Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 713, Mon Oct 23 23:38:38 2000 UTC	revision 844, Wed Jun 20 20:39:15 2001 UTC
#	Line 13	Line 13
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2001/06/20 16:40:00 EDT
18		Tag: blume-20010620-cml
19		Description:
20
21		CML now compiles and works again.
22
23		Libraries (for version > 110.33):
24
25		$cml/cml.cm            Main CML library.
26		$cml/basis.cm          CML's version of $/basis.cm.
27		$cml/cml-internal.cm   Internal helper library.
28		$cml/core-cml.cm       Internal helper library.
29		$cml-lib/trace-cml.cm  Tracing facility.
30		$cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
31
32		The installer (config/install.sh) has been taught how to properly
33		install this stuff.
34
35		----------------------------------------------------------------------
36		Name: Matthias Blume
37		Date: 2001/06/19 17:55:00 EDT
38		Tag: blume-20010619-instantiate
39		Description:
40
41		This un-breaks the fix for bug 1432.
42		(The bug was originally fixed in 110.9 but I broke it again some
43		time after that.)
44
45		----------------------------------------------------------------------
46		Name: Matthias Blume
47		Date: 2001/06/19 17:25:00 EDT
48		Tag: blume-20010619-signals
49		Description:
50
51		This should (hopefully) fix the long-standing signal handling bug.
52		(The runtime system was constructing a continuation record with an
53		incorrect descriptor which would cause the GC to drop data on the floor...)
54
55		----------------------------------------------------------------------
56		Name: Matthias Blume
57		Date: 2001/06/15 15:05:00 EDT
58		Tag: blume-20010615-moresparc
59		Description:
60
61		Here is a short late-hour update related to Sparc c-calls:
62
63		-- made handling of double-word arguments a bit smarter
64
65		-- instruction selection phase tries to collapse certain clumsily
66		constructed ML-Trees; typical example:
67
68		ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
69
70		This currently has no further impact on SML/NJ since mlriscGen does
71		not seem to generate such patterns in the first place, and c-calls
72		(which did generate them in the beginning) has meanwhile been fixed
73		so as to avoid them as well.
74
75		----------------------------------------------------------------------
76		Name: Matthias Blume
77		Date: 2001/06/15 15:05:00 EDT
78		Tag: blume-20010615-sparc
79		Description:
80
81		The purpose of this update is to provide an implementation of NLFFI
82		on Sparc machines.
83
84		Here are the changes in detail:
85
86		* src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
87		the Sparc implementation of the c-calls API.
88		* The Sparc backend of SML/NJ has been modified to uniformely use %fp
89		for accessing the ML frame.  Thus, we have a real frame pointer and
90		can freely modify %sp without need for an omit-frame-ptr phase.
91		The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
92		this case.
93		* ml-nlffigen has been taught to produce code for different architectures
94		and calling conventions.
95		* In a way similar to what was done in the x86 case, the Sparc
96		backend uses its own specific extension to mltree.  (For example,
97		it needs to be able to generate UNIMP instructions which are part
98		of the calling convention.)
99		* ml-nlffi-lib was reorganized to make it more modular (in particular,
100		to make it easier to plug in new machine- and os-dependent parts).
101
102		There are some other fairly unrelated bug fixes and cleanups as well:
103
104		* I further hacked the .cm files for MLRISC tools (like MDLGen) so
105		that they properly share their libraries with existing SML/NJ libraries.
106		* I fixed a minor cosmetic bug in CM, supressing certain spurious
107		follow-up error messages.
108		* Updates to CM/CMB documentation.
109
110		TODO items:
111
112		* MLRISC should use a different register as its asmTemp on the Sparc.
113		(The current %o2 is a really bad choice because it is part of the
114		calling conventions, so things might interfere in unexpected ways.)
115
116		----------------------------------------------------------------------
117		Name: Matthias Blume
118		Date: 2001/06/07
119		Tag: blume-20010607-calls
120		Description:
121
122		A number of internal changes related to C calls and calling conventions:
123
124		1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
125		number of bytes popped implicitly (by the callee).  In most cases
126		this field is 0 but on x86/win32 it is some non-zero value.  This
127		is information provided for the benefit of the "omit-frameptr" pass.
128		2. The CALL instruction on the x86 carries a similar "pops" field.
129		The instruction selection phase copies its value from the ML-Tree
130		CALL statement.
131		3. On all other architectures, the instruction selection phase checks
132		whether "pops=0" and complains if not.
133		4. The c-calls implementation for x86 now accepts two calling conventions:
134		"ccall" and "stdcall".  When "ccall" is selected, the caller cleans
135		up after the call and pops is set to 0.  For "stdcall", the caller
136		does nothing, leaving the cleanup to the callee; pops is set to
137		the number of bytes that were pushed onto the stack.
138		5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
139		distinguish between "ccall" and "stdcall".
140		6. The UNIMP instruction has been added to the supported Sparc instruction
141		set. (This is needed for implementing the official C calling convention
142		on this architecture.)
143		7. I fixed some of the .cm files under src/MLRISC/Tools to make them
144		work with the latest CM.
145
146		----------------------------------------------------------------------
147		Name: Matthias Blume
148		Date: 2001/06/05 15:10:00 EDT
149		Tag: blume-20010605-cm-index
150		Description:
151
152		0. The "lambdasplit" parameter for class "sml" in CM has been documented.
153
154		1. CM can now generate "index files".  These are human-readable files
155		that list on a per-.cm-file basis each toplevel symbol defined or
156		imported.  The location of the index file for
157		<p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
158		To enable index-file generation, set CM.Control.generate_index to true
159		or export an environment-symbol: export CM_GENERATE_INDEX=true.
160
161		The CM manual has been updated accordingly.
162
163		2. I made some slight modifications to the c-calls API in MLRISC.
164
165		a) There is now a callback to support saving/restoring of
166		dedicated but caller-save registers around the actual call
167		instruction.
168		b) One can optionally specify a comment-annotation for the
169		call instruction.
170
171		3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
172		(For example, the comment annotation shows the C prototype of
173		the function being called.)
174
175		----------------------------------------------------------------------
176		Name: Matthias Blume
177		Date: 2001/06/01 13:30:00 EDT
178		Tag: blume-20010601-nlffi-cleanup
179		Description:
180
181		This is mostly a cleanup of MLFFI stuff:
182
183		- some signature files have been put into a more exposed place
184		- the ugly 'f type parameter is gone (simplifies types tremendously!)
185		- ml-nlffigen changed accordingly
186		- tutorial updated
187
188		Other changes:
189
190		- author's affiliation in CM manual(s) updated
191		- some more recognized keywords added to Allen's sml.sty
192
193		----------------------------------------------------------------------
194		Name: Matthias Blume
195		Date: 2001/05/25 15:30:00 EDT
196		Tag: blume-20010525-iptr
197		Description:
198
199		- put the official 110.33-README (as it appears on the ftp server) under
200		CVS
201		- fixed a small bug related to incomplete pointer types in
202		ml-nlffigen
203		- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
204		(it does not need the 'f type parameter)
205
206		----------------------------------------------------------------------
207		Name: Matthias Blume
208		Date: 2001/05/23 14:30:00 EDT
209		Tag: Release_110_33
210		Description:
211
212		New version number (110.33).  New bootfiles.
213
214		----------------------------------------------------------------------
215		Name: Matthias Blume
216		Date: 2001/05/22 18:06:00 EDT
217		Tag: blume-20010522-targets
218		Description:
219
220		Made install.sh use file config/targets.customized if it exists, falling
221		back to config/targets if it doesn't.  This way one can have a customized
222		version of the targets file without touching the "real thing", thus
223		eliminating the constant fear of accidentally checking something bogus
224		back into the CVS repository...  (File config/targets.customized must
225		not be added to the repository!)
226
227		----------------------------------------------------------------------
228		Name: Matthias Blume
229		Date: 2001/05/22 16:30:00 EDT
230		Tag: blume-20010522-minitut
231		Description:
232
233		1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
234		struct returns.
235		2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
236		incomplete, preliminary documentation for NLFFI.
237
238		----------------------------------------------------------------------
239		Name: Matthias Blume
240		Date: 2001/05/14 11:30:00 EDT
241		Tag: blume-20010514-script
242		Description:
243
244		Some bugs in install script fixed.
245
246		In addition to that I also made a slight change to the NLFFI API:
247		Functors generated by ml-nlffigen now take the dynamic library as a
248		straight functor argument, not as a suspended one.  (The original
249		functor code used to force the suspension right away anyway, so there
250		was nothing gained by this complication of the interface.)
251
252		----------------------------------------------------------------------
253		Name: Matthias Blume
254		Date: 2001/05/11 14:35:00 EDT
255		Tag: blume-20010511-ml-nlffi
256		Description:
257
258		I finally took the plunge and added my new FFI code to the main
259		repository.  For x86-linux it is now ready for prime-time.
260
261		There are two new subdirectories of "src":
262
263		- ml-nlffi-lib:
264		The utility library for programs using the FFI interface.
265		Here is the implementation of $/c.cm and its associated low-level
266		partners $/c-int.cm and $/memory.cm.
267		- ml-nlffigen:
268		A stand-alone program for generating ML glue code from C source
269		code.
270
271		Building ml-nlffigen requires $/ckit-lib.cm.
272
273		The config/install.sh script has been updates to do the Right Thing
274		(hopefully).
275
276		Notice that the source tree for the C-Kit will not be put under "src"
277		but directly under the installation root directory.  (This is the
278		structure that currently exists on the CVS server when you check out
279		module "sml".)  Fortunately, config/install.sh knows about this oddity.
280
281		Bugs: No documentation yet.
282
283		----------------------------------------------------------------------
284		Name: Matthias Blume
285		Date: 2001/05/09 16:35:00 EDT
286		Tag: blume-20010509-cpscontract
287		Description:
288
289		Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
290		wrapper/unwrapper elimination did not decrement usage counts and some
291		dead variables got overlooked by the dead-up logic.)
292
293		----------------------------------------------------------------------
294		Name: Lal George
295		Date: 2001/05/08  17:26:09 EDT
296		Tag: george-20010508-omit-frameptr
297		Description:
298
299		Changes to implement the omit-frame-pointer optimization to support
300		raw C calls. For now, there is only support on the Intel x86, but
301		other architectures will follow as more experience is gained with this.
302
303
304		----------------------------------------------------------------------
305		Name: Matthias Blume
306		Date: 2001/05/07 14:40:00 EDT
307		Tag: blume-20010507-proxies
308		Description:
309
310		I made into "proxy libraries" all libraries that qualify for such a
311		change.  (A qualifying library is a library that has another library or
312		groups as its sole member and repeats that member's export list
313		verbatim.  A proxy library avoids this repetition by omitting its export
314		list, effectively inheriting the list that its (only) member exports.
315		See the CM manual for more explanation.)
316		The main effect is that explicit export lists for these libraries
317		do not have to be kepts in sync, making maintenance a bit easier.
318
319		I also added copyright notices to many .cm-files.
320
321		Last but not least, I made a new set of bootfiles.
322
323		----------------------------------------------------------------------
324		Name: Matthias Blume
325		Date: 2001/05/04 17:00:00 EDT
326		Tag: blume-20010504-cm-lsplit
327		Description:
328
329		0. John merged pending changes to $/smlnj-lib.cm
330
331		1. Allen's previous change accidentally backed out of one of Lal's
332		earlier changes.  I undid this mistake (re-introducing Lal's change).
333
334		2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
335		within the compiler where applicable.  There is some code simplification
336		because of that.
337
338		3. The "split" phase (in FLINT) is now part of the default list of phases.
339		Compiler.Control.LambdaSplitting.* can be used to globally control the
340		lambda-splitting (cross-module-inlining) engine.  In addition to that,
341		it can now also be controlled on a per-source basis: CM has been taught
342		a new tool parameter applicable to ML source files.
343
344		- To turn lambda-splitting off completely:
345		local open Compiler.Control.LambdaSplitting in
346		val _ = set Off
347		end
348		- To make "no lambda-splitting" the global default (but allow per-source
349		overriding); this is the initial setting:
350		local open Compiler.Control.LambdaSplitting in
351		val _ = set (Default NONE)
352		end
353		- To make "lambda-splitting with aggressiveness a" the global default
354		(and allow per-source overriding):
355		local open Compiler.Control.LambdaSplitting in
356		val _ = set (Default (SOME a))
357		end
358
359		- To turn lambda-splitting off for a given ML souce file (say: a.sml)
360		write (in the respective .cm-file):
361		a.sml (lambdasplitting:off)
362		- To turn lambda-splitting for a.sml on with minimal aggressiveness:
363		a.sml (lambdasplitting:on)
364		- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
365		<a> is a decimal non-negative integer):
366		a.sml (lambdasplitting:<a>)
367		- To turn lambda-splitting for a.sml on with maximal aggressiveness:
368		a.sml (lambdasplitting:infinity)
369		- To use the global default for a.sml:
370		a.sml (lambdasplitting:default)
371		or simply
372		a.sml
373
374		----------------------------------------------------------------------
375		Name: Allen Leung
376		Date: 2001/05/04 01:57:00 EDT
377		Tag: leunga-20010504-sync
378		Description:
379
380		MLRISC features.
381
382		1. Fix to CMPXCHG instructions.
383		2. Changed RA interface to allow annotations in callbacks.
384		3. Added a new method to the stream interface to allow annotations updates.
385
386		----------------------------------------------------------------------
387		Name: Matthias Blume
388		Date: 2001/05/01 11:45:00 EDT
389		Tag: blume-20010501-pcedittmp
390		Description:
391
392		Changed install.sh to use the current working directory instead of
393		/usr/tmp for a temporary file (pcedittmp).  The previous choice
394		of /usr/tmp caused trouble with MacOS X because of file premission
395		problems.
396
397		----------------------------------------------------------------------
398		Name: Matthias Blume
399		Date: 2001/04/20 11:10:00 EDT
400		Tag: blume-20010420-inMLflag
401		Description:
402
403		- added vp_limitPtrMask to vproc-state.h
404		(for use by the raw-C-calls mechanism to implement proper interrupt
405		handling)
406		- made the ML compiler aware of various data-structure offsets so it
407		can generate code for accessing the vp_inML flag and vp_limitPtrMask
408		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
409		raw C-calls
410
411		----------------------------------------------------------------------
412		Name: Lal George
413		Date: 2001/04/20 09:15:28 EDT
414		Tag: george-20010420-macosX
415		Description:
416
417		- Changes to port to Mac OS X; Darwin.
418
419		- In the process I found that sqrt was broken on the PPC, because the
420		fsqrt instruction is not implemented.
421
422		----------------------------------------------------------------------
423		Name: Matthias Blume
424		Date: 2001/04/18 12:45:00 EDT
425		Tag: blume-20010418-ccalls
426		Description:
427
428		- fixed two off-by-4 errors in the x86-specific c-calls implementation
429		(this bug prevented structure arguments containing pointers from being
430		passed correctly)
431		- changed the raw-C-call code in mlriscGen.sml in such a way that
432		structure arguments are represented as a pointer to the beginning
433		of the structure (instead of having a series of synthesized arguments,
434		one for each structure member)
435
436		- made makeml script's verbosity level configurable via environment
437		variable (MAKEML_VERBOSITY)
438
439		- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
440		in rawmem-x86.sml; we are now using the real thing
441
442		----------------------------------------------------------------------
443		Name: Matthias Blume
444		Date: 2001/03/22 16:25:00 EST
445		Tag: blume-20010322-bootfiles
446		Description:
447
448		Created a new set of bootfiles (for your automatic installation convenience).
449
450		----------------------------------------------------------------------
451		Name: Matthias Blume
452		Date: 2001/03/22 15:10:00 EST
453		Tag: blume-20010322-rawmem-parcm
454		Description:
455
456		1. All "raw memory access" primitives for the new FFI are implemented now
457		(at least on the x86).
458		2. Some further cleanup of CM's parallel make mechanism.
459
460		----------------------------------------------------------------------
461		Name: Matthias Blume
462		Date: 2001/03/19 17:53:00 EST
463		Tag: blume-20010319-parallel
464		Description:
465
466		Parallel make (using compile servers) now works again.
467
468		To this end, CM.stabilize and CMB.make have been modified to work in
469		two passes when compile servers are attached:
470		1. Compile everything, do not perform stabilization; this pass
471		uses compile servers
472		2. Stabilize everything; this pass does not use compile servers
473		If there are no compile servers, the two passes are combined into one
474		(as before).  Splitting the passes increases the inherent parallelism
475		in the dependency graph because the entire graph including all
476		libraries is available at the same time.  This, in turn, improves
477		server utilization.  The downside is that the master process will
478		have to do some extra work after compilation is done (because for
479		technical reasons it must re-read all the binfiles during stabilization).
480
481		----------------------------------------------------------------------
482		Name: Matthias Blume
483		Date: 2001/03/16 12:22:00 EST
484		Tag: blume-20010316-bootfiles
485		Description:
486
487		Created a new set of bootfiles (for your automatic installation convenience).
488
489		----------------------------------------------------------------------
490		Name: Matthias Blume
491		Date: 2001/03/16 11:00:00 EST
492		Tag: blume-20010316-MLTREE-fixup
493		Description:
494
495		This is a minor fixup for an (untagged) earlier commit by Allen.
496		(A file was missing).
497
498		----------------------------------------------------------------------
499		Name: Allen Leung
500		Date: Mon Mar  5 18:54:57 EST 2001
501		Tag: leunga-20010305-cut-support
502
503		1. New support for alternative control-flow in MLTREE.
504		Currently we support
505
506		FLOW_TO(CALL ...., [k1,...,kn])
507
508		This is needed for 'cuts to' in C-- and try/handle-like constructs
509		in Moby
510
511		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
512
513		2. Register Allocator
514
515		Changes in interface [from Fermin, John]
516
517		3. Alpha 8-bit SLL support [Fermin]
518
519		4. All architectures
520
521		A new module (ClusterExpandCopies) for expanding parallel copies.
522
523		----------------------------------------------------------------------
524		Name: Allen Leung
525		Date: 2001/02/27 23:07:00 EST
526		Tag: leunga-20010227-minor-stuff
527
528		1. Alpha bug fix for CMOVNE
529		2. Handle mltree COND(..,FCMP ...,...)
530		3. Bug fix in simplifier
531
532		----------------------------------------------------------------------
533		Name: Matthias Blume
534		Date: 2001/01/30 17:50:00 EST
535		Tag: blume-20010130-sync
536		Description:
537
538		This is just a minor update to sync my devel branch with the main brach.
539		The only visible change is the addition of some README files.
540
541		----------------------------------------------------------------------
542		Name: Matthias Blume
543		Date: 2001/01/12 23:30:00 JST
544		Tag: blume-20010112-bootfiles
545		Description:
546
547		Made a new set of bootfiles that goes with the current state of the
548		repository.
549
550		----------------------------------------------------------------------
551		Name: Matthias Blume
552		Date: 2001/01/12 21:20:00 JST
553		Tag: blume-20010112-sync
554		Description:
555
556		I am just flushing out some minor changes that had accumulated in
557		my private branch in order to sync with the main tree.  (This is
558		mainly because I had CVS trouble when trying to merge _into_ my
559		private branch.)
560
561		Most people should be completely unaffected by this.
562
563		----------------------------------------------------------------------
564		Name: Allen Leung
565		Date: Thu Jan 11 21:03:00 EST 2001
566		Tag: leunga-20010111-labexp=mltree
567		Description:
568
569		1.  Removed the type LabelExp and replace it by MLTree.
570		2.  Rewritten mltree-simplify with the pattern matcher tool.
571		3.  There were some bugs in alpha code generator which would break
572		64-bit code generation.
573		4.  Redo the tools to generate code with the
574		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
575		are now generated by perl scripts.
576
577		----------------------------------------------------------------------
578		Name: Matthias Blume
579		Date: 2001/01/10 21:55:00 JST
580		Tag: blume-20010110-rcc
581		Description:
582
583		The RCC stuff now seems to work (but only on the x86).
584		This required hacking of the c-calls interface (and -implementation) in
585		MLRISC.
586
587		Normal compiler users should be unaffected.
588
589		----------------------------------------------------------------------
590		Name: Matthias Blume
591		Date: 2001/01/09 01:20:00 JST
592		Tag: blume-20010109-rcc
593		Description:
594
595		This is a fairly big patch, flushing out a large number of pending
596		changes that I made to my development copy over the last couple of days.
597
598		Of practical relevance at this moment is a workaround for a pickling
599		bug that Allen ran into the other day.  The cause of the bug itself is
600		still unknown and it might be hard to fix it properly, but the
601		workaround has some merits of its own (namely somewhat reducing pickling
602		overhead for certain libraries).  Therefore, I think this solution should
603		be satisfactory at this time.
604
605		The rest of the changes (i.e., the vast majority) has to do with my
606		ongoing efforts of providing direct support for C function calls from
607		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
608		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
609		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
610		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
611		handle RCC.
612
613		The last part (the changes to mlriscGen.sml) are still known to be
614		wrong on the x86 and not implemented on all other architectures.  But
615		the infrastructure is in place. I had to change a few functor
616		signatures in the backend to be able to route the CCalls interface
617		from MLRISC there, and I had to specialize the mltree type (on the
618		x86) to include the necessary extensions. (The extensions themselves
619		were already there and redy to go in MLRISC/x86).
620
621		Everything should be very happy as soon as someone helps me with
622		mlriscGen.sml...
623
624		In any case, nothing of this should matter to anyone as long as the
625		new primop is not being used (which is going to be the case unless you
626		find it where I hid it :). The rest of the compiler is completely
627		unaffected.
628
629		----------------------------------------------------------------------
630		Name: Matthias Blume
631		Date: 2001/01/05 00:30:00 JST
632		Tag: blume-20010105-primops
633		Description:
634
635		Added some experimental support for work that I am doing right now.
636		These changes mostly concern added primops, but there is also a new
637		experimental C library in the runtime system (but currently not enabled
638		anywhere except on Linux/X86).
639
640		In the course of adding primops (and playing with them), I discovered that
641		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
642		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
643		To recover, I made type information in INL_PRIM mandatory and changed
644		prim.sml as well as built-in.sml accordingly.  The InLine structure now
645		has complete, correct type information (i.e., no bottom types).
646
647		Since all these changes mean that we need new binfiles, I also bumped the
648		version number to 110.32.1.
649
650		----------------------------------------------------------------------
651		Name: Matthias Blume
652		Date: 2000/12/30 22:10:00 JST
653		Tag: blume-20001230-various
654		Description:
655
656		Added proxy libraries for MLRISC and let MLRISC libraries refer
657		to each other using path anchors.  (See CM manual for explanation.)
658
659		Updated CM documentation.
660
661		Fixed some bugs in CM.
662
663		Implemented "proxy" libraries (= syntactic sugar for CM).
664
665		Added "-quiet" option to makeml and changed runtime system accordingly.
666
667		Added cleanup handler for exportML to reset timers and compiler stats.
668
669		----------------------------------------------------------------------
670		Name: Lal George
671		Date: 2000/12/22 22:22:58 EST 2000
672		Tag: Release_110_32
673		Description:
674
675		Infinite precision used throughout MLRISC.
676		see MLRISC/mltree/machine-int.sig
677
678		----------------------------------------------------------------------
679		Name: Matthias Blume
680		Date: 2000/12/22 23:16:00 JST
681		Tag: blume-20001222-warn
682		Description:
683
684		Corrected wording and formatting of some CM warning message which I
685		broke in my previous patch.
686
687		----------------------------------------------------------------------
688		Name: Matthias Blume
689		Date: 2000/12/22 21:20:00 JST
690		Tag: blume-20001222-anchorenv
691		Description:
692
693		Fixed CM's handling of anchor environments in connection with CMB.make.
694
695		----------------------------------------------------------------------
696		Name: Matthias Blume
697		Date: 2000/12/22 13:15:00 JST
698		Tag: blume-20001222-cleanup
699		Description:
700
701		Removed src/cm/ffi which does not (and did not) belong here.
702
703		----------------------------------------------------------------------
704		Name: Matthias Blume
705		Date: 2000/12/21 23:55:00 JST
706		Tag: blume-20001221-exn
707		Description:
708
709		Probably most important: CM no longer silently swallows all exceptions
710		in the compiler.
711		Plus: some other minor CM changes.  For example, CM now reports some
712		sizes for generated binfiles (code, data, envpickle, lambdapickle).
713
714		----------------------------------------------------------------------
715		Name: Matthias Blume
716		Date: 2000/12/15 00:01:05 JST
717		Tag: blume-20001215-dirtool
718		Description:
719
720		- "dir" tool added.
721		- improvements and cleanup to Tools structure
722		- documentation updates
723
724		----------------------------------------------------------------------
725		Name: Allen Leung
726		Date: Thu Dec 14 03:45:24 EST 2000
727		Description:
728		Tag:  leunga-20001214-int-inf
729		Description:
730
731		In IntInf, added these standard functions, which are missing from our
732		implementation:
733
734		andb : int * int -> int
735		xorb : int * int -> int
736		orb  : int * int -> int
737		notb : int -> int
738		<<   : int * word -> int
739		~>>  : int * word -> int
740
741		Not tested, I hope they are correct.
742
743		----------------------------------------------------------------------
744		Name: Allen Leung
745		Date: Fri Dec  8 19:23:26 EST 2000
746		Description:
747		Tag:  leunga-20001208-nowhere
748		Description:
749
750		Slight improvements to the 'nowhere' tool to handle OR-patterns,
751		to generate better error messages etc.  Plus a brief manual.
752
753		----------------------------------------------------------------------
754		Name: Lal George
755		Date: 2000/12/08 09:54:02 EST 2000
756		Tag: Release_110_31
757		Description:
758
759		- Version 110.31
760		----------------------------------------------------------------------
761		Name: Allen Leung
762		Date: Thu Dec  7 22:01:04 EST 2000
763		Tag:  leunga-20001207-cell-monster-hack
764		Description:
765
766		Major MLRISC internal changes.  Affect all clients.
767		Summary:
768
769		1.  Type CELLS.cell = int is now replaced by a datatype.
770		As a result, the old regmap is now gone.  Almost all interfaces
771		in MLRISC change as a consequence.
772
773		2.  A new brand version of machine description tool (v3.0) that generates
774		modules expecting the new interface.  The old version is removed.
775
776		3.  The RA interface has been further abstracted into two new functors.
777		RISC_RA and X86RA.  These functors have much simpler interfaces.
778		[See also directory MLRISC/demo.]
779
780		4.  Some other new source->source code generation tools are available:
781
782		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
783		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
784		I use this tool to generate the peephole optimizers---with the new
785		cell type changes, peephole rules are becoming difficult to write
786		without conditional pattern matching.
787
788		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
789		cover the entire MLRISC source tree so many things broke.
790
791		6.  CM files have been moved to the subdirectory MLRISC/cm.
792		They are moved because there are a lot of them and they clutter up the
793		root dir.
794
795		7.  More detailed documentation to come...
796
797		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
798		a makeml -rebuild first.  This is because of other other
799		changes that Matthias has made (see below).
800
801
802		----------------------------------------------------------------------
803		Name: Matthias Blume
804		Date: 2000/11/30 23:12:00 JST
805		Tag: blume-20001130-filereorg
806		Description:
807
808		Some manual updates and some file reorganizations in CM.
809
810		----------------------------------------------------------------------
811		Name: Matthias Blume
812		Date: 2000/11/24 17:45:00 JST
813		Tag: blume-20001124-link
814		Description:
815
816		Drastically improved link traversal code for the case that the dynamic
817		value was already loaded at bootstrap time.  As a result, CM and CMB
818		now both load blazingly fast -- even on a very slow machine.  Also,
819		memory consumption has been further reduced by this.
820
821		Warning: The format of the PIDMAP file has changed.  THerefore, to
822		bootstrap you have to do this:
823
824		1. Run CMB.make
825		2. Make a symbolic link for the boot directory:
826		ln -s sml.boot.ARCH-OS xxx
827		3. "Rebuild" the boot directory:
828		./makeml -boot xxx -rebuild sml ; rm xxx
829		4. Boot normally:
830		./makeml
831
832		----------------------------------------------------------------------
833		Name: Matthias Blume
834		Date: 2000/11/21 21:20:00 JST
835		Tag: blume-20001121-tools
836		Description:
837
838		Continued hacking on autoloading problem -- with success this time.
839		Also changed tool-plugin mechanism.  See new CM manual.
840
841		----------------------------------------------------------------------
842		Name: Matthias Blume
843		Date: 2000/11/19 14:30:00 JST
844		Tag:  blume-20001119-autoload
845		Description:
846
847		Some hacking to make autoloading faster.  Success for CMB, no success
848		so far for CM.  There is a reduced structure CM' that autoloads faster.
849		(This is a temporary, non-documented hack to be eliminated again when
850		the general problem is solved.)
851
852		----------------------------------------------------------------------
853		Name: Matthias Blume
854		Date: 2000/11/17 14:10:00 JST
855		Tag: blume-20001117-pickle-lib
856		Description:
857
858		1. Eliminated comp-lib.cm
859		2. Made pickle-lib.cm
860		3. Eliminated all uses of intset.sml (from comp-lib.cm)
861		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
862		equivalent constructs from smlnj-lib.cm (INtHashTable).
863		5. Point 4. also goes for those uses of intmap.* in MLRISC.
864		Duplicated intmap modules thrown out.
865		6. Hunted down all duplicated SCC code and replaced it with
866		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
867		7. Rewrote Feedback module.
868		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
869		should be thrown out and equivalent modules from smlnj-lib.cm
870		should be used (IntRedBlackSet, IntListSet, ...).
871
872		Confirmed that compiler compiles to fixpoint.
873
874		----------------------------------------------------------------------
875		Name: Allen Leung
876		Date: 2000/11/10 18:00:00
877		Tag: leunga-20001110-new-x86-fp
878
879		A new x86 floating point code generator has been added.
880		By default this is turned off.  To turn this on, do:
881
882		CM.autoload "$smlnj/compiler.cm";
883		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
884
885		Changes:
886
887		1.  Changed FTAN to FPTAN so that the assembly output is correct.
888		2.  Changed the extension callback for FTANGENT to generate:
889
890		fptan
891		fstp  %st(0)
892		instead of
893		fptan
894		fstpl ftempmem
895
896		3.  Numerous assembly fixes for x86.
897
898		5.  Cleaned up the machine code output module x86/x86MC.sml and added
899		support for a whole bunch of instructions and addressing modes:
900
901		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
902		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
903		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
904		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
905		fxch %st(n)
906		fld %st(n)
907		fst %st(n)
908		fst mem
909		fstp %st(n)
910		fucom %st(n)
911		fucomp %st(n)
912
913		All these are now generated when the fast fp mode is turned on.
914
915		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
916
917		----------------------------------------------------------------------
918		Name: Matthias Blume
919		Date: 2000/11/09 11:20:00 JST
920		Tag: blume-20001109-scc
921		Description:
922
923		Eliminated some code duplication:
924
925		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
926		(Otherwise the functor is useless.)
927		2. Used GraphSCCFn where SCCUtilFun was used previously.
928		3. Got rid of SCCUtilFun (in comp-lib.cm).
929
930		----------------------------------------------------------------------
931		Name: Lal George
932		Date: 2000/11/06 09:02:21 EST 2000
933		Tag: Release_110_30
934		Description:
935
936		- Version 110.30
937		----------------------------------------------------------------------
938		Name: Matthias Blume
939		Date: 2000/11/04 14:45:00
940		Tag: blume-20001104-mlbuild
941		Description:
942
943		- Made ml-build faster on startup.
944		- Documentation fixes.
945
946		----------------------------------------------------------------------
947		Name: Matthias Blume
948		Date: 2000/11/02 17:00:00 JST
949		Tag: blume-20001102-condcomp
950		Description:
951
952		- Small tweaks to pickler -- new BOOTFILES!
953		- Version bumped to 110.29.2.
954		- Added conditional compilation facility to init.cmi (see comment there).
955		----------------------------------------------------------------------
956	Name: Allen Leung	Name: Allen Leung
957	Date: 2000/10/23 19:31:00	Date: 2000/10/23 19:31:00
958	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.844

---

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