Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 705, Wed Sep 27 18:44:44 2000 UTC	revision 812, Fri Apr 20 15:07:30 2001 UTC
#	Line 13	Line 13
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2001/04/20 11:10:00 EDT
18		Tag: blume-20010420-inMLflag
19		Description:
20
21		- added vp_limitPtrMask to vproc-state.h
22		(for use by the raw-C-calls mechanism to implement proper interrupt
23		handling)
24		- made the ML compiler aware of various data-structure offsets so it
25		can generate code for accessing the vp_inML flag and vp_limitPtrMask
26		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
27		raw C-calls
28
29		----------------------------------------------------------------------
30		Name: Lal George
31		Date: 2001/04/20 09:15:28 EDT
32		Tag: george-20010420-macosX
33		Description:
34
35		- Changes to port to Mac OS X; Darwin.
36
37		- In the process I found that sqrt was broken on the PPC, because the
38		fsqrt instruction is not implemented.
39
40		----------------------------------------------------------------------
41		Name: Matthias Blume
42		Date: 2001/04/18 12:45:00 EDT
43		Tag: blume-20010418-ccalls
44		Description:
45
46		- fixed two off-by-4 errors in the x86-specific c-calls implementation
47		(this bug prevented structure arguments containing pointers from being
48		passed correctly)
49		- changed the raw-C-call code in mlriscGen.sml in such a way that
50		structure arguments are represented as a pointer to the beginning
51		of the structure (instead of having a series of synthesized arguments,
52		one for each structure member)
53
54		- made makeml script's verbosity level configurable via environment
55		variable (MAKEML_VERBOSITY)
56
57		- eliminteded placeholder implementations for f32l, w16s, i16s, and f32s
58		in rawmem-x86.sml; we are now using the real thing
59
60		----------------------------------------------------------------------
61		Name: Matthias Blume
62		Date: 2001/03/22 16:25:00 EST
63		Tag: blume-20010322-bootfiles
64		Description:
65
66		Created a new set of bootfiles (for your automatic installation convenience).
67
68		----------------------------------------------------------------------
69		Name: Matthias Blume
70		Date: 2001/03/22 15:10:00 EST
71		Tag: blume-20010322-rawmem-parcm
72		Description:
73
74		1. All "raw memory access" primitives for the new FFI are implemented now
75		(at least on the x86).
76		2. Some further cleanup of CM's parallel make mechanism.
77
78		----------------------------------------------------------------------
79		Name: Matthias Blume
80		Date: 2001/03/19 17:53:00 EST
81		Tag: blume-20010319-parallel
82		Description:
83
84		Parallel make (using compile servers) now works again.
85
86		To this end, CM.stabilize and CMB.make have been modified to work in
87		two passes when compile servers are attached:
88		1. Compile everything, do not perform stabilization; this pass
89		uses compile servers
90		2. Stabilize everything; this pass does not use compile servers
91		If there are no compile servers, the two passes are combined into one
92		(as before).  Splitting the passes increases the inherent parallelism
93		in the dependency graph because the entire graph including all
94		libraries is available at the same time.  This, in turn, improves
95		server utilization.  The downside is that the master process will
96		have to do some extra work after compilation is done (because for
97		technical reasons it must re-read all the binfiles during stabilization).
98
99		----------------------------------------------------------------------
100		Name: Matthias Blume
101		Date: 2001/03/16 12:22:00 EST
102		Tag: blume-20010316-bootfiles
103		Description:
104
105		Created a new set of bootfiles (for your automatic installation convenience).
106
107		----------------------------------------------------------------------
108		Name: Matthias Blume
109		Date: 2001/03/16 11:00:00 EST
110		Tag: blume-20010316-MLTREE-fixup
111		Description:
112
113		This is a minor fixup for an (untagged) earlier commit by Allen.
114		(A file was missing).
115
116		----------------------------------------------------------------------
117		Name: Allen Leung
118		Date: Mon Mar  5 18:54:57 EST 2001
119		Tag: leunga-20010305-cut-support
120
121		1. New support for alternative control-flow in MLTREE.
122		Currently we support
123
124		FLOW_TO(CALL ...., [k1,...,kn])
125
126		This is needed for 'cuts to' in C-- and try/handle-like constructs
127		in Moby
128
129		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
130
131		2. Register Allocator
132
133		Changes in interface [from Fermin, John]
134
135		3. Alpha 8-bit SLL support [Fermin]
136
137		4. All architectures
138
139		A new module (ClusterExpandCopies) for expanding parallel copies.
140
141		----------------------------------------------------------------------
142		Name: Allen Leung
143		Date: 2001/02/27 23:07:00 EST
144		Tag: leunga-20010227-minor-stuff
145
146		1. Alpha bug fix for CMOVNE
147		2. Handle mltree COND(..,FCMP ...,...)
148		3. Bug fix in simplifier
149
150		----------------------------------------------------------------------
151		Name: Matthias Blume
152		Date: 2001/01/30 17:50:00 EST
153		Tag: blume-20010130-sync
154		Description:
155
156		This is just a minor update to sync my devel branch with the main brach.
157		The only visible change is the addition of some README files.
158
159		----------------------------------------------------------------------
160		Name: Matthias Blume
161		Date: 2001/01/12 23:30:00 JST
162		Tag: blume-20010112-bootfiles
163		Description:
164
165		Made a new set of bootfiles that goes with the current state of the
166		repository.
167
168		----------------------------------------------------------------------
169		Name: Matthias Blume
170		Date: 2001/01/12 21:20:00 JST
171		Tag: blume-20010112-sync
172		Description:
173
174		I am just flushing out some minor changes that had accumulated in
175		my private branch in order to sync with the main tree.  (This is
176		mainly because I had CVS trouble when trying to merge _into_ my
177		private branch.)
178
179		Most people should be completely unaffected by this.
180
181		----------------------------------------------------------------------
182		Name: Allen Leung
183		Date: Thu Jan 11 21:03:00 EST 2001
184		Tag: leunga-20010111-labexp=mltree
185		Description:
186
187		1.  Removed the type LabelExp and replace it by MLTree.
188		2.  Rewritten mltree-simplify with the pattern matcher tool.
189		3.  There were some bugs in alpha code generator which would break
190		64-bit code generation.
191		4.  Redo the tools to generate code with the
192		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
193		are now generated by perl scripts.
194
195		----------------------------------------------------------------------
196		Name: Matthias Blume
197		Date: 2001/01/10 21:55:00 JST
198		Tag: blume-20010110-rcc
199		Description:
200
201		The RCC stuff now seems to work (but only on the x86).
202		This required hacking of the c-calls interface (and -implementation) in
203		MLRISC.
204
205		Normal compiler users should be unaffected.
206
207		----------------------------------------------------------------------
208		Name: Matthias Blume
209		Date: 2001/01/09 01:20:00 JST
210		Tag: blume-20010109-rcc
211		Description:
212
213		This is a fairly big patch, flushing out a large number of pending
214		changes that I made to my development copy over the last couple of days.
215
216		Of practical relevance at this moment is a workaround for a pickling
217		bug that Allen ran into the other day.  The cause of the bug itself is
218		still unknown and it might be hard to fix it properly, but the
219		workaround has some merits of its own (namely somewhat reducing pickling
220		overhead for certain libraries).  Therefore, I think this solution should
221		be satisfactory at this time.
222
223		The rest of the changes (i.e., the vast majority) has to do with my
224		ongoing efforts of providing direct support for C function calls from
225		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
226		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
227		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
228		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
229		handle RCC.
230
231		The last part (the changes to mlriscGen.sml) are still known to be
232		wrong on the x86 and not implemented on all other architectures.  But
233		the infrastructure is in place. I had to change a few functor
234		signatures in the backend to be able to route the CCalls interface
235		from MLRISC there, and I had to specialize the mltree type (on the
236		x86) to include the necessary extensions. (The extensions themselves
237		were already there and redy to go in MLRISC/x86).
238
239		Everything should be very happy as soon as someone helps me with
240		mlriscGen.sml...
241
242		In any case, nothing of this should matter to anyone as long as the
243		new primop is not being used (which is going to be the case unless you
244		find it where I hid it :). The rest of the compiler is completely
245		unaffected.
246
247		----------------------------------------------------------------------
248		Name: Matthias Blume
249		Date: 2001/01/05 00:30:00 JST
250		Tag: blume-20010105-primops
251		Description:
252
253		Added some experimental support for work that I am doing right now.
254		These changes mostly concern added primops, but there is also a new
255		experimental C library in the runtime system (but currently not enabled
256		anywhere except on Linux/X86).
257
258		In the course of adding primops (and playing with them), I discovered that
259		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
260		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
261		To recover, I made type information in INL_PRIM mandatory and changed
262		prim.sml as well as built-in.sml accordingly.  The InLine structure now
263		has complete, correct type information (i.e., no bottom types).
264
265		Since all these changes mean that we need new binfiles, I also bumped the
266		version number to 110.32.1.
267
268		----------------------------------------------------------------------
269		Name: Matthias Blume
270		Date: 2000/12/30 22:10:00 JST
271		Tag: blume-20001230-various
272		Description:
273
274		Added proxy libraries for MLRISC and let MLRISC libraries refer
275		to each other using path anchors.  (See CM manual for explanation.)
276
277		Updated CM documentation.
278
279		Fixed some bugs in CM.
280
281		Implemented "proxy" libraries (= syntactic sugar for CM).
282
283		Added "-quiet" option to makeml and changed runtime system accordingly.
284
285		Added cleanup handler for exportML to reset timers and compiler stats.
286
287		----------------------------------------------------------------------
288		Name: Lal George
289		Date: 2000/12/22 22:22:58 EST 2000
290		Tag: Release_110_32
291		Description:
292
293		Infinite precision used throughout MLRISC.
294		see MLRISC/mltree/machine-int.sig
295
296		----------------------------------------------------------------------
297		Name: Matthias Blume
298		Date: 2000/12/22 23:16:00 JST
299		Tag: blume-20001222-warn
300		Description:
301
302		Corrected wording and formatting of some CM warning message which I
303		broke in my previous patch.
304
305		----------------------------------------------------------------------
306		Name: Matthias Blume
307		Date: 2000/12/22 21:20:00 JST
308		Tag: blume-20001222-anchorenv
309		Description:
310
311		Fixed CM's handling of anchor environments in connection with CMB.make.
312
313		----------------------------------------------------------------------
314		Name: Matthias Blume
315		Date: 2000/12/22 13:15:00 JST
316		Tag: blume-20001222-cleanup
317		Description:
318
319		Removed src/cm/ffi which does not (and did not) belong here.
320
321		----------------------------------------------------------------------
322		Name: Matthias Blume
323		Date: 2000/12/21 23:55:00 JST
324		Tag: blume-20001221-exn
325		Description:
326
327		Probably most important: CM no longer silently swallows all exceptions
328		in the compiler.
329		Plus: some other minor CM changes.  For example, CM now reports some
330		sizes for generated binfiles (code, data, envpickle, lambdapickle).
331
332		----------------------------------------------------------------------
333		Name: Matthias Blume
334		Date: 2000/12/15 00:01:05 JST
335		Tag: blume-20001215-dirtool
336		Description:
337
338		- "dir" tool added.
339		- improvements and cleanup to Tools structure
340		- documentation updates
341
342		----------------------------------------------------------------------
343		Name: Allen Leung
344		Date: Thu Dec 14 03:45:24 EST 2000
345		Description:
346		Tag:  leunga-20001214-int-inf
347		Description:
348
349		In IntInf, added these standard functions, which are missing from our
350		implementation:
351
352		andb : int * int -> int
353		xorb : int * int -> int
354		orb  : int * int -> int
355		notb : int -> int
356		<<   : int * word -> int
357		~>>  : int * word -> int
358
359		Not tested, I hope they are correct.
360
361		----------------------------------------------------------------------
362		Name: Allen Leung
363		Date: Fri Dec  8 19:23:26 EST 2000
364		Description:
365		Tag:  leunga-20001208-nowhere
366		Description:
367
368		Slight improvements to the 'nowhere' tool to handle OR-patterns,
369		to generate better error messages etc.  Plus a brief manual.
370
371		----------------------------------------------------------------------
372		Name: Lal George
373		Date: 2000/12/08 09:54:02 EST 2000
374		Tag: Release_110_31
375		Description:
376
377		- Version 110.31
378		----------------------------------------------------------------------
379		Name: Allen Leung
380		Date: Thu Dec  7 22:01:04 EST 2000
381		Tag:  leunga-20001207-cell-monster-hack
382		Description:
383
384		Major MLRISC internal changes.  Affect all clients.
385		Summary:
386
387		1.  Type CELLS.cell = int is now replaced by a datatype.
388		As a result, the old regmap is now gone.  Almost all interfaces
389		in MLRISC change as a consequence.
390
391		2.  A new brand version of machine description tool (v3.0) that generates
392		modules expecting the new interface.  The old version is removed.
393
394		3.  The RA interface has been further abstracted into two new functors.
395		RISC_RA and X86RA.  These functors have much simpler interfaces.
396		[See also directory MLRISC/demo.]
397
398		4.  Some other new source->source code generation tools are available:
399
400		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
401		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
402		I use this tool to generate the peephole optimizers---with the new
403		cell type changes, peephole rules are becoming difficult to write
404		without conditional pattern matching.
405
406		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
407		cover the entire MLRISC source tree so many things broke.
408
409		6.  CM files have been moved to the subdirectory MLRISC/cm.
410		They are moved because there are a lot of them and they clutter up the
411		root dir.
412
413		7.  More detailed documentation to come...
414
415		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
416		a makeml -rebuild first.  This is because of other other
417		changes that Matthias has made (see below).
418
419
420		----------------------------------------------------------------------
421		Name: Matthias Blume
422		Date: 2000/11/30 23:12:00 JST
423		Tag: blume-20001130-filereorg
424		Description:
425
426		Some manual updates and some file reorganizations in CM.
427
428		----------------------------------------------------------------------
429		Name: Matthias Blume
430		Date: 2000/11/24 17:45:00 JST
431		Tag: blume-20001124-link
432		Description:
433
434		Drastically improved link traversal code for the case that the dynamic
435		value was already loaded at bootstrap time.  As a result, CM and CMB
436		now both load blazingly fast -- even on a very slow machine.  Also,
437		memory consumption has been further reduced by this.
438
439		Warning: The format of the PIDMAP file has changed.  THerefore, to
440		bootstrap you have to do this:
441
442		1. Run CMB.make
443		2. Make a symbolic link for the boot directory:
444		ln -s sml.boot.ARCH-OS xxx
445		3. "Rebuild" the boot directory:
446		./makeml -boot xxx -rebuild sml ; rm xxx
447		4. Boot normally:
448		./makeml
449
450		----------------------------------------------------------------------
451		Name: Matthias Blume
452		Date: 2000/11/21 21:20:00 JST
453		Tag: blume-20001121-tools
454		Description:
455
456		Continued hacking on autoloading problem -- with success this time.
457		Also changed tool-plugin mechanism.  See new CM manual.
458
459		----------------------------------------------------------------------
460		Name: Matthias Blume
461		Date: 2000/11/19 14:30:00 JST
462		Tag:  blume-20001119-autoload
463		Description:
464
465		Some hacking to make autoloading faster.  Success for CMB, no success
466		so far for CM.  There is a reduced structure CM' that autoloads faster.
467		(This is a temporary, non-documented hack to be eliminated again when
468		the general problem is solved.)
469
470		----------------------------------------------------------------------
471		Name: Matthias Blume
472		Date: 2000/11/17 14:10:00 JST
473		Tag: blume-20001117-pickle-lib
474		Description:
475
476		1. Eliminated comp-lib.cm
477		2. Made pickle-lib.cm
478		3. Eliminated all uses of intset.sml (from comp-lib.cm)
479		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
480		equivalent constructs from smlnj-lib.cm (INtHashTable).
481		5. Point 4. also goes for those uses of intmap.* in MLRISC.
482		Duplicated intmap modules thrown out.
483		6. Hunted down all duplicated SCC code and replaced it with
484		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
485		7. Rewrote Feedback module.
486		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
487		should be thrown out and equivalent modules from smlnj-lib.cm
488		should be used (IntRedBlackSet, IntListSet, ...).
489
490		Confirmed that compiler compiles to fixpoint.
491
492		----------------------------------------------------------------------
493		Name: Allen Leung
494		Date: 2000/11/10 18:00:00
495		Tag: leunga-20001110-new-x86-fp
496
497		A new x86 floating point code generator has been added.
498		By default this is turned off.  To turn this on, do:
499
500		CM.autoload "$smlnj/compiler.cm";
501		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
502
503		Changes:
504
505		1.  Changed FTAN to FPTAN so that the assembly output is correct.
506		2.  Changed the extension callback for FTANGENT to generate:
507
508		fptan
509		fstp  %st(0)
510		instead of
511		fptan
512		fstpl ftempmem
513
514		3.  Numerous assembly fixes for x86.
515
516		5.  Cleaned up the machine code output module x86/x86MC.sml and added
517		support for a whole bunch of instructions and addressing modes:
518
519		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
520		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
521		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
522		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
523		fxch %st(n)
524		fld %st(n)
525		fst %st(n)
526		fst mem
527		fstp %st(n)
528		fucom %st(n)
529		fucomp %st(n)
530
531		All these are now generated when the fast fp mode is turned on.
532
533		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
534
535		----------------------------------------------------------------------
536		Name: Matthias Blume
537		Date: 2000/11/09 11:20:00 JST
538		Tag: blume-20001109-scc
539		Description:
540
541		Eliminated some code duplication:
542
543		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
544		(Otherwise the functor is useless.)
545		2. Used GraphSCCFn where SCCUtilFun was used previously.
546		3. Got rid of SCCUtilFun (in comp-lib.cm).
547
548		----------------------------------------------------------------------
549		Name: Lal George
550		Date: 2000/11/06 09:02:21 EST 2000
551		Tag: Release_110_30
552		Description:
553
554		- Version 110.30
555		----------------------------------------------------------------------
556		Name: Matthias Blume
557		Date: 2000/11/04 14:45:00
558		Tag: blume-20001104-mlbuild
559		Description:
560
561		- Made ml-build faster on startup.
562		- Documentation fixes.
563
564		----------------------------------------------------------------------
565		Name: Matthias Blume
566		Date: 2000/11/02 17:00:00 JST
567		Tag: blume-20001102-condcomp
568		Description:
569
570		- Small tweaks to pickler -- new BOOTFILES!
571		- Version bumped to 110.29.2.
572		- Added conditional compilation facility to init.cmi (see comment there).
573		----------------------------------------------------------------------
574		Name: Allen Leung
575		Date: 2000/10/23 19:31:00
576		Tag: leunga-20001023-demo-ra
577
578		1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
579		2. Test programs for the graph library updated
580		3. Some new MLRISC demo programs added
581
582		----------------------------------------------------------------------
583		Name: Matthias Blume
584		Date: 2000/08/31 22:15:00 JST
585		Tag: blume-20001017-errmsg
586		Description:
587
588		More error message grief: Where there used to be no messages, there
589		now were some that had bogus error regions.  Fixed.
590
591		----------------------------------------------------------------------
592		Name: Matthias Blume
593		Date: 2000/08/31 17:30:00 JST
594		Tag: blume-20001017-v110p29p1
595		Description:
596
597		I made a version 110.29.1 with new bootfiles.
598
599		Changes:  Modified pickler/unpickler for faster and leaner unpickling.
600		CM documentation changes and a small bugfix in CM's error reporting.
601
602		----------------------------------------------------------------------
603	Name: Lal George	Name: Lal George
604	Date: 2000/09/27 14:42:35 EDT	Date: 2000/09/27 14:42:35 EDT
605	Tag: george-20000927-nodestatus	Tag: george-20000927-nodestatus

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.705   changed lines   Added in v.812

---

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