blume-20001230-various
624		Description:
625
626		Added proxy libraries for MLRISC and let MLRISC libraries refer
627		to each other using path anchors.  (See CM manual for explanation.)
628
629		Updated CM documentation.
630
631		Fixed some bugs in CM.
632
633		Implemented "proxy" libraries (= syntactic sugar for CM).
634
635		Added "-quiet" option to makeml and changed runtime system accordingly.
636
637		Added cleanup handler for exportML to reset timers and compiler stats.
638
639		----------------------------------------------------------------------
640		Name: Lal George
641		Date: 2000/12/22 22:22:58 EST 2000
642		Tag: Release_110_32
643		Description:
644
645		Infinite precision used throughout MLRISC.
646		see MLRISC/mltree/machine-int.sig
647
648		----------------------------------------------------------------------
649		Name: Matthias Blume
650		Date: 2000/12/22 23:16:00 JST
651		Tag: blume-20001222-warn
652		Description:
653
654		Corrected wording and formatting of some CM warning message which I
655		broke in my previous patch.
656
657		----------------------------------------------------------------------
658		Name: Matthias Blume
659		Date: 2000/12/22 21:20:00 JST
660		Tag: blume-20001222-anchorenv
661		Description:
662
663		Fixed CM's handling of anchor environments in connection with CMB.make.
664
665		----------------------------------------------------------------------
666		Name: Matthias Blume
667		Date: 2000/12/22 13:15:00 JST
668		Tag: blume-20001222-cleanup
669		Description:
670
671		Removed src/cm/ffi which does not (and did not) belong here.
672
673		----------------------------------------------------------------------
674		Name: Matthias Blume
675		Date: 2000/12/21 23:55:00 JST
676		Tag: blume-20001221-exn
677		Description:
678
679		Probably most important: CM no longer silently swallows all exceptions
680		in the compiler.
681		Plus: some other minor CM changes.  For example, CM now reports some
682		sizes for generated binfiles (code, data, envpickle, lambdapickle).
683
684		----------------------------------------------------------------------
685		Name: Matthias Blume
686		Date: 2000/12/15 00:01:05 JST
687		Tag: blume-20001215-dirtool
688		Description:
689
690		- "dir" tool added.
691		- improvements and cleanup to Tools structure
692		- documentation updates
693
694		----------------------------------------------------------------------
695		Name: Allen Leung
696		Date: Thu Dec 14 03:45:24 EST 2000
697		Description:
698		Tag:  leunga-20001214-int-inf
699		Description:
700
701		In IntInf, added these standard functions, which are missing from our
702		implementation:
703
704		andb : int * int -> int
705		xorb : int * int -> int
706		orb  : int * int -> int
707		notb : int -> int
708		<<   : int * word -> int
709		~>>  : int * word -> int
710
711		Not tested, I hope they are correct.
712
713		----------------------------------------------------------------------
714		Name: Allen Leung
715		Date: Fri Dec  8 19:23:26 EST 2000
716		Description:
717		Tag:  leunga-20001208-nowhere
718		Description:
719
720		Slight improvements to the 'nowhere' tool to handle OR-patterns,
721		to generate better error messages etc.  Plus a brief manual.
722
723		----------------------------------------------------------------------
724		Name: Lal George
725		Date: 2000/12/08 09:54:02 EST 2000
726		Tag: Release_110_31
727		Description:
728
729		- Version 110.31
730		----------------------------------------------------------------------
731		Name: Allen Leung
732		Date: Thu Dec  7 22:01:04 EST 2000
733		Tag:  leunga-20001207-cell-monster-hack
734		Description:
735
736		Major MLRISC internal changes.  Affect all clients.
737		Summary:
738
739		1.  Type CELLS.cell = int is now replaced by a datatype.
740		As a result, the old regmap is now gone.  Almost all interfaces
741		in MLRISC change as a consequence.
742
743		2.  A new brand version of machine description tool (v3.0) that generates
744		modules expecting the new interface.  The old version is removed.
745
746		3.  The RA interface has been further abstracted into two new functors.
747		RISC_RA and X86RA.  These functors have much simpler interfaces.
748		[See also directory MLRISC/demo.]
749
750		4.  Some other new source->source code generation tools are available:
751
752		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
753		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
754		I use this tool to generate the peephole optimizers---with the new
755		cell type changes, peephole rules are becoming difficult to write
756		without conditional pattern matching.
757
758		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
759		cover the entire MLRISC source tree so many things broke.
760
761		6.  CM files have been moved to the subdirectory MLRISC/cm.
762		They are moved because there are a lot of them and they clutter up the
763		root dir.
764
765		7.  More detailed documentation to come...
766
767		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
768		a makeml -rebuild first.  This is because of other other
769		changes that Matthias has made (see below).
770
771
772		----------------------------------------------------------------------
773		Name: Matthias Blume
774		Date: 2000/11/30 23:12:00 JST
775		Tag: blume-20001130-filereorg
776		Description:
777
778		Some manual updates and some file reorganizations in CM.
779
780		----------------------------------------------------------------------
781		Name: Matthias Blume
782		Date: 2000/11/24 17:45:00 JST
783		Tag: blume-20001124-link
784		Description:
785
786		Drastically improved link traversal code for the case that the dynamic
787		value was already loaded at bootstrap time.  As a result, CM and CMB
788		now both load blazingly fast -- even on a very slow machine.  Also,
789		memory consumption has been further reduced by this.
790
791		Warning: The format of the PIDMAP file has changed.  THerefore, to
792		bootstrap you have to do this:
793
794		1. Run CMB.make
795		2. Make a symbolic link for the boot directory:
796		ln -s sml.boot.ARCH-OS xxx
797		3. "Rebuild" the boot directory:
798		./makeml -boot xxx -rebuild sml ; rm xxx
799		4. Boot normally:
800		./makeml
801
802		----------------------------------------------------------------------
803		Name: Matthias Blume
804		Date: 2000/11/21 21:20:00 JST
805		Tag: blume-20001121-tools
806		Description:
807
808		Continued hacking on autoloading problem -- with success this time.
809		Also changed tool-plugin mechanism.  See new CM manual.
810
811		----------------------------------------------------------------------
812		Name: Matthias Blume
813		Date: 2000/11/19 14:30:00 JST
814		Tag:  blume-20001119-autoload
815		Description:
816
817		Some hacking to make autoloading faster.  Success for CMB, no success
818		so far for CM.  There is a reduced structure CM' that autoloads faster.
819		(This is a temporary, non-documented hack to be eliminated again when
820		the general problem is solved.)
821
822		----------------------------------------------------------------------
823		Name: Matthias Blume
824		Date: 2000/11/17 14:10:00 JST
825		Tag: blume-20001117-pickle-lib
826		Description:
827
828		1. Eliminated comp-lib.cm
829		2. Made pickle-lib.cm
830		3. Eliminated all uses of intset.sml (from comp-lib.cm)
831		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
832		equivalent constructs from smlnj-lib.cm (INtHashTable).
833		5. Point 4. also goes for those uses of intmap.* in MLRISC.
834		Duplicated intmap modules thrown out.
835		6. Hunted down all duplicated SCC code and replaced it with
836		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
837		7. Rewrote Feedback module.
838		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
839		should be thrown out and equivalent modules from smlnj-lib.cm
840		should be used (IntRedBlackSet, IntListSet, ...).
841
842		Confirmed that compiler compiles to fixpoint.
843
844		----------------------------------------------------------------------
845		Name: Allen Leung
846		Date: 2000/11/10 18:00:00
847		Tag: leunga-20001110-new-x86-fp
848
849		A new x86 floating point code generator has been added.
850		By default this is turned off.  To turn this on, do:
851
852		CM.autoload "$smlnj/compiler.cm";
853		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
854
855		Changes:
856
857		1.  Changed FTAN to FPTAN so that the assembly output is correct.
858		2.  Changed the extension callback for FTANGENT to generate:
859
860		fptan
861		fstp  %st(0)
862		instead of
863		fptan
864		fstpl ftempmem
865
866		3.  Numerous assembly fixes for x86.
867
868		5.  Cleaned up the machine code output module x86/x86MC.sml and added
869		support for a whole bunch of instructions and addressing modes:
870
871		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
872		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
873		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
874		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
875		fxch %st(n)
876		fld %st(n)
877		fst %st(n)
878		fst mem
879		fstp %st(n)
880		fucom %st(n)
881		fucomp %st(n)
882
883		All these are now generated when the fast fp mode is turned on.
884
885		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
886
887		----------------------------------------------------------------------
888		Name: Matthias Blume
889		Date: 2000/11/09 11:20:00 JST
890		Tag: blume-20001109-scc
891		Description:
892
893		Eliminated some code duplication:
894
895		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
896		(Otherwise the functor is useless.)
897		2. Used GraphSCCFn where SCCUtilFun was used previously.
898		3. Got rid of SCCUtilFun (in comp-lib.cm).
899
900		----------------------------------------------------------------------
901		Name: Lal George
902		Date: 2000/11/06 09:02:21 EST 2000
903		Tag: Release_110_30
904		Description:
905
906		- Version 110.30
907		----------------------------------------------------------------------
908		Name: Matthias Blume
909		Date: 2000/11/04 14:45:00
910		Tag: blume-20001104-mlbuild
911		Description:
912
913		- Made ml-build faster on startup.
914		- Documentation fixes.
915
916		----------------------------------------------------------------------
917		Name: Matthias Blume
918		Date: 2000/11/02 17:00:00 JST
919		Tag: blume-20001102-condcomp
920		Description:
921
922		- Small tweaks to pickler -- new BOOTFILES!
923		- Version bumped to 110.29.2.
924		- Added conditional compilation facility to init.cmi (see comment there).
925		----------------------------------------------------------------------
926	Name: Allen Leung	Name: Allen Leung
927	Date: 2000/10/23 19:31:00	Date: 2000/10/23 19:31:00
928	Tag: leunga-20001023-demo-ra	Tag: leunga-20001023-demo-ra

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.713   changed lines   Added in v.842

---

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