Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

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

---

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