Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 675, Fri Jun 23 09:18:18 2000 UTC	revision 763, Sat Dec 23 09:20:02 2000 UTC
#	Line 12	Line 12
12	Tag: <post-commit CVS tag>	Tag: <post-commit CVS tag>
13	Description:	Description:
14	----------------------------------------------------------------------	----------------------------------------------------------------------
15		Name: Lal George
16		Date: 2000/12/22 22:22:58 EST 2000
17		Tag: Release_110_32
18		Description:
19
20		Infinite precision used throughout MLRISC.
21		see MLRISC/mltree/machine-int.sig
22
23		----------------------------------------------------------------------
24		Name: Matthias Blume
25		Date: 2000/12/22 23:16:00 JST
26		Tag: blume-20001222-warn
27		Description:
28
29		Corrected wording and formatting of some CM warning message which I
30		broke in my previous patch.
31
32		----------------------------------------------------------------------
33		Name: Matthias Blume
34		Date: 2000/12/22 21:20:00 JST
35		Tag: blume-20001222-anchorenv
36		Description:
37
38		Fixed CM's handling of anchor environments in connection with CMB.make.
39
40		----------------------------------------------------------------------
41		Name: Matthias Blume
42		Date: 2000/12/22 13:15:00 JST
43		Tag: blume-20001222-cleanup
44		Description:
45
46		Removed src/cm/ffi which does not (and did not) belong here.
47
48		----------------------------------------------------------------------
49		Name: Matthias Blume
50		Date: 2000/12/21 23:55:00 JST
51		Tag: blume-20001221-exn
52		Description:
53
54		Probably most important: CM no longer silently swallows all exceptions
55		in the compiler.
56		Plus: some other minor CM changes.  For example, CM now reports some
57		sizes for generated binfiles (code, data, envpickle, lambdapickle).
58
59		----------------------------------------------------------------------
60		Name: Matthias Blume
61		Date: 2000/12/15 00:01:05 JST
62		Tag: blume-20001215-dirtool
63		Description:
64
65		- "dir" tool added.
66		- improvements and cleanup to Tools structure
67		- documentation updates
68
69		----------------------------------------------------------------------
70		Name: Allen Leung
71		Date: Thu Dec 14 03:45:24 EST 2000
72		Description:
73		Tag:  leunga-20001214-int-inf
74		Description:
75
76		In IntInf, added these standard functions, which are missing from our
77		implementation:
78
79		andb : int * int -> int
80		xorb : int * int -> int
81		orb  : int * int -> int
82		notb : int -> int
83		<<   : int * word -> int
84		~>>  : int * word -> int
85
86		Not tested, I hope they are correct.
87
88		----------------------------------------------------------------------
89		Name: Allen Leung
90		Date: Fri Dec  8 19:23:26 EST 2000
91		Description:
92		Tag:  leunga-20001208-nowhere
93		Description:
94
95		Slight improvements to the 'nowhere' tool to handle OR-patterns,
96		to generate better error messages etc.  Plus a brief manual.
97
98		----------------------------------------------------------------------
99		Name: Lal George
100		Date: 2000/12/08 09:54:02 EST 2000
101		Tag: Release_110_31
102		Description:
103
104		- Version 110.31
105		----------------------------------------------------------------------
106		Name: Allen Leung
107		Date: Thu Dec  7 22:01:04 EST 2000
108		Tag:  leunga-20001207-cell-monster-hack
109		Description:
110
111		Major MLRISC internal changes.  Affect all clients.
112		Summary:
113
114		1.  Type CELLS.cell = int is now replaced by a datatype.
115		As a result, the old regmap is now gone.  Almost all interfaces
116		in MLRISC change as a consequence.
117
118		2.  A new brand version of machine description tool (v3.0) that generates
119		modules expecting the new interface.  The old version is removed.
120
121		3.  The RA interface has been further abstracted into two new functors.
122		RISC_RA and X86RA.  These functors have much simpler interfaces.
123		[See also directory MLRISC/demo.]
124
125		4.  Some other new source->source code generation tools are available:
126
127		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
128		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
129		I use this tool to generate the peephole optimizers---with the new
130		cell type changes, peephole rules are becoming difficult to write
131		without conditional pattern matching.
132
133		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
134		cover the entire MLRISC source tree so many things broke.
135
136		6.  CM files have been moved to the subdirectory MLRISC/cm.
137		They are moved because there are a lot of them and they clutter up the
138		root dir.
139
140		7.  More detailed documentation to come...
141
142		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
143		a makeml -rebuild first.  This is because of other other
144		changes that Matthias has made (see below).
145
146
147		----------------------------------------------------------------------
148		Name: Matthias Blume
149		Date: 2000/11/30 23:12:00 JST
150		Tag: blume-20001130-filereorg
151		Description:
152
153		Some manual updates and some file reorganizations in CM.
154
155		----------------------------------------------------------------------
156		Name: Matthias Blume
157		Date: 2000/11/24 17:45:00 JST
158		Tag: blume-20001124-link
159		Description:
160
161		Drastically improved link traversal code for the case that the dynamic
162		value was already loaded at bootstrap time.  As a result, CM and CMB
163		now both load blazingly fast -- even on a very slow machine.  Also,
164		memory consumption has been further reduced by this.
165
166		Warning: The format of the PIDMAP file has changed.  THerefore, to
167		bootstrap you have to do this:
168
169		1. Run CMB.make
170		2. Make a symbolic link for the boot directory:
171		ln -s sml.boot.ARCH-OS xxx
172		3. "Rebuild" the boot directory:
173		./makeml -boot xxx -rebuild sml ; rm xxx
174		4. Boot normally:
175		./makeml
176
177		----------------------------------------------------------------------
178		Name: Matthias Blume
179		Date: 2000/11/21 21:20:00 JST
180		Tag: blume-20001121-tools
181		Description:
182
183		Continued hacking on autoloading problem -- with success this time.
184		Also changed tool-plugin mechanism.  See new CM manual.
185
186		----------------------------------------------------------------------
187		Name: Matthias Blume
188		Date: 2000/11/19 14:30:00 JST
189		Tag:  blume-20001119-autoload
190		Description:
191
192		Some hacking to make autoloading faster.  Success for CMB, no success
193		so far for CM.  There is a reduced structure CM' that autoloads faster.
194		(This is a temporary, non-documented hack to be eliminated again when
195		the general problem is solved.)
196
197		----------------------------------------------------------------------
198		Name: Matthias Blume
199		Date: 2000/11/17 14:10:00 JST
200		Tag: blume-20001117-pickle-lib
201		Description:
202
203		1. Eliminated comp-lib.cm
204		2. Made pickle-lib.cm
205		3. Eliminated all uses of intset.sml (from comp-lib.cm)
206		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
207		equivalent constructs from smlnj-lib.cm (INtHashTable).
208		5. Point 4. also goes for those uses of intmap.* in MLRISC.
209		Duplicated intmap modules thrown out.
210		6. Hunted down all duplicated SCC code and replaced it with
211		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
212		7. Rewrote Feedback module.
213		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
214		should be thrown out and equivalent modules from smlnj-lib.cm
215		should be used (IntRedBlackSet, IntListSet, ...).
216
217		Confirmed that compiler compiles to fixpoint.
218
219		----------------------------------------------------------------------
220		Name: Allen Leung
221		Date: 2000/11/10 18:00:00
222		Tag: leunga-20001110-new-x86-fp
223
224		A new x86 floating point code generator has been added.
225		By default this is turned off.  To turn this on, do:
226
227		CM.autoload "$smlnj/compiler.cm";
228		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
229
230		Changes:
231
232		1.  Changed FTAN to FPTAN so that the assembly output is correct.
233		2.  Changed the extension callback for FTANGENT to generate:
234
235		fptan
236		fstp  %st(0)
237		instead of
238		fptan
239		fstpl ftempmem
240
241		3.  Numerous assembly fixes for x86.
242
243		5.  Cleaned up the machine code output module x86/x86MC.sml and added
244		support for a whole bunch of instructions and addressing modes:
245
246		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
247		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
248		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
249		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
250		fxch %st(n)
251		fld %st(n)
252		fst %st(n)
253		fst mem
254		fstp %st(n)
255		fucom %st(n)
256		fucomp %st(n)
257
258		All these are now generated when the fast fp mode is turned on.
259
260		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
261
262		----------------------------------------------------------------------
263		Name: Matthias Blume
264		Date: 2000/11/09 11:20:00 JST
265		Tag: blume-20001109-scc
266		Description:
267
268		Eliminated some code duplication:
269
270		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
271		(Otherwise the functor is useless.)
272		2. Used GraphSCCFn where SCCUtilFun was used previously.
273		3. Got rid of SCCUtilFun (in comp-lib.cm).
274
275		----------------------------------------------------------------------
276		Name: Lal George
277		Date: 2000/11/06 09:02:21 EST 2000
278		Tag: Release_110_30
279		Description:
280
281		- Version 110.30
282		----------------------------------------------------------------------
283		Name: Matthias Blume
284		Date: 2000/11/04 14:45:00
285		Tag: blume-20001104-mlbuild
286		Description:
287
288		- Made ml-build faster on startup.
289		- Documentation fixes.
290
291		----------------------------------------------------------------------
292		Name: Matthias Blume
293		Date: 2000/11/02 17:00:00 JST
294		Tag: blume-20001102-condcomp
295		Description:
296
297		- Small tweaks to pickler -- new BOOTFILES!
298		- Version bumped to 110.29.2.
299		- Added conditional compilation facility to init.cmi (see comment there).
300		----------------------------------------------------------------------
301		Name: Allen Leung
302		Date: 2000/10/23 19:31:00
303		Tag: leunga-20001023-demo-ra
304
305		1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
306		2. Test programs for the graph library updated
307		3. Some new MLRISC demo programs added
308
309		----------------------------------------------------------------------
310		Name: Matthias Blume
311		Date: 2000/08/31 22:15:00 JST
312		Tag: blume-20001017-errmsg
313		Description:
314
315		More error message grief: Where there used to be no messages, there
316		now were some that had bogus error regions.  Fixed.
317
318		----------------------------------------------------------------------
319		Name: Matthias Blume
320		Date: 2000/08/31 17:30:00 JST
321		Tag: blume-20001017-v110p29p1
322		Description:
323
324		I made a version 110.29.1 with new bootfiles.
325
326		Changes:  Modified pickler/unpickler for faster and leaner unpickling.
327		CM documentation changes and a small bugfix in CM's error reporting.
328
329		----------------------------------------------------------------------
330		Name: Lal George
331		Date: 2000/09/27 14:42:35 EDT
332		Tag: george-20000927-nodestatus
333		Description:
334
335		Changed the type of the nodestatus, so that:
336
337		SPILLED(~1)             is now SPILLED
338		SPILLED(m) where m>=0   is now MEMREG(m)
339		SPILLED(s) where s<~1   is now SPILL_LOC(~s)
340
341		----------------------------------------------------------------------
342		Name: Matthias Blume
343		Date: 2000/09/07 14:45:00 JST
344		Tag: blume-20000907-cmerrmsg
345		Description:
346
347		Small tweak to CM to avoid getting ML syntax error messages twice.
348
349		----------------------------------------------------------------------
350		Name: Matthias Blume
351		Date: 2000/08/31 18:00:00 JST
352		Tag: blume-20000831-cvsbootfiles
353		Description:
354
355		New URL for boot files (because the 110.29 files on the BL server do
356		now work correctly with my updated install scripts for yacc and lex).
357
358		----------------------------------------------------------------------
359		Name: Matthias Blume
360		Date: 2000/08/08 12:33:00 JST
361		Tag: blume-20000808-manual
362		Description:
363
364		Tiny update to CM manual.
365
366		----------------------------------------------------------------------
367		Name: Allen Leung
368		Date: 2000/08/7 19:31:00
369		Tag: leunga-20000807-a-whole-bunch-of-stuff
370
371		Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
372		and Mobdy.
373
374		1.  x86
375
376		a.  Fixes to peephole module by John and Dan.
377		b.  Assembly fix to SETcc by Allen.
378		c.  Fix to c-call by John.
379		d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
380		e.  Instruction selection optimization to SETcc as suggested by John.
381
382		For example,
383
384		MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
385
386		should generate:
387
388		MOVL a, x
389		SUBL b, x
390		SHRL 31, x
391
392		2.  IR stuff
393
394		A bunch of new DJ-graph related algorithms added.  These
395		speed up SSA construction.
396
397		3.  SSA + Scheduling
398
399		Added code for SSA and scheduling to the repository
400
401		----------------------------------------------------------------------
402		Name: Lal George
403		Date: 2000/07/27 11:53:14 EDT
404
405		Tag: lal-20000727-linux-ppc
406		Description:
407
408		Made changes to support Linux PPC.
409		p.s. I have confirmation that the 110.29 boot files work fine.
410
411		----------------------------------------------------------------------
412		Name: Matthias Blume
413		Date: 2000/07/27 17:40:00 JST
414		Tag: blume-20000727-scripts
415		Description:
416
417		!!!! WARNING !!!!
418		You must recompile the runtime system!
419		!!!! WARNING !!!!
420
421		This is basically another round of script-enhancements:
422
423		1. sml, ml-build, and ml-makedepend accept options -D and -U to define
424		and undefine CM preprocessor symbols.
425
426		2. ml-build avoids generating a new heap image if it finds that the
427		existing one is still ok.  (The condition is that no ML file had to
428		be recompiled and all ML files are found to be older that the heap
429		file.)
430
431		To make this work smoothly, I also hacked the runtime system as
432		well as SMLofNJ.SysInfo to get access to the heap image suffix
433		(.sparc-solaris, ...) that is currently being used.
434
435		Moreover, the signature of CM.mk_standalone has changed.  See the
436		CM manual.
437
438		3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
439		CM manual for details.)
440
441		4. More CM manual updates:
442		- all of the above has been documented.
443		- there is now a section describing the (CM-related) command line
444		arguments that are accepted by the "sml" command
445
446		----------------------------------------------------------------------
447		Name: Matthias Blume
448		Date: 2000/07/25 16:20:00 JST
449		Tag: blume-20000725-makedepend
450		Description:
451
452		Added a script called ml-makedepend.  This can be used in makefiles
453		for Unix' make in a way very similar to the "makedepend" command for
454		C.
455
456		The script internally uses function CM.sources.
457
458		Synopsis:
459
460		ml-makedepend [-f makefile] cmfile targetname
461
462		The default for the makefile is "makefile" (or "Makefile" should
463		"makefile" not exist).
464
465		ml-makedepend adds a cmfile/targetname-specific section to this
466		makefile (after removing the previous version of this section).  The
467		section contains a single dependency specification with targetname on
468		the LHS (targetname is an arbitrary name), and a list of files derived
469		from the cmfile on the RHS.  Some of the files on the RHS are
470		ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
471		"make" variables $(ARCH) and $(OPSYS) in place of the corresponding
472		path names.  The makefile writer is responsible for making sure that
473		these variables have correct at the time "make" is invoked.
474
475		----------------------------------------------------------------------
476		Name: Matthias Blume
477		Date: 2000/07/22 23:30:00 JST
478		Tag: blume-20000722-urlupdate
479		Description:
480
481		Changed BOOT and config/srcarchiveurl to point to BL server:
482
483		ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
484
485		----------------------------------------------------------------------
486		Name: Matthias Blume
487		Date: 2000/07/18 18:00:00 JST
488		Tag: blume-20000718-Version_110_29
489		Description:
490
491		1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
492
493		2. Updated config/version to 110.29
494
495		3. Updated config/srcarchiveurl
496
497		3. New boot files!
498		ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
499
500		----------------------------------------------------------------------
501		Name: Matthias Blume
502		Date: 2000/07/11 13:58:00 JST
503		Tag: blume-20000711-doctypo
504		Description:
505
506		Fixed a few typos in CM manual.
507
508		----------------------------------------------------------------------
509		Name: Allen Leung
510		Date: 2000/06/15 00:38:00
511		Tag: leunga-20000704-sparc-x86
512
513		1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
514		2. fix to x86 RET bug [found by Dan Grossman]
515		3. sparc assembly bug fix for ticc instructions [found by Fermin]
516
517		Affects c-- and moby only
518
519		----------------------------------------------------------------------
520		Name: Matthias Blume
521		Date: 2000/07/04 15:26:00
522		Tag: blume-20000704-trigger
523		Description:
524
525		1. Improvements to CM manual.
526		2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
527		of getting a back-trace.  The function, when called, raises an
528		internal exception which explicitly carries the full back-trace history,
529		so it is unaffected by any intervening handle-raise pairs ("trivial"
530		or not).  The interactive loop will print that history once it arrives
531		at top level.
532		Short of having all exceptions implicitly carry the full history, the
533		recommended way of using this facility is:
534		- compile your program with instrumentation "on"
535		- run it, when it raises an exception, look at the history
536		- if the history is "cut off" because of some handler, go and modify
537		your program so that it explicitly calls BTrace.trigger
538		- recompile (still instrumented), and rerun; look at the full history
539
540		----------------------------------------------------------------------
541		Name: Matthias Blume
542		Date: 2000/07/03 15:36:00 JST
543		Tag: blume-20000702-manual
544		Description:
545
546		Small corrections and updates to CM manual.
547
548		----------------------------------------------------------------------
549		Name: Matthias Blume
550		Date: 2000/06/29 16:04:00 JST
551		Tag: blume-20000629-yacctool
552		Description:
553
554		Changes:
555
556		1. Class "mlyacc" now takes separate arguments to pass options to
557		generated .sml- and .sig-files independently.
558		2. Corresponding CM manual updates.
559		3. BTrace module now also reports call sites.  (However, for loop clusters
560		it only shows from where the cluster was entered.)  There are associated
561		modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
562
563		----------------------------------------------------------------------
564		Name: Matthias Blume
565		Date: 2000/06/27 16:51:00 JST
566		Tag: blume-20000627-noweb
567		Description:
568
569		Changes:
570
571		1. Implemented "subdir" and "witness" options for noweb tool.
572		This caused some slight internal changes in CM's tool implementation.
573		2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
574		some remaining issues from earlier path anchor changes.
575		3. Updated CM manual accordingly.
576
577		4. Changed implementation of back-tracing so that I now consider it
578		ready for prime-time.
579
580		In particular, you don't have to explicitly trigger the back-trace
581		anymore.  Instead, if you are running BTrace-instrumented code and
582		there is an uncaught exception (regardless of whether or not it was
583		raised in instrumented code), the top-level evalloop will print
584		the back-trace.
585
586		Features:
587
588		- Instrumented and uninstrumented code work together seemlessly.
589		(Of course, uninstrumented code is never mentioned in actual
590		back-traces.)
591
592		- Asymptotic time- and space-complexity of instrumented code is
593		equal to that of uninstrumented code.  (This means that
594		tail-recursion is preserved by the instrumentation phase.)
595
596		- Modules whose code has been instrumented in different sessions
597		work together without problem.
598
599		- There is no penalty whatsoever on uninstrumented code.
600
601		- There is no penalty on "raise" expressions, even in
602		instrumented code.
603
604		A potential bug (or perhaps it is a feature, too):
605
606		A back-trace reaches no further than the outermost instrumented
607		non-trivial "raise".  Here, a "trivial" raise is one that is the
608		sole RHS of a "handle" rule.  Thus, back-traces reach trough
609
610		<exp> handle e => raise e
611
612		and even
613
614		<exp> handle Foo => raise Bar
615
616		and, of course, through
617
618		<exp> handle Foo => ...
619
620		if the exception was not Foo.
621
622		Back-traces always reach right through any un-instrumented code
623		including any of its "handle" expressions, trivial or not.
624
625		To try this out, do the following:
626
627		- Erase all existing binfiles for your program.
628		(You may keep binfiles for those modules where you think you
629		definitely don't need back-tracing.)
630		- Turn on back-trace instrumentation:
631		SMLofNJ.Internals.BTrace.mode (SOME true);
632		- Recompile your program.  (I.e., run "CM.make" or "use".)
633		- You may now turn instrumentation off again (if you want):
634		SMLofNJ.Internals.BTrace.mode (SOME false);
635		- Run your program as usual.  If it raises an exception that
636		reaches the interactive toplevel, then a back-trace will
637		automatically be printed.  After that, the toplevel loop
638		will print the exception history as usual.
639
640		----------------------------------------------------------------------
641		Name: Matthias Blume
642		Date: 2000/06/26 09:56:46 JST
643		Tag: blume-20000626-setup
644		Description:
645
646		CM: - setup-parameter to "sml" added; this can be used to run arbitrary
647		ML code before and after compiling a file (e.g., to set compiler
648		flags)
649
650		Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
651		- associated changes to btrace.sml (BTrace instrumentation pass)
652		- cleaner implementation of btimp.sml (BTrace tracing and report
653		module)
654
655		CM manual: * new path encoding documented
656		* description of setup-parameter to "sml" added
657
658		The biggest user-visible change to back-tracing is that it is no
659		longer necessary to compile all traced modules within the same
660		session.  (This was a real limitation.)
661
662		----------------------------------------------------------------------
663		Name: Matthias Blume
664		Date: 2000/06/24 12:40:00 JST
665		Tag: blume-20000624-startup
666		Description:
667
668		Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
669		bit too often -- to put it mildly. :)
670
671		----------------------------------------------------------------------
672	Name: Matthias Blume	Name: Matthias Blume
673	Date: 2000/06/23 18:20:00 JST	Date: 2000/06/23 18:20:00 JST
674	Tag: blume-20000623-btrace	Tag: blume-20000623-btrace

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.675   changed lines   Added in v.763

---

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