Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 652, Tue Jun 6 02:14:56 2000 UTC	revision 826, Wed May 9 20:32:17 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: 2001/05/09 16:35:00 EDT
18		Tag: blume-20010509-cpscontract
19		Description:
20
21		Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
22		wrapper/unwrapper elimination did not decrement usage counts and some
23		dead variables got overlooked by the dead-up logic.)
24
25		----------------------------------------------------------------------
26		Name: Lal George
27		Date: 2001/05/08  17:26:09 EDT
28		Tag: george-20010508-omit-frameptr
29		Description:
30
31		Changes to implement the omit-frame-pointer optimization to support
32		raw C calls. For now, there is only support on the Intel x86, but
33		other architectures will follow as more experience is gained with this.
34
35
36		----------------------------------------------------------------------
37		Name: Matthias Blume
38		Date: 2001/05/07 14:40:00 EDT
39		Tag: blume-20010507-proxies
40		Description:
41
42		I made into "proxy libraries" all libraries that qualify for such a
43		change.  (A qualifying library is a library that has another library or
44		groups as its sole member and repeats that member's export list
45		verbatim.  A proxy library avoids this repetition by omitting its export
46		list, effectively inheriting the list that its (only) member exports.
47		See the CM manual for more explanation.)
48		The main effect is that explicit export lists for these libraries
49		do not have to be kepts in sync, making maintenance a bit easier.
50
51		I also added copyright notices to many .cm-files.
52
53		Last but not least, I made a new set of bootfiles.
54
55		----------------------------------------------------------------------
56		Name: Matthias Blume
57		Date: 2001/05/04 17:00:00 EDT
58		Tag: blume-20010504-cm-lsplit
59		Description:
60
61		0. John merged pending changes to $/smlnj-lib.cm
62
63		1. Allen's previous change accidentally backed out of one of Lal's
64		earlier changes.  I undid this mistake (re-introducing Lal's change).
65
66		2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
67		within the compiler where applicable.  There is some code simplification
68		because of that.
69
70		3. The "split" phase (in FLINT) is now part of the default list of phases.
71		Compiler.Control.LambdaSplitting.* can be used to globally control the
72		lambda-splitting (cross-module-inlining) engine.  In addition to that,
73		it can now also be controlled on a per-source basis: CM has been taught
74		a new tool parameter applicable to ML source files.
75
76		- To turn lambda-splitting off completely:
77		local open Compiler.Control.LambdaSplitting in
78		val _ = set Off
79		end
80		- To make "no lambda-splitting" the global default (but allow per-source
81		overriding); this is the initial setting:
82		local open Compiler.Control.LambdaSplitting in
83		val _ = set (Default NONE)
84		end
85		- To make "lambda-splitting with aggressiveness a" the global default
86		(and allow per-source overriding):
87		local open Compiler.Control.LambdaSplitting in
88		val _ = set (Default (SOME a))
89		end
90
91		- To turn lambda-splitting off for a given ML souce file (say: a.sml)
92		write (in the respective .cm-file):
93		a.sml (lambdasplitting:off)
94		- To turn lambda-splitting for a.sml on with minimal aggressiveness:
95		a.sml (lambdasplitting:on)
96		- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
97		<a> is a decimal non-negative integer):
98		a.sml (lambdasplitting:<a>)
99		- To turn lambda-splitting for a.sml on with maximal aggressiveness:
100		a.sml (lambdasplitting:infinity)
101		- To use the global default for a.sml:
102		a.sml (lambdasplitting:default)
103		or simply
104		a.sml
105
106		----------------------------------------------------------------------
107		Name: Allen Leung
108		Date: 2001/05/04 01:57:00 EDT
109		Tag: leunga-20010504-sync
110		Description:
111
112		MLRISC features.
113
114		1. Fix to CMPXCHG instructions.
115		2. Changed RA interface to allow annotations in callbacks.
116		3. Added a new method to the stream interface to allow annotations updates.
117
118		----------------------------------------------------------------------
119		Name: Matthias Blume
120		Date: 2001/05/01 11:45:00 EDT
121		Tag: blume-20010501-pcedittmp
122		Description:
123
124		Changed install.sh to use the current working directory instead of
125		/usr/tmp for a temporary file (pcedittmp).  The previous choice
126		of /usr/tmp caused trouble with MacOS X because of file premission
127		problems.
128
129		----------------------------------------------------------------------
130		Name: Matthias Blume
131		Date: 2001/04/20 11:10:00 EDT
132		Tag: blume-20010420-inMLflag
133		Description:
134
135		- added vp_limitPtrMask to vproc-state.h
136		(for use by the raw-C-calls mechanism to implement proper interrupt
137		handling)
138		- made the ML compiler aware of various data-structure offsets so it
139		can generate code for accessing the vp_inML flag and vp_limitPtrMask
140		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
141		raw C-calls
142
143		----------------------------------------------------------------------
144		Name: Lal George
145		Date: 2001/04/20 09:15:28 EDT
146		Tag: george-20010420-macosX
147		Description:
148
149		- Changes to port to Mac OS X; Darwin.
150
151		- In the process I found that sqrt was broken on the PPC, because the
152		fsqrt instruction is not implemented.
153
154		----------------------------------------------------------------------
155		Name: Matthias Blume
156		Date: 2001/04/18 12:45:00 EDT
157		Tag: blume-20010418-ccalls
158		Description:
159
160		- fixed two off-by-4 errors in the x86-specific c-calls implementation
161		(this bug prevented structure arguments containing pointers from being
162		passed correctly)
163		- changed the raw-C-call code in mlriscGen.sml in such a way that
164		structure arguments are represented as a pointer to the beginning
165		of the structure (instead of having a series of synthesized arguments,
166		one for each structure member)
167
168		- made makeml script's verbosity level configurable via environment
169		variable (MAKEML_VERBOSITY)
170
171		- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
172		in rawmem-x86.sml; we are now using the real thing
173
174		----------------------------------------------------------------------
175		Name: Matthias Blume
176		Date: 2001/03/22 16:25:00 EST
177		Tag: blume-20010322-bootfiles
178		Description:
179
180		Created a new set of bootfiles (for your automatic installation convenience).
181
182		----------------------------------------------------------------------
183		Name: Matthias Blume
184		Date: 2001/03/22 15:10:00 EST
185		Tag: blume-20010322-rawmem-parcm
186		Description:
187
188		1. All "raw memory access" primitives for the new FFI are implemented now
189		(at least on the x86).
190		2. Some further cleanup of CM's parallel make mechanism.
191
192		----------------------------------------------------------------------
193		Name: Matthias Blume
194		Date: 2001/03/19 17:53:00 EST
195		Tag: blume-20010319-parallel
196		Description:
197
198		Parallel make (using compile servers) now works again.
199
200		To this end, CM.stabilize and CMB.make have been modified to work in
201		two passes when compile servers are attached:
202		1. Compile everything, do not perform stabilization; this pass
203		uses compile servers
204		2. Stabilize everything; this pass does not use compile servers
205		If there are no compile servers, the two passes are combined into one
206		(as before).  Splitting the passes increases the inherent parallelism
207		in the dependency graph because the entire graph including all
208		libraries is available at the same time.  This, in turn, improves
209		server utilization.  The downside is that the master process will
210		have to do some extra work after compilation is done (because for
211		technical reasons it must re-read all the binfiles during stabilization).
212
213		----------------------------------------------------------------------
214		Name: Matthias Blume
215		Date: 2001/03/16 12:22:00 EST
216		Tag: blume-20010316-bootfiles
217		Description:
218
219		Created a new set of bootfiles (for your automatic installation convenience).
220
221		----------------------------------------------------------------------
222		Name: Matthias Blume
223		Date: 2001/03/16 11:00:00 EST
224		Tag: blume-20010316-MLTREE-fixup
225		Description:
226
227		This is a minor fixup for an (untagged) earlier commit by Allen.
228		(A file was missing).
229
230		----------------------------------------------------------------------
231		Name: Allen Leung
232		Date: Mon Mar  5 18:54:57 EST 2001
233		Tag: leunga-20010305-cut-support
234
235		1. New support for alternative control-flow in MLTREE.
236		Currently we support
237
238		FLOW_TO(CALL ...., [k1,...,kn])
239
240		This is needed for 'cuts to' in C-- and try/handle-like constructs
241		in Moby
242
243		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
244
245		2. Register Allocator
246
247		Changes in interface [from Fermin, John]
248
249		3. Alpha 8-bit SLL support [Fermin]
250
251		4. All architectures
252
253		A new module (ClusterExpandCopies) for expanding parallel copies.
254
255		----------------------------------------------------------------------
256		Name: Allen Leung
257		Date: 2001/02/27 23:07:00 EST
258		Tag: leunga-20010227-minor-stuff
259
260		1. Alpha bug fix for CMOVNE
261		2. Handle mltree COND(..,FCMP ...,...)
262		3. Bug fix in simplifier
263
264		----------------------------------------------------------------------
265		Name: Matthias Blume
266		Date: 2001/01/30 17:50:00 EST
267		Tag: blume-20010130-sync
268		Description:
269
270		This is just a minor update to sync my devel branch with the main brach.
271		The only visible change is the addition of some README files.
272
273		----------------------------------------------------------------------
274		Name: Matthias Blume
275		Date: 2001/01/12 23:30:00 JST
276		Tag: blume-20010112-bootfiles
277		Description:
278
279		Made a new set of bootfiles that goes with the current state of the
280		repository.
281
282		----------------------------------------------------------------------
283		Name: Matthias Blume
284		Date: 2001/01/12 21:20:00 JST
285		Tag: blume-20010112-sync
286		Description:
287
288		I am just flushing out some minor changes that had accumulated in
289		my private branch in order to sync with the main tree.  (This is
290		mainly because I had CVS trouble when trying to merge _into_ my
291		private branch.)
292
293		Most people should be completely unaffected by this.
294
295		----------------------------------------------------------------------
296		Name: Allen Leung
297		Date: Thu Jan 11 21:03:00 EST 2001
298		Tag: leunga-20010111-labexp=mltree
299		Description:
300
301		1.  Removed the type LabelExp and replace it by MLTree.
302		2.  Rewritten mltree-simplify with the pattern matcher tool.
303		3.  There were some bugs in alpha code generator which would break
304		64-bit code generation.
305		4.  Redo the tools to generate code with the
306		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
307		are now generated by perl scripts.
308
309		----------------------------------------------------------------------
310		Name: Matthias Blume
311		Date: 2001/01/10 21:55:00 JST
312		Tag: blume-20010110-rcc
313		Description:
314
315		The RCC stuff now seems to work (but only on the x86).
316		This required hacking of the c-calls interface (and -implementation) in
317		MLRISC.
318
319		Normal compiler users should be unaffected.
320
321		----------------------------------------------------------------------
322		Name: Matthias Blume
323		Date: 2001/01/09 01:20:00 JST
324		Tag: blume-20010109-rcc
325		Description:
326
327		This is a fairly big patch, flushing out a large number of pending
328		changes that I made to my development copy over the last couple of days.
329
330		Of practical relevance at this moment is a workaround for a pickling
331		bug that Allen ran into the other day.  The cause of the bug itself is
332		still unknown and it might be hard to fix it properly, but the
333		workaround has some merits of its own (namely somewhat reducing pickling
334		overhead for certain libraries).  Therefore, I think this solution should
335		be satisfactory at this time.
336
337		The rest of the changes (i.e., the vast majority) has to do with my
338		ongoing efforts of providing direct support for C function calls from
339		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
340		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
341		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
342		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
343		handle RCC.
344
345		The last part (the changes to mlriscGen.sml) are still known to be
346		wrong on the x86 and not implemented on all other architectures.  But
347		the infrastructure is in place. I had to change a few functor
348		signatures in the backend to be able to route the CCalls interface
349		from MLRISC there, and I had to specialize the mltree type (on the
350		x86) to include the necessary extensions. (The extensions themselves
351		were already there and redy to go in MLRISC/x86).
352
353		Everything should be very happy as soon as someone helps me with
354		mlriscGen.sml...
355
356		In any case, nothing of this should matter to anyone as long as the
357		new primop is not being used (which is going to be the case unless you
358		find it where I hid it :). The rest of the compiler is completely
359		unaffected.
360
361		----------------------------------------------------------------------
362		Name: Matthias Blume
363		Date: 2001/01/05 00:30:00 JST
364		Tag: blume-20010105-primops
365		Description:
366
367		Added some experimental support for work that I am doing right now.
368		These changes mostly concern added primops, but there is also a new
369		experimental C library in the runtime system (but currently not enabled
370		anywhere except on Linux/X86).
371
372		In the course of adding primops (and playing with them), I discovered that
373		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
374		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
375		To recover, I made type information in INL_PRIM mandatory and changed
376		prim.sml as well as built-in.sml accordingly.  The InLine structure now
377		has complete, correct type information (i.e., no bottom types).
378
379		Since all these changes mean that we need new binfiles, I also bumped the
380		version number to 110.32.1.
381
382		----------------------------------------------------------------------
383		Name: Matthias Blume
384		Date: 2000/12/30 22:10:00 JST
385		Tag: blume-20001230-various
386		Description:
387
388		Added proxy libraries for MLRISC and let MLRISC libraries refer
389		to each other using path anchors.  (See CM manual for explanation.)
390
391		Updated CM documentation.
392
393		Fixed some bugs in CM.
394
395		Implemented "proxy" libraries (= syntactic sugar for CM).
396
397		Added "-quiet" option to makeml and changed runtime system accordingly.
398
399		Added cleanup handler for exportML to reset timers and compiler stats.
400
401		----------------------------------------------------------------------
402		Name: Lal George
403		Date: 2000/12/22 22:22:58 EST 2000
404		Tag: Release_110_32
405		Description:
406
407		Infinite precision used throughout MLRISC.
408		see MLRISC/mltree/machine-int.sig
409
410		----------------------------------------------------------------------
411		Name: Matthias Blume
412		Date: 2000/12/22 23:16:00 JST
413		Tag: blume-20001222-warn
414		Description:
415
416		Corrected wording and formatting of some CM warning message which I
417		broke in my previous patch.
418
419		----------------------------------------------------------------------
420		Name: Matthias Blume
421		Date: 2000/12/22 21:20:00 JST
422		Tag: blume-20001222-anchorenv
423		Description:
424
425		Fixed CM's handling of anchor environments in connection with CMB.make.
426
427		----------------------------------------------------------------------
428		Name: Matthias Blume
429		Date: 2000/12/22 13:15:00 JST
430		Tag: blume-20001222-cleanup
431		Description:
432
433		Removed src/cm/ffi which does not (and did not) belong here.
434
435		----------------------------------------------------------------------
436		Name: Matthias Blume
437		Date: 2000/12/21 23:55:00 JST
438		Tag: blume-20001221-exn
439		Description:
440
441		Probably most important: CM no longer silently swallows all exceptions
442		in the compiler.
443		Plus: some other minor CM changes.  For example, CM now reports some
444		sizes for generated binfiles (code, data, envpickle, lambdapickle).
445
446		----------------------------------------------------------------------
447		Name: Matthias Blume
448		Date: 2000/12/15 00:01:05 JST
449		Tag: blume-20001215-dirtool
450		Description:
451
452		- "dir" tool added.
453		- improvements and cleanup to Tools structure
454		- documentation updates
455
456		----------------------------------------------------------------------
457		Name: Allen Leung
458		Date: Thu Dec 14 03:45:24 EST 2000
459		Description:
460		Tag:  leunga-20001214-int-inf
461		Description:
462
463		In IntInf, added these standard functions, which are missing from our
464		implementation:
465
466		andb : int * int -> int
467		xorb : int * int -> int
468		orb  : int * int -> int
469		notb : int -> int
470		<<   : int * word -> int
471		~>>  : int * word -> int
472
473		Not tested, I hope they are correct.
474
475		----------------------------------------------------------------------
476		Name: Allen Leung
477		Date: Fri Dec  8 19:23:26 EST 2000
478		Description:
479		Tag:  leunga-20001208-nowhere
480		Description:
481
482		Slight improvements to the 'nowhere' tool to handle OR-patterns,
483		to generate better error messages etc.  Plus a brief manual.
484
485		----------------------------------------------------------------------
486		Name: Lal George
487		Date: 2000/12/08 09:54:02 EST 2000
488		Tag: Release_110_31
489		Description:
490
491		- Version 110.31
492		----------------------------------------------------------------------
493		Name: Allen Leung
494		Date: Thu Dec  7 22:01:04 EST 2000
495		Tag:  leunga-20001207-cell-monster-hack
496		Description:
497
498		Major MLRISC internal changes.  Affect all clients.
499		Summary:
500
501		1.  Type CELLS.cell = int is now replaced by a datatype.
502		As a result, the old regmap is now gone.  Almost all interfaces
503		in MLRISC change as a consequence.
504
505		2.  A new brand version of machine description tool (v3.0) that generates
506		modules expecting the new interface.  The old version is removed.
507
508		3.  The RA interface has been further abstracted into two new functors.
509		RISC_RA and X86RA.  These functors have much simpler interfaces.
510		[See also directory MLRISC/demo.]
511
512		4.  Some other new source->source code generation tools are available:
513
514		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
515		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
516		I use this tool to generate the peephole optimizers---with the new
517		cell type changes, peephole rules are becoming difficult to write
518		without conditional pattern matching.
519
520		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
521		cover the entire MLRISC source tree so many things broke.
522
523		6.  CM files have been moved to the subdirectory MLRISC/cm.
524		They are moved because there are a lot of them and they clutter up the
525		root dir.
526
527		7.  More detailed documentation to come...
528
529		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
530		a makeml -rebuild first.  This is because of other other
531		changes that Matthias has made (see below).
532
533
534		----------------------------------------------------------------------
535		Name: Matthias Blume
536		Date: 2000/11/30 23:12:00 JST
537		Tag: blume-20001130-filereorg
538		Description:
539
540		Some manual updates and some file reorganizations in CM.
541
542		----------------------------------------------------------------------
543		Name: Matthias Blume
544		Date: 2000/11/24 17:45:00 JST
545		Tag: blume-20001124-link
546		Description:
547
548		Drastically improved link traversal code for the case that the dynamic
549		value was already loaded at bootstrap time.  As a result, CM and CMB
550		now both load blazingly fast -- even on a very slow machine.  Also,
551		memory consumption has been further reduced by this.
552
553		Warning: The format of the PIDMAP file has changed.  THerefore, to
554		bootstrap you have to do this:
555
556		1. Run CMB.make
557		2. Make a symbolic link for the boot directory:
558		ln -s sml.boot.ARCH-OS xxx
559		3. "Rebuild" the boot directory:
560		./makeml -boot xxx -rebuild sml ; rm xxx
561		4. Boot normally:
562		./makeml
563
564		----------------------------------------------------------------------
565		Name: Matthias Blume
566		Date: 2000/11/21 21:20:00 JST
567		Tag: blume-20001121-tools
568		Description:
569
570		Continued hacking on autoloading problem -- with success this time.
571		Also changed tool-plugin mechanism.  See new CM manual.
572
573		----------------------------------------------------------------------
574		Name: Matthias Blume
575		Date: 2000/11/19 14:30:00 JST
576		Tag:  blume-20001119-autoload
577		Description:
578
579		Some hacking to make autoloading faster.  Success for CMB, no success
580		so far for CM.  There is a reduced structure CM' that autoloads faster.
581		(This is a temporary, non-documented hack to be eliminated again when
582		the general problem is solved.)
583
584		----------------------------------------------------------------------
585		Name: Matthias Blume
586		Date: 2000/11/17 14:10:00 JST
587		Tag: blume-20001117-pickle-lib
588		Description:
589
590		1. Eliminated comp-lib.cm
591		2. Made pickle-lib.cm
592		3. Eliminated all uses of intset.sml (from comp-lib.cm)
593		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
594		equivalent constructs from smlnj-lib.cm (INtHashTable).
595		5. Point 4. also goes for those uses of intmap.* in MLRISC.
596		Duplicated intmap modules thrown out.
597		6. Hunted down all duplicated SCC code and replaced it with
598		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
599		7. Rewrote Feedback module.
600		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
601		should be thrown out and equivalent modules from smlnj-lib.cm
602		should be used (IntRedBlackSet, IntListSet, ...).
603
604		Confirmed that compiler compiles to fixpoint.
605
606		----------------------------------------------------------------------
607		Name: Allen Leung
608		Date: 2000/11/10 18:00:00
609		Tag: leunga-20001110-new-x86-fp
610
611		A new x86 floating point code generator has been added.
612		By default this is turned off.  To turn this on, do:
613
614		CM.autoload "$smlnj/compiler.cm";
615		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
616
617		Changes:
618
619		1.  Changed FTAN to FPTAN so that the assembly output is correct.
620		2.  Changed the extension callback for FTANGENT to generate:
621
622		fptan
623		fstp  %st(0)
624		instead of
625		fptan
626		fstpl ftempmem
627
628		3.  Numerous assembly fixes for x86.
629
630		5.  Cleaned up the machine code output module x86/x86MC.sml and added
631		support for a whole bunch of instructions and addressing modes:
632
633		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
634		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
635		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
636		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
637		fxch %st(n)
638		fld %st(n)
639		fst %st(n)
640		fst mem
641		fstp %st(n)
642		fucom %st(n)
643		fucomp %st(n)
644
645		All these are now generated when the fast fp mode is turned on.
646
647		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
648
649		----------------------------------------------------------------------
650		Name: Matthias Blume
651		Date: 2000/11/09 11:20:00 JST
652		Tag: blume-20001109-scc
653		Description:
654
655		Eliminated some code duplication:
656
657		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
658		(Otherwise the functor is useless.)
659		2. Used GraphSCCFn where SCCUtilFun was used previously.
660		3. Got rid of SCCUtilFun (in comp-lib.cm).
661
662		----------------------------------------------------------------------
663		Name: Lal George
664		Date: 2000/11/06 09:02:21 EST 2000
665		Tag: Release_110_30
666		Description:
667
668		- Version 110.30
669		----------------------------------------------------------------------
670		Name: Matthias Blume
671		Date: 2000/11/04 14:45:00
672		Tag: blume-20001104-mlbuild
673		Description:
674
675		- Made ml-build faster on startup.
676		- Documentation fixes.
677
678		----------------------------------------------------------------------
679		Name: Matthias Blume
680		Date: 2000/11/02 17:00:00 JST
681		Tag: blume-20001102-condcomp
682		Description:
683
684		- Small tweaks to pickler -- new BOOTFILES!
685		- Version bumped to 110.29.2.
686		- Added conditional compilation facility to init.cmi (see comment there).
687		----------------------------------------------------------------------
688		Name: Allen Leung
689		Date: 2000/10/23 19:31:00
690		Tag: leunga-20001023-demo-ra
691
692		1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
693		2. Test programs for the graph library updated
694		3. Some new MLRISC demo programs added
695
696		----------------------------------------------------------------------
697		Name: Matthias Blume
698		Date: 2000/08/31 22:15:00 JST
699		Tag: blume-20001017-errmsg
700		Description:
701
702		More error message grief: Where there used to be no messages, there
703		now were some that had bogus error regions.  Fixed.
704
705		----------------------------------------------------------------------
706		Name: Matthias Blume
707		Date: 2000/08/31 17:30:00 JST
708		Tag: blume-20001017-v110p29p1
709		Description:
710
711		I made a version 110.29.1 with new bootfiles.
712
713		Changes:  Modified pickler/unpickler for faster and leaner unpickling.
714		CM documentation changes and a small bugfix in CM's error reporting.
715
716		----------------------------------------------------------------------
717		Name: Lal George
718		Date: 2000/09/27 14:42:35 EDT
719		Tag: george-20000927-nodestatus
720		Description:
721
722		Changed the type of the nodestatus, so that:
723
724		SPILLED(~1)             is now SPILLED
725		SPILLED(m) where m>=0   is now MEMREG(m)
726		SPILLED(s) where s<~1   is now SPILL_LOC(~s)
727
728		----------------------------------------------------------------------
729		Name: Matthias Blume
730		Date: 2000/09/07 14:45:00 JST
731		Tag: blume-20000907-cmerrmsg
732		Description:
733
734		Small tweak to CM to avoid getting ML syntax error messages twice.
735
736		----------------------------------------------------------------------
737		Name: Matthias Blume
738		Date: 2000/08/31 18:00:00 JST
739		Tag: blume-20000831-cvsbootfiles
740		Description:
741
742		New URL for boot files (because the 110.29 files on the BL server do
743		now work correctly with my updated install scripts for yacc and lex).
744
745		----------------------------------------------------------------------
746		Name: Matthias Blume
747		Date: 2000/08/08 12:33:00 JST
748		Tag: blume-20000808-manual
749		Description:
750
751		Tiny update to CM manual.
752
753		----------------------------------------------------------------------
754		Name: Allen Leung
755		Date: 2000/08/7 19:31:00
756		Tag: leunga-20000807-a-whole-bunch-of-stuff
757
758		Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
759		and Mobdy.
760
761		1.  x86
762
763		a.  Fixes to peephole module by John and Dan.
764		b.  Assembly fix to SETcc by Allen.
765		c.  Fix to c-call by John.
766		d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
767		e.  Instruction selection optimization to SETcc as suggested by John.
768
769		For example,
770
771		MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
772
773		should generate:
774
775		MOVL a, x
776		SUBL b, x
777		SHRL 31, x
778
779		2.  IR stuff
780
781		A bunch of new DJ-graph related algorithms added.  These
782		speed up SSA construction.
783
784		3.  SSA + Scheduling
785
786		Added code for SSA and scheduling to the repository
787
788		----------------------------------------------------------------------
789		Name: Lal George
790		Date: 2000/07/27 11:53:14 EDT
791
792		Tag: lal-20000727-linux-ppc
793		Description:
794
795		Made changes to support Linux PPC.
796		p.s. I have confirmation that the 110.29 boot files work fine.
797
798		----------------------------------------------------------------------
799		Name: Matthias Blume
800		Date: 2000/07/27 17:40:00 JST
801		Tag: blume-20000727-scripts
802		Description:
803
804		!!!! WARNING !!!!
805		You must recompile the runtime system!
806		!!!! WARNING !!!!
807
808		This is basically another round of script-enhancements:
809
810		1. sml, ml-build, and ml-makedepend accept options -D and -U to define
811		and undefine CM preprocessor symbols.
812
813		2. ml-build avoids generating a new heap image if it finds that the
814		existing one is still ok.  (The condition is that no ML file had to
815		be recompiled and all ML files are found to be older that the heap
816		file.)
817
818		To make this work smoothly, I also hacked the runtime system as
819		well as SMLofNJ.SysInfo to get access to the heap image suffix
820		(.sparc-solaris, ...) that is currently being used.
821
822		Moreover, the signature of CM.mk_standalone has changed.  See the
823		CM manual.
824
825		3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
826		CM manual for details.)
827
828		4. More CM manual updates:
829		- all of the above has been documented.
830		- there is now a section describing the (CM-related) command line
831		arguments that are accepted by the "sml" command
832
833		----------------------------------------------------------------------
834		Name: Matthias Blume
835		Date: 2000/07/25 16:20:00 JST
836		Tag: blume-20000725-makedepend
837		Description:
838
839		Added a script called ml-makedepend.  This can be used in makefiles
840		for Unix' make in a way very similar to the "makedepend" command for
841		C.
842
843		The script internally uses function CM.sources.
844
845		Synopsis:
846
847		ml-makedepend [-f makefile] cmfile targetname
848
849		The default for the makefile is "makefile" (or "Makefile" should
850		"makefile" not exist).
851
852		ml-makedepend adds a cmfile/targetname-specific section to this
853		makefile (after removing the previous version of this section).  The
854		section contains a single dependency specification with targetname on
855		the LHS (targetname is an arbitrary name), and a list of files derived
856		from the cmfile on the RHS.  Some of the files on the RHS are
857		ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
858		"make" variables $(ARCH) and $(OPSYS) in place of the corresponding
859		path names.  The makefile writer is responsible for making sure that
860		these variables have correct at the time "make" is invoked.
861
862		----------------------------------------------------------------------
863		Name: Matthias Blume
864		Date: 2000/07/22 23:30:00 JST
865		Tag: blume-20000722-urlupdate
866		Description:
867
868		Changed BOOT and config/srcarchiveurl to point to BL server:
869
870		ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
871
872		----------------------------------------------------------------------
873		Name: Matthias Blume
874		Date: 2000/07/18 18:00:00 JST
875		Tag: blume-20000718-Version_110_29
876		Description:
877
878		1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
879
880		2. Updated config/version to 110.29
881
882		3. Updated config/srcarchiveurl
883
884		3. New boot files!
885		ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
886
887		----------------------------------------------------------------------
888		Name: Matthias Blume
889		Date: 2000/07/11 13:58:00 JST
890		Tag: blume-20000711-doctypo
891		Description:
892
893		Fixed a few typos in CM manual.
894
895		----------------------------------------------------------------------
896		Name: Allen Leung
897		Date: 2000/06/15 00:38:00
898		Tag: leunga-20000704-sparc-x86
899
900		1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
901		2. fix to x86 RET bug [found by Dan Grossman]
902		3. sparc assembly bug fix for ticc instructions [found by Fermin]
903
904		Affects c-- and moby only
905
906		----------------------------------------------------------------------
907		Name: Matthias Blume
908		Date: 2000/07/04 15:26:00
909		Tag: blume-20000704-trigger
910		Description:
911
912		1. Improvements to CM manual.
913		2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
914		of getting a back-trace.  The function, when called, raises an
915		internal exception which explicitly carries the full back-trace history,
916		so it is unaffected by any intervening handle-raise pairs ("trivial"
917		or not).  The interactive loop will print that history once it arrives
918		at top level.
919		Short of having all exceptions implicitly carry the full history, the
920		recommended way of using this facility is:
921		- compile your program with instrumentation "on"
922		- run it, when it raises an exception, look at the history
923		- if the history is "cut off" because of some handler, go and modify
924		your program so that it explicitly calls BTrace.trigger
925		- recompile (still instrumented), and rerun; look at the full history
926
927		----------------------------------------------------------------------
928		Name: Matthias Blume
929		Date: 2000/07/03 15:36:00 JST
930		Tag: blume-20000702-manual
931		Description:
932
933		Small corrections and updates to CM manual.
934
935		----------------------------------------------------------------------
936		Name: Matthias Blume
937		Date: 2000/06/29 16:04:00 JST
938		Tag: blume-20000629-yacctool
939		Description:
940
941		Changes:
942
943		1. Class "mlyacc" now takes separate arguments to pass options to
944		generated .sml- and .sig-files independently.
945		2. Corresponding CM manual updates.
946		3. BTrace module now also reports call sites.  (However, for loop clusters
947		it only shows from where the cluster was entered.)  There are associated
948		modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
949
950		----------------------------------------------------------------------
951		Name: Matthias Blume
952		Date: 2000/06/27 16:51:00 JST
953		Tag: blume-20000627-noweb
954		Description:
955
956		Changes:
957
958		1. Implemented "subdir" and "witness" options for noweb tool.
959		This caused some slight internal changes in CM's tool implementation.
960		2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
961		some remaining issues from earlier path anchor changes.
962		3. Updated CM manual accordingly.
963
964		4. Changed implementation of back-tracing so that I now consider it
965		ready for prime-time.
966
967		In particular, you don't have to explicitly trigger the back-trace
968		anymore.  Instead, if you are running BTrace-instrumented code and
969		there is an uncaught exception (regardless of whether or not it was
970		raised in instrumented code), the top-level evalloop will print
971		the back-trace.
972
973		Features:
974
975		- Instrumented and uninstrumented code work together seemlessly.
976		(Of course, uninstrumented code is never mentioned in actual
977		back-traces.)
978
979		- Asymptotic time- and space-complexity of instrumented code is
980		equal to that of uninstrumented code.  (This means that
981		tail-recursion is preserved by the instrumentation phase.)
982
983		- Modules whose code has been instrumented in different sessions
984		work together without problem.
985
986		- There is no penalty whatsoever on uninstrumented code.
987
988		- There is no penalty on "raise" expressions, even in
989		instrumented code.
990
991		A potential bug (or perhaps it is a feature, too):
992
993		A back-trace reaches no further than the outermost instrumented
994		non-trivial "raise".  Here, a "trivial" raise is one that is the
995		sole RHS of a "handle" rule.  Thus, back-traces reach trough
996
997		<exp> handle e => raise e
998
999		and even
1000
1001		<exp> handle Foo => raise Bar
1002
1003		and, of course, through
1004
1005		<exp> handle Foo => ...
1006
1007		if the exception was not Foo.
1008
1009		Back-traces always reach right through any un-instrumented code
1010		including any of its "handle" expressions, trivial or not.
1011
1012		To try this out, do the following:
1013
1014		- Erase all existing binfiles for your program.
1015		(You may keep binfiles for those modules where you think you
1016		definitely don't need back-tracing.)
1017		- Turn on back-trace instrumentation:
1018		SMLofNJ.Internals.BTrace.mode (SOME true);
1019		- Recompile your program.  (I.e., run "CM.make" or "use".)
1020		- You may now turn instrumentation off again (if you want):
1021		SMLofNJ.Internals.BTrace.mode (SOME false);
1022		- Run your program as usual.  If it raises an exception that
1023		reaches the interactive toplevel, then a back-trace will
1024		automatically be printed.  After that, the toplevel loop
1025		will print the exception history as usual.
1026
1027		----------------------------------------------------------------------
1028		Name: Matthias Blume
1029		Date: 2000/06/26 09:56:46 JST
1030		Tag: blume-20000626-setup
1031		Description:
1032
1033		CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1034		ML code before and after compiling a file (e.g., to set compiler
1035		flags)
1036
1037		Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1038		- associated changes to btrace.sml (BTrace instrumentation pass)
1039		- cleaner implementation of btimp.sml (BTrace tracing and report
1040		module)
1041
1042		CM manual: * new path encoding documented
1043		* description of setup-parameter to "sml" added
1044
1045		The biggest user-visible change to back-tracing is that it is no
1046		longer necessary to compile all traced modules within the same
1047		session.  (This was a real limitation.)
1048
1049		----------------------------------------------------------------------
1050		Name: Matthias Blume
1051		Date: 2000/06/24 12:40:00 JST
1052		Tag: blume-20000624-startup
1053		Description:
1054
1055		Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1056		bit too often -- to put it mildly. :)
1057
1058		----------------------------------------------------------------------
1059		Name: Matthias Blume
1060		Date: 2000/06/23 18:20:00 JST
1061		Tag: blume-20000623-btrace
1062		Description:
1063
1064		This updates adds a backtrace facility to aid programmers in debugging
1065		their programs.  This involves the following changes:
1066
1067		1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1068		keeping track of the current call stack.  When programs are compiled
1069		in a special mode, the compiler will insert calls to these hooks
1070		into the user program.
1071		"Hook" means that it is possible for different implementations of
1072		back-tracing to register themselves (at different times).
1073
1074		2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1075		as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1076		It can be turned on using this call:
1077		SMLofNJ.Internals.BTrace.mode (SOME true);
1078		Turning it off again:
1079		SMLofNJ.Internals.BTrace.mode (SOME false);
1080		Querying the current status:
1081		SMLofNJ.Internals.BTrace.mode NONE;
1082		Annotated programs are about twice as big as normal ones, and they
1083		run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1084		where all hooks do nothing).  The slowdown with a plugin that is
1085		actually useful (such as the one supplied by default) is even greater,
1086		but in the case of the default plugin it is still only an constant
1087		factor (amortized).
1088
1089		3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1090		with a sub-structure BTrace for controlling back-tracing.  In particular,
1091		the above-mentioned function "mode" controls whether the annotation
1092		phase is invoked by the compiler.  Another important function is
1093		"trigger": when called it aborts the current execution and causes
1094		the top-level loop to print a full back-trace.
1095
1096		4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1097		for back-tracing.  It keeps track of the dynamic call stack and in
1098		addition to that it keeps a partial history at each "level" of that
1099		stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1100		d and b (at separate times, dynamically), then the report will show:
1101
1102		GOTO   d
1103		/c
1104		GOTO  \b
1105		CALL   a
1106
1107		This shows that there was an initial non-tail call of a, then a
1108		tail-call to b or c, looping behavior in a cluster of functions that
1109		consist of b and c, and then a goto from that cluster (i.e., either from
1110		b or from c) to d.
1111
1112		Note that (depending on the user program) the amount of information
1113		that the back-trace module has to keep track of at each level is bounded
1114		by a constant.  Thus, the whole implementation has the same asymptotical
1115		complexity as the original program (both in space and in time).
1116
1117		5. compiler/TopLevel/interact/evalloop.sml has been modified to
1118		handle the special exception SMLofNJ.Internals.BTrace.BTrace
1119		which is raised by the "trigger" function mentioned above.
1120
1121		Notes on usage:
1122
1123		- Annotated code works well together with unannotated code:
1124		Unannotated calls simply do not show up at all in the backtrace.
1125
1126		- It is not a good idea to let modules that were annotated during
1127		different sessions run at the same time.  This is because the compiler
1128		chooses small integers to identify individual functions, and there
1129		will be clashes if different modules were compiled in separate sessions.
1130		(Nothing will crash, and you will even be told about the clashes, but
1131		back-trace information will in general not be useful.)
1132
1133		- Back-tracing can be confused by callcc and capture.
1134
1135		- The only way of getting a back-trace right now is to explicitly
1136		invoke the "trigger" function from your user program.  Eventually, we
1137		should make every exception carry back-trace information (if
1138		available).  But since this creates more overhead at "raise"-time
1139		(similar to the current exnHistory overhead), I have not yet
1140		implemented this.  (The implementation will be rather easy.)  With
1141		exceptions carrying back-trace information, this facility will be even
1142		more useful because users don't need to modify their programs...
1143
1144		- While it is possible to compile the compiler with back-trace
1145		annotations turned on (I did it to get some confidence in
1146		correctness), you must make absolutely sure that core.sml and
1147		btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1148		be compiled with annotation because there is no core access yet, but
1149		if you compile btimp.sml with annotation, then the system will go into
1150		an infinite recursion and crash.)
1151		Since CM currently does not know about BTrace, the only way to turn
1152		annotations on and off for different modules of the compiler is to
1153		interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1154		this is awkward and clumsy.
1155
1156		Sample sessions:
1157
1158		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1159		- SMLofNJ.Internals.BTrace.mode (SOME true);
1160		[autoloading]
1161		[autoloading done]
1162		val it = false : bool
1163		- structure X = struct
1164		-     fun main n = let
1165		-         fun a (x, 0) = d x
1166		-           | a (x, n) = b (x, n - 1)
1167		-         and b (x, n) = c (x, n)
1168		-         and c (x, n) = a (x, n)
1169		-         and d x = e (x, 3)
1170		-         and e (x, 0) = f x
1171		-           | e (x, n) = e (x, n - 1)
1172		-         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1173		-           | f n = n * g (n - 1)
1174		-         and g n = a (n, 3)
1175		-     in
1176		-         f n
1177		-     end
1178		- end;
1179		structure X : sig val main : int -> int end
1180		- X.main 3;
1181		*** BACK-TRACE ***
1182		GOTO   stdIn:4.2-13.20: X.main[2].f
1183		GOTO-( stdIn:4.2-13.20: X.main[2].e
1184		GOTO   stdIn:4.2-13.20: X.main[2].d
1185		/ stdIn:4.2-13.20: X.main[2].a
1186		| stdIn:4.2-13.20: X.main[2].b
1187		GOTO-\ stdIn:4.2-13.20: X.main[2].c
1188		CALL   stdIn:4.2-13.20: X.main[2].g
1189		GOTO   stdIn:4.2-13.20: X.main[2].f
1190		GOTO-( stdIn:4.2-13.20: X.main[2].e
1191		GOTO   stdIn:4.2-13.20: X.main[2].d
1192		/ stdIn:4.2-13.20: X.main[2].a
1193		| stdIn:4.2-13.20: X.main[2].b
1194		GOTO-\ stdIn:4.2-13.20: X.main[2].c
1195		CALL   stdIn:4.2-13.20: X.main[2].g
1196		GOTO   stdIn:4.2-13.20: X.main[2].f
1197		GOTO-( stdIn:4.2-13.20: X.main[2].e
1198		GOTO   stdIn:4.2-13.20: X.main[2].d
1199		/ stdIn:4.2-13.20: X.main[2].a
1200		| stdIn:4.2-13.20: X.main[2].b
1201		GOTO-\ stdIn:4.2-13.20: X.main[2].c
1202		CALL   stdIn:4.2-13.20: X.main[2].g
1203		GOTO   stdIn:4.2-13.20: X.main[2].f
1204		CALL   stdIn:2.15-17.4: X.main[2]
1205		-
1206
1207		(Note that because of a FLINt bug the above code currently does not
1208		compile without BTrace turned on.)
1209
1210		Here is another example, using my modified Tiger compiler:
1211
1212		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1213		- SMLofNJ.Internals.BTrace.mode (SOME true);
1214		[autoloading]
1215		[autoloading done]
1216		val it = false : bool
1217		- CM.make "sources.cm";
1218		[autoloading]
1219		...
1220		[autoloading done]
1221		[scanning sources.cm]
1222		[parsing (sources.cm):parse.sml]
1223		[creating directory CM/SKEL ...]
1224		[parsing (sources.cm):tiger.lex.sml]
1225		...
1226		[wrote CM/sparc-unix/semant.sml]
1227		[compiling (sources.cm):main.sml]
1228		[wrote CM/sparc-unix/main.sml]
1229		[New bindings added.]
1230		val it = true : bool
1231		- Main.compile ("../testcases/merge.tig", "foo.out");
1232		*** BACK-TRACE ***
1233		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1234		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1235		CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1236		GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1237		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1238		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1239		CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1240		/ lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1241		CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1242		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1243		CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1244		CALL   main.sml:1.18-118.4: Main.compile[2]
1245		-
1246
1247		----------------------------------------------------------------------
1248		Name: Matthias Blumen
1249		Date: 2000/06/21 18:00:00 JST
1250		Tag: blume-20000621-manual
1251		Description:
1252
1253		CM manual update: Path environments documented.
1254
1255		----------------------------------------------------------------------
1256		Name: Matthias Blume
1257		Date: 2000/06/19 13:40:00
1258		Tag: blume-20000619-manual
1259		Description:
1260
1261		CM manual and system/README update.  This only covers the fact that
1262		there are no more implicit anchors.  (Path environments and the "bind"
1263		option to "cm" have yet to be documented.)
1264
1265		----------------------------------------------------------------------
1266		Name: Matthias Blume
1267		Date: 2000/06/19 11:05:00 JST
1268		Tag: blume-20000619-chdir-bugfix
1269		Description:
1270
1271		Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1272
1273		----------------------------------------------------------------------
1274		Name: Matthias Blume
1275		Date: 2000/06/18 22:00:10 JST
1276		Tag: blume-20000618-implicit-anchors-really-gone
1277		Description:
1278
1279		I updates the previous HISTORY entry where I forgot to mention that
1280		implicit anchors are no longer with us.
1281
1282		The current update also gets rid of the (now useless) controller
1283		CM.Control.implicit_anchors.
1284
1285		----------------------------------------------------------------------
1286		Name: Matthias Blume
1287		Date: 2000/06/16 17:30:00 JST
1288		Tag: blume-20000616-anchorenv
1289		Description:
1290
1291		This patch implements the long anticipated (just kidding :) "anchor
1292		environment" mechanism.  In the course of doing this, I also
1293		re-implemented CM's internal "SrcPath" module from scratch.  The new
1294		one should be more robust in certain boundary cases.  In any case, it
1295		is a lot cleaner than its predecessor (IMHO).
1296
1297		This time, although there is yet another boot file format change, I
1298		kept the unpickler backward-compatible.  As a result, no new bootfiles
1299		are necessary and bootstrapping is straightforward.  (You cannot read
1300		new bootfiles into an old system, but the other way around is no
1301		problem.)
1302
1303		Visible changes:
1304
1305		** 0. Implicit path anchors (without the leading $-symbol) are no
1306		longer recognized at all. This means that such path names are not
1307		illegal either.  For example, the name basis.cm simply refers to a
1308		local file called "basis.cm" (i.e, the name is an ordinary path
1309		relative to .cm-files directory).  Or, to put it differently, only
1310		names that start with $ are anchored paths.
1311
1312		** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1313		vanished.
1314
1315		John (Reppy) had critizised this as soon as I originally proposed and
1316		implemented it, but at that time I did not really deeply believe
1317		him. :) Now I came full-circle because I need the $<singlearc> syntax
1318		in another place where it cannot be seen as an abbreviation for
1319		$/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1320		seems to mean (i.e., it "expands" into the corresponding anchor
1321		value).
1322
1323		However, when paths are used as members in CM description files, it
1324		continues to be true that there must be at least another arc after the
1325		anchor.  This is now enforced separately during semantic analysis
1326		(i.e., from a lexical/syntactical point of view, the notation is ok.)
1327
1328		** 2. The "cm" class now accepts an option "bind".  The option's value
1329		is a sub-option list of precisely two items -- one labeled "anchor"
1330		and the other one labeled "value".  As you might expect, "anchor" is
1331		used to specify an anchor name to be bound, and "value" specifies what
1332		the anchor is being bound to.
1333
1334		The value must be a directory name and can be given in either standard
1335		syntax (including the possibility that it is itself an anchored path)
1336		or native syntax.
1337
1338		Examples:
1339
1340		foo.cm (bind:(anchor:bar value:$mystuff/bar))
1341		lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1342
1343		and so on.
1344
1345		The meaning of this is that the .cm-file will be processed with an
1346		augmented anchor environment where the given anchor(s) is/are bound to
1347		the given values(s).
1348
1349		The rationale for having this feature is this: Suppose you are trying
1350		to use two different (already stable) libraries a.cm and b.cm (that
1351		you perhaps didn't write yourself).  Further, suppose each of these
1352		two libraries internally uses its own auxiliary library $aux/lib.cm.
1353		Normally you would now have a problem because the anchor "lib" can not
1354		be bound to more than one value globally.  Therefore, the project that
1355		uses both a.cm and b.cm must locally redirect the anchor to some other
1356		place:
1357
1358		a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1359		b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1360
1361		This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1362		/usr/lib/smlnj/b-stuff/aux.cm, respectively.
1363
1364		Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1365		you when you do so at the time of CM.stabilize).  Therefore, you can
1366		also use an anchored path as the value:
1367
1368		a.cm (bind:(anchor:lib value:$a-lib))
1369		b.cm (bind:(anchor:lib value:$b-lib))
1370
1371		Now you can globally configure (using the usual CM.Anchor.anchor or
1372		pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1373		itself is always locally bound, setting it globally is no longer
1374		meaningful or necessary (but it does not hurt either).  In fact, "lib"
1375		can still be used as a global anchor for separate purposes.  As a
1376		matter of fact, one can locally define "lib" in terms of a global
1377		"lib":
1378
1379		a.cm (bind:(anchor:lib value:$lib/a))
1380		b.cm (bind:(anchor:lib value:$lib/b))
1381
1382		** 3: The encoding of path names has changed.  This affects the way
1383		path names are shown in CM's progress report and also the internal
1384		protocol encoding used for parallel make.
1385
1386		The encoding now uses one or more ':'-separated segments.  Each
1387		segments corresponds to a file that has been specified relative to the
1388		file given by its preceding segment.  The first segment is either
1389		relative to the CWD, absolute, or anchored.  Each segment itself is
1390		basically a Unix pathname; all segments but the first are relative.
1391
1392		Example:
1393
1394		$foo/bar/baz.cm:a/b/c.sml
1395
1396		This path denotes the file bar/a/b/c.sml relative to the directory
1397		denoted by anchor "foo".  Notice that the encoding also includes
1398		baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1399		paths are resolved relative to the .cm-files directory, so baz.cm must
1400		be ignored to get the "real" pathname.
1401
1402		To make this fact more obvious, CM puts the names of such "virtual
1403		arcs" into parentheses when they appear in progress reports. (No
1404		parentheses will appear in the internal protocol encoding.)  Thus,
1405		what you really see is:
1406
1407		$foo/bar/(baz.cm):a/b/c.sml
1408
1409		I find this notation to be much more informative than before.
1410
1411		Another new feature of the encoding is that special characters
1412		including parentheses, colons, (back)slashes, and white space are
1413		written as \ddd (where ddd is the decimal encoding of the character).
1414
1415		*** The CM manual still needs to be updated.
1416
1417		----------------------------------------------------------------------
1418		Name: Allen Leung
1419		Date: 2000/06/15 00:38:00
1420		Tag: leunga-20000615-x86-peephole
1421
1422		x86 Peephole fix by Fermin.  Affects c-- and moby only.
1423
1424		----------------------------------------------------------------------
1425		Name: Matthias Blume
1426		Date: 2000/06/12 11:40:00
1427		Tag: blume-20000612-parmakefix
1428		Description:
1429
1430		More cleanup after changing the file naming scheme: This time I
1431		repaired the parallel make mechanism for CMB.make which I broke earlier.
1432
1433		----------------------------------------------------------------------
1434		Name: Allen Leung
1435		Date: 2000/06/09 01:25:00
1436		Tag: leunga-20000609-various
1437
1438		None of these things should affect normal SML/NJ operations
1439
1440		1. Peephole improvements provided by Fermin (c--)
1441		2. New annotation DEFUSE for adding extra dependence (moby)
1442		3. New X86 LOCK instructions (moby)
1443		4. New machine description language for reservation tables (scheduling)
1444		5. Fixes to various optimization/analysis modules (branch chaining, dominator
1445		trees etc.)
1446		6. I've changed the CM files so that they can work with versions
1447		110.0.6, 110.25 and 110.28
1448
1449		----------------------------------------------------------------------
1450		Name: Matthias Blume
1451		Date: 2000/06/09 12:40:00
1452		Tag: blume-20000609-log
1453		Description:
1454
1455		- Removed all(?) remaining RCS Log entries from sources.
1456
1457		- Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1458		anchored paths).
1459
1460		----------------------------------------------------------------------
1461		Name: Matthias Blume
1462		Date: 2000/06/07 17:00:00 JST
1463		Tag: blume-20000607-no-implicit-anchors
1464		Description:
1465
1466		1. This update changes the default setting for
1467		CM.Control.implicit_anchors from true to false.  This means that
1468		implicit anchors are no longer permitted by default.  I also tried to
1469		make sure that nothing else still relies on implicit anchors.
1470		(This is the next step on the schedule towards a CM that does not even
1471		have the notion of implicit anchors anymore.)
1472
1473		2. More CM manual updates.
1474
1475		3. I managed to track down and fix the pickling bug I mentioned last
1476		time.  Because of the previously existing workaround, this entails no
1477		immediate practical changes.
1478
1479	----------------------------------------------------------------------	----------------------------------------------------------------------
1480	Name: Matthias Blume	Name: Matthias Blume
1481	Date: 2000/06/06 11:15:00 JST	Date: 2000/06/06 11:15:00 JST
#	Line 544	Line 2009
2009
2010	3.  Assembly	3.  Assembly
2011
2012	When generating assemby, resolve the value of client defined constants,	When generating assembly, resolve the value of client defined constants,
2013	instead of generating symbolic values.  This is controlled by the	instead of generating symbolic values.  This is controlled by the
2014	new flag "asm-resolve-constants", which is default to true.	new flag "asm-resolve-constants", which is default to true.
2015
#	Line 567	Line 2032
2032
2033	To this end, I arranged that instead of "structure Core" as "structure	To this end, I arranged that instead of "structure Core" as "structure
2034	_Core" is bound in the pervasive environment.  Core access is done via	_Core" is bound in the pervasive environment.  Core access is done via
2035	_Core (which can never be accidentially rebound because _Core is not a	_Core (which can never be accidentally rebound because _Core is not a
2036	legal surface-syntax symbol).	legal surface-syntax symbol).
2037
2038	The current solution is much cleaner because the core environment is	The current solution is much cleaner because the core environment is
#	Line 577	Line 2042
2042	with dynamic and symbolic parts of the core environment.	with dynamic and symbolic parts of the core environment.
2043
2044	Remaining hackery (to bind the "magic" symbol _Core) is localized in the	Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2045	compilation mananger's bootstrap compiler (actually: in the "init group"	compilation manager's bootstrap compiler (actually: in the "init group"
2046	handling).  See the comments in src/system/smlnj/init/init.cmi for	handling).  See the comments in src/system/smlnj/init/init.cmi for
2047	more details.	more details.
2048
#	Line 692	Line 2157
2157	(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
2158	used).	used).
2159
2160	This was done to accomodate the new "make" and "shell" tools which	This was done to accommodate the new "make" and "shell" tools which
2161	facilitate fairly seemless hookup to portions of code managed using	facilitate fairly seamless hookup to portions of code managed using
2162	Makefiles or Shell scripts.	Makefiles or Shell scripts.
2163
2164	There are no classes "shared" or "private" anymore.  Instead, the	There are no classes "shared" or "private" anymore.  Instead, the
#	Line 705	Line 2170
2170
2171	All existing tools are described in the CM manual.	All existing tools are described in the CM manual.
2172
2173	- Slightly better error handling.  (CM now surpresses many followup	- Slightly better error handling.  (CM now suppresses many followup
2174	error messages that tended to be more annoying than helpful.)	error messages that tended to be more annoying than helpful.)
2175
2176	2. Major changes to the compiler's static environment data structures.	2. Major changes to the compiler's static environment data structures.
#	Line 839	Line 2304
2304
2305	I've changed andl to testl in the floating point test sequence	I've changed andl to testl in the floating point test sequence
2306	whenever appropriate.  The Intel optimization guide states that	whenever appropriate.  The Intel optimization guide states that
2307	testl is perferable to andl.	testl is preferable to andl.
2308
2309	7. RA (x86 only)	7. RA (x86 only)
2310
#	Line 1021	Line 2486
2486
2487	1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an	1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2488	additional argument called "template" which is an optional string that	additional argument called "template" which is an optional string that
2489	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
2490	explanation.	explanation.
2491
2492	2. A special-purpose tool can be "regisitered" by simply dropping the	2. A special-purpose tool can be "registered" by simply dropping the
2493	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2494	directory where the .cm file lives that uses this tool.  (The	directory where the .cm file lives that uses this tool.  (The
2495	behavior/misfeature until now was to look for the tool description	behavior/misfeature until now was to look for the tool description
#	Line 1068	Line 2533
2533	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
2534	<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,
2535	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
2536	perhaps got accidentially killed).  In this case, fixpt will quickly	perhaps got accidentally killed).  In this case, fixpt will quickly
2537	move through what exists before continuing where it left off earlier,	move through what exists before continuing where it left off earlier,
2538	and, thus, saves a lot of time.	and, thus, saves a lot of time.
2539
#	Line 1118	Line 2583
2583	it from that remote directory.	it from that remote directory.
2584	This should simplify installation further:  For machines that have	This should simplify installation further:  For machines that have
2585	access to the internet, just fetch <version>-config.tgz, unpack it,	access to the internet, just fetch <version>-config.tgz, unpack it,
2586	edit config/targets, and go (run config/install.sh).  The scipt will	edit config/targets, and go (run config/install.sh).  The script will
2587	fetch everything else that it might need all by itself.	fetch everything else that it might need all by itself.
2588
2589	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.652   changed lines   Added in v.826

---

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