Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 676, Sat Jun 24 03:37:03 2000 UTC	revision 836, Fri May 25 19:28:51 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: 2001/05/25 15:30:00 EDT
18		Tag: blume-20010525-iptr
19		Description:
20
21		- put the official 110.33-README (as it appears on the ftp server) under
22		CVS
23		- fixed a small bug related to incomplete pointer types in
24		ml-nlffigen
25		- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
26		(it does not need the 'f type parameter)
27
28		----------------------------------------------------------------------
29		Name: Matthias Blume
30		Date: 2001/05/23 14:30:00 EDT
31		Tag: Release_110_33
32		Description:
33
34		New version number (110.33).  New bootfiles.
35
36		----------------------------------------------------------------------
37		Name: Matthias Blume
38		Date: 2001/05/22 18:06:00 EDT
39		Tag: blume-20010522-targets
40		Description:
41
42		Made install.sh use file config/targets.customized if it exists, falling
43		back to config/targets if it doesn't.  This way one can have a customized
44		version of the targets file without touching the "real thing", thus
45		eliminating the constant fear of accidentally checking something bogus
46		back into the CVS repository...  (File config/targets.customized must
47		not be added to the repository!)
48
49		----------------------------------------------------------------------
50		Name: Matthias Blume
51		Date: 2001/05/22 16:30:00 EDT
52		Tag: blume-20010522-minitut
53		Description:
54
55		1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
56		struct returns.
57		2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
58		incomplete, preliminary documentation for NLFFI.
59
60		----------------------------------------------------------------------
61		Name: Matthias Blume
62		Date: 2001/05/14 11:30:00 EDT
63		Tag: blume-20010514-script
64		Description:
65
66		Some bugs in install script fixed.
67
68		In addition to that I also made a slight change to the NLFFI API:
69		Functors generated by ml-nlffigen now take the dynamic library as a
70		straight functor argument, not as a suspended one.  (The original
71		functor code used to force the suspension right away anyway, so there
72		was nothing gained by this complication of the interface.)
73
74		----------------------------------------------------------------------
75		Name: Matthias Blume
76		Date: 2001/05/11 14:35:00 EDT
77		Tag: blume-20010511-ml-nlffi
78		Description:
79
80		I finally took the plunge and added my new FFI code to the main
81		repository.  For x86-linux it is now ready for prime-time.
82
83		There are two new subdirectories of "src":
84
85		- ml-nlffi-lib:
86		The utility library for programs using the FFI interface.
87		Here is the implementation of $/c.cm and its associated low-level
88		partners $/c-int.cm and $/memory.cm.
89		- ml-nlffigen:
90		A stand-alone program for generating ML glue code from C source
91		code.
92
93		Building ml-nlffigen requires $/ckit-lib.cm.
94
95		The config/install.sh script has been updates to do the Right Thing
96		(hopefully).
97
98		Notice that the source tree for the C-Kit will not be put under "src"
99		but directly under the installation root directory.  (This is the
100		structure that currently exists on the CVS server when you check out
101		module "sml".)  Fortunately, config/install.sh knows about this oddity.
102
103		Bugs: No documentation yet.
104
105		----------------------------------------------------------------------
106		Name: Matthias Blume
107		Date: 2001/05/09 16:35:00 EDT
108		Tag: blume-20010509-cpscontract
109		Description:
110
111		Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
112		wrapper/unwrapper elimination did not decrement usage counts and some
113		dead variables got overlooked by the dead-up logic.)
114
115		----------------------------------------------------------------------
116		Name: Lal George
117		Date: 2001/05/08  17:26:09 EDT
118		Tag: george-20010508-omit-frameptr
119		Description:
120
121		Changes to implement the omit-frame-pointer optimization to support
122		raw C calls. For now, there is only support on the Intel x86, but
123		other architectures will follow as more experience is gained with this.
124
125
126		----------------------------------------------------------------------
127		Name: Matthias Blume
128		Date: 2001/05/07 14:40:00 EDT
129		Tag: blume-20010507-proxies
130		Description:
131
132		I made into "proxy libraries" all libraries that qualify for such a
133		change.  (A qualifying library is a library that has another library or
134		groups as its sole member and repeats that member's export list
135		verbatim.  A proxy library avoids this repetition by omitting its export
136		list, effectively inheriting the list that its (only) member exports.
137		See the CM manual for more explanation.)
138		The main effect is that explicit export lists for these libraries
139		do not have to be kepts in sync, making maintenance a bit easier.
140
141		I also added copyright notices to many .cm-files.
142
143		Last but not least, I made a new set of bootfiles.
144
145		----------------------------------------------------------------------
146		Name: Matthias Blume
147		Date: 2001/05/04 17:00:00 EDT
148		Tag: blume-20010504-cm-lsplit
149		Description:
150
151		0. John merged pending changes to $/smlnj-lib.cm
152
153		1. Allen's previous change accidentally backed out of one of Lal's
154		earlier changes.  I undid this mistake (re-introducing Lal's change).
155
156		2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
157		within the compiler where applicable.  There is some code simplification
158		because of that.
159
160		3. The "split" phase (in FLINT) is now part of the default list of phases.
161		Compiler.Control.LambdaSplitting.* can be used to globally control the
162		lambda-splitting (cross-module-inlining) engine.  In addition to that,
163		it can now also be controlled on a per-source basis: CM has been taught
164		a new tool parameter applicable to ML source files.
165
166		- To turn lambda-splitting off completely:
167		local open Compiler.Control.LambdaSplitting in
168		val _ = set Off
169		end
170		- To make "no lambda-splitting" the global default (but allow per-source
171		overriding); this is the initial setting:
172		local open Compiler.Control.LambdaSplitting in
173		val _ = set (Default NONE)
174		end
175		- To make "lambda-splitting with aggressiveness a" the global default
176		(and allow per-source overriding):
177		local open Compiler.Control.LambdaSplitting in
178		val _ = set (Default (SOME a))
179		end
180
181		- To turn lambda-splitting off for a given ML souce file (say: a.sml)
182		write (in the respective .cm-file):
183		a.sml (lambdasplitting:off)
184		- To turn lambda-splitting for a.sml on with minimal aggressiveness:
185		a.sml (lambdasplitting:on)
186		- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
187		<a> is a decimal non-negative integer):
188		a.sml (lambdasplitting:<a>)
189		- To turn lambda-splitting for a.sml on with maximal aggressiveness:
190		a.sml (lambdasplitting:infinity)
191		- To use the global default for a.sml:
192		a.sml (lambdasplitting:default)
193		or simply
194		a.sml
195
196		----------------------------------------------------------------------
197		Name: Allen Leung
198		Date: 2001/05/04 01:57:00 EDT
199		Tag: leunga-20010504-sync
200		Description:
201
202		MLRISC features.
203
204		1. Fix to CMPXCHG instructions.
205		2. Changed RA interface to allow annotations in callbacks.
206		3. Added a new method to the stream interface to allow annotations updates.
207
208		----------------------------------------------------------------------
209		Name: Matthias Blume
210		Date: 2001/05/01 11:45:00 EDT
211		Tag: blume-20010501-pcedittmp
212		Description:
213
214		Changed install.sh to use the current working directory instead of
215		/usr/tmp for a temporary file (pcedittmp).  The previous choice
216		of /usr/tmp caused trouble with MacOS X because of file premission
217		problems.
218
219		----------------------------------------------------------------------
220		Name: Matthias Blume
221		Date: 2001/04/20 11:10:00 EDT
222		Tag: blume-20010420-inMLflag
223		Description:
224
225		- added vp_limitPtrMask to vproc-state.h
226		(for use by the raw-C-calls mechanism to implement proper interrupt
227		handling)
228		- made the ML compiler aware of various data-structure offsets so it
229		can generate code for accessing the vp_inML flag and vp_limitPtrMask
230		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
231		raw C-calls
232
233		----------------------------------------------------------------------
234		Name: Lal George
235		Date: 2001/04/20 09:15:28 EDT
236		Tag: george-20010420-macosX
237		Description:
238
239		- Changes to port to Mac OS X; Darwin.
240
241		- In the process I found that sqrt was broken on the PPC, because the
242		fsqrt instruction is not implemented.
243
244		----------------------------------------------------------------------
245		Name: Matthias Blume
246		Date: 2001/04/18 12:45:00 EDT
247		Tag: blume-20010418-ccalls
248		Description:
249
250		- fixed two off-by-4 errors in the x86-specific c-calls implementation
251		(this bug prevented structure arguments containing pointers from being
252		passed correctly)
253		- changed the raw-C-call code in mlriscGen.sml in such a way that
254		structure arguments are represented as a pointer to the beginning
255		of the structure (instead of having a series of synthesized arguments,
256		one for each structure member)
257
258		- made makeml script's verbosity level configurable via environment
259		variable (MAKEML_VERBOSITY)
260
261		- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
262		in rawmem-x86.sml; we are now using the real thing
263
264		----------------------------------------------------------------------
265		Name: Matthias Blume
266		Date: 2001/03/22 16:25:00 EST
267		Tag: blume-20010322-bootfiles
268		Description:
269
270		Created a new set of bootfiles (for your automatic installation convenience).
271
272		----------------------------------------------------------------------
273		Name: Matthias Blume
274		Date: 2001/03/22 15:10:00 EST
275		Tag: blume-20010322-rawmem-parcm
276		Description:
277
278		1. All "raw memory access" primitives for the new FFI are implemented now
279		(at least on the x86).
280		2. Some further cleanup of CM's parallel make mechanism.
281
282		----------------------------------------------------------------------
283		Name: Matthias Blume
284		Date: 2001/03/19 17:53:00 EST
285		Tag: blume-20010319-parallel
286		Description:
287
288		Parallel make (using compile servers) now works again.
289
290		To this end, CM.stabilize and CMB.make have been modified to work in
291		two passes when compile servers are attached:
292		1. Compile everything, do not perform stabilization; this pass
293		uses compile servers
294		2. Stabilize everything; this pass does not use compile servers
295		If there are no compile servers, the two passes are combined into one
296		(as before).  Splitting the passes increases the inherent parallelism
297		in the dependency graph because the entire graph including all
298		libraries is available at the same time.  This, in turn, improves
299		server utilization.  The downside is that the master process will
300		have to do some extra work after compilation is done (because for
301		technical reasons it must re-read all the binfiles during stabilization).
302
303		----------------------------------------------------------------------
304		Name: Matthias Blume
305		Date: 2001/03/16 12:22:00 EST
306		Tag: blume-20010316-bootfiles
307		Description:
308
309		Created a new set of bootfiles (for your automatic installation convenience).
310
311		----------------------------------------------------------------------
312		Name: Matthias Blume
313		Date: 2001/03/16 11:00:00 EST
314		Tag: blume-20010316-MLTREE-fixup
315		Description:
316
317		This is a minor fixup for an (untagged) earlier commit by Allen.
318		(A file was missing).
319
320		----------------------------------------------------------------------
321		Name: Allen Leung
322		Date: Mon Mar  5 18:54:57 EST 2001
323		Tag: leunga-20010305-cut-support
324
325		1. New support for alternative control-flow in MLTREE.
326		Currently we support
327
328		FLOW_TO(CALL ...., [k1,...,kn])
329
330		This is needed for 'cuts to' in C-- and try/handle-like constructs
331		in Moby
332
333		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
334
335		2. Register Allocator
336
337		Changes in interface [from Fermin, John]
338
339		3. Alpha 8-bit SLL support [Fermin]
340
341		4. All architectures
342
343		A new module (ClusterExpandCopies) for expanding parallel copies.
344
345		----------------------------------------------------------------------
346		Name: Allen Leung
347		Date: 2001/02/27 23:07:00 EST
348		Tag: leunga-20010227-minor-stuff
349
350		1. Alpha bug fix for CMOVNE
351		2. Handle mltree COND(..,FCMP ...,...)
352		3. Bug fix in simplifier
353
354		----------------------------------------------------------------------
355		Name: Matthias Blume
356		Date: 2001/01/30 17:50:00 EST
357		Tag: blume-20010130-sync
358		Description:
359
360		This is just a minor update to sync my devel branch with the main brach.
361		The only visible change is the addition of some README files.
362
363		----------------------------------------------------------------------
364		Name: Matthias Blume
365		Date: 2001/01/12 23:30:00 JST
366		Tag: blume-20010112-bootfiles
367		Description:
368
369		Made a new set of bootfiles that goes with the current state of the
370		repository.
371
372		----------------------------------------------------------------------
373		Name: Matthias Blume
374		Date: 2001/01/12 21:20:00 JST
375		Tag: blume-20010112-sync
376		Description:
377
378		I am just flushing out some minor changes that had accumulated in
379		my private branch in order to sync with the main tree.  (This is
380		mainly because I had CVS trouble when trying to merge _into_ my
381		private branch.)
382
383		Most people should be completely unaffected by this.
384
385		----------------------------------------------------------------------
386		Name: Allen Leung
387		Date: Thu Jan 11 21:03:00 EST 2001
388		Tag: leunga-20010111-labexp=mltree
389		Description:
390
391		1.  Removed the type LabelExp and replace it by MLTree.
392		2.  Rewritten mltree-simplify with the pattern matcher tool.
393		3.  There were some bugs in alpha code generator which would break
394		64-bit code generation.
395		4.  Redo the tools to generate code with the
396		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
397		are now generated by perl scripts.
398
399		----------------------------------------------------------------------
400		Name: Matthias Blume
401		Date: 2001/01/10 21:55:00 JST
402		Tag: blume-20010110-rcc
403		Description:
404
405		The RCC stuff now seems to work (but only on the x86).
406		This required hacking of the c-calls interface (and -implementation) in
407		MLRISC.
408
409		Normal compiler users should be unaffected.
410
411		----------------------------------------------------------------------
412		Name: Matthias Blume
413		Date: 2001/01/09 01:20:00 JST
414		Tag: blume-20010109-rcc
415		Description:
416
417		This is a fairly big patch, flushing out a large number of pending
418		changes that I made to my development copy over the last couple of days.
419
420		Of practical relevance at this moment is a workaround for a pickling
421		bug that Allen ran into the other day.  The cause of the bug itself is
422		still unknown and it might be hard to fix it properly, but the
423		workaround has some merits of its own (namely somewhat reducing pickling
424		overhead for certain libraries).  Therefore, I think this solution should
425		be satisfactory at this time.
426
427		The rest of the changes (i.e., the vast majority) has to do with my
428		ongoing efforts of providing direct support for C function calls from
429		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
430		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
431		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
432		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
433		handle RCC.
434
435		The last part (the changes to mlriscGen.sml) are still known to be
436		wrong on the x86 and not implemented on all other architectures.  But
437		the infrastructure is in place. I had to change a few functor
438		signatures in the backend to be able to route the CCalls interface
439		from MLRISC there, and I had to specialize the mltree type (on the
440		x86) to include the necessary extensions. (The extensions themselves
441		were already there and redy to go in MLRISC/x86).
442
443		Everything should be very happy as soon as someone helps me with
444		mlriscGen.sml...
445
446		In any case, nothing of this should matter to anyone as long as the
447		new primop is not being used (which is going to be the case unless you
448		find it where I hid it :). The rest of the compiler is completely
449		unaffected.
450
451		----------------------------------------------------------------------
452		Name: Matthias Blume
453		Date: 2001/01/05 00:30:00 JST
454		Tag: blume-20010105-primops
455		Description:
456
457		Added some experimental support for work that I am doing right now.
458		These changes mostly concern added primops, but there is also a new
459		experimental C library in the runtime system (but currently not enabled
460		anywhere except on Linux/X86).
461
462		In the course of adding primops (and playing with them), I discovered that
463		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
464		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
465		To recover, I made type information in INL_PRIM mandatory and changed
466		prim.sml as well as built-in.sml accordingly.  The InLine structure now
467		has complete, correct type information (i.e., no bottom types).
468
469		Since all these changes mean that we need new binfiles, I also bumped the
470		version number to 110.32.1.
471
472		----------------------------------------------------------------------
473		Name: Matthias Blume
474		Date: 2000/12/30 22:10:00 JST
475		Tag: blume-20001230-various
476		Description:
477
478		Added proxy libraries for MLRISC and let MLRISC libraries refer
479		to each other using path anchors.  (See CM manual for explanation.)
480
481		Updated CM documentation.
482
483		Fixed some bugs in CM.
484
485		Implemented "proxy" libraries (= syntactic sugar for CM).
486
487		Added "-quiet" option to makeml and changed runtime system accordingly.
488
489		Added cleanup handler for exportML to reset timers and compiler stats.
490
491		----------------------------------------------------------------------
492		Name: Lal George
493		Date: 2000/12/22 22:22:58 EST 2000
494		Tag: Release_110_32
495		Description:
496
497		Infinite precision used throughout MLRISC.
498		see MLRISC/mltree/machine-int.sig
499
500		----------------------------------------------------------------------
501		Name: Matthias Blume
502		Date: 2000/12/22 23:16:00 JST
503		Tag: blume-20001222-warn
504		Description:
505
506		Corrected wording and formatting of some CM warning message which I
507		broke in my previous patch.
508
509		----------------------------------------------------------------------
510		Name: Matthias Blume
511		Date: 2000/12/22 21:20:00 JST
512		Tag: blume-20001222-anchorenv
513		Description:
514
515		Fixed CM's handling of anchor environments in connection with CMB.make.
516
517		----------------------------------------------------------------------
518		Name: Matthias Blume
519		Date: 2000/12/22 13:15:00 JST
520		Tag: blume-20001222-cleanup
521		Description:
522
523		Removed src/cm/ffi which does not (and did not) belong here.
524
525		----------------------------------------------------------------------
526		Name: Matthias Blume
527		Date: 2000/12/21 23:55:00 JST
528		Tag: blume-20001221-exn
529		Description:
530
531		Probably most important: CM no longer silently swallows all exceptions
532		in the compiler.
533		Plus: some other minor CM changes.  For example, CM now reports some
534		sizes for generated binfiles (code, data, envpickle, lambdapickle).
535
536		----------------------------------------------------------------------
537		Name: Matthias Blume
538		Date: 2000/12/15 00:01:05 JST
539		Tag: blume-20001215-dirtool
540		Description:
541
542		- "dir" tool added.
543		- improvements and cleanup to Tools structure
544		- documentation updates
545
546		----------------------------------------------------------------------
547		Name: Allen Leung
548		Date: Thu Dec 14 03:45:24 EST 2000
549		Description:
550		Tag:  leunga-20001214-int-inf
551		Description:
552
553		In IntInf, added these standard functions, which are missing from our
554		implementation:
555
556		andb : int * int -> int
557		xorb : int * int -> int
558		orb  : int * int -> int
559		notb : int -> int
560		<<   : int * word -> int
561		~>>  : int * word -> int
562
563		Not tested, I hope they are correct.
564
565		----------------------------------------------------------------------
566		Name: Allen Leung
567		Date: Fri Dec  8 19:23:26 EST 2000
568		Description:
569		Tag:  leunga-20001208-nowhere
570		Description:
571
572		Slight improvements to the 'nowhere' tool to handle OR-patterns,
573		to generate better error messages etc.  Plus a brief manual.
574
575		----------------------------------------------------------------------
576		Name: Lal George
577		Date: 2000/12/08 09:54:02 EST 2000
578		Tag: Release_110_31
579		Description:
580
581		- Version 110.31
582		----------------------------------------------------------------------
583		Name: Allen Leung
584		Date: Thu Dec  7 22:01:04 EST 2000
585		Tag:  leunga-20001207-cell-monster-hack
586		Description:
587
588		Major MLRISC internal changes.  Affect all clients.
589		Summary:
590
591		1.  Type CELLS.cell = int is now replaced by a datatype.
592		As a result, the old regmap is now gone.  Almost all interfaces
593		in MLRISC change as a consequence.
594
595		2.  A new brand version of machine description tool (v3.0) that generates
596		modules expecting the new interface.  The old version is removed.
597
598		3.  The RA interface has been further abstracted into two new functors.
599		RISC_RA and X86RA.  These functors have much simpler interfaces.
600		[See also directory MLRISC/demo.]
601
602		4.  Some other new source->source code generation tools are available:
603
604		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
605		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
606		I use this tool to generate the peephole optimizers---with the new
607		cell type changes, peephole rules are becoming difficult to write
608		without conditional pattern matching.
609
610		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
611		cover the entire MLRISC source tree so many things broke.
612
613		6.  CM files have been moved to the subdirectory MLRISC/cm.
614		They are moved because there are a lot of them and they clutter up the
615		root dir.
616
617		7.  More detailed documentation to come...
618
619		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
620		a makeml -rebuild first.  This is because of other other
621		changes that Matthias has made (see below).
622
623
624		----------------------------------------------------------------------
625		Name: Matthias Blume
626		Date: 2000/11/30 23:12:00 JST
627		Tag: blume-20001130-filereorg
628		Description:
629
630		Some manual updates and some file reorganizations in CM.
631
632		----------------------------------------------------------------------
633		Name: Matthias Blume
634		Date: 2000/11/24 17:45:00 JST
635		Tag: blume-20001124-link
636		Description:
637
638		Drastically improved link traversal code for the case that the dynamic
639		value was already loaded at bootstrap time.  As a result, CM and CMB
640		now both load blazingly fast -- even on a very slow machine.  Also,
641		memory consumption has been further reduced by this.
642
643		Warning: The format of the PIDMAP file has changed.  THerefore, to
644		bootstrap you have to do this:
645
646		1. Run CMB.make
647		2. Make a symbolic link for the boot directory:
648		ln -s sml.boot.ARCH-OS xxx
649		3. "Rebuild" the boot directory:
650		./makeml -boot xxx -rebuild sml ; rm xxx
651		4. Boot normally:
652		./makeml
653
654		----------------------------------------------------------------------
655		Name: Matthias Blume
656		Date: 2000/11/21 21:20:00 JST
657		Tag: blume-20001121-tools
658		Description:
659
660		Continued hacking on autoloading problem -- with success this time.
661		Also changed tool-plugin mechanism.  See new CM manual.
662
663		----------------------------------------------------------------------
664		Name: Matthias Blume
665		Date: 2000/11/19 14:30:00 JST
666		Tag:  blume-20001119-autoload
667		Description:
668
669		Some hacking to make autoloading faster.  Success for CMB, no success
670		so far for CM.  There is a reduced structure CM' that autoloads faster.
671		(This is a temporary, non-documented hack to be eliminated again when
672		the general problem is solved.)
673
674		----------------------------------------------------------------------
675		Name: Matthias Blume
676		Date: 2000/11/17 14:10:00 JST
677		Tag: blume-20001117-pickle-lib
678		Description:
679
680		1. Eliminated comp-lib.cm
681		2. Made pickle-lib.cm
682		3. Eliminated all uses of intset.sml (from comp-lib.cm)
683		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
684		equivalent constructs from smlnj-lib.cm (INtHashTable).
685		5. Point 4. also goes for those uses of intmap.* in MLRISC.
686		Duplicated intmap modules thrown out.
687		6. Hunted down all duplicated SCC code and replaced it with
688		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
689		7. Rewrote Feedback module.
690		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
691		should be thrown out and equivalent modules from smlnj-lib.cm
692		should be used (IntRedBlackSet, IntListSet, ...).
693
694		Confirmed that compiler compiles to fixpoint.
695
696		----------------------------------------------------------------------
697		Name: Allen Leung
698		Date: 2000/11/10 18:00:00
699		Tag: leunga-20001110-new-x86-fp
700
701		A new x86 floating point code generator has been added.
702		By default this is turned off.  To turn this on, do:
703
704		CM.autoload "$smlnj/compiler.cm";
705		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
706
707		Changes:
708
709		1.  Changed FTAN to FPTAN so that the assembly output is correct.
710		2.  Changed the extension callback for FTANGENT to generate:
711
712		fptan
713		fstp  %st(0)
714		instead of
715		fptan
716		fstpl ftempmem
717
718		3.  Numerous assembly fixes for x86.
719
720		5.  Cleaned up the machine code output module x86/x86MC.sml and added
721		support for a whole bunch of instructions and addressing modes:
722
723		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
724		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
725		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
726		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
727		fxch %st(n)
728		fld %st(n)
729		fst %st(n)
730		fst mem
731		fstp %st(n)
732		fucom %st(n)
733		fucomp %st(n)
734
735		All these are now generated when the fast fp mode is turned on.
736
737		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
738
739		----------------------------------------------------------------------
740		Name: Matthias Blume
741		Date: 2000/11/09 11:20:00 JST
742		Tag: blume-20001109-scc
743		Description:
744
745		Eliminated some code duplication:
746
747		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
748		(Otherwise the functor is useless.)
749		2. Used GraphSCCFn where SCCUtilFun was used previously.
750		3. Got rid of SCCUtilFun (in comp-lib.cm).
751
752		----------------------------------------------------------------------
753		Name: Lal George
754		Date: 2000/11/06 09:02:21 EST 2000
755		Tag: Release_110_30
756		Description:
757
758		- Version 110.30
759		----------------------------------------------------------------------
760		Name: Matthias Blume
761		Date: 2000/11/04 14:45:00
762		Tag: blume-20001104-mlbuild
763		Description:
764
765		- Made ml-build faster on startup.
766		- Documentation fixes.
767
768		----------------------------------------------------------------------
769		Name: Matthias Blume
770		Date: 2000/11/02 17:00:00 JST
771		Tag: blume-20001102-condcomp
772		Description:
773
774		- Small tweaks to pickler -- new BOOTFILES!
775		- Version bumped to 110.29.2.
776		- Added conditional compilation facility to init.cmi (see comment there).
777		----------------------------------------------------------------------
778		Name: Allen Leung
779		Date: 2000/10/23 19:31:00
780		Tag: leunga-20001023-demo-ra
781
782		1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
783		2. Test programs for the graph library updated
784		3. Some new MLRISC demo programs added
785
786		----------------------------------------------------------------------
787		Name: Matthias Blume
788		Date: 2000/08/31 22:15:00 JST
789		Tag: blume-20001017-errmsg
790		Description:
791
792		More error message grief: Where there used to be no messages, there
793		now were some that had bogus error regions.  Fixed.
794
795		----------------------------------------------------------------------
796		Name: Matthias Blume
797		Date: 2000/08/31 17:30:00 JST
798		Tag: blume-20001017-v110p29p1
799		Description:
800
801		I made a version 110.29.1 with new bootfiles.
802
803		Changes:  Modified pickler/unpickler for faster and leaner unpickling.
804		CM documentation changes and a small bugfix in CM's error reporting.
805
806		----------------------------------------------------------------------
807		Name: Lal George
808		Date: 2000/09/27 14:42:35 EDT
809		Tag: george-20000927-nodestatus
810		Description:
811
812		Changed the type of the nodestatus, so that:
813
814		SPILLED(~1)             is now SPILLED
815		SPILLED(m) where m>=0   is now MEMREG(m)
816		SPILLED(s) where s<~1   is now SPILL_LOC(~s)
817
818		----------------------------------------------------------------------
819		Name: Matthias Blume
820		Date: 2000/09/07 14:45:00 JST
821		Tag: blume-20000907-cmerrmsg
822		Description:
823
824		Small tweak to CM to avoid getting ML syntax error messages twice.
825
826		----------------------------------------------------------------------
827		Name: Matthias Blume
828		Date: 2000/08/31 18:00:00 JST
829		Tag: blume-20000831-cvsbootfiles
830		Description:
831
832		New URL for boot files (because the 110.29 files on the BL server do
833		now work correctly with my updated install scripts for yacc and lex).
834
835		----------------------------------------------------------------------
836		Name: Matthias Blume
837		Date: 2000/08/08 12:33:00 JST
838		Tag: blume-20000808-manual
839		Description:
840
841		Tiny update to CM manual.
842
843		----------------------------------------------------------------------
844		Name: Allen Leung
845		Date: 2000/08/7 19:31:00
846		Tag: leunga-20000807-a-whole-bunch-of-stuff
847
848		Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
849		and Mobdy.
850
851		1.  x86
852
853		a.  Fixes to peephole module by John and Dan.
854		b.  Assembly fix to SETcc by Allen.
855		c.  Fix to c-call by John.
856		d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
857		e.  Instruction selection optimization to SETcc as suggested by John.
858
859		For example,
860
861		MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
862
863		should generate:
864
865		MOVL a, x
866		SUBL b, x
867		SHRL 31, x
868
869		2.  IR stuff
870
871		A bunch of new DJ-graph related algorithms added.  These
872		speed up SSA construction.
873
874		3.  SSA + Scheduling
875
876		Added code for SSA and scheduling to the repository
877
878		----------------------------------------------------------------------
879		Name: Lal George
880		Date: 2000/07/27 11:53:14 EDT
881
882		Tag: lal-20000727-linux-ppc
883		Description:
884
885		Made changes to support Linux PPC.
886		p.s. I have confirmation that the 110.29 boot files work fine.
887
888		----------------------------------------------------------------------
889		Name: Matthias Blume
890		Date: 2000/07/27 17:40:00 JST
891		Tag: blume-20000727-scripts
892		Description:
893
894		!!!! WARNING !!!!
895		You must recompile the runtime system!
896		!!!! WARNING !!!!
897
898		This is basically another round of script-enhancements:
899
900		1. sml, ml-build, and ml-makedepend accept options -D and -U to define
901		and undefine CM preprocessor symbols.
902
903		2. ml-build avoids generating a new heap image if it finds that the
904		existing one is still ok.  (The condition is that no ML file had to
905		be recompiled and all ML files are found to be older that the heap
906		file.)
907
908		To make this work smoothly, I also hacked the runtime system as
909		well as SMLofNJ.SysInfo to get access to the heap image suffix
910		(.sparc-solaris, ...) that is currently being used.
911
912		Moreover, the signature of CM.mk_standalone has changed.  See the
913		CM manual.
914
915		3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
916		CM manual for details.)
917
918		4. More CM manual updates:
919		- all of the above has been documented.
920		- there is now a section describing the (CM-related) command line
921		arguments that are accepted by the "sml" command
922
923		----------------------------------------------------------------------
924		Name: Matthias Blume
925		Date: 2000/07/25 16:20:00 JST
926		Tag: blume-20000725-makedepend
927		Description:
928
929		Added a script called ml-makedepend.  This can be used in makefiles
930		for Unix' make in a way very similar to the "makedepend" command for
931		C.
932
933		The script internally uses function CM.sources.
934
935		Synopsis:
936
937		ml-makedepend [-f makefile] cmfile targetname
938
939		The default for the makefile is "makefile" (or "Makefile" should
940		"makefile" not exist).
941
942		ml-makedepend adds a cmfile/targetname-specific section to this
943		makefile (after removing the previous version of this section).  The
944		section contains a single dependency specification with targetname on
945		the LHS (targetname is an arbitrary name), and a list of files derived
946		from the cmfile on the RHS.  Some of the files on the RHS are
947		ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
948		"make" variables $(ARCH) and $(OPSYS) in place of the corresponding
949		path names.  The makefile writer is responsible for making sure that
950		these variables have correct at the time "make" is invoked.
951
952		----------------------------------------------------------------------
953		Name: Matthias Blume
954		Date: 2000/07/22 23:30:00 JST
955		Tag: blume-20000722-urlupdate
956		Description:
957
958		Changed BOOT and config/srcarchiveurl to point to BL server:
959
960		ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
961
962		----------------------------------------------------------------------
963		Name: Matthias Blume
964		Date: 2000/07/18 18:00:00 JST
965		Tag: blume-20000718-Version_110_29
966		Description:
967
968		1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
969
970		2. Updated config/version to 110.29
971
972		3. Updated config/srcarchiveurl
973
974		3. New boot files!
975		ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
976
977		----------------------------------------------------------------------
978		Name: Matthias Blume
979		Date: 2000/07/11 13:58:00 JST
980		Tag: blume-20000711-doctypo
981		Description:
982
983		Fixed a few typos in CM manual.
984
985		----------------------------------------------------------------------
986		Name: Allen Leung
987		Date: 2000/06/15 00:38:00
988		Tag: leunga-20000704-sparc-x86
989
990		1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
991		2. fix to x86 RET bug [found by Dan Grossman]
992		3. sparc assembly bug fix for ticc instructions [found by Fermin]
993
994		Affects c-- and moby only
995
996		----------------------------------------------------------------------
997		Name: Matthias Blume
998		Date: 2000/07/04 15:26:00
999		Tag: blume-20000704-trigger
1000		Description:
1001
1002		1. Improvements to CM manual.
1003		2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1004		of getting a back-trace.  The function, when called, raises an
1005		internal exception which explicitly carries the full back-trace history,
1006		so it is unaffected by any intervening handle-raise pairs ("trivial"
1007		or not).  The interactive loop will print that history once it arrives
1008		at top level.
1009		Short of having all exceptions implicitly carry the full history, the
1010		recommended way of using this facility is:
1011		- compile your program with instrumentation "on"
1012		- run it, when it raises an exception, look at the history
1013		- if the history is "cut off" because of some handler, go and modify
1014		your program so that it explicitly calls BTrace.trigger
1015		- recompile (still instrumented), and rerun; look at the full history
1016
1017		----------------------------------------------------------------------
1018		Name: Matthias Blume
1019		Date: 2000/07/03 15:36:00 JST
1020		Tag: blume-20000702-manual
1021		Description:
1022
1023		Small corrections and updates to CM manual.
1024
1025		----------------------------------------------------------------------
1026		Name: Matthias Blume
1027		Date: 2000/06/29 16:04:00 JST
1028		Tag: blume-20000629-yacctool
1029		Description:
1030
1031		Changes:
1032
1033		1. Class "mlyacc" now takes separate arguments to pass options to
1034		generated .sml- and .sig-files independently.
1035		2. Corresponding CM manual updates.
1036		3. BTrace module now also reports call sites.  (However, for loop clusters
1037		it only shows from where the cluster was entered.)  There are associated
1038		modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1039
1040		----------------------------------------------------------------------
1041		Name: Matthias Blume
1042		Date: 2000/06/27 16:51:00 JST
1043		Tag: blume-20000627-noweb
1044		Description:
1045
1046		Changes:
1047
1048		1. Implemented "subdir" and "witness" options for noweb tool.
1049		This caused some slight internal changes in CM's tool implementation.
1050		2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1051		some remaining issues from earlier path anchor changes.
1052		3. Updated CM manual accordingly.
1053
1054		4. Changed implementation of back-tracing so that I now consider it
1055		ready for prime-time.
1056
1057		In particular, you don't have to explicitly trigger the back-trace
1058		anymore.  Instead, if you are running BTrace-instrumented code and
1059		there is an uncaught exception (regardless of whether or not it was
1060		raised in instrumented code), the top-level evalloop will print
1061		the back-trace.
1062
1063		Features:
1064
1065		- Instrumented and uninstrumented code work together seemlessly.
1066		(Of course, uninstrumented code is never mentioned in actual
1067		back-traces.)
1068
1069		- Asymptotic time- and space-complexity of instrumented code is
1070		equal to that of uninstrumented code.  (This means that
1071		tail-recursion is preserved by the instrumentation phase.)
1072
1073		- Modules whose code has been instrumented in different sessions
1074		work together without problem.
1075
1076		- There is no penalty whatsoever on uninstrumented code.
1077
1078		- There is no penalty on "raise" expressions, even in
1079		instrumented code.
1080
1081		A potential bug (or perhaps it is a feature, too):
1082
1083		A back-trace reaches no further than the outermost instrumented
1084		non-trivial "raise".  Here, a "trivial" raise is one that is the
1085		sole RHS of a "handle" rule.  Thus, back-traces reach trough
1086
1087		<exp> handle e => raise e
1088
1089		and even
1090
1091		<exp> handle Foo => raise Bar
1092
1093		and, of course, through
1094
1095		<exp> handle Foo => ...
1096
1097		if the exception was not Foo.
1098
1099		Back-traces always reach right through any un-instrumented code
1100		including any of its "handle" expressions, trivial or not.
1101
1102		To try this out, do the following:
1103
1104		- Erase all existing binfiles for your program.
1105		(You may keep binfiles for those modules where you think you
1106		definitely don't need back-tracing.)
1107		- Turn on back-trace instrumentation:
1108		SMLofNJ.Internals.BTrace.mode (SOME true);
1109		- Recompile your program.  (I.e., run "CM.make" or "use".)
1110		- You may now turn instrumentation off again (if you want):
1111		SMLofNJ.Internals.BTrace.mode (SOME false);
1112		- Run your program as usual.  If it raises an exception that
1113		reaches the interactive toplevel, then a back-trace will
1114		automatically be printed.  After that, the toplevel loop
1115		will print the exception history as usual.
1116
1117		----------------------------------------------------------------------
1118		Name: Matthias Blume
1119		Date: 2000/06/26 09:56:46 JST
1120		Tag: blume-20000626-setup
1121		Description:
1122
1123		CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1124		ML code before and after compiling a file (e.g., to set compiler
1125		flags)
1126
1127		Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1128		- associated changes to btrace.sml (BTrace instrumentation pass)
1129		- cleaner implementation of btimp.sml (BTrace tracing and report
1130		module)
1131
1132		CM manual: * new path encoding documented
1133		* description of setup-parameter to "sml" added
1134
1135		The biggest user-visible change to back-tracing is that it is no
1136		longer necessary to compile all traced modules within the same
1137		session.  (This was a real limitation.)
1138
1139	----------------------------------------------------------------------	----------------------------------------------------------------------
1140	Name: Matthias Blume	Name: Matthias Blume
1141	Date: 2000/06/24 12:40:00 JST	Date: 2000/06/24 12:40:00 JST

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.676   changed lines   Added in v.836

---

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