Diff of /sml/branches/primop-branch-3/NOTES/HISTORY

Parent Directory | Revision Log | Patch

	revision 1171, Fri Mar 22 20:41:43 2002 UTC	revision 1620, Mon Sep 13 20:10:38 2004 UTC
#	Line 13	Line 13
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: John Reppy
17		Date: 2004/09/13
18		Tag: jhr-20040913-signals
19		Description:
20
21		Fixed the signal masking code to properly nest mask/unmask operations
22		on a per-signal basis.
23
24		----------------------------------------------------------------------
25		Name: Matthias Blume (blume (at) tti - c (dot) org)
26		Date: 2004/09/08 13:20:00 CDT
27		Tag: blume-20040908-heap-magic
28		Description:
29
30		Bumped the heap macig number to 0x09082004 to account for the changed
31		layout of the ML frame under MacOS X.
32
33		----------------------------------------------------------------------
34		Name: Allen Leung (leunga (at) reservoir (dot) com)
35		Date: 2004/09/03 11:26:00 EST
36		Tag: leunga-20040903-cygwin-install
37		Description:
38
39		Added a patch to _arch-n-opsys to enable the Cygwin runtime.
40		The Cygwin runtime is turned on by setting the environment
41		variable SMLNJ_CYGWIN_RUNTIME to 1.
42
43		----------------------------------------------------------------------
44		Name: Matthias Blume (blume (at) tti - c (dot) org)
45		Date: 2004/08/31 17:15:00 CDT
46		Tag: blume-20040831-core
47		Description:
48
49		Added some exports to src/compiler/core.cm upon request by J. Joemann.
50
51		----------------------------------------------------------------------
52		Name: Matthias Blume (blume (at) tti - c (dot) org)
53		Date: 2004/08/30 17:55:00 CDT
54		Tag: blume-20040830-installer
55		Description:
56
57		Upon request by Johannes Joemann:
58
59		- improved ML code of installer to fall back to coping when renaming
60		fails (i.e., when source and target are on different file systems);
61		the code compiles but has yet to be tested in anger
62		- removed mlrisc from list of default targets (config/targets)
63
64		----------------------------------------------------------------------
65		Name: Matthias Blume (blume (at) tti - c (dot) org)
66		Date: 2004/08/27 17:20:00 CDT
67		Tag: blume-20040827-ptreql
68		Description:
69
70		Added ptreql primop to structure InlineT (upon request from Larry
71		Paulson).
72
73		----------------------------------------------------------------------
74		Name: Allen Leung (leunga (at) reservoir (dot) com)
75		Date: 2004/08/15 21:21:00 EST
76		Tag: leunga-110_48-udgraph
77		Description:
78
79		Another bug fix from Carl Hauser:
80
81		diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
82		> 48c48
83		> <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
84		> e::L)
85		> ---
86		> >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
87		> Without this, any deletion of an edge in an undirected graph does severe
88		> violence to the graph.
89
90		----------------------------------------------------------------------
91		Name: Allen Leung (leunga (at) reservoir (dot) com)
92		Date: 2004/08/10 23:35:00 EST
93		Tag: leunga-110_48-ppc
94		Description:
95
96		The IBM/MacOS syntax switch on PPC was incorrectly swapped.
97
98		----------------------------------------------------------------------
99		Name: Matthias Blume (blume (at) tti - c (dot) org)
100		Date: 2004/08/10 12:00:00 CDT
101		Tag: Release_110_48
102		Description:
103
104		New working version (110.48).  NEW BOOTFILES!
105
106		----------------------------------------------------------------------
107		Name: Allen Leung (leunga (at) reservoir (dot) com)
108		Date: 2004/08/09 12:21:00 EST
109		Tag: leunga-110_47-dijsktra
110		Description:
111
112		Bug fix from Carl Hauser:
113
114		single_source_shortest_paths in dijkstra.sml was observed to get wrong
115		answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
116
117		The problem is that following the expression A.update(dist,s,Num.zero)
118		it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
119
120		----------------------------------------------------------------------
121		Name: Matthias Blume (blume (at) tti - c (dot) org)
122		Date: 2004/08/06 18:10:00 CDT
123		Tag: blume-20040806-cmdline
124		Description:
125
126		Fiddled with handling of command-line options:
127
128		* sml now quits after processing the command line
129		if -H, -S, -h<n>, or -s<n> appears as the last
130		command-line argument
131		* a new option -q terminates the session when encountered on
132		the command line; subsequent arguments will be ignored
133		* bug fixes: short (erroneous) arguments are no longer ignored
134		completely
135
136		----------------------------------------------------------------------
137		Name: Allen Leung (leunga (at) reservoir (dot) com)
138		Date: 2004/08/04 18:17:00 EST
139		Tag: leunga-110_47-ppc-ibm-asm
140		Description:
141
142		- Added minimal IBM assembly syntax support for PowerPC.
143
144		- Cygwin: manually changed the file cygwin.def.  Some exported symbols have
145		been altered in the runtime.  We need an automatic way to keep the file
146		in sync.
147
148		----------------------------------------------------------------------
149		Name: Matthias Blume (blume (at) tti - c (dot) org)
150		Date: 2004/08/04 14:00:00 CDT
151		Tag: Release_110_47
152		Description:
153
154		New working version (110.47).  NEW BOOTFILES!
155
156		----------------------------------------------------------------------
157		Name: Matthias Blume (blume (at) tti - c (dot) org)
158		Date: 2004/08/03 14:25:00 CDT
159		Tag: blume-20040803-callingconv
160		Description:
161
162		Added low-level support for choosing C calling conventions by
163		twiddling the type of rawccall.  (See
164		src/compiler/Semant/types/cproto.sml for details.)
165
166		----------------------------------------------------------------------
167		Name: Matthias Blume (blume (at) tti - c (dot) org)
168		Date: 2004/08/02 15:55:00 CDT
169		Tag: blume-20040802-backout
170		Description:
171
172		Backed out of change to win32-filesys.c.  The earlier patch to
173		get_file_time caused CM to produce files with the wrong time stamp.
174
175		----------------------------------------------------------------------
176		Name: Matthias Blume (blume (at) tti - c (dot) org)
177		Date: 2004/08/02 14:45:00 CDT
178		Tag: blume-20040802-nlffi-win32
179		Description:
180
181		Added NLFFI support for Win32, adapted from a patch provided by David
182		Hansel.  This is currently completely untested.  Also, the issue
183		concerning stdcall vs. ccall is still unresolved.
184
185		----------------------------------------------------------------------
186		Name: Matthias Blume (blume (at) tti - c (dot) org)
187		Date: 2004/07/30 17:55:00 CDT
188		Tag: blume-20040730-various
189		Description:
190
191		Gearing up towards 110.47...
192
193		- various minor bugfixes to ml-nlffigen
194		- a beginning of a manual for nlffi
195
196		- eliminated 'export name=value' in config/install.sh as this does
197		not work with certain versions of /bin/sh
198		(Thanks to David King at Motorola for catching this.)
199
200		- several bugfixes provided or suggested by David Hansel at Reactive Systems:
201		- added a test for tm==NULL to gmtime.c and localtime.c
202		- applied patch for incorrect GetFileTime under win32
203		- toSeconds -> toMilliseconds in Win32/win32-process.sml
204
205		----------------------------------------------------------------------
206		Name: Matthias Blume (blume (at) tti - c (dot) org)
207		Date: 2004/07/21 18:20:00 CDT
208		Tag: blume-20040721-nlffigen
209		Description:
210
211		- Fixed minor issue in ml-nlffigen: Now generate structure T_foo
212		for a typedef to an incomplete type, but leave out the "typ" member.
213		(This is just for consistency.)
214		- Started to produce what is supposed to become better (i.e., comprehensive)
215		documentation of what ml-nlffigen does and produces.
216
217		----------------------------------------------------------------------
218		Name: Matthias Blume (blume (at) tti - c (dot) org)
219		Date: 2004/07/14 16:25:00 CDT
220		Tag: blume-20040714-union
221		Description:
222
223		Added C_UNION to c-calls/c-types.sml and updated the machinery
224		(ml-nlffigen, cproto.sml) that conveys C function interface
225		information to the code generator.
226
227		However, the actual architecture-specific implementation of function
228		arguments and results that are C unions is still not implemented.
229
230		----------------------------------------------------------------------
231		Name: Allen Leung (leunga (at) reservoir (dot) com)
232		Date: 2004/07/14 14:38:00 EST
233		Tag: leunga-110_46_1-ppc-lwzu
234		Description:
235
236		Added these instructions to the PowerPC architecture:
237		LBZU(X), LHZU(X), LWZU(X),
238		STWU(X), STFDU, STFSU
239
240		etc...
241
242		Note: I haven't added their instruction encoding into the description.
243
244		----------------------------------------------------------------------
245		Name: Allen Leung (leunga (at) reservoir (dot) com)
246		Date: 2004/07/13 15:04:00 EST
247		Tag: leunga-110_46_1-ppc-lwarx
248		Description:
249
250		Added the two instructions LWARX and STWCX to the PowerPC
251		instruction set.
252
253		A (untested) rewrite of loop-structure.sml.  The old version
254		is completely broken.
255
256		----------------------------------------------------------------------
257		Name: Matthias Blume (blume (at) tti - c (dot) org)
258		Date: 2004/07/13 13:50:00 CDT
259		Tag: blume-20040713-nlffi
260		Description:
261
262		- use paramAlloc to report c-calls with too many arguments
263		(for PPC version where parameter area is pre-allocated)
264		- added ccall_maxargspace to machspec (to implement the above)
265		- made "make" commend in CM's "make" tool configurable
266		- added option (default: on) for passing the name of the SML/NJ's "bin"
267		directory to "make"; the call looks like this:
268
269		make <options> SMLNJ_BINDIR=<dir> <target>
270
271		This can be used by the Makefile to, e.g., pick the "right" version
272		of ml-nlffigen.
273		- minor code tweaks
274
275		----------------------------------------------------------------------
276		Name: Matthias Blume (blume (at) tti - c (dot) org)
277		Date: 2004/07/12 22:50:00 CDT
278		Tag: blume-110_46_1-macosx-nlffi
279		Description:
280
281		NLFFI under Mac OS X now working (sort of).  This is largely untested,
282		though.
283
284		Note:  1. You have to make a new, clean build of the runtime system.
285		2. There are new BOOTFILES, you have to use them!
286		(Doing the bootstrap process yourself would be *very* painful!
287		If you absolutely have to do it, build the system under
288		a different architecture and then cross-compile.)
289
290		Version bumped to 110.46.1 to account for runtime data format changes.
291
292		----------------------------------------------------------------------
293		Name: Matthias Blume (blume (at) tti - c (dot) org)
294		Date: 2004/06/18 14:30:00 CDT
295		Tag: blume-20040618-unix
296		Description:
297
298		Changed the implementation of structure Unix so that the same stream
299		is returned every time one of the {text,bin}{In,Out}streamOf functions
300		is invoked on the same proc.  This is not what the spec currently
301		says -- although IMO it arguably should.  (See discussion below.)
302
303		----------------------------------------------------------------------
304		Name: Matthias Blume (blume (at) tti - c (dot) org)
305		Date: 2004/06/17 18:15:00 CDT
306		Tag: Release_110_46
307		Description:
308
309		New working version (110.46).  NEW BOOTFILES!
310
311		----------------------------------------------------------------------
312		Name: Matthias Blume (blume (at) tti - c (dot) org)
313		Date: 2004/06/17 17:20:00 CDT
314		Tag: blume-20040617-timer-unix
315		Description:
316
317		Changed the interface of structures Timer and Unix to match the most
318		recent Basis spec.
319
320		In the case of Unix there still seems to be an open/weird issue:
321
322		The {text,bin}{In,Out}streamOf functions are supposed to create
323		fresh streams whenever they are called -- as opposed to have them
324		return the same stream every time.  This design is supposed to
325		prevent space leaks caused by proc values hanging on to streams.
326
327		The reap function, on the other hand, is supposed to close the
328		streams.  This cannot be done without having a handle on the
329		stream in proc after all...
330
331		I took the liberty to implement the following stopgap solution:
332
333		The proc value hangs on to the most recently created stream(s).
334		Reap closes those.  If either or both of the two streams hadn't
335		been created at all yet, then reap will close the corresponding
336		file descriptors directly.
337
338		PS: I don't understand the original space leak argument anymore.  If
339		a proc hangs on to the imperative stream, then I/O operations on those
340		will advance the state of the cached stream and avoid the space leak.
341
342		----------------------------------------------------------------------
343		Name: Matthias Blume (blume (at) tti - c (dot) org)
344		Date: 2004/05/28 16:45:00 CDT
345		Tag: blume-20040528-basis
346		Description:
347
348		Added signature PACK_REAL and exported functor PrimIO.
349
350		----------------------------------------------------------------------
351		Name: Matthias Blume (blume (at) tti - c (dot) org)
352		Date: 2004/05/25 16:00:00 CDT
353		Tag: blume-20040525-group-owner
354		Description:
355
356		CM now ignores (but still accepts) the "owner" information in group
357		descriptions.  The owner of a group is its next enclosing
358		library. Each group must have a unique owner.  (There is a virtual
359		"toplevel" library that owns groups which are not nested within a real
360		library.)  Previously, each group had to explicitly declare its owner,
361		and CM would check that such a declaration is correct.  The new scheme
362		is to have CM check that for each group there is precisely one owning
363		library.
364
365		The advantage of the new scheme is that the programmer no longer needs
366		to maintain the somewhat annoying owner information.  The downside is
367		that CM cannot enforce the ownership rule across multiple runs of
368		CM.make.  Fortunately, enclosing the same group in two different
369		libraries A and B which are not part of the same program does not
370		cause real problems.
371
372		----------------------------------------------------------------------
373		Name: Matthias Blume (blume (at) tti - c (dot) org)
374		Date: 2004/05/20 16:00:00 CDT
375		Tag: blume-20040520-win32
376		Description:
377
378		Made the win32 version work again.  (Strangely, a misplaced comma had
379		slipped into win32-process.c which prevented the runtime from being
380		compiled correctly.)
381
382		Also, included a minor addition to ml-build.bat analogous to what was
383		done in blume-20040519-ml-build.
384
385		----------------------------------------------------------------------
386		Name: Matthias Blume (blume (at) tti - c (dot) org)
387		Date: 2004/05/19 22:10:00 CDT
388		Tag: blume-20040519-ml-build
389		Description:
390
391		Arranged for ml-build to clean up after itself a little bit better.
392		The script generates a temporary SML source file and compiles it using
393		CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
394		gets rid of those at the end, so they don't accumulate under .cm.
395
396		This required a minor change to install.sh because the name of the
397		metadata directory (default: .cm) is actually configurable at
398		installation time.
399
400		----------------------------------------------------------------------
401		Name: Matthias Blume (blume (at) tti - c (dot) org)
402		Date: 2004/05/18 15:50:00 CDT
403		Tag: blume-20040518-mkreader
404		Description:
405
406		Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
407		implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
408
409		----------------------------------------------------------------------
410		Name: Matthias Blume (blume (at) tti - c (dot) org)
411		Date: 2004/05/11 14:35:00 CDT
412		Tag: blume-20040511-win32sock
413		Description:
414
415		Added previously missing support for many socket-related functions
416		under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
417		for the voluminous patch!
418
419		(I have not tested this patch under win32 yet.)
420
421		Here is David's e-mail:
422
423		Hi,
424
425		Attached to this email you find a diff against sml/nj 110.45
426		that will enable socket support under Windows.
427
428		To apply the patch (using unix or cygwin)
429		1) gunzip runtime.diff.gz
430		2) "cd" into "src/runtime" in the source tree of a fresh
431		110.45 installation.
432		3) patch -p 1 < [your/path/to]runtime.diff
433
434		The code compiles fine but has NOT yet been extensively tested.
435		I only ran a few tests for basic socket client functionality
436		(which worked fine).  Especially the functions that use ioctl
437		are not tested at all and might not work (see below).
438
439		I implemented this since we want to move to a newer version of sml/nj
440		but need socket support in order to use it.  This is the first time I
441		even had a look at the sml/nj source,  so please review my changes
442		before making this part of the distribution!  Here are a few issues
443		that I think might be better for someone to solve who is more
444		familiar with the sml/nj source (and socket programming):
445
446		- getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
447		exception since I could not figure out what the windows equivalent
448		of these functions is
449
450		- In sockets-osdep.h there are a some #include statements that are
451		only used in a few files that include sockets-osdep.h
452
453		- In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
454		does not process its return value since I don't know how to
455		report an error upwards.
456
457		- It would probably be good to have a call to WSACleanup() when
458		the library is unloaded (if there is such a possibility).
459		Otherwise I think Windows will take care of this automatically
460		when the process finishes.
461
462		- I used ioctlsocket() as a replacement for ioctl() but I have
463		no idea if that is actually the proper replacement on Windows.
464
465		- All these issues are marked in the code by "FIXME" comments.
466
467		We use sml/nj extensively in our products and are quite happy
468		with it.  I hope this contribution will help you.
469
470		Keep up the good work!
471
472		David
473
474		----------------------------------------------------------------------
475		Name: Matthias Blume (blume (at) tti - c (dot) org)
476		Date: 2004/05/11 14:20:00 CDT
477		Tag: blume-20040511-installml
478		Description:
479
480		Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
481		the patch.)
482
483		----------------------------------------------------------------------
484		Name: Matthias Blume (blume (at) tti - c (dot) org)
485		Date: 2004/05/11 14:05:00 CDT
486		Tag: blume-20040511-nlffi-netbsd
487		Description:
488
489		Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
490		the patch.)
491
492		----------------------------------------------------------------------
493		Name: Matthias Blume (blume (at) tti - c (dot) org)
494		Date: 2004/05/11 12:05:00 CDT
495		Tag: blume-20040511-exports
496		Description:
497
498		As per request by Adam Chlipala <adam@hcoop.net>, extended various
499		export lists in compiler-related .cm-files.
500
501		----------------------------------------------------------------------
502		Name: Matthias Blume (blume (at) tti - c (dot) org)
503		Date: 2004/05/11 11:35:00 CDT
504		Tag: blume-20040511-allsource
505		Description:
506
507		The installer now honors the "src-smlnj" target again, although its meaning
508		has changed from "all sources required for the compiler" to "all sources
509		the installer knows about".  In other words, if you enable "src-smlnj"
510		in the "targets" file, then the installer will pull in sources for
511		everything.  (Notice that this refers to source code only.  Compiled
512		code is still only installed for modules that were requested explicitly
513		or which are required for other modules that were requested explicitly.)
514
515		----------------------------------------------------------------------
516		Name: Matthias Blume (blume (at) tti - c (dot) org)
517		Date: 2004/04/23 17:40:00 CDT
518		Tag: blume-20040423-ieee-scan
519		Description:
520
521		Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
522		in the exponent calculation we get INF or ZERO (depending on the mantissa
523		and the sign of the exponent).
524
525		----------------------------------------------------------------------
526		Name: Matthias Blume (blume (at) tti - c (dot) org)
527		Date: 2004/04/23 10:40:00 CDT
528		Tag: blume-20040423-ml-build
529		Description:
530
531		The ml-build script now terminates with a non-0 status when something
532		goes wrong.
533
534		----------------------------------------------------------------------
535		Name: Matthias Blume (blume (at) tti - c (dot) org)
536		Date: 2004/04/22 16:35:00 CDT
537		Tag: blume-20040422-Option
538		Description:
539
540		Made exception Option to be the same as exception Option.Option
541		(as it should be).
542
543		----------------------------------------------------------------------
544		Name: Allen Leung (leunga (at) reservoir (dot) com)
545		Date: 2004/03/19 14:40:00 EST
546		Tag: leunga-20040319-cygwin-nlffi
547		Description:
548
549		Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
550		of SML/NJ.  The problem is that
551
552		lib = dlopen(NULL, ...)
553		f   = dlsym(lib, "malloc");
554
555		does not work on Windows unless we explicitly export symbols
556		such as 'malloc' during linking.  We fixed this by explicitly
557		exporting the required symbols with the magic gcc incantation:
558
559		-Wl,--export-all cygwin.def
560
561		where cygwin.def is a file containing all the symbols that
562		we wish to export.
563
564		I suspect this is a Windows problem and we'll have to
565		do the same (somehow with windows compilers) when
566		we build the native win32 version with the system
567		calls LoadLibrary/GetProcAddress.
568
569		----------------------------------------------------------------------
570		Name: Matthias Blume (blume (at) tti - c (dot) org)
571		Date: 2004/03/04 16:35:00 CST
572		Tag: blume-20040304-intinf-fmt
573		Description:
574
575		Fixed problem with IntInf.fmt (sign would show up on the right instead
576		of on the left for BIN, OCT, and HEX).
577
578		----------------------------------------------------------------------
579		Name: Matthias Blume (blume (at) tti - c (dot) org)
580		Date: 2004/03/04 11:25:00 CST
581		Tag: blume-20040304-symlinks
582		Description:
583
584		Fixed problem with installer script (unix only) where bin/ml-yacc and
585		friends pointed (via symlinks) to absolute locations instead of just
586		.run-sml.  This was reported by Vesa A Norrman.
587
588		----------------------------------------------------------------------
589		Name: Matthias Blume (blume (at) tti - c (dot) org)
590		Date: 2004/02/13 14:50:00 CST
591		Tag: Release_110_45
592		Description:
593
594		New working version (110.45). New bootfiles.
595
596		----------------------------------------------------------------------
597		Name: Matthias Blume (blume (at) tti - c (dot) org)
598		Date: 2004/01/26 15:15:15 CST
599		Tag: blume-20040126-toplevel
600		Description:
601
602		Improved handling of exceptions at the interactive toplevel.
603
604		----------------------------------------------------------------------
605		Name: Matthias Blume (blume (at) tti - c (dot) org)
606		Date: 2004/01/26 11:25:00
607		Tag: blume-20040126-app
608		Description:
609
610		Type of top-level "app" corrected.
611		Added code for setting vp_limitPtrMask to Win32-specific runtime.
612
613		----------------------------------------------------------------------
614		Name: Matthias Blume (blume (at) tti - c (dot) org)
615		Date: 2003/11/18 17:10 CST
616		Tag: blume-20031118-basis-fiddle
617		Description:
618
619		- changed Timer interface to what might become the spec
620		- POSIX_FLAGS -> BIT_FLAGS according to spec
621		- some other minor discrepancies wrt. spec eliminated
622
623		----------------------------------------------------------------------
624		Name: Matthias Blume (blume (at) tti - c (dot) org)
625		Date: 2003/11/06 12:00:00 CST
626		Tag: Release_110_44
627		Description:
628
629		New working version (110.44). New bootfiles.
630
631		----------------------------------------------------------------------
632		Name: Matthias Blume (blume (at) tti - c (dot) org)
633		Date: 2003/11/04 11:50:00 CST
634		Tag: blume-20031104-move-libraries
635		Description:
636
637		Eliminated the "dont_move_libraries" directive in config/targets.
638		(The mechanism was broken and could not be fixed easily.  Moreover,
639		there does not seem to be any reason not to move all libraries into
640		lib during installation.  I originally implemented this directive as a
641		backward-compatibility feature when I first introduced the new CM.
642		Now that things have been stable for a long time and going back to the
643		old CM is not an option, there is no reason to keep it around.)
644
645		----------------------------------------------------------------------
646		Name: Matthias Blume (blume (at) tti - c (dot) org)
647		Date: 2003/11/03 16:00:00 CST
648		Tag: blume-20031103-installdir
649		Description:
650
651		Made installer honor INSTALLDIR variable again.  (Thanks to Chris
652		Richards for pointing out the problem and providing the solution.)
653
654		----------------------------------------------------------------------
655		Name: Matthias Blume (blume (at) tti - c (dot) org)
656		Date: 2003/10/01 17:05:00 CDT
657		Tag: blume-20031001-lal-mlrisc
658		Description:
659
660		MLRISC bug fix from Lal.
661
662		----------------------------------------------------------------------
663		Name: Matthias Blume (blume (at) tti - c (dot) org)
664		Date: 2003/09/30 16:10:00 CDT
665		Tag: blume-20030930-primio-bat
666		Description:
667
668		1. Added openVector, nullRd, and nullWr to PRIM_IO.
669		2. Improved .bat files (for Win32 port) to make things work under Win95.
670		(thanks to Aaron S. Hawley for this one)
671
672		----------------------------------------------------------------------
673		Name: Matthias Blume (blume (at) tti - c (dot) org)
674		Date: 2003/09/26 16:05:00 CDT
675		Tag: blume-20030926-wrappriv
676		Description:
677
678		Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
679
680		----------------------------------------------------------------------
681		Name: Matthias Blume (blume (at) tti - c (dot) org)
682		Date: 2003/09/26 15:00:00 CDT
683		Tag: blume-20030926-110_43_3
684		Description:
685
686		- additional cleanup
687		- version number bump, NEW BOOTFILES
688
689		----------------------------------------------------------------------
690		Name: Matthias Blume (blume (at) tti - c (dot) org)
691		Date: 2003/09/26 12:00:00 CDT
692		Tag: blume-20030926-ppautoload
693		Description:
694
695		I modified the read-eval-print loop so that the autoloader gets
696		invoked whenever the prettyprinter tries to look up a symbol that
697		is not currently defined in the toplevel environment but which
698		appears in CM's autoload registry.  As a result, we see far fewer of
699		those ?.Foo.Bar.xxx names in the prettyprinter's output.
700
701		In addition to this I tried to clean up some pieces of the Basis
702		implementation (e.g., Socket, Word8Array) in order to prevent other
703		instances of these ?.Foo.Bar.xxx names from being printed.
704
705		The mechanism that picks names for types still needs some work, though.
706		(Right now it seems that if there is a type A.t which is defined to
707		be B.u, but B is unavailable at toplevel, then A.t gets printed as
708		"?.B.u" although the perhaps more sensible solution would be to use
709		"A.t" in this case.  In other words, the prettyprinter should follow
710		a chain of DEFtycs not farther than there are corresponding toplevel
711		names in the current environment.)
712
713		----------------------------------------------------------------------
714		Name: Matthias Blume (blume (at) tti - c (dot) org)
715		Date: 2003/09/24 16:31:00 CDT
716		Tag: blume-20030924-installer
717		Description:
718
719		Another installer tweak:  All the ML code for the installer is now
720		compiled during CMB.make and put into a little library called
721		$smlnj/installer.cm.  The installation then simply invokes
722
723		sml -m $smlnj/installer.cm
724
725		and everything happens automagically.
726
727		Win32:  ML code senses value of environment variable SMLNJ_HOME.
728		Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
729		and BINDIR.
730
731		The new scheme guarantees that the ML code responsible for the installation
732		is in sync with the APIs of the main system.  Also, the installer is
733		somewhat faster because the installer script is precompiled.
734
735		----------------------------------------------------------------------
736		Name: Matthias Blume (blume (at) tti - c (dot) org)
737		Date: 2003/09/24 15:35:00 CDT
738		Tag: blume-20030924-synsock
739		Description:
740
741		Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
742		but excludes all non-blocking operations.  Defined SOCKET (in Basis)
743		and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
744		implementations of non-blocking operations from CML's Socket
745		structure.
746
747		----------------------------------------------------------------------
748		Name: Matthias Blume (blume (at) tti - c (dot) org)
749		Date: 2003/09/24 15:10:05 CDT
750		Tag: blume-20030924-sockets
751		Description:
752
753		1. Fixed SOCKET API and implementation to match Basis spec.
754		This required changing the internal representation of sockets to one
755		that remembers (for each socket file descriptor) whether it is currently
756		blocking or non-blocking.  This state is maintained lazily (i.e., a system
757		call is made only if the state actually needs to change).
758
759		2. OS-specific details of sockets were moved into separate files, thus
760		making it possible to unify the bulk of the socket implementations
761		between Unix and Win32.
762
763		3. CML's socket API changed accordingly.
764		(Note that we need to remove non-blocking functions from this API
765		since they are redundant in the case of CML!)
766
767		4. CML's socket implementation now makes use of non-blocking functions
768		provided by Basis, thus removing all OS-dependent code from this part
769		of CML.
770
771		5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
772
773		----------------------------------------------------------------------
774		Name: Matthias Blume (blume (at) tti - c (dot) org)
775		Date: 2003/09/22 12:10:00 CDT
776		Tag: blume-20030922-110_43_2
777		Description:
778
779		Made a new interim version and bootfiles for developer's bootstrapping
780		convenience.
781
782		110.43.2 -- NEW BOOTFILES
783
784		----------------------------------------------------------------------
785		Name: Matthias Blume (blume (at) tti - c (dot) org)
786		Date: 2003/09/19 15:55:00 CDT
787		Tag: blume-20030919-cmdir
788		Description:
789
790		1. new-install.sh -> install.sh
791		2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
792		3. tweaked installer so that another name instead of .cm can be chosen
793		at install time (by setting the CM_DIR_ARC environment variable
794		during installation); once installation is complete, the name is
795		fixed
796
797		----------------------------------------------------------------------
798		Name: Matthias Blume (blume (at) tti - c (dot) org)
799		Date: 2003/09/18 16:00:00 CDT
800		Tag: blume-20030918-110_43_1
801		Description:
802
803		Made a new interim version and bootfiles for developer's bootstrapping
804		convenience.
805
806		110.43.1 -- NEW BOOTFILES
807
808		----------------------------------------------------------------------
809		Name: Matthias Blume (blume (at) tti - c (dot) org)
810		Date: 2003/09/18 15:20:00 CDT
811		Tag: blume-20030918-misc
812		Description:
813
814		1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
815		this seems to be controversial at the moment) and used those in
816		Posix.ProcEnv.times.
817
818		2. Added Time.{from,to}Nanoseconds to Time.
819
820		3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
821		For example, fromLargeInt never needs to look at more than 3 "big
822		digits" to get its 53 bits of precision.
823
824		----------------------------------------------------------------------
825		Name: Matthias Blume (blume (at) tti - c (dot) org)
826		Date: 2003/09/17 16:30:00 CDT
827		Tag: blume-20030917-real32-slices
828		Description:
829
830		Added an entry to the primitive environment
831		(compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
832		added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
833
834		Removed some of the "magic" constants in real64.sml and replaced them
835		with code that generates these values from their corresponding
836		integer counterparts.
837
838		Made all(?) the slice-related changes to the Basis and made everything
839		compile again...
840
841		----------------------------------------------------------------------
842		Name: Matthias Blume (blume (at) tti - c (dot) org)
843		Date: 2003/09/15 17:45:00 CDT
844		Tag: blume-20030915-rbase
845		Description:
846
847		Fixed bug in Real.fromLargeInt.
848
849		----------------------------------------------------------------------
850		Name: Matthias Blume (blume (at) tti - c (dot) org)
851		Date: 2003/09/13 18:11:00 CDT
852		Tag: blume-20030913-libinstall
853		Description:
854
855		Minor bugfix in config/libinstall (set anchor with path to
856		standalone tool after installing it, otherwise libraries that
857		need ml-lex or ml-yacc won't compile the first time the installer
858		runs).
859
860		----------------------------------------------------------------------
861		Name: Matthias Blume (blume (at) tti - c (dot) org)
862		Date: 2003/09/12 11:45:00 CDT
863		Tag: blume-20030912-various
864		Description:
865
866		- fixed bug in Real.toLargeInt
867		- fixed bug in Posix.ProcEnv.times
868		- changed inputLine functions to return an option
869		- minor installer improvements / bugfixes
870		- changed default @SMLalloc parameter for x86/celeron to 64k
871
872		----------------------------------------------------------------------
873		Name: Matthias Blume (blume (at) tti - c (dot) org)
874		Date: 2003/09/09 22:00:00 CDT
875		Tag: Release_110_43
876		Description:
877
878		New working release 110.43.  New bootfiles.
879
880		----------------------------------------------------------------------
881		Name: Matthias Blume (blume (at) tti - c (dot) org)
882		Date: 2003/09/09 19:20:00 CDT
883		Tag: blume-20030909-installer
884		Description:
885
886		Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
887		Modified config/install.bat to take advantage of it.  Also modified
888		config/install.sh (and called it config/new-install.sh) to take advantage
889		of it on Unix systems.  (The SML code is (supposed to be) platform-
890		independent.)
891
892		The installer can now install everything under Win32
893		as well as under *nix as long as it compiles.
894
895		Other changes:
896
897		- made CML compile again under Win32
898		- made eXene compile under Win32 (by providing a fake structure UnixSock
899		and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
900		- fixed a bug in nowhere: it assumed that type OS.Process.status is the
901		same as type int;  under Win32 it isn't
902		- fixed some slice-related problems in the win32-specific parts of CML
903		- added a functor argument "sameVol" to os-path-fn.sml in the Basis
904		(under Win32, the volume name is case-insensitive, and the
905		OS.Path code compares volume names for equality)
906
907		----------------------------------------------------------------------
908		Name: Matthias Blume (blume (at) tti - c (dot) org)
909		Date: 2003/09/08 11:55:00 CDT
910		Tag: blume-20030908-fullpath
911		Description:
912
913		Made Win32 version of OS.FileSys.fullPath return current directory
914		when given an empty string.  This is what the spec says, and incidentally,
915		CM depends on it.  (CM otherwise goes into an infinite loop in certain
916		cases when presented with the name of a non-existing .cm file.)
917
918		----------------------------------------------------------------------
919		Name: Matthias Blume (blume (at) tti - c (dot) org)
920		Date: 2003/09/04 16:30:00 CDT
921		Tag: blume-20030905-slices-etc
922		Description:
923
924		1. Changed interface to vectors and arrays in Basis to match
925		(draft) Basis spec.
926		2. Added signatures and implementations of slices according to
927		Basis spec.
928		3. Edited source code throughout the system to make it compile again
929		under 1. and 2.  (In some cases code had to be added to have it
930		match the new signatures.)
931		4. MLRISC should be backward-compatible: the copies of the originals
932		of files that needed to change under 3. were retained, the .cm files
933		check the compiler version number and use old versions when
934		appropriate.
935		5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
936		dirstream -> string option (in accordance with Basis spec).
937		6. When generating code that counts lines, ml-lex used function
938		CharVector.foldli, taking advantage of its old interface.
939		This has been replaced with the corresponding code from
940		CharVectorSlice.  (html-lex must be re-lexed!)
941		7. BitArray in smlnj-lib/Util has been extended/modified to match the
942		new MONO_ARRAY signature.  (Do we need BitArraySlice?)
943		8. Removed temporary additions (fromInternal, toInternal) from the
944		(now obsolete) IntInf in smlnj-lib/Util.
945		9. Cleaned up structure Byte.
946		10. Added localOffset, scan, and fromString to Date (according to spec).
947		Cleaned/corrected implementation of Date.
948		(Still need to check for correctness; implement better canonicalizeDate.)
949		11. Added "scan" to signature IEEE_REAL.
950		12. Some improvements to IntInf [in particular: efficiency-hack for
951		mod and rem when second operand is 2 (for parity checks).]
952		13. Changed representation of type Time.time, using a single IntInf.int
953		value counting microseconds.  This considerably simplified the
954		implementation of structure Time.  We now support negative time
955		values; scan and fromString handle signs.
956		14. Functor PrimIO now takes two additional arguments (VectorSlice and
957		ArraySlice).
958
959		----------------------------------------------------------------------
960		Name: Matthias Blume (blume (at) tti - c (dot) org)
961		Date: 2003/08/28 17:00:00 CDT
962		Tag: blume-20030828-intinf
963		Description:
964
965		This is a major update which comes with a version number bump
966		(110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
967		and an implementation of IntInf in the Basis.
968
969		There are a fairly large number of related changes and updates throughout
970		the system:
971
972		Basis:
973		- Implemented IntInf.
974		- Made LargeInt a projection of IntInf (by filtering through INTEGER).
975		- Added some missing Real64 operations, most notably Real.toLargeInt.
976		- Added FixedInt as a synonym for Int32.
977
978		compiler:
979		* Added support for a built-in intinf type.
980		- literals
981		- pattern matching
982		- conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
983		- overloading on literals and operations
984
985		This required adding a primitive type intinf, some additional
986		primops, and implementations for several non-trivial intinf
987		operations in Core.  (The intinf type is completely abstract
988		to the compiler; all operations get delegated back to the Core.)
989
990		* Intinf equality is handled by polyequal.  However, the compiler
991		does not print its usual warning in this case (since polyequal
992		is the right thing to do there).
993
994		* Improved the organization of structure InlineT.
995
996		* A word about conversion primops:
997		If conversions involving intinf do not cancel out during
998		CPS contract, then the compiler must insert calls to Core functions.
999		Since all core access must be resolved already during the FLINT
1000		translate phase, it would be too late a the time of CPS contract
1001		to add new Core calls.  For this reason, conversion primops
1002		for intinf carry two arguments:  1. the numeric argument that
1003		they are supposed to convert, and 2. the Core function that
1004		can help with this conversion if necessary.  If CPS contract
1005		eliminates a primop, then the associated Core function becomes
1006		dead and goes away.  Intinf conversion primops that do not get
1007		eliminated by CPS contract get rewritten into calls of their
1008		core functions by a separate, new phase.
1009
1010		interactive system:
1011		- Control.Print.intinfDepth controls max length of intinf constants
1012		being printed. (Analogous to Control.Print.stringDepth.)
1013		- Cleanup in printutil and pputil: got rid of unused stuff and
1014		duplicates; replaced some of the code with code that makes better
1015		use of library functionality.
1016
1017		CM:
1018		Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1019		are no longer silent.
1020
1021		CKIT:
1022		Fixed mismatched uses of Int32 and LargeInt.  I always decided
1023		in favor of LargeInt -- which is now the same as IntInf.
1024		CKIT-knowledgable people should check whether this is what's
1025		intended and otherwise change things back to using Int32 or
1026		FixedInt.
1027
1028		Throughout the code:
1029		Started using IntInf.int literals and built-in operations
1030		(e.g., comparison with 0) where this seems appropriate.
1031
1032
1033		----------------------------------------------------------------------
1034		Name: Dave MacQueen (dbm@cs.uchicago.edu)
1035		Date: 2003/08/13 11:36:00 CDT
1036		Tag: dbm-20030813-mcz-merge1
1037		Description:
1038
1039		Merging changes from the mcz-branch development branch into trunk.
1040		These changes involve replacement of the emulated old prettyprinter
1041		interface with direct use of the SML/NJ Lib PP library, and fixing
1042		of a couple of bugs (895, 1186) relating to error messages.  A new
1043		prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1044		has been added.
1045
1046		----------------------------------------------------------------------
1047		Name: Matthias Blume (blume (at) tti - c (dot) org)
1048		Date: 2003/08/11 15:45:00 CDT
1049		Tag: blume-20030811-windows
1050		Description:
1051
1052		Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1053
1054		http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1055
1056		This patch restores SML/NJ's ability to run under win32.  There are a
1057		number of changes, including fixes for several bugs that had gone
1058		unnoticed until now:
1059
1060		- uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1061		trying to run the win32 version from within cygwin.)
1062
1063		- There are a number of simple .bat scripts that substitute for their
1064		corresponding Unix shell-scripts.  (See below.)
1065
1066		- The internals of ml-build have been modified slightly.  The main
1067		difference is that instead of calling ".link-sml" (or link-sml.bat)
1068		using OS.Process.system, the ML process delegates this task back
1069		to the script.  Otherwise problems arise in mixed environments such
1070		as Cygwin where scripts look and work like Unix scripts, but
1071		where OS.Process.system cannot run them.
1072
1073		- In CM, the srcpath pickler used native pathname syntax -- which
1074		is incorrect in the case of cross-compilation.  The new pickle format
1075		is independent of platform-specific naming conventions.
1076
1077		- Path configuration files (such as lib/pathconfig) can now choose
1078		between native and standard syntax.  Placing a line of the form
1079
1080		standard!
1081
1082		into the file causes all subsequent paths to be interpreted using
1083		CM standard pathname syntax (= Unix conventions); a line
1084
1085		native!
1086
1087		switches back to native style.  This was needed so that
1088		path config files can be written portably, see src/system/pathconfig.
1089
1090		- Runtime system:
1091
1092		- win32-filesys.c:  get_file_time and set_file_time now
1093		access modification time, not creation time.
1094
1095		- I/O code made aware of new array representation.
1096
1097		- Bug fixes in X86.prim.masm.
1098
1099		- src/system/makeml made aware of win32. (For use under cygwin
1100		and other Unix-environments for windows.)
1101
1102		- In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1103		which caused BinIO.inputAll to fail consistently.
1104
1105		.bat scripts:
1106
1107		Windows .bat scripts assume that SMLNJ_HOME is defined.
1108
1109		- sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1110		applications (sml, ml-yacc, ml-lex).
1111		- ml-build.bat: analogous to ml-build.
1112		- config\install.bat:  Analogous to config/install.sh.  This requires
1113		that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1114		(nmake etc. must be on the path, and vcvars32 must have been run.)
1115		Moreover, sources for ml-lex and ml-yacc need to exist under src,
1116		and the bootfile hierarchy must have been unpacked under
1117		sml.boot.x86-win32.
1118		The script is very primitive and does a poor job at error checking.
1119		It only installs the base system, ml-lex, and ml-yacc.  No other
1120		libraries are being installed (i.e., you get only those that
1121		are part of the compiler.)
1122		- link-sml.bat:  analogous to .link-sml, but not currently used
1123
1124		Unrelated bug fixes:
1125
1126		- ml-nlffigen now exports structures ST_* corresponding to incomplete
1127		types.
1128		- Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1129		otherwise.)
1130
1131		----------------------------------------------------------------------
1132		Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1133		Date: 2003/06/17
1134		Tag: macqueen-20030617-bug895
1135		Description:
1136
1137		Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1138		Tag will be used for new development branch (mcz-branch) for
1139		use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1140
1141		----------------------------------------------------------------------
1142		Name: Matthias Blume (blume (at) tti - c (dot) org)
1143		Date: 2003/05/27 16:55:00 CDT
1144		Tag: blume-20030527-polyeq
1145		Description:
1146
1147		Tried to eliminated most cases of polymorphic equality.
1148
1149		----------------------------------------------------------------------
1150		Name: Matthias Blume (blume (at) tti - c (dot) org)
1151		Date: 2003/05/21 17:45:00 CDT
1152		Tag: blume-20030517-complete
1153		Description:
1154
1155		Two changes:
1156
1157		1. Added a flag for controlling whether non-exhaustive bindings will
1158		be treated as errors (default is false).
1159		2. Cleaned up the *entire* source tree so that CMB.make goes through
1160		without a single non-exhaustive match- or bind warning.
1161
1162		----------------------------------------------------------------------
1163		Name: Matthias Blume (blume (at) tti - c (dot) org)
1164		Date: 2003/05/17 10:20:00 CDT
1165		Tag: blume-20030517-absyn
1166		Description:
1167
1168		1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1169
1170		This mainly affects the quality of error messages.  However, some
1171		of the code is now more straightforward than before. (Treatment of
1172		the above four constructs in translate.sml is much simpler than
1173		the "macro-expansion" that was going on before.  Plus, the mach-
1174		compiler no longer gets invoked just to be able to compile an
1175		if-expression.)
1176
1177		2. The ErrorMsg.Error exception is now caught and absorbed by the
1178		interactive loop.
1179
1180		----------------------------------------------------------------------
1181		Name: Allen Leung
1182		Date: 2003/05/16 13:05:00 CDT
1183		Tag: leunga-20030516-cygwin-runtime
1184		Description:
1185
1186		Ported the runtime system to cygwin, which uses the unix
1187		x86-unix bin files.  Missing/buggy features:
1188
1189		o getnetbyname, getnetbyaddr: these functions seem to be missing in
1190		the Cygwin library.
1191		o Ctrl-C handling may be flaky.
1192		o Windows system calls and Windows I/O are not supported.
1193
1194		A new set of binfiles is located at:
1195
1196		http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1197
1198		This is only needed for bootstrapping the cygwin version of smlnj.
1199		Other x86 versions can use the existing binfiles.
1200
1201		----------------------------------------------------------------------
1202		Name: Matthias Blume
1203		Date: 2003/04/08 15:42:00 CDT
1204		Tag: blume-20030408-listpair
1205		Description:
1206
1207		1. Added a target 'mlrisc' to installer.
1208
1209		2. Added missing elements to structure ListPair.
1210
1211		----------------------------------------------------------------------
1212		Name: Allen Leung
1213		Date: 2003/01/07 10:40:00 EST
1214		Tag: leunga-20030107-int-rem
1215		Description:
1216
1217		Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1218		The arguments to the SUBL instruction were swapped.
1219
1220		----------------------------------------------------------------------
1221		Name: Matthias Blume
1222		Date: 2002/12/12 16:25:00 EST
1223		Tag: blume-20021212-risc-ra
1224		Description:
1225
1226		Fixed a serious bug in the rewrite code for FP spilling/reloading that
1227		sent the RA into an infinite loop when floating point registers get
1228		spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1229		110.37 and 110.38.)
1230		There was another set of potential problems related to the handling of
1231		MLRISC annotations (but those did not yet cause real problems, apparently).
1232
1233		----------------------------------------------------------------------
1234		Name: Matthias Blume
1235		Date: 2002/12/06 22:40:00 EST
1236		Tag: blume-20021206-cm-fileid
1237		Description:
1238
1239		Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1240		This fixes the problem of CM getting confused by files that suddenly
1241		change their identity (e.g., by getting unlinked and recreated by some
1242		text editor such as vi).  There might be a better/cheaper/cleaner way
1243		of doing this, but for now this will have to do.
1244
1245		----------------------------------------------------------------------
1246		Name: Matthias Blume
1247		Date: 2002/10/28 09:50:00 EST
1248		Tag: blume-20021028-typecheck
1249		Description:
1250
1251		Exported structure Typecheck from $smlnj/viscomp/core.cm.
1252
1253		----------------------------------------------------------------------
1254		Name: Matthias Blume
1255		Date: 2002/10/17 09:10:00 EDT
1256		Tag: Release_110_42
1257		Description:
1258
1259		In good old tradition, there has been a slight hiccup so that we have
1260		to patch 110.42 after the fact.   The old release tag has been replaced
1261		(see below).
1262
1263		The change solves a problem with two competing approaches the
1264		configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1265		each other's way.
1266
1267		This change only affects the runtime system code and the installer script.
1268		(No new bootfiles.)
1269
1270		----------------------------------------------------------------------
1271		Name: Matthias Blume
1272		Date: 2002/10/16 12:00:00 EDT
1273		Tag: Release_110_42_removed
1274		Description:
1275
1276		New working release.  New bootfiles.
1277
1278		----------------------------------------------------------------------
1279		Name: Matthias Blume
1280		Date: 2002/10/10 13:10:00 EDT
1281		Tag: blume-20021010-ppc-divs
1282		Description:
1283
1284		The mltree operator DIVS must be implemented with an overflow check on
1285		the PPC because the hardware indicates divide-by-zero using "overflow" as
1286		well.
1287
1288		----------------------------------------------------------------------
1289		Name: Matthias Blume
1290		Date: 2002/07/23 11:45:00
1291		Tag: blume-20020723-smlnj-home
1292		Description:
1293
1294		Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1295		then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1296		CM_PATHCONFIG is also set), the path configuration file is assumed
1297		to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1298		the entire tree to some other place and everything will "just work".
1299
1300		(Companion commands such as ml-build and ml-makedepend also sense this
1301		variable.)
1302
1303		----------------------------------------------------------------------
1304		Name: Matthias Blume
1305		Date: 2002/07/12 21:19:00 EDT
1306		Tag: blume-20020712-liveness
1307		Description:
1308
1309		Exported two useful "step" functions from liveness module (MLRISC).
1310
1311		----------------------------------------------------------------------
1312		Name: Matthias Blume
1313		Date: 2002/07/05 16:00 EDT
1314		Tag: Release_110_41
1315		Description:
1316
1317		New working release.  New bootfiles.
1318
1319		----------------------------------------------------------------------
1320		Name: Matthias Blume
1321		Date: 2002/07/05 10:25:00 EDT
1322		Tag: blume-20020705-btimp
1323		Description:
1324
1325		Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1326		other clients can set up backtracing support.
1327
1328		----------------------------------------------------------------------
1329		Name: Matthias Blume
1330		Date: 2002/06/25 17:23:00 EDT
1331		Tag: blume-20020625-fpmax
1332		Description:
1333
1334		Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1335		case.  (The sense of the isNaN test was reversed -- which made min and
1336		max always return their first argument.)
1337
1338		----------------------------------------------------------------------
1339		Name: Matthias Blume
1340		Date: 2002/06/11
1341		Tag: blume-20020611-unixpath
1342		Description:
1343
1344		Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1345
1346		----------------------------------------------------------------------
1347		Name: Matthias Blume
1348		Date: 2002/06/10 16:35:00 EDT
1349		Tag: blume-20020610-ieeereal
1350		Description:
1351
1352		I back-ported my implementation of IEEEReal.fromString from the
1353		idlbasis-devel branch so that we can test it.
1354
1355		Another small change is that ppDec tries to give more information
1356		than just "<sig>" in the case of functors.  However, this code is
1357		broken in some mysterious way if the functor's body's signature
1358		has not been declared by ascription but gets inferred from the
1359		implementation.  This needs fixing...
1360
1361		----------------------------------------------------------------------
1362		Name: Matthias Blume
1363		Date: 2002/05/31
1364		Tag: blume-20020531-btrace-mode
1365		Description:
1366
1367		Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1368		the wayside when I switched over to using Controls everywhere.)
1369
1370		----------------------------------------------------------------------
1371		Name: Lal George
1372		Date: 2002/05/23 12:21:40 EDT
1373		Tag: george-20020523-visual-labels
1374		Description:
1375
1376		Labels are now displayed in the graphical output to make
1377		the fall-through and target blocks obvious.
1378
1379		----------------------------------------------------------------------
1380		Name: Matthias Blume
1381		Date: 2002/05/22 11:03:00 EDT
1382		Tag: blume-20020522-shrink
1383		Description:
1384
1385		John tweaked yesterday's fix for 1131 to handle an out-of-memory
1386		situation that comes up when allocating huge arrays.
1387
1388		----------------------------------------------------------------------
1389		Name: Matthias Blume
1390		Date: 2002/05/21 16:00:00 EDT
1391		Tag: Release_110_40
1392		Description:
1393
1394		New working release (110.40).  New bootfiles.
1395
1396		[Also: John Reppy fixed GC bug 1131.]
1397
1398		----------------------------------------------------------------------
1399		Name: Matthias Blume
1400		Date: 2002/05/21 12:35:00 EDT
1401		Tag: blume-20020521-cmdoc
1402		Description:
1403
1404		CM documentation update.
1405
1406		----------------------------------------------------------------------
1407		Name: Matthias Blume
1408		Date: 2002/05/21 10:55:00 EDT
1409		Tag: blume-20020521-misc
1410		Description:
1411
1412		- John tweaked runtime to be silent on heap export (except when
1413		GC messages are on).
1414		- I added a few more things (cross-compiling versions of CMB) to
1415		config/preloads (as suggestions).
1416
1417		----------------------------------------------------------------------
1418		Name: Matthias Blume
1419		Date: 2002/05/20 22:25:00 EDT
1420		Tag: blume-20020520-controls
1421		Description:
1422
1423		- Added ControlUtil structure to control-lib.cm.
1424		- Use it throughout.
1425		- Used Controls facility to define MLRISC controls (as opposed to
1426		registering MLRISC control ref cells with Controls after the fact)
1427		- Fixed messed-up controls priorities.
1428
1429		* Removed again all the stuff from config/preloads that one wouldn't
1430		be able to preload at the time the initial heap image is built.
1431		(Many libraries, e.g., CML, do not exist yet at this time.  The
1432		only libraries that can be preloaded via config/preloads are those
1433		that come bundled with the bootfiles.)
1434
1435		----------------------------------------------------------------------
1436		Name: Matthias Blume
1437		Date: 2002/05/20 10:59:00 EDT
1438		Tag: blume-20020520-preloads
1439		Description:
1440
1441		Added a lot of commented-out suggestions for things to be included
1442		in config/preloads.
1443
1444		----------------------------------------------------------------------
1445		Name: Allen Leung
1446		Date: 2002/05/18 14:20:00 EDT
1447		Tag: leunga-20020518-mdl
1448		Description:
1449
1450		o Made the mdl tool stuff compile and run again.
1451		o I've disabled all the stuff that depends on RTL specifications; they
1452		are all badly broken anyway.
1453
1454		----------------------------------------------------------------------
1455		Name: Matthias Blume
1456		Date: 2002/05/17 16:49:00 EDT
1457		Tag: blume-20020517-controls
1458		Description:
1459
1460		0. John Reppy made several modifications to the SML/NJ library.
1461		In particular, there is a shiny new controls-lib.cm.
1462
1463		1. Pushed new controls interface through compiler so that everything
1464		compiles again.
1465
1466		2. Added FormatComb and FORMAT_COMB to the CML version of the
1467		SML/NJ library (so that CML compiles again).
1468
1469		3. Modified init scripts because XXX_DEFAULT environment variables
1470		are no longer with us.  (Boot-time initialization is now done
1471		using the same environment variables that are also used for
1472		startup-time initialization of controls.)
1473
1474		----------------------------------------------------------------------
1475		Name: Lal George
1476		Date: 2002/05/15 09:20:10 EDT
1477		Tag: george-20020515-pseudo-op-decls
1478		Description:
1479
1480		All pseudo-ops emitted before the first segment declaration
1481		such as TEXT, DATA, and BSS directives are assumed to be global
1482		declarations and are emitted first in the assembly file. This is
1483		useful in a number of situations where one has pseudo-ops that are not
1484		specific to any segment, and also works around the constraint that one
1485		cannot have client pseudo-ops in the TEXT segment.
1486
1487		Because no segment is associated with these declarations it is
1488		an error to allocate any space or objects before the first segment
1489		directive and an exception will be raised. However, we cannot make
1490		this check for client pseudo-ops.
1491
1492		These top level declarations are a field in the CFG graph_info.
1493		In theory you can continue to add to this field after the CFG has been
1494		built -- provided you know what you are doing;-)
1495
1496		----------------------------------------------------------------------
1497		Name: Matthias Blume
1498		Date: 2002/05/13 16:40:00 EDT
1499		Tag: blume-20020513-pp-etc
1500		Description:
1501
1502		A few minor bugfixes:
1503
1504		- Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1505		(Bogus printouts for redefined bindings still occur.  Compiler
1506		bug should no longer occur now.  We need to redo the prettyprinter
1507		from scratch.)
1508
1509		- CM pathname printer now also adds escape sequences for ( and )
1510
1511		- commend and docu fixes for ml-nlffi
1512
1513		----------------------------------------------------------------------
1514		Name: Matthias Blume
1515		Date: 2002/05/10 16:40:00 EDT
1516		Tag: blume-20020510-erg-textio
1517		Description:
1518
1519		Applied the following bugfix provided by Emden Gansner:
1520
1521		Output is corrupted when outputSubstr is used rather than output.
1522
1523		The problem occurs when a substring
1524
1525		ss = (s, dataStart, dataLen)
1526
1527		where dataStart > 0, fills a stream buffer with avail bytes left.
1528		avail bytes of s, starting at index dataStart, are copied into the
1529		buffer, the buffer is flushed, and then the remaining dataLen-avail
1530		bytes of ss are copied into the beginning of the buffer. Instead of
1531		starting this copy at index dataStart+avail in s, the current code
1532		starts the copy at index avail.
1533
1534		Fix:
1535		In text-io-fn.sml, change line 695 from
1536		val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1537		to
1538		val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1539
1540		----------------------------------------------------------------------
1541		Name: Matthias Blume
1542		Date: 2002/04/12 13:55:00 EDT
1543		Tag: blume-20020412-assyntax
1544		Description:
1545
1546		1. Grabbed newer assyntax.h from the XFree86 project.
1547		2. Fiddled with how to compile X86.prim.asm without warnings.
1548		3. (Very) Minor cleanup in CM.
1549
1550		----------------------------------------------------------------------
1551		Name: Matthias Blume
1552		Date: 2002/04/01 (no joke!) 17:07:00 EST
1553		Tag: blume-20020401-x86div
1554		Description:
1555
1556		Added full support for div/mod/rem/quot on the x86, using the machine
1557		instruction's two results (without clumsily recomputing the remainder)
1558		directly where appropriate.
1559
1560		Some more extensive power-of-two support was added to the x86 instruction
1561		selector (avoiding expensive divs, mods, and muls where they can be
1562		replaced with cheaper shifts and masks).  However, this sort of thing
1563		ought to be done earlier, e.g., within the CPS optimizer so that
1564		all architectures benefit from it.
1565
1566		The compiler compiles to a fixed point, but changes might be somewhat
1567		fragile nevertheless.  Please, report any strange things that you might
1568		see wrt. div/mod/quot/rem...
1569
1570		----------------------------------------------------------------------
1571		Name: Matthias Blume
1572		Date: 2002/03/29 17:22:00
1573		Tag: blume-20020329-div
1574		Description:
1575
1576		Fixed my broken div/mod logic.  Unfortunately, this means that the
1577		inline code for div/mod now has one more comparison than before.
1578		Fast paths (quotient > 0 or remainder = 0) are not affected, though.
1579		The problem was with quotient = 0, because that alone does not tell
1580		us which way the rounding went.  One then has to look at whether
1581		remainder and divisor have the same sign...  :(
1582
1583		Anyway, I replaced the bootfiles with fresh ones...
1584
1585		----------------------------------------------------------------------
1586		Name: Matthias Blume
1587		Date: 2002/03/29 14:10:00 EST
1588		Tag: blume-20020329-inlprims
1589		Description:
1590
1591		NEW BOOTFILES!!!    Version number bumped to 110.39.3.
1592
1593		Primops have changed. This means that the bin/boot-file formats have
1594		changed as well.
1595
1596		To make sure that there is no confusion, I made a new version.
1597
1598
1599		CHANGES:
1600
1601		* removed REMT from mltree (remainder should never overflow).
1602
1603		* added primops to deal with divisions of all flavors to the frontend
1604
1605		* handled these primops all the way through so they map to their respective
1606		MLRISC support
1607
1608		* used these primops in the implementation of Int, Int32, Word, Word32
1609
1610		* removed INLDIV, INLMOD, and INLREM as they are no longer necessary
1611
1612		* parameterized INLMIN, INLMAX, and INLABS by a numkind
1613
1614		* translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
1615		floating point
1616
1617		* used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
1618		and Real (but Real.abs maps to a separate floating-point-only primop)
1619
1620
1621		TODO items:
1622
1623		* Hacked Alpha32 instruction selection, disabling the selection of REMx
1624		instructions because the machine instruction encoder cannot handle
1625		them.  (Hppa, PPC, and Sparc instruction selection did not handle
1626		REM in the first place, and REM is supported by the x86 machine coder.)
1627
1628		* Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
1629		selection phase.  (The two can be streamlined because the hardware
1630		delivers both quotient and remainder at the same time anyway.)
1631
1632		* Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
1633		(Currently the behavior is inconsistent both across architectures and
1634		wrt. the draft Basis spec.)
1635
1636		* Word8 should eventually be handled natively, too.
1637
1638		* There seems to be one serious bug in mltree-gen.sml.  It appears, though,
1639		as if there currently is no execution path that could trigger it in
1640		SML/NJ.  (The assumptions underlying functions arith and promotable do not
1641		hold for things like multiplication and division.)
1642
1643		----------------------------------------------------------------------
1644		Name: Matthias Blume
1645		Date: 2002/03/27 16:27:00 EST
1646		Tag: blume-20020327-mlrisc-divisions
1647		Description:
1648
1649		Added support for all four division operations (ML's div, mod, quot,
1650		and rem) to MLRISC.  In the course of doing so, I also rationalized
1651		the naming (no more annoying switch-around of DIV and QUOT), by
1652		parameterizing the operation by div_rounding_mode (which can be either
1653		DIV_TO_ZERO or DIV_TO_NEGINF).
1654
1655		The generic MLTreeGen functor takes care of compiling all four
1656		operations down to only round-to-zero div.
1657
1658		Missing pieces:
1659
1660		* Doing something smarter than relying on MLTreeGen on architectures
1661		like, e.g., the x86 where hardware division delivers both quotient and
1662		remainder at the same time.  With this, the implementation of the
1663		round-to-neginf operations could be further streamlined.
1664
1665		* Remove inlining support for div/mod/rem from the frontend and replace it
1666		with primops that get carried through to the backend.  Do this for all
1667		int and word types.
1668
1669		----------------------------------------------------------------------
1670		Name: Matthias Blume
1671		Date: 2002/03/25 17:25:00 EST
1672		Tag: blume-20020325-divmod
1673		Description:
1674
1675		I improved (hopefully without breaking them) the implementation of Int.div,
1676		Int.mod, and Int.rem.   For this, the code in translate.sml now takes
1677		advantage of the following observations:
1678
1679		Let  q = x quot y      r = x rem y
1680		d = x div  y      m = x mod y
1681
1682		where "quot" is the round-to-zero version of integer division that
1683		hardware usually provides.  Then we have:
1684
1685		r = x - q * y        where neither the * nor the - will overflow
1686		d = if q >= 0 orelse x = q * y then q else q - 1
1687		where neither the * nor the - will overflow
1688		m = if q >= 0 orelse r = 0 then r else r + y
1689		where the + will not overflow
1690
1691		This results in substantial simplification of the generated code.
1692		The following table shows the number of CFG nodes and edges generated
1693		for
1694		fun f (x, y) = x OPER y
1695		(* with OPER \in div, mod, quot, rem *)
1696
1697
1698		OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
1699		--------------------------------------------------------
1700		div |         24 |         39 |         12 |         16
1701		mod |         41 |         71 |         12 |         16
1702		quot |          8 |         10 |          8 |         10
1703		rem |         10 |         14 |          8 |         10
1704
1705
1706		----------------------------------------------------------------------
1707		Name: Matthias Blume
1708		Date: 2002/03/25 22:06:00 EST
1709		Tag: blume-20020325-cprotobug
1710		Description:
1711
1712		Fixed a bug in cproto (c prototype decoder).
1713
1714		----------------------------------------------------------------------
1715		Name: Matthias Blume
1716		Date: 2002/03/25 16:00:00 EST
1717		Tag: blume-20020325-raw-primops
1718		Description:
1719
1720		I did some cleanup to Allen's new primop code and
1721		replaced yesterday's bootfiles with new ones.
1722		(But they are stored in the same place.)
1723
1724		----------------------------------------------------------------------
1725		Name: Matthias Blume
1726		Date: 2002/03/24 22:40:00 EST
1727		Tag: blume-20020324-bootfiles
1728		Description:
1729
1730		Made the bootfiles that Allen asked for.
1731
1732		----------------------------------------------------------------------
1733		Name: Allen Leung
1734		Date: 2002/03/23 15:50:00 EST
1735		Tag: leunga-20020323-flint-cps-rcc-primops
1736		Description:
1737
1738		1. Changes to FLINT primops:
1739
1740		(* make a call to a C-function;
1741		* The primop carries C function prototype information and specifies
1742		* which of its (ML-) arguments are floating point. C prototype
1743		* information is for use by the backend, ML information is for
1744		* use by the CPS converter. *)
1745		| RAW_CCALL of { c_proto: CTypes.c_proto,
1746		ml_args: ccall_type list,
1747		ml_res_opt: ccall_type option,
1748		reentrant : bool
1749		} option
1750		(* Allocate uninitialized storage on the heap.
1751		* The record is meant to hold short-lived C objects, i.e., they
1752		* are not ML pointers.  With the tag, the representation is
1753		* the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1754		*)
1755		| RAW_RECORD of {tag:bool,sz:int}
1756		and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1757
1758		2.  These CPS primops are now overloaded:
1759
1760		rawload of {kind:numkind}
1761		rawstore of {kind:numkind}
1762
1763		The one argument form is:
1764
1765		rawload {kind} address
1766
1767		The two argument form is:
1768
1769		rawload {kind} [ml object, byte-offset]
1770
1771		3. RAW_CCALL/RCC now takes two extra arguments:
1772
1773		a. The first is whether the C call is reentrant, i.e., whether
1774		ML state should be saved and restored.
1775		b. The second argument is a string argument specifying the name of
1776		library and the C function.
1777
1778		These things are currently not handled in the code generator, yet.
1779
1780		4. In CProto,
1781
1782		An encoding type of "bool" means "ml object" and is mapped into
1783		C prototype of PTR.  Note that "bool" is different than "string",
1784		even though "string" is also mapped into PTR, because "bool"
1785		is assigned an CPS type of BOGt, while "string" is assigned INT32t.
1786
1787		5. Pickler/unpicker
1788
1789		Changed to handle RAW_RECORD and newest RAW_CCALL
1790
1791		6. MLRiscGen,
1792
1793		1. Changed to handle the new rawload/rawstore/rawrecord operators.
1794		2. Code for handling C Calls has been moved to a new module CPSCCalls,
1795		in the file CodeGen/cpscompile/cps-c-calls.sml
1796
1797		7. Added the conditional move operator
1798
1799		condmove of branch
1800
1801		to cps.  Generation of this is still buggy so it is currently
1802		disabled.
1803
1804		----------------------------------------------------------------------
1805	Name: Lal George	Name: Lal George
1806	Date: 2002/03/22 14:18:25 EST	Date: 2002/03/22 14:18:25 EST
1807	Tag: george-20020322-cps-branch-prob	Tag: george-20020322-cps-branch-prob

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.1171   changed lines   Added in v.1620

---

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