Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 633, Sun Apr 30 16:44:09 2000 UTC	revision 814, Tue May 1 15:44:45 2001 UTC
#	Line 11	Line 11
11	Date:	Date:
12	Tag: <post-commit CVS tag>	Tag: <post-commit CVS tag>
13	Description:	Description:
14
15		----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2001/05/01 11:45:00 EDT
18		Tag: blume-20010501-pcedittmp
19		Description:
20
21		Changed install.sh to use the current working directory instead of
22		/usr/tmp for a temporary file (pcedittmp).  The previous choice
23		of /usr/tmp caused trouble with MacOS X because of file premission
24		problems.
25
26		----------------------------------------------------------------------
27		Name: Matthias Blume
28		Date: 2001/04/20 11:10:00 EDT
29		Tag: blume-20010420-inMLflag
30		Description:
31
32		- added vp_limitPtrMask to vproc-state.h
33		(for use by the raw-C-calls mechanism to implement proper interrupt
34		handling)
35		- made the ML compiler aware of various data-structure offsets so it
36		can generate code for accessing the vp_inML flag and vp_limitPtrMask
37		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
38		raw C-calls
39
40		----------------------------------------------------------------------
41		Name: Lal George
42		Date: 2001/04/20 09:15:28 EDT
43		Tag: george-20010420-macosX
44		Description:
45
46		- Changes to port to Mac OS X; Darwin.
47
48		- In the process I found that sqrt was broken on the PPC, because the
49		fsqrt instruction is not implemented.
50
51		----------------------------------------------------------------------
52		Name: Matthias Blume
53		Date: 2001/04/18 12:45:00 EDT
54		Tag: blume-20010418-ccalls
55		Description:
56
57		- fixed two off-by-4 errors in the x86-specific c-calls implementation
58		(this bug prevented structure arguments containing pointers from being
59		passed correctly)
60		- changed the raw-C-call code in mlriscGen.sml in such a way that
61		structure arguments are represented as a pointer to the beginning
62		of the structure (instead of having a series of synthesized arguments,
63		one for each structure member)
64
65		- made makeml script's verbosity level configurable via environment
66		variable (MAKEML_VERBOSITY)
67
68		- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
69		in rawmem-x86.sml; we are now using the real thing
70
71		----------------------------------------------------------------------
72		Name: Matthias Blume
73		Date: 2001/03/22 16:25:00 EST
74		Tag: blume-20010322-bootfiles
75		Description:
76
77		Created a new set of bootfiles (for your automatic installation convenience).
78
79		----------------------------------------------------------------------
80		Name: Matthias Blume
81		Date: 2001/03/22 15:10:00 EST
82		Tag: blume-20010322-rawmem-parcm
83		Description:
84
85		1. All "raw memory access" primitives for the new FFI are implemented now
86		(at least on the x86).
87		2. Some further cleanup of CM's parallel make mechanism.
88
89		----------------------------------------------------------------------
90		Name: Matthias Blume
91		Date: 2001/03/19 17:53:00 EST
92		Tag: blume-20010319-parallel
93		Description:
94
95		Parallel make (using compile servers) now works again.
96
97		To this end, CM.stabilize and CMB.make have been modified to work in
98		two passes when compile servers are attached:
99		1. Compile everything, do not perform stabilization; this pass
100		uses compile servers
101		2. Stabilize everything; this pass does not use compile servers
102		If there are no compile servers, the two passes are combined into one
103		(as before).  Splitting the passes increases the inherent parallelism
104		in the dependency graph because the entire graph including all
105		libraries is available at the same time.  This, in turn, improves
106		server utilization.  The downside is that the master process will
107		have to do some extra work after compilation is done (because for
108		technical reasons it must re-read all the binfiles during stabilization).
109
110		----------------------------------------------------------------------
111		Name: Matthias Blume
112		Date: 2001/03/16 12:22:00 EST
113		Tag: blume-20010316-bootfiles
114		Description:
115
116		Created a new set of bootfiles (for your automatic installation convenience).
117
118		----------------------------------------------------------------------
119		Name: Matthias Blume
120		Date: 2001/03/16 11:00:00 EST
121		Tag: blume-20010316-MLTREE-fixup
122		Description:
123
124		This is a minor fixup for an (untagged) earlier commit by Allen.
125		(A file was missing).
126
127		----------------------------------------------------------------------
128		Name: Allen Leung
129		Date: Mon Mar  5 18:54:57 EST 2001
130		Tag: leunga-20010305-cut-support
131
132		1. New support for alternative control-flow in MLTREE.
133		Currently we support
134
135		FLOW_TO(CALL ...., [k1,...,kn])
136
137		This is needed for 'cuts to' in C-- and try/handle-like constructs
138		in Moby
139
140		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
141
142		2. Register Allocator
143
144		Changes in interface [from Fermin, John]
145
146		3. Alpha 8-bit SLL support [Fermin]
147
148		4. All architectures
149
150		A new module (ClusterExpandCopies) for expanding parallel copies.
151
152		----------------------------------------------------------------------
153		Name: Allen Leung
154		Date: 2001/02/27 23:07:00 EST
155		Tag: leunga-20010227-minor-stuff
156
157		1. Alpha bug fix for CMOVNE
158		2. Handle mltree COND(..,FCMP ...,...)
159		3. Bug fix in simplifier
160
161		----------------------------------------------------------------------
162		Name: Matthias Blume
163		Date: 2001/01/30 17:50:00 EST
164		Tag: blume-20010130-sync
165		Description:
166
167		This is just a minor update to sync my devel branch with the main brach.
168		The only visible change is the addition of some README files.
169
170		----------------------------------------------------------------------
171		Name: Matthias Blume
172		Date: 2001/01/12 23:30:00 JST
173		Tag: blume-20010112-bootfiles
174		Description:
175
176		Made a new set of bootfiles that goes with the current state of the
177		repository.
178
179		----------------------------------------------------------------------
180		Name: Matthias Blume
181		Date: 2001/01/12 21:20:00 JST
182		Tag: blume-20010112-sync
183		Description:
184
185		I am just flushing out some minor changes that had accumulated in
186		my private branch in order to sync with the main tree.  (This is
187		mainly because I had CVS trouble when trying to merge _into_ my
188		private branch.)
189
190		Most people should be completely unaffected by this.
191
192		----------------------------------------------------------------------
193		Name: Allen Leung
194		Date: Thu Jan 11 21:03:00 EST 2001
195		Tag: leunga-20010111-labexp=mltree
196		Description:
197
198		1.  Removed the type LabelExp and replace it by MLTree.
199		2.  Rewritten mltree-simplify with the pattern matcher tool.
200		3.  There were some bugs in alpha code generator which would break
201		64-bit code generation.
202		4.  Redo the tools to generate code with the
203		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
204		are now generated by perl scripts.
205
206		----------------------------------------------------------------------
207		Name: Matthias Blume
208		Date: 2001/01/10 21:55:00 JST
209		Tag: blume-20010110-rcc
210		Description:
211
212		The RCC stuff now seems to work (but only on the x86).
213		This required hacking of the c-calls interface (and -implementation) in
214		MLRISC.
215
216		Normal compiler users should be unaffected.
217
218		----------------------------------------------------------------------
219		Name: Matthias Blume
220		Date: 2001/01/09 01:20:00 JST
221		Tag: blume-20010109-rcc
222		Description:
223
224		This is a fairly big patch, flushing out a large number of pending
225		changes that I made to my development copy over the last couple of days.
226
227		Of practical relevance at this moment is a workaround for a pickling
228		bug that Allen ran into the other day.  The cause of the bug itself is
229		still unknown and it might be hard to fix it properly, but the
230		workaround has some merits of its own (namely somewhat reducing pickling
231		overhead for certain libraries).  Therefore, I think this solution should
232		be satisfactory at this time.
233
234		The rest of the changes (i.e., the vast majority) has to do with my
235		ongoing efforts of providing direct support for C function calls from
236		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
237		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
238		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
239		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
240		handle RCC.
241
242		The last part (the changes to mlriscGen.sml) are still known to be
243		wrong on the x86 and not implemented on all other architectures.  But
244		the infrastructure is in place. I had to change a few functor
245		signatures in the backend to be able to route the CCalls interface
246		from MLRISC there, and I had to specialize the mltree type (on the
247		x86) to include the necessary extensions. (The extensions themselves
248		were already there and redy to go in MLRISC/x86).
249
250		Everything should be very happy as soon as someone helps me with
251		mlriscGen.sml...
252
253		In any case, nothing of this should matter to anyone as long as the
254		new primop is not being used (which is going to be the case unless you
255		find it where I hid it :). The rest of the compiler is completely
256		unaffected.
257
258		----------------------------------------------------------------------
259		Name: Matthias Blume
260		Date: 2001/01/05 00:30:00 JST
261		Tag: blume-20010105-primops
262		Description:
263
264		Added some experimental support for work that I am doing right now.
265		These changes mostly concern added primops, but there is also a new
266		experimental C library in the runtime system (but currently not enabled
267		anywhere except on Linux/X86).
268
269		In the course of adding primops (and playing with them), I discovered that
270		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
271		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
272		To recover, I made type information in INL_PRIM mandatory and changed
273		prim.sml as well as built-in.sml accordingly.  The InLine structure now
274		has complete, correct type information (i.e., no bottom types).
275
276		Since all these changes mean that we need new binfiles, I also bumped the
277		version number to 110.32.1.
278
279		----------------------------------------------------------------------
280		Name: Matthias Blume
281		Date: 2000/12/30 22:10:00 JST
282		Tag: blume-20001230-various
283		Description:
284
285		Added proxy libraries for MLRISC and let MLRISC libraries refer
286		to each other using path anchors.  (See CM manual for explanation.)
287
288		Updated CM documentation.
289
290		Fixed some bugs in CM.
291
292		Implemented "proxy" libraries (= syntactic sugar for CM).
293
294		Added "-quiet" option to makeml and changed runtime system accordingly.
295
296		Added cleanup handler for exportML to reset timers and compiler stats.
297
298		----------------------------------------------------------------------
299		Name: Lal George
300		Date: 2000/12/22 22:22:58 EST 2000
301		Tag: Release_110_32
302		Description:
303
304		Infinite precision used throughout MLRISC.
305		see MLRISC/mltree/machine-int.sig
306
307		----------------------------------------------------------------------
308		Name: Matthias Blume
309		Date: 2000/12/22 23:16:00 JST
310		Tag: blume-20001222-warn
311		Description:
312
313		Corrected wording and formatting of some CM warning message which I
314		broke in my previous patch.
315
316		----------------------------------------------------------------------
317		Name: Matthias Blume
318		Date: 2000/12/22 21:20:00 JST
319		Tag: blume-20001222-anchorenv
320		Description:
321
322		Fixed CM's handling of anchor environments in connection with CMB.make.
323
324		----------------------------------------------------------------------
325		Name: Matthias Blume
326		Date: 2000/12/22 13:15:00 JST
327		Tag: blume-20001222-cleanup
328		Description:
329
330		Removed src/cm/ffi which does not (and did not) belong here.
331
332		----------------------------------------------------------------------
333		Name: Matthias Blume
334		Date: 2000/12/21 23:55:00 JST
335		Tag: blume-20001221-exn
336		Description:
337
338		Probably most important: CM no longer silently swallows all exceptions
339		in the compiler.
340		Plus: some other minor CM changes.  For example, CM now reports some
341		sizes for generated binfiles (code, data, envpickle, lambdapickle).
342
343		----------------------------------------------------------------------
344		Name: Matthias Blume
345		Date: 2000/12/15 00:01:05 JST
346		Tag: blume-20001215-dirtool
347		Description:
348
349		- "dir" tool added.
350		- improvements and cleanup to Tools structure
351		- documentation updates
352
353		----------------------------------------------------------------------
354		Name: Allen Leung
355		Date: Thu Dec 14 03:45:24 EST 2000
356		Description:
357		Tag:  leunga-20001214-int-inf
358		Description:
359
360		In IntInf, added these standard functions, which are missing from our
361		implementation:
362
363		andb : int * int -> int
364		xorb : int * int -> int
365		orb  : int * int -> int
366		notb : int -> int
367		<<   : int * word -> int
368		~>>  : int * word -> int
369
370		Not tested, I hope they are correct.
371
372		----------------------------------------------------------------------
373		Name: Allen Leung
374		Date: Fri Dec  8 19:23:26 EST 2000
375		Description:
376		Tag:  leunga-20001208-nowhere
377		Description:
378
379		Slight improvements to the 'nowhere' tool to handle OR-patterns,
380		to generate better error messages etc.  Plus a brief manual.
381
382		----------------------------------------------------------------------
383		Name: Lal George
384		Date: 2000/12/08 09:54:02 EST 2000
385		Tag: Release_110_31
386		Description:
387
388		- Version 110.31
389		----------------------------------------------------------------------
390		Name: Allen Leung
391		Date: Thu Dec  7 22:01:04 EST 2000
392		Tag:  leunga-20001207-cell-monster-hack
393		Description:
394
395		Major MLRISC internal changes.  Affect all clients.
396		Summary:
397
398		1.  Type CELLS.cell = int is now replaced by a datatype.
399		As a result, the old regmap is now gone.  Almost all interfaces
400		in MLRISC change as a consequence.
401
402		2.  A new brand version of machine description tool (v3.0) that generates
403		modules expecting the new interface.  The old version is removed.
404
405		3.  The RA interface has been further abstracted into two new functors.
406		RISC_RA and X86RA.  These functors have much simpler interfaces.
407		[See also directory MLRISC/demo.]
408
409		4.  Some other new source->source code generation tools are available:
410
411		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
412		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
413		I use this tool to generate the peephole optimizers---with the new
414		cell type changes, peephole rules are becoming difficult to write
415		without conditional pattern matching.
416
417		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
418		cover the entire MLRISC source tree so many things broke.
419
420		6.  CM files have been moved to the subdirectory MLRISC/cm.
421		They are moved because there are a lot of them and they clutter up the
422		root dir.
423
424		7.  More detailed documentation to come...
425
426		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
427		a makeml -rebuild first.  This is because of other other
428		changes that Matthias has made (see below).
429
430
431		----------------------------------------------------------------------
432		Name: Matthias Blume
433		Date: 2000/11/30 23:12:00 JST
434		Tag: blume-20001130-filereorg
435		Description:
436
437		Some manual updates and some file reorganizations in CM.
438
439		----------------------------------------------------------------------
440		Name: Matthias Blume
441		Date: 2000/11/24 17:45:00 JST
442		Tag: blume-20001124-link
443		Description:
444
445		Drastically improved link traversal code for the case that the dynamic
446		value was already loaded at bootstrap time.  As a result, CM and CMB
447		now both load blazingly fast -- even on a very slow machine.  Also,
448		memory consumption has been further reduced by this.
449
450		Warning: The format of the PIDMAP file has changed.  THerefore, to
451		bootstrap you have to do this:
452
453		1. Run CMB.make
454		2. Make a symbolic link for the boot directory:
455		ln -s sml.boot.ARCH-OS xxx
456		3. "Rebuild" the boot directory:
457		./makeml -boot xxx -rebuild sml ; rm xxx
458		4. Boot normally:
459		./makeml
460
461		----------------------------------------------------------------------
462		Name: Matthias Blume
463		Date: 2000/11/21 21:20:00 JST
464		Tag: blume-20001121-tools
465		Description:
466
467		Continued hacking on autoloading problem -- with success this time.
468		Also changed tool-plugin mechanism.  See new CM manual.
469
470		----------------------------------------------------------------------
471		Name: Matthias Blume
472		Date: 2000/11/19 14:30:00 JST
473		Tag:  blume-20001119-autoload
474		Description:
475
476		Some hacking to make autoloading faster.  Success for CMB, no success
477		so far for CM.  There is a reduced structure CM' that autoloads faster.
478		(This is a temporary, non-documented hack to be eliminated again when
479		the general problem is solved.)
480
481		----------------------------------------------------------------------
482		Name: Matthias Blume
483		Date: 2000/11/17 14:10:00 JST
484		Tag: blume-20001117-pickle-lib
485		Description:
486
487		1. Eliminated comp-lib.cm
488		2. Made pickle-lib.cm
489		3. Eliminated all uses of intset.sml (from comp-lib.cm)
490		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
491		equivalent constructs from smlnj-lib.cm (INtHashTable).
492		5. Point 4. also goes for those uses of intmap.* in MLRISC.
493		Duplicated intmap modules thrown out.
494		6. Hunted down all duplicated SCC code and replaced it with
495		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
496		7. Rewrote Feedback module.
497		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
498		should be thrown out and equivalent modules from smlnj-lib.cm
499		should be used (IntRedBlackSet, IntListSet, ...).
500
501		Confirmed that compiler compiles to fixpoint.
502
503		----------------------------------------------------------------------
504		Name: Allen Leung
505		Date: 2000/11/10 18:00:00
506		Tag: leunga-20001110-new-x86-fp
507
508		A new x86 floating point code generator has been added.
509		By default this is turned off.  To turn this on, do:
510
511		CM.autoload "$smlnj/compiler.cm";
512		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
513
514		Changes:
515
516		1.  Changed FTAN to FPTAN so that the assembly output is correct.
517		2.  Changed the extension callback for FTANGENT to generate:
518
519		fptan
520		fstp  %st(0)
521		instead of
522		fptan
523		fstpl ftempmem
524
525		3.  Numerous assembly fixes for x86.
526
527		5.  Cleaned up the machine code output module x86/x86MC.sml and added
528		support for a whole bunch of instructions and addressing modes:
529
530		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
531		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
532		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
533		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
534		fxch %st(n)
535		fld %st(n)
536		fst %st(n)
537		fst mem
538		fstp %st(n)
539		fucom %st(n)
540		fucomp %st(n)
541
542		All these are now generated when the fast fp mode is turned on.
543
544		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
545
546		----------------------------------------------------------------------
547		Name: Matthias Blume
548		Date: 2000/11/09 11:20:00 JST
549		Tag: blume-20001109-scc
550		Description:
551
552		Eliminated some code duplication:
553
554		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
555		(Otherwise the functor is useless.)
556		2. Used GraphSCCFn where SCCUtilFun was used previously.
557		3. Got rid of SCCUtilFun (in comp-lib.cm).
558
559		----------------------------------------------------------------------
560		Name: Lal George
561		Date: 2000/11/06 09:02:21 EST 2000
562		Tag: Release_110_30
563		Description:
564
565		- Version 110.30
566		----------------------------------------------------------------------
567		Name: Matthias Blume
568		Date: 2000/11/04 14:45:00
569		Tag: blume-20001104-mlbuild
570		Description:
571
572		- Made ml-build faster on startup.
573		- Documentation fixes.
574
575		----------------------------------------------------------------------
576		Name: Matthias Blume
577		Date: 2000/11/02 17:00:00 JST
578		Tag: blume-20001102-condcomp
579		Description:
580
581		- Small tweaks to pickler -- new BOOTFILES!
582		- Version bumped to 110.29.2.
583		- Added conditional compilation facility to init.cmi (see comment there).
584		----------------------------------------------------------------------
585		Name: Allen Leung
586		Date: 2000/10/23 19:31:00
587		Tag: leunga-20001023-demo-ra
588
589		1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
590		2. Test programs for the graph library updated
591		3. Some new MLRISC demo programs added
592
593		----------------------------------------------------------------------
594		Name: Matthias Blume
595		Date: 2000/08/31 22:15:00 JST
596		Tag: blume-20001017-errmsg
597		Description:
598
599		More error message grief: Where there used to be no messages, there
600		now were some that had bogus error regions.  Fixed.
601
602		----------------------------------------------------------------------
603		Name: Matthias Blume
604		Date: 2000/08/31 17:30:00 JST
605		Tag: blume-20001017-v110p29p1
606		Description:
607
608		I made a version 110.29.1 with new bootfiles.
609
610		Changes:  Modified pickler/unpickler for faster and leaner unpickling.
611		CM documentation changes and a small bugfix in CM's error reporting.
612
613		----------------------------------------------------------------------
614		Name: Lal George
615		Date: 2000/09/27 14:42:35 EDT
616		Tag: george-20000927-nodestatus
617		Description:
618
619		Changed the type of the nodestatus, so that:
620
621		SPILLED(~1)             is now SPILLED
622		SPILLED(m) where m>=0   is now MEMREG(m)
623		SPILLED(s) where s<~1   is now SPILL_LOC(~s)
624
625		----------------------------------------------------------------------
626		Name: Matthias Blume
627		Date: 2000/09/07 14:45:00 JST
628		Tag: blume-20000907-cmerrmsg
629		Description:
630
631		Small tweak to CM to avoid getting ML syntax error messages twice.
632
633		----------------------------------------------------------------------
634		Name: Matthias Blume
635		Date: 2000/08/31 18:00:00 JST
636		Tag: blume-20000831-cvsbootfiles
637		Description:
638
639		New URL for boot files (because the 110.29 files on the BL server do
640		now work correctly with my updated install scripts for yacc and lex).
641
642		----------------------------------------------------------------------
643		Name: Matthias Blume
644		Date: 2000/08/08 12:33:00 JST
645		Tag: blume-20000808-manual
646		Description:
647
648		Tiny update to CM manual.
649
650		----------------------------------------------------------------------
651		Name: Allen Leung
652		Date: 2000/08/7 19:31:00
653		Tag: leunga-20000807-a-whole-bunch-of-stuff
654
655		Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
656		and Mobdy.
657
658		1.  x86
659
660		a.  Fixes to peephole module by John and Dan.
661		b.  Assembly fix to SETcc by Allen.
662		c.  Fix to c-call by John.
663		d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
664		e.  Instruction selection optimization to SETcc as suggested by John.
665
666		For example,
667
668		MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
669
670		should generate:
671
672		MOVL a, x
673		SUBL b, x
674		SHRL 31, x
675
676		2.  IR stuff
677
678		A bunch of new DJ-graph related algorithms added.  These
679		speed up SSA construction.
680
681		3.  SSA + Scheduling
682
683		Added code for SSA and scheduling to the repository
684
685		----------------------------------------------------------------------
686		Name: Lal George
687		Date: 2000/07/27 11:53:14 EDT
688
689		Tag: lal-20000727-linux-ppc
690		Description:
691
692		Made changes to support Linux PPC.
693		p.s. I have confirmation that the 110.29 boot files work fine.
694
695		----------------------------------------------------------------------
696		Name: Matthias Blume
697		Date: 2000/07/27 17:40:00 JST
698		Tag: blume-20000727-scripts
699		Description:
700
701		!!!! WARNING !!!!
702		You must recompile the runtime system!
703		!!!! WARNING !!!!
704
705		This is basically another round of script-enhancements:
706
707		1. sml, ml-build, and ml-makedepend accept options -D and -U to define
708		and undefine CM preprocessor symbols.
709
710		2. ml-build avoids generating a new heap image if it finds that the
711		existing one is still ok.  (The condition is that no ML file had to
712		be recompiled and all ML files are found to be older that the heap
713		file.)
714
715		To make this work smoothly, I also hacked the runtime system as
716		well as SMLofNJ.SysInfo to get access to the heap image suffix
717		(.sparc-solaris, ...) that is currently being used.
718
719		Moreover, the signature of CM.mk_standalone has changed.  See the
720		CM manual.
721
722		3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
723		CM manual for details.)
724
725		4. More CM manual updates:
726		- all of the above has been documented.
727		- there is now a section describing the (CM-related) command line
728		arguments that are accepted by the "sml" command
729
730		----------------------------------------------------------------------
731		Name: Matthias Blume
732		Date: 2000/07/25 16:20:00 JST
733		Tag: blume-20000725-makedepend
734		Description:
735
736		Added a script called ml-makedepend.  This can be used in makefiles
737		for Unix' make in a way very similar to the "makedepend" command for
738		C.
739
740		The script internally uses function CM.sources.
741
742		Synopsis:
743
744		ml-makedepend [-f makefile] cmfile targetname
745
746		The default for the makefile is "makefile" (or "Makefile" should
747		"makefile" not exist).
748
749		ml-makedepend adds a cmfile/targetname-specific section to this
750		makefile (after removing the previous version of this section).  The
751		section contains a single dependency specification with targetname on
752		the LHS (targetname is an arbitrary name), and a list of files derived
753		from the cmfile on the RHS.  Some of the files on the RHS are
754		ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
755		"make" variables $(ARCH) and $(OPSYS) in place of the corresponding
756		path names.  The makefile writer is responsible for making sure that
757		these variables have correct at the time "make" is invoked.
758
759		----------------------------------------------------------------------
760		Name: Matthias Blume
761		Date: 2000/07/22 23:30:00 JST
762		Tag: blume-20000722-urlupdate
763		Description:
764
765		Changed BOOT and config/srcarchiveurl to point to BL server:
766
767		ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
768
769		----------------------------------------------------------------------
770		Name: Matthias Blume
771		Date: 2000/07/18 18:00:00 JST
772		Tag: blume-20000718-Version_110_29
773		Description:
774
775		1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
776
777		2. Updated config/version to 110.29
778
779		3. Updated config/srcarchiveurl
780
781		3. New boot files!
782		ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
783
784		----------------------------------------------------------------------
785		Name: Matthias Blume
786		Date: 2000/07/11 13:58:00 JST
787		Tag: blume-20000711-doctypo
788		Description:
789
790		Fixed a few typos in CM manual.
791
792		----------------------------------------------------------------------
793		Name: Allen Leung
794		Date: 2000/06/15 00:38:00
795		Tag: leunga-20000704-sparc-x86
796
797		1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
798		2. fix to x86 RET bug [found by Dan Grossman]
799		3. sparc assembly bug fix for ticc instructions [found by Fermin]
800
801		Affects c-- and moby only
802
803		----------------------------------------------------------------------
804		Name: Matthias Blume
805		Date: 2000/07/04 15:26:00
806		Tag: blume-20000704-trigger
807		Description:
808
809		1. Improvements to CM manual.
810		2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
811		of getting a back-trace.  The function, when called, raises an
812		internal exception which explicitly carries the full back-trace history,
813		so it is unaffected by any intervening handle-raise pairs ("trivial"
814		or not).  The interactive loop will print that history once it arrives
815		at top level.
816		Short of having all exceptions implicitly carry the full history, the
817		recommended way of using this facility is:
818		- compile your program with instrumentation "on"
819		- run it, when it raises an exception, look at the history
820		- if the history is "cut off" because of some handler, go and modify
821		your program so that it explicitly calls BTrace.trigger
822		- recompile (still instrumented), and rerun; look at the full history
823
824		----------------------------------------------------------------------
825		Name: Matthias Blume
826		Date: 2000/07/03 15:36:00 JST
827		Tag: blume-20000702-manual
828		Description:
829
830		Small corrections and updates to CM manual.
831
832		----------------------------------------------------------------------
833		Name: Matthias Blume
834		Date: 2000/06/29 16:04:00 JST
835		Tag: blume-20000629-yacctool
836		Description:
837
838		Changes:
839
840		1. Class "mlyacc" now takes separate arguments to pass options to
841		generated .sml- and .sig-files independently.
842		2. Corresponding CM manual updates.
843		3. BTrace module now also reports call sites.  (However, for loop clusters
844		it only shows from where the cluster was entered.)  There are associated
845		modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
846
847		----------------------------------------------------------------------
848		Name: Matthias Blume
849		Date: 2000/06/27 16:51:00 JST
850		Tag: blume-20000627-noweb
851		Description:
852
853		Changes:
854
855		1. Implemented "subdir" and "witness" options for noweb tool.
856		This caused some slight internal changes in CM's tool implementation.
857		2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
858		some remaining issues from earlier path anchor changes.
859		3. Updated CM manual accordingly.
860
861		4. Changed implementation of back-tracing so that I now consider it
862		ready for prime-time.
863
864		In particular, you don't have to explicitly trigger the back-trace
865		anymore.  Instead, if you are running BTrace-instrumented code and
866		there is an uncaught exception (regardless of whether or not it was
867		raised in instrumented code), the top-level evalloop will print
868		the back-trace.
869
870		Features:
871
872		- Instrumented and uninstrumented code work together seemlessly.
873		(Of course, uninstrumented code is never mentioned in actual
874		back-traces.)
875
876		- Asymptotic time- and space-complexity of instrumented code is
877		equal to that of uninstrumented code.  (This means that
878		tail-recursion is preserved by the instrumentation phase.)
879
880		- Modules whose code has been instrumented in different sessions
881		work together without problem.
882
883		- There is no penalty whatsoever on uninstrumented code.
884
885		- There is no penalty on "raise" expressions, even in
886		instrumented code.
887
888		A potential bug (or perhaps it is a feature, too):
889
890		A back-trace reaches no further than the outermost instrumented
891		non-trivial "raise".  Here, a "trivial" raise is one that is the
892		sole RHS of a "handle" rule.  Thus, back-traces reach trough
893
894		<exp> handle e => raise e
895
896		and even
897
898		<exp> handle Foo => raise Bar
899
900		and, of course, through
901
902		<exp> handle Foo => ...
903
904		if the exception was not Foo.
905
906		Back-traces always reach right through any un-instrumented code
907		including any of its "handle" expressions, trivial or not.
908
909		To try this out, do the following:
910
911		- Erase all existing binfiles for your program.
912		(You may keep binfiles for those modules where you think you
913		definitely don't need back-tracing.)
914		- Turn on back-trace instrumentation:
915		SMLofNJ.Internals.BTrace.mode (SOME true);
916		- Recompile your program.  (I.e., run "CM.make" or "use".)
917		- You may now turn instrumentation off again (if you want):
918		SMLofNJ.Internals.BTrace.mode (SOME false);
919		- Run your program as usual.  If it raises an exception that
920		reaches the interactive toplevel, then a back-trace will
921		automatically be printed.  After that, the toplevel loop
922		will print the exception history as usual.
923
924		----------------------------------------------------------------------
925		Name: Matthias Blume
926		Date: 2000/06/26 09:56:46 JST
927		Tag: blume-20000626-setup
928		Description:
929
930		CM: - setup-parameter to "sml" added; this can be used to run arbitrary
931		ML code before and after compiling a file (e.g., to set compiler
932		flags)
933
934		Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
935		- associated changes to btrace.sml (BTrace instrumentation pass)
936		- cleaner implementation of btimp.sml (BTrace tracing and report
937		module)
938
939		CM manual: * new path encoding documented
940		* description of setup-parameter to "sml" added
941
942		The biggest user-visible change to back-tracing is that it is no
943		longer necessary to compile all traced modules within the same
944		session.  (This was a real limitation.)
945
946		----------------------------------------------------------------------
947		Name: Matthias Blume
948		Date: 2000/06/24 12:40:00 JST
949		Tag: blume-20000624-startup
950		Description:
951
952		Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
953		bit too often -- to put it mildly. :)
954
955		----------------------------------------------------------------------
956		Name: Matthias Blume
957		Date: 2000/06/23 18:20:00 JST
958		Tag: blume-20000623-btrace
959		Description:
960
961		This updates adds a backtrace facility to aid programmers in debugging
962		their programs.  This involves the following changes:
963
964		1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
965		keeping track of the current call stack.  When programs are compiled
966		in a special mode, the compiler will insert calls to these hooks
967		into the user program.
968		"Hook" means that it is possible for different implementations of
969		back-tracing to register themselves (at different times).
970
971		2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
972		as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
973		It can be turned on using this call:
974		SMLofNJ.Internals.BTrace.mode (SOME true);
975		Turning it off again:
976		SMLofNJ.Internals.BTrace.mode (SOME false);
977		Querying the current status:
978		SMLofNJ.Internals.BTrace.mode NONE;
979		Annotated programs are about twice as big as normal ones, and they
980		run a factor of 2 to 4 slower with a dummy back-trace plugin (one
981		where all hooks do nothing).  The slowdown with a plugin that is
982		actually useful (such as the one supplied by default) is even greater,
983		but in the case of the default plugin it is still only an constant
984		factor (amortized).
985
986		3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
987		with a sub-structure BTrace for controlling back-tracing.  In particular,
988		the above-mentioned function "mode" controls whether the annotation
989		phase is invoked by the compiler.  Another important function is
990		"trigger": when called it aborts the current execution and causes
991		the top-level loop to print a full back-trace.
992
993		4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
994		for back-tracing.  It keeps track of the dynamic call stack and in
995		addition to that it keeps a partial history at each "level" of that
996		stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
997		d and b (at separate times, dynamically), then the report will show:
998
999		GOTO   d
1000		/c
1001		GOTO  \b
1002		CALL   a
1003
1004		This shows that there was an initial non-tail call of a, then a
1005		tail-call to b or c, looping behavior in a cluster of functions that
1006		consist of b and c, and then a goto from that cluster (i.e., either from
1007		b or from c) to d.
1008
1009		Note that (depending on the user program) the amount of information
1010		that the back-trace module has to keep track of at each level is bounded
1011		by a constant.  Thus, the whole implementation has the same asymptotical
1012		complexity as the original program (both in space and in time).
1013
1014		5. compiler/TopLevel/interact/evalloop.sml has been modified to
1015		handle the special exception SMLofNJ.Internals.BTrace.BTrace
1016		which is raised by the "trigger" function mentioned above.
1017
1018		Notes on usage:
1019
1020		- Annotated code works well together with unannotated code:
1021		Unannotated calls simply do not show up at all in the backtrace.
1022
1023		- It is not a good idea to let modules that were annotated during
1024		different sessions run at the same time.  This is because the compiler
1025		chooses small integers to identify individual functions, and there
1026		will be clashes if different modules were compiled in separate sessions.
1027		(Nothing will crash, and you will even be told about the clashes, but
1028		back-trace information will in general not be useful.)
1029
1030		- Back-tracing can be confused by callcc and capture.
1031
1032		- The only way of getting a back-trace right now is to explicitly
1033		invoke the "trigger" function from your user program.  Eventually, we
1034		should make every exception carry back-trace information (if
1035		available).  But since this creates more overhead at "raise"-time
1036		(similar to the current exnHistory overhead), I have not yet
1037		implemented this.  (The implementation will be rather easy.)  With
1038		exceptions carrying back-trace information, this facility will be even
1039		more useful because users don't need to modify their programs...
1040
1041		- While it is possible to compile the compiler with back-trace
1042		annotations turned on (I did it to get some confidence in
1043		correctness), you must make absolutely sure that core.sml and
1044		btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1045		be compiled with annotation because there is no core access yet, but
1046		if you compile btimp.sml with annotation, then the system will go into
1047		an infinite recursion and crash.)
1048		Since CM currently does not know about BTrace, the only way to turn
1049		annotations on and off for different modules of the compiler is to
1050		interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1051		this is awkward and clumsy.
1052
1053		Sample sessions:
1054
1055		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1056		- SMLofNJ.Internals.BTrace.mode (SOME true);
1057		[autoloading]
1058		[autoloading done]
1059		val it = false : bool
1060		- structure X = struct
1061		-     fun main n = let
1062		-         fun a (x, 0) = d x
1063		-           | a (x, n) = b (x, n - 1)
1064		-         and b (x, n) = c (x, n)
1065		-         and c (x, n) = a (x, n)
1066		-         and d x = e (x, 3)
1067		-         and e (x, 0) = f x
1068		-           | e (x, n) = e (x, n - 1)
1069		-         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1070		-           | f n = n * g (n - 1)
1071		-         and g n = a (n, 3)
1072		-     in
1073		-         f n
1074		-     end
1075		- end;
1076		structure X : sig val main : int -> int end
1077		- X.main 3;
1078		*** BACK-TRACE ***
1079		GOTO   stdIn:4.2-13.20: X.main[2].f
1080		GOTO-( stdIn:4.2-13.20: X.main[2].e
1081		GOTO   stdIn:4.2-13.20: X.main[2].d
1082		/ stdIn:4.2-13.20: X.main[2].a
1083		| stdIn:4.2-13.20: X.main[2].b
1084		GOTO-\ stdIn:4.2-13.20: X.main[2].c
1085		CALL   stdIn:4.2-13.20: X.main[2].g
1086		GOTO   stdIn:4.2-13.20: X.main[2].f
1087		GOTO-( stdIn:4.2-13.20: X.main[2].e
1088		GOTO   stdIn:4.2-13.20: X.main[2].d
1089		/ stdIn:4.2-13.20: X.main[2].a
1090		| stdIn:4.2-13.20: X.main[2].b
1091		GOTO-\ stdIn:4.2-13.20: X.main[2].c
1092		CALL   stdIn:4.2-13.20: X.main[2].g
1093		GOTO   stdIn:4.2-13.20: X.main[2].f
1094		GOTO-( stdIn:4.2-13.20: X.main[2].e
1095		GOTO   stdIn:4.2-13.20: X.main[2].d
1096		/ stdIn:4.2-13.20: X.main[2].a
1097		| stdIn:4.2-13.20: X.main[2].b
1098		GOTO-\ stdIn:4.2-13.20: X.main[2].c
1099		CALL   stdIn:4.2-13.20: X.main[2].g
1100		GOTO   stdIn:4.2-13.20: X.main[2].f
1101		CALL   stdIn:2.15-17.4: X.main[2]
1102		-
1103
1104		(Note that because of a FLINt bug the above code currently does not
1105		compile without BTrace turned on.)
1106
1107		Here is another example, using my modified Tiger compiler:
1108
1109		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1110		- SMLofNJ.Internals.BTrace.mode (SOME true);
1111		[autoloading]
1112		[autoloading done]
1113		val it = false : bool
1114		- CM.make "sources.cm";
1115		[autoloading]
1116		...
1117		[autoloading done]
1118		[scanning sources.cm]
1119		[parsing (sources.cm):parse.sml]
1120		[creating directory CM/SKEL ...]
1121		[parsing (sources.cm):tiger.lex.sml]
1122		...
1123		[wrote CM/sparc-unix/semant.sml]
1124		[compiling (sources.cm):main.sml]
1125		[wrote CM/sparc-unix/main.sml]
1126		[New bindings added.]
1127		val it = true : bool
1128		- Main.compile ("../testcases/merge.tig", "foo.out");
1129		*** BACK-TRACE ***
1130		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1131		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1132		CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1133		GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1134		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1135		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1136		CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1137		/ lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1138		CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1139		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1140		CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1141		CALL   main.sml:1.18-118.4: Main.compile[2]
1142		-
1143
1144		----------------------------------------------------------------------
1145		Name: Matthias Blumen
1146		Date: 2000/06/21 18:00:00 JST
1147		Tag: blume-20000621-manual
1148		Description:
1149
1150		CM manual update: Path environments documented.
1151
1152		----------------------------------------------------------------------
1153		Name: Matthias Blume
1154		Date: 2000/06/19 13:40:00
1155		Tag: blume-20000619-manual
1156		Description:
1157
1158		CM manual and system/README update.  This only covers the fact that
1159		there are no more implicit anchors.  (Path environments and the "bind"
1160		option to "cm" have yet to be documented.)
1161
1162		----------------------------------------------------------------------
1163		Name: Matthias Blume
1164		Date: 2000/06/19 11:05:00 JST
1165		Tag: blume-20000619-chdir-bugfix
1166		Description:
1167
1168		Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1169
1170		----------------------------------------------------------------------
1171		Name: Matthias Blume
1172		Date: 2000/06/18 22:00:10 JST
1173		Tag: blume-20000618-implicit-anchors-really-gone
1174		Description:
1175
1176		I updates the previous HISTORY entry where I forgot to mention that
1177		implicit anchors are no longer with us.
1178
1179		The current update also gets rid of the (now useless) controller
1180		CM.Control.implicit_anchors.
1181
1182		----------------------------------------------------------------------
1183		Name: Matthias Blume
1184		Date: 2000/06/16 17:30:00 JST
1185		Tag: blume-20000616-anchorenv
1186		Description:
1187
1188		This patch implements the long anticipated (just kidding :) "anchor
1189		environment" mechanism.  In the course of doing this, I also
1190		re-implemented CM's internal "SrcPath" module from scratch.  The new
1191		one should be more robust in certain boundary cases.  In any case, it
1192		is a lot cleaner than its predecessor (IMHO).
1193
1194		This time, although there is yet another boot file format change, I
1195		kept the unpickler backward-compatible.  As a result, no new bootfiles
1196		are necessary and bootstrapping is straightforward.  (You cannot read
1197		new bootfiles into an old system, but the other way around is no
1198		problem.)
1199
1200		Visible changes:
1201
1202		** 0. Implicit path anchors (without the leading $-symbol) are no
1203		longer recognized at all. This means that such path names are not
1204		illegal either.  For example, the name basis.cm simply refers to a
1205		local file called "basis.cm" (i.e, the name is an ordinary path
1206		relative to .cm-files directory).  Or, to put it differently, only
1207		names that start with $ are anchored paths.
1208
1209		** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1210		vanished.
1211
1212		John (Reppy) had critizised this as soon as I originally proposed and
1213		implemented it, but at that time I did not really deeply believe
1214		him. :) Now I came full-circle because I need the $<singlearc> syntax
1215		in another place where it cannot be seen as an abbreviation for
1216		$/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1217		seems to mean (i.e., it "expands" into the corresponding anchor
1218		value).
1219
1220		However, when paths are used as members in CM description files, it
1221		continues to be true that there must be at least another arc after the
1222		anchor.  This is now enforced separately during semantic analysis
1223		(i.e., from a lexical/syntactical point of view, the notation is ok.)
1224
1225		** 2. The "cm" class now accepts an option "bind".  The option's value
1226		is a sub-option list of precisely two items -- one labeled "anchor"
1227		and the other one labeled "value".  As you might expect, "anchor" is
1228		used to specify an anchor name to be bound, and "value" specifies what
1229		the anchor is being bound to.
1230
1231		The value must be a directory name and can be given in either standard
1232		syntax (including the possibility that it is itself an anchored path)
1233		or native syntax.
1234
1235		Examples:
1236
1237		foo.cm (bind:(anchor:bar value:$mystuff/bar))
1238		lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1239
1240		and so on.
1241
1242		The meaning of this is that the .cm-file will be processed with an
1243		augmented anchor environment where the given anchor(s) is/are bound to
1244		the given values(s).
1245
1246		The rationale for having this feature is this: Suppose you are trying
1247		to use two different (already stable) libraries a.cm and b.cm (that
1248		you perhaps didn't write yourself).  Further, suppose each of these
1249		two libraries internally uses its own auxiliary library $aux/lib.cm.
1250		Normally you would now have a problem because the anchor "lib" can not
1251		be bound to more than one value globally.  Therefore, the project that
1252		uses both a.cm and b.cm must locally redirect the anchor to some other
1253		place:
1254
1255		a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1256		b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1257
1258		This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1259		/usr/lib/smlnj/b-stuff/aux.cm, respectively.
1260
1261		Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1262		you when you do so at the time of CM.stabilize).  Therefore, you can
1263		also use an anchored path as the value:
1264
1265		a.cm (bind:(anchor:lib value:$a-lib))
1266		b.cm (bind:(anchor:lib value:$b-lib))
1267
1268		Now you can globally configure (using the usual CM.Anchor.anchor or
1269		pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1270		itself is always locally bound, setting it globally is no longer
1271		meaningful or necessary (but it does not hurt either).  In fact, "lib"
1272		can still be used as a global anchor for separate purposes.  As a
1273		matter of fact, one can locally define "lib" in terms of a global
1274		"lib":
1275
1276		a.cm (bind:(anchor:lib value:$lib/a))
1277		b.cm (bind:(anchor:lib value:$lib/b))
1278
1279		** 3: The encoding of path names has changed.  This affects the way
1280		path names are shown in CM's progress report and also the internal
1281		protocol encoding used for parallel make.
1282
1283		The encoding now uses one or more ':'-separated segments.  Each
1284		segments corresponds to a file that has been specified relative to the
1285		file given by its preceding segment.  The first segment is either
1286		relative to the CWD, absolute, or anchored.  Each segment itself is
1287		basically a Unix pathname; all segments but the first are relative.
1288
1289		Example:
1290
1291		$foo/bar/baz.cm:a/b/c.sml
1292
1293		This path denotes the file bar/a/b/c.sml relative to the directory
1294		denoted by anchor "foo".  Notice that the encoding also includes
1295		baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1296		paths are resolved relative to the .cm-files directory, so baz.cm must
1297		be ignored to get the "real" pathname.
1298
1299		To make this fact more obvious, CM puts the names of such "virtual
1300		arcs" into parentheses when they appear in progress reports. (No
1301		parentheses will appear in the internal protocol encoding.)  Thus,
1302		what you really see is:
1303
1304		$foo/bar/(baz.cm):a/b/c.sml
1305
1306		I find this notation to be much more informative than before.
1307
1308		Another new feature of the encoding is that special characters
1309		including parentheses, colons, (back)slashes, and white space are
1310		written as \ddd (where ddd is the decimal encoding of the character).
1311
1312		*** The CM manual still needs to be updated.
1313
1314		----------------------------------------------------------------------
1315		Name: Allen Leung
1316		Date: 2000/06/15 00:38:00
1317		Tag: leunga-20000615-x86-peephole
1318
1319		x86 Peephole fix by Fermin.  Affects c-- and moby only.
1320
1321		----------------------------------------------------------------------
1322		Name: Matthias Blume
1323		Date: 2000/06/12 11:40:00
1324		Tag: blume-20000612-parmakefix
1325		Description:
1326
1327		More cleanup after changing the file naming scheme: This time I
1328		repaired the parallel make mechanism for CMB.make which I broke earlier.
1329
1330		----------------------------------------------------------------------
1331		Name: Allen Leung
1332		Date: 2000/06/09 01:25:00
1333		Tag: leunga-20000609-various
1334
1335		None of these things should affect normal SML/NJ operations
1336
1337		1. Peephole improvements provided by Fermin (c--)
1338		2. New annotation DEFUSE for adding extra dependence (moby)
1339		3. New X86 LOCK instructions (moby)
1340		4. New machine description language for reservation tables (scheduling)
1341		5. Fixes to various optimization/analysis modules (branch chaining, dominator
1342		trees etc.)
1343		6. I've changed the CM files so that they can work with versions
1344		110.0.6, 110.25 and 110.28
1345
1346		----------------------------------------------------------------------
1347		Name: Matthias Blume
1348		Date: 2000/06/09 12:40:00
1349		Tag: blume-20000609-log
1350		Description:
1351
1352		- Removed all(?) remaining RCS Log entries from sources.
1353
1354		- Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1355		anchored paths).
1356
1357		----------------------------------------------------------------------
1358		Name: Matthias Blume
1359		Date: 2000/06/07 17:00:00 JST
1360		Tag: blume-20000607-no-implicit-anchors
1361		Description:
1362
1363		1. This update changes the default setting for
1364		CM.Control.implicit_anchors from true to false.  This means that
1365		implicit anchors are no longer permitted by default.  I also tried to
1366		make sure that nothing else still relies on implicit anchors.
1367		(This is the next step on the schedule towards a CM that does not even
1368		have the notion of implicit anchors anymore.)
1369
1370		2. More CM manual updates.
1371
1372		3. I managed to track down and fix the pickling bug I mentioned last
1373		time.  Because of the previously existing workaround, this entails no
1374		immediate practical changes.
1375
1376		----------------------------------------------------------------------
1377		Name: Matthias Blume
1378		Date: 2000/06/06 11:15:00 JST
1379		Tag: blume-20000606-lazierpickle
1380		Description:
1381
1382		!!!! NEW BOOT FILES !!!!
1383
1384		* The main purpose of this update is to make library pickles lazier in
1385		order to reduce the initial space penalty for autoloading a library.
1386		As a result, it is now possible to have $smlnj/compiler.cm
1387		pre-registered.  This should take care of the many complaints or
1388		inquiries about missing structure Compiler.  This required changes to
1389		CM's internal data structures and small tweaks to some algorithms.
1390
1391		As a neat additional effect, it is no longer necessary (for the sake
1392		of lean heap image files) to distinguish between a "minimal" CM and a
1393		"full" CM.  Now, there is only one CM (i.e., the "full" version:
1394		$smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1395		interactive top level. ($smlnj/cm/minimal.cm is gone.)
1396
1397		To make the life of compiler-hackers easier, "makeml" now also
1398		pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1399		words, after you bootstrap a new sml for the first time, you will not
1400		have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1401		around you will still have to do it, though.)
1402
1403		* A second change consists of major updates to the CM manual.  There
1404		are now several appendices with summary information and also a full
1405		specification of the CM description file syntax.
1406
1407		* In directory src/system I added the script "allcross".  This script
1408		invokes sml and cross-compiles the compiler for all supported
1409		architectures.  (Useful when providing a new set of boot files.)
1410
1411		* There seems to be a latent bug in my "lazy pickles" mechanism.  I
1412		added a small tweak to pickle-util.sml to work around this problem,
1413		but it is not a proper fix yet.  I will investigate further.  (The
1414		effect of the bug was an inflation of library pickle size.)
1415
1416		* Version number increased to 110.28.1 (to avoid compatibility problems).
1417
1418		----------------------------------------------------------------------
1419		Name: Allen Leung
1420		Date: 2000/05/25 17:28 EDT
1421		Tag: leunga-20000525-ra
1422		Description:
1423
1424		Fixed a bug in freezing phase of the register allocator.
1425
1426		----------------------------------------------------------------------
1427		Name: Allen Leung
1428		Date: 2000/05/15 22:53 EDT
1429		Tag: leunga-20000515-alpha-x86-ra
1430		Description:
1431
1432		1. Alpha
1433
1434		Slight cleanup.  Removed the instruction SGNXL
1435
1436		2. X86
1437
1438		Added the following instructions to the instruction set:
1439
1440		ROLx, RORx,
1441		BTx, BTSx, BTLx, BTRx,
1442		XCHGx, and variants with the LOCK prefix
1443
1444		3. Register Allocation
1445
1446		The module ra-rewrite-with-renaming has been improved.
1447
1448		These have no effect on SML/NJ.
1449
1450		----------------------------------------------------------------------
1451		Name: Matthias Blume
1452		Date: 2000/05/15 16:20:00 JST
1453		Tag: blume-20000515-lightrebuild
1454		Description:
1455
1456		1. I added an alternative to "-rebuild" to "makeml".  The difference is
1457		that prior to calling CMB.make' the CM-variable "LIGHT" will be
1458		defined.  In effect, the command will not build any cross-compiler
1459		backends and therefore finish more quickly.
1460
1461		The "fixpt" script also takes a "-light" switch to be able to use
1462		this new facility while compiling for a fixpoint.
1463
1464		2. I replaced all mentions of anchored paths in group owner specifications
1465		with simple relative paths (usually starting with "..").
1466		The rationale is that a library's internal workings should not be
1467		compromised by the lack of some anchor.  (An anchor is necessary
1468		for someone who wants to refer to the library by an anchored path,
1469		but it should not be necessary to build the same library in the first
1470		place.)
1471
1472		3. I changed the way CM's tool mechanism determines the shell command
1473		string used for things like ml-yacc etc. so that it does not break
1474		when CM.Control.implicit_anchors is turned off.
1475
1476		----------------------------------------------------------------------
1477		Name: Matthias Blume
1478		Date: 2000/05/12 18:20:00 JST
1479		Tag: blume-20000512-ml-build
1480		Description:
1481
1482		Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1483		getting installed properly (by config/install.sh).
1484
1485		----------------------------------------------------------------------
1486		Name: Matthias Blume
1487		Date: 2000/05/12 17:30:00 JST
1488		Tag: blume-20000512-anchors
1489		Description:
1490
1491		!!! NEW BOOT FILES !!!
1492
1493		This change is in preparation of fading out support for "implicitly
1494		anchored path names".  I went through all sources and used the
1495		explicit (and relatively new) $-notation.  See system/README and the
1496		CM manual for more info on this.
1497
1498		I also modified the anchoring scheme for some things such as "smlnj",
1499		"MLRISC", "cm", etc. to take advantage of the fact that explicit
1500		anchors are more expressive: anchor name and first arc do not have to
1501		coincide.  This entails the following user-visible change:
1502
1503		You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1504		particular, when you fire up sml with a command-line argument, say,
1505		e.g.:
1506
1507		sml '$smlnj/cmb.cm'
1508
1509		At the ML toplevel prompt:
1510
1511		CM.autoload "$smlnj/cmb.cm";
1512
1513		There is also a new controller in CM.Control that can be used to turn
1514		off all remaining support for implicit anchors by saying:
1515
1516		CM.autoload "$smlnj/
1517		#set CM.Control.implicit_anchors false;
1518
1519		This causes CM to reject implicitly anchored paths.  This is (for the
1520		time being) less permissive than the "final" version where there will
1521		be no more such implicit anchors and relative paths will be just that:
1522		relative.
1523
1524		The next step (version after next version?) will be to make the
1525		default for CM.Control.implicit_anchors false.  After the dust has
1526		settled, I can then produce the "final" version of this...
1527
1528		Note: Since bootstrapping is a bit tricky, I provided new boot files.
1529
1530		----------------------------------------------------------------------
1531		Name: Matthias Blume
1532		Date: 2000/05/11 16:30:00 JST
1533		Tag: blume-20000511-sources
1534		Description:
1535
1536		The main change is that I added function CM.sources as a generalized
1537		version of the earlier CM.makedepend.  This entails the following
1538		additional changes:
1539
1540		- CM.makedepend has been dropped.
1541
1542		- CM manual has been updated.
1543
1544		- TOOLS signature and API have been changed.
1545
1546		----------------------------------------------------------------------
1547		Name: Allen Leung
1548		Date: 2000/05/10 21:17 EDT
1549		Tag: leunga-20000510-moby-c--ssa
1550		Description:
1551
1552		Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1553		None of these affect SML/NJ.
1554
1555		1. Register Allocation
1556
1557		a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1558		This module tries to remove local (i.e. basic block level) redundancies
1559		during spilling.
1560
1561		b. A new framework for performing region based register allocation.
1562		Not yet entirely functional.
1563
1564		2. X86
1565
1566		a. DefUse for POP was missing the stack pointer [found by Lal]
1567		b. Reload for CALL was incorrect in X86Spill [found by John]
1568		c. Various fixes in X86Spill so that it can be used correctly for
1569		the new spilling module.
1570
1571		3. SSA/IR
1572
1573		a. New module ir/dj-dataflow.sml implements elimination based
1574		data flow analysis.
1575
1576		4. MLRiscGen
1577
1578		a. Fix for gc type annotation
1579
1580		5. MDGen
1581
1582		Various fixes for machine description -> ml code translation.  For ssa
1583		only.
1584
1585		----------------------------------------------------------------------
1586		Name: Allen Leung
1587		Date: 2000/05/08 22:17 EDT
1588		Tag: leunga-20000508-labexp
1589		Description:
1590
1591		Fermin has found a few assembly problems with constant expressions
1592		generated in LabelExp.  Mostly, the problems involve extra parentheses,
1593		which choke on dumb assemblers.  This is his fix.
1594
1595		----------------------------------------------------------------------
1596		Name: Dave MacQueen
1597		Date: 2000/04/09 14:00 EDT
1598		Tag: dbm-20000502-Version_110_28
1599		Description:
1600
1601		1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1602
1603		2. Updated config/version to 110.28
1604
1605		3. Updated config/srcarchiveurl
1606
1607		3. New boot files!
1608		ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1609
1610		----------------------------------------------------------------------
1611		Name: Matthias Blume
1612		Date: 2000/05/01 19:05:00 JST
1613		Tag: blume-20000501-noweb
1614		Description:
1615
1616		A new noweb tool has been added.  The existing system is entirely
1617		unaffected by this, but some CM users have asked for renewed noweb
1618		support.  Everything is documented in the CM manual.
1619
1620		New (plugin) libraries:
1621
1622		noweb-tool.cm
1623		nw-ext.cm
1624
1625	----------------------------------------------------------------------	----------------------------------------------------------------------
1626	Name: Dave MacQueen	Name: Dave MacQueen
1627	Date: 2000/04/30 12:40PM EDT	Date: 2000/04/30 12:40PM EDT
#	Line 295	Line 1906
1906
1907	3.  Assembly	3.  Assembly
1908
1909	When generating assemby, resolve the value of client defined constants,	When generating assembly, resolve the value of client defined constants,
1910	instead of generating symbolic values.  This is controlled by the	instead of generating symbolic values.  This is controlled by the
1911	new flag "asm-resolve-constants", which is default to true.	new flag "asm-resolve-constants", which is default to true.
1912
#	Line 318	Line 1929
1929
1930	To this end, I arranged that instead of "structure Core" as "structure	To this end, I arranged that instead of "structure Core" as "structure
1931	_Core" is bound in the pervasive environment.  Core access is done via	_Core" is bound in the pervasive environment.  Core access is done via
1932	_Core (which can never be accidentially rebound because _Core is not a	_Core (which can never be accidentally rebound because _Core is not a
1933	legal surface-syntax symbol).	legal surface-syntax symbol).
1934
1935	The current solution is much cleaner because the core environment is	The current solution is much cleaner because the core environment is
#	Line 328	Line 1939
1939	with dynamic and symbolic parts of the core environment.	with dynamic and symbolic parts of the core environment.
1940
1941	Remaining hackery (to bind the "magic" symbol _Core) is localized in the	Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1942	compilation mananger's bootstrap compiler (actually: in the "init group"	compilation manager's bootstrap compiler (actually: in the "init group"
1943	handling).  See the comments in src/system/smlnj/init/init.cmi for	handling).  See the comments in src/system/smlnj/init/init.cmi for
1944	more details.	more details.
1945
#	Line 443	Line 2054
2054	(specified in the .cm file at each instance where the tool's class is	(specified in the .cm file at each instance where the tool's class is
2055	used).	used).
2056
2057	This was done to accomodate the new "make" and "shell" tools which	This was done to accommodate the new "make" and "shell" tools which
2058	facilitate fairly seemless hookup to portions of code managed using	facilitate fairly seamless hookup to portions of code managed using
2059	Makefiles or Shell scripts.	Makefiles or Shell scripts.
2060
2061	There are no classes "shared" or "private" anymore.  Instead, the	There are no classes "shared" or "private" anymore.  Instead, the
#	Line 456	Line 2067
2067
2068	All existing tools are described in the CM manual.	All existing tools are described in the CM manual.
2069
2070	- Slightly better error handling.  (CM now surpresses many followup	- Slightly better error handling.  (CM now suppresses many followup
2071	error messages that tended to be more annoying than helpful.)	error messages that tended to be more annoying than helpful.)
2072
2073	2. Major changes to the compiler's static environment data structures.	2. Major changes to the compiler's static environment data structures.
#	Line 590	Line 2201
2201
2202	I've changed andl to testl in the floating point test sequence	I've changed andl to testl in the floating point test sequence
2203	whenever appropriate.  The Intel optimization guide states that	whenever appropriate.  The Intel optimization guide states that
2204	testl is perferable to andl.	testl is preferable to andl.
2205
2206	7. RA (x86 only)	7. RA (x86 only)
2207
#	Line 772	Line 2383
2383
2384	1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an	1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2385	additional argument called "template" which is an optional string that	additional argument called "template" which is an optional string that
2386	specifiel the layout of the tool command line.  See the CM manual for	specifies the layout of the tool command line.  See the CM manual for
2387	explanation.	explanation.
2388
2389	2. A special-purpose tool can be "regisitered" by simply dropping the	2. A special-purpose tool can be "registered" by simply dropping the
2390	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2391	directory where the .cm file lives that uses this tool.  (The	directory where the .cm file lives that uses this tool.  (The
2392	behavior/misfeature until now was to look for the tool description	behavior/misfeature until now was to look for the tool description
#	Line 819	Line 2430
2430	rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for	rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
2431	<k>=1,2,... is normally not a good idea when invoking fixpt.  However,	<k>=1,2,... is normally not a good idea when invoking fixpt.  However,
2432	they might be the result of an earlier partial run of fixpt (which	they might be the result of an earlier partial run of fixpt (which
2433	perhaps got accidentially killed).  In this case, fixpt will quickly	perhaps got accidentally killed).  In this case, fixpt will quickly
2434	move through what exists before continuing where it left off earlier,	move through what exists before continuing where it left off earlier,
2435	and, thus, saves a lot of time.	and, thus, saves a lot of time.
2436
#	Line 869	Line 2480
2480	it from that remote directory.	it from that remote directory.
2481	This should simplify installation further:  For machines that have	This should simplify installation further:  For machines that have
2482	access to the internet, just fetch <version>-config.tgz, unpack it,	access to the internet, just fetch <version>-config.tgz, unpack it,
2483	edit config/targets, and go (run config/install.sh).  The scipt will	edit config/targets, and go (run config/install.sh).  The script will
2484	fetch everything else that it might need all by itself.	fetch everything else that it might need all by itself.
2485
2486	For CVS users, this mechanism is not relevant for source archives, but	For CVS users, this mechanism is not relevant for source archives, but

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.633   changed lines   Added in v.814

---

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