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

Parent Directory | Revision Log

---

Revision 1791 - (view) (download)
Original Path: sml/trunk/HISTORY

1 :	dbm	570	This is the HISTORY file for the Yale SML/NJ CVS repository.
2 :
3 :			An entry should be made for _every_ commit to the repository.
4 :			The entries in this file will be used when creating the README
5 :			for new versions, so keep that in mind when writing the
6 :			description.
7 :
8 :			The form of an entry should be:
9 :
10 :	mblume	1615	Name:
11 :	mblume	1480	Date: yyyy/mm/dd
12 :			Tag: <post-commit CVS tag>
13 :			Description:
14 :
15 :			----------------------------------------------------------------------
16 :	mblume	1791	Name: Matthias Blume (blume (at) tti - c (dot) org)
17 :			Date: 2005/05/09 21:55:00 CDT
18 :			Tag: blume-20050509-word64
19 :			Description:
20 :
21 :			Added a hack to the existing hack known as Word64 to make fromString
22 :			behave correctly. I am still not sure whether Word64.scan will work
23 :			as specified with respect to the interaction of radix and prefix.
24 :
25 :			----------------------------------------------------------------------
26 :	allenleung	1785	Name: Allen Leung (leunga (at) reservoir (dot) com)
27 :			Date: 2005/05/04 11:50:00 EST
28 :			Tag: leunga-20050504-checkgc
29 :			Description:
30 :
31 :			Added a gc protocol checking phase. This phase is enabled with
32 :			the flag "check-gc". "debug-check-gc" turns on the verbose mode.
33 :
34 :			----------------------------------------------------------------------
35 :	mblume	1741	Name: Matthias Blume (blume (at) tti - c (dot) org)
36 :	mblume	1784	Date: 2005/05/04 10:45:00 CDT
37 :			Tag: blume-20050504-intinf
38 :			Description:
39 :
40 :			Fixed a bug in the implementation of div and mod for IntInf.
41 :			Thanks to Neophytos Michael for reporting the problem.
42 :
43 :			----------------------------------------------------------------------
44 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
45 :	mblume	1783	Date: 2005/05/04 10:35:00 CDT
46 :			Tag: blume-20050504-join
47 :			Description:
48 :
49 :			Added a "join" combinator to the ParserComb module in smlnj-lib.cm.
50 :
51 :			----------------------------------------------------------------------
52 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
53 :	mblume	1770	Date: 2005/02/28 23:40:00 CST
54 :			Tag: blume-20050228-mVar
55 :			Description:
56 :
57 :			Fixed serious bug (brown paper bag variety) in new implementation of
58 :			structure Atom in CML. (I had accidentally used a mailbox instead of
59 :			an mvar, leaving the door open for races.)
60 :
61 :			----------------------------------------------------------------------
62 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
63 :	mblume	1761	Date: 2005/02/25 15:00:00 CST
64 :			Tag: Release_110_53
65 :			Description:
66 :
67 :			New working version (110.53). NEW BOOTFILES!
68 :
69 :			----------------------------------------------------------------------
70 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
71 :	mblume	1760	Date: 2005/02/25 14:50:00 CST
72 :			Tag: blume-20050225-susp
73 :			Description:
74 :
75 :			Brought back SMLofNJ.Susp. The underlying suspension type is the one
76 :			implemented in Core -- which means that it is the same as the one used
77 :			by the lazy extension.
78 :
79 :			----------------------------------------------------------------------
80 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
81 :	mblume	1759	Date: 2005/02/24 16:50:00 CST
82 :			Tag: blume-20050224-cml-atom
83 :			Description:
84 :
85 :			Simpler and at the same time more general implementation of structure
86 :			Atom in CML.
87 :
88 :			----------------------------------------------------------------------
89 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
90 :	mblume	1758	Date: 2005/02/15 17:35:00 CST
91 :			Tag: blume-20050215-tools
92 :			Description:
93 :
94 :			Created new "tools" directory under "src" and moved "TraceDebugProf"
95 :			there.
96 :
97 :			----------------------------------------------------------------------
98 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
99 :	mblume	1755	Date: 2005/02/10 17:55:00 CST
100 :			Tag: blume-20050210-longlong
101 :			Description:
102 :
103 :			Implemented "long long" arguments and results for NLFFI. (Only the
104 :			PPC/MacOS implementation is complete, the other backends still need to
105 :			be updated.)
106 :
107 :			----------------------------------------------------------------------
108 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
109 :	mblume	1751	Date: 2005/01/24 17:40:00 CST
110 :			Tag: blume-20050124-mlyacc
111 :			Description:
112 :
113 :			Minor cleanup in ML-Yacc rule printing mechanism. This should fix a
114 :			problem with certain "as" patterns which previously got rendered
115 :			using incorrect syntax.
116 :
117 :			----------------------------------------------------------------------
118 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
119 :	mblume	1750	Date: 2005/01/18 12:00:00 CST
120 :			Tag: blume-20050118-profile
121 :			Description:
122 :
123 :			Made time profiling code (interrupt handler) in runtime system aware
124 :			of new array representation.
125 :
126 :			----------------------------------------------------------------------
127 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
128 :	mblume	1747	Date: 2005/01/14 18:00:00 CST
129 :			Tag: blume-20050114-heap2exec
130 :			Description:
131 :
132 :			Implemented new (but still experimental) heap2exec facility. This is
133 :			tested under Mac OS X and should work under Linux (will test shortly).
134 :			It will probably also work on the Sparc (will test some time later).
135 :
136 :			- removed old "HACKED_STANDALONE" hack from runtime
137 :
138 :			To be able to test this, uncomment the request for "heap2asm" in
139 :			config/targets prior to installation. (Notice that this is different
140 :			from "heap2exec" mentioned below. Not a typo.)
141 :
142 :			To perform an actual test, run the command
143 :
144 :			$ bin/heap2exec heapfile execfile
145 :
146 :			(You can put heap2exec on your shell's path.)
147 :
148 :			For example, run
149 :
150 :			$ bin/heap2exec bin/.heap/ml-yacc.ppc-darwin mly
151 :
152 :			This will create a standalone executable called "mly" which you can
153 :			then invoke directly as a command.
154 :
155 :			----------------------------------------------------------------------
156 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
157 :	mblume	1746	Date: 2005/01/07 11:44:00 CST
158 :			Tag: blume-20050107-mlstring
159 :			Description:
160 :
161 :			fixed off-by-one error in ML_STRING macro (globals.c)
162 :
163 :			----------------------------------------------------------------------
164 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
165 :	mblume	1744	Date: 2004/12/23 18:00:00 CST
166 :			Tag: blume-20041223-santa
167 :			Description:
168 :
169 :			Made ml-build script "smarter" (but only very little).
170 :
171 :			----------------------------------------------------------------------
172 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
173 :	mblume	1742	Date: 2004/12/21 15:05:00 CST
174 :			Tag: blume-20041221-longlong
175 :			Description:
176 :
177 :			* Implemented access to signed and unsigned long long data in NLFFI.
178 :			(The parameter-passing part of the picture has not complete. But
179 :			data structure access seems to work.)
180 :
181 :			* Fixed CM's incorrect assumption that the PPC is little-endian.
182 :			(On the Mac, it is big-endian. And that's currently our only
183 :			PPC platform.)
184 :
185 :			----------------------------------------------------------------------
186 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
187 :	mblume	1741	Date: 2004/12/21 12:50:00 CST
188 :			Tag: blume-20041221-memory
189 :			Description:
190 :
191 :			Some cleanup in the $c/memory.cm library: separated some concerns by
192 :			moving allocation code and memory access code each into their own
193 :			files.
194 :
195 :			----------------------------------------------------------------------
196 :	allenleung	1740	Name: Allen Leung (leunga (at) reservoir (dot) com)
197 :			Date: 2004/12/17 16:12:00 EST
198 :			Tag: leunga-20041217-cygwin-smlnj-home
199 :			Description:
200 :
201 :			The Unix I/O library of SML/NJ on cygwin does not understand
202 :			Windows style pathname, so problems arise when SMLNJ_HOME is set
203 :			to a Windows style pathname. _run-sml now converts SMLNJ_HOME
204 :			to a POSIX pathname on cygwin.
205 :
206 :			----------------------------------------------------------------------
207 :	mblume	1737	Name: Matthias Blume (blume (at) tti - c (dot) org)
208 :	mblume	1736	Date: 2004/12/16 13:35:00 CST
209 :			Tag: Release_110_52
210 :			Description:
211 :
212 :			Last-minute changes incorporated into 110.52. Release tag moved.
213 :
214 :			The changes:
215 :			- HashString.hashString' -> HashString.hashSubstring
216 :			- bug fix in UnivariateStats
217 :
218 :			----------------------------------------------------------------------
219 :	mblume	1702	Name: Matthias Blume (blume (at) tti - c (dot) org)
220 :	mblume	1733	Date: 2004/12/15 23:40:00 CST
221 :			Tag: blume-20041215-hashSubstring
222 :			Description:
223 :
224 :			- HashString.hashString' -> HashString.hashSubstring
225 :			- corresponding changes in atom.sml
226 :			- "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
227 :			added some comments
228 :
229 :			----------------------------------------------------------------------
230 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
231 :	mblume	1728	Date: 2004/12/15 15:30:00 CST
232 :	mblume	1736	Tag: (Release_110_52)
233 :	mblume	1728	Description:
234 :
235 :			New working version (110.52). NEW BOOTFILES!
236 :
237 :			----------------------------------------------------------------------
238 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
239 :	mblume	1727	Date: 2004/12/15 12:45:00 CST
240 :			Tag: blume-20041215-spaces
241 :			Description:
242 :
243 :			More on the space problem (this time for Win32).
244 :
245 :			----------------------------------------------------------------------
246 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
247 :	mblume	1725	Date: 2004/12/14 17:30:00 CST
248 :			Tag: blume-20041214-spaces
249 :			Description:
250 :
251 :			Hacked some of the scripts (in particular: the installer) to cope with
252 :			spaces in filenames a bit better. But beware: the current "solution"
253 :			is likely still full of bugs and inherently incomplete. (We need to
254 :			do away with those shell scripts for a comprehensive solution.)
255 :
256 :			----------------------------------------------------------------------
257 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
258 :	mblume	1724	Date: 2004/12/13 14:45:00 CST
259 :			Tag: blume-20041213-ml-makedepend
260 :			Description:
261 :
262 :			Fixed bug in code for ml-makedepend.
263 :
264 :			----------------------------------------------------------------------
265 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
266 :	mblume	1720	Date: 2004/12/09 16:30:00 CST
267 :			Tag: blume-20041209-statistics
268 :			Description:
269 :
270 :			Added two simple but potentially useful statistics modules to SML/NJ Library.
271 :			(See CHANGES file there.)
272 :
273 :			----------------------------------------------------------------------
274 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
275 :	mblume	1718	Date: 2004/12/01 16:50:00 CST
276 :			Tag: blume-20041201-atom
277 :			Description:
278 :
279 :			smlnj-lib:
280 :
281 :			Added function HashString.hashString' for substrings.
282 :			Hand-inlined CharVector.foldl into HashString (for speed).
283 :			Modified implementation of structure Atom to avoid extracting
284 :			strings from substrings unless necessary.
285 :			(Also see CHANGES file for smlnj-lib.)
286 :
287 :			----------------------------------------------------------------------
288 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
289 :	mblume	1717	Date: 2004/11/24 22:15:00 CST
290 :			Tag: blume-20041124-cml
291 :			Description:
292 :
293 :			Made sure CML compiles when Position = Int64.
294 :
295 :			----------------------------------------------------------------------
296 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
297 :	mblume	1715	Date: 2004/11/24 14:30:00 CST
298 :			Tag: blume-20041124-position
299 :			Description:
300 :
301 :			The compiler can now be compiled in a mode that makes structure Position
302 :			equal to Int64. The default, however, is unchanged (Position = Int31)
303 :			for the time being.
304 :
305 :			To enable 64-bit positions, use the following procedure:
306 :
307 :			1. Start sml
308 :			2. Autoload $smlnj/cmb.cm (if not already autoloaded)
309 :			3. Type
310 :			#set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
311 :			4. Run CMB.make() as usual.
312 :
313 :			This is barely tested. The only test so far was a little SML program
314 :			counting the number of characters in an 8-gigabyte file by
315 :			reading it character-by-character. That test was successful.
316 :
317 :			In support of 64-bit positions, a number of new functions have been
318 :			added to the runtime system.
319 :
320 :			----------------------------------------------------------------------
321 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
322 :	mblume	1716	Date: 2004/11/23 14:45:00 CST
323 :			Tag: blume-20041123-useFile
324 :	mblume	1712	Description:
325 :
326 :			Fixed a problem with unhelpful error messages related to problems with
327 :			.cm- or .sml files that appear as part of the sml command line.
328 :
329 :			----------------------------------------------------------------------
330 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
331 :	mblume	1703	Date: 2004/11/18 15:40:00 CST
332 :			Tag: Release_110_51
333 :			Description:
334 :
335 :			New working version (110.51). NEW BOOTFILES!
336 :
337 :			----------------------------------------------------------------------
338 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
339 :	mblume	1702	Date: 2004/11/18 15:35:00 CST
340 :			Tag: <none>
341 :			Description:
342 :
343 :			Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
344 :			Joemann).
345 :
346 :			----------------------------------------------------------------------
347 :	allenleung	1701	Name: Allen Leung (leunga (at) reservoir (dot) com)
348 :			Date: 2004/11/17 16:05:21 EST 2004
349 :			Tag: leunga-20041117-mlrisc-live-kill
350 :			Description:
351 :
352 :			Added support for MLTree constructs LIVE and KILL
353 :			to all the architectures.
354 :
355 :			----------------------------------------------------------------------
356 :	mblume	1687	Name: Matthias Blume (blume (at) tti - c (dot) org)
357 :	mblume	1689	Date: 2004/11/13 00:20:00 CST
358 :			Tag: blume-20041113-versiontool
359 :			Description:
360 :
361 :			- Stripped down the versiontool: It now only handles the version number.
362 :			The date string is generated at bootstrap time (during makeml).
363 :
364 :			- In a previous commit, fixed a minor issue with how polyequal is being
365 :			translated. In particular, the code now "looks through" abstractions.
366 :			This results in slightly fewer polyEqual warnings and hopefully slightly
367 :			more efficient code. Important examples for where this matters are
368 :			the new int64 and word64 types.
369 :
370 :			----------------------------------------------------------------------
371 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
372 :	mblume	1687	Date: 2004/11/12 00:30:00 CST
373 :			Tag: blume-20041112-int64
374 :			Description:
375 :
376 :			Structure Int64 fully hooked in. (The implementation is not very
377 :			efficient, though.)
378 :
379 :			----------------------------------------------------------------------
380 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
381 :	mblume	1686	Date: 2004/11/11 17:30:00 CST
382 :			Tag: blume-20041111-more64
383 :			Description:
384 :
385 :			All the pieces of Word64 are now there, with the exception of the
386 :			conversions from and to LargeWord. (Eventually these need to be identities,
387 :			but for the time being they don't even make sense because LargeWord is
388 :			32-bit wide.)
389 :
390 :			Also started to add similar support for Int64, but major pieces of that
391 :			are still missing.
392 :
393 :			----------------------------------------------------------------------
394 :	mblume	1668	Name: Matthias Blume (blume (at) tti - c (dot) org)
395 :	mblume	1684	Date: 2004/11/11 00:15:00 CST
396 :			Tag: blume-20041111-word64
397 :			Description:
398 :
399 :			Structure Word64 is now (almost) complete, word literals and patterns
400 :			seem to work. There are a few odd pieces missing. In particular,
401 :			I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
402 :			at the moment.
403 :
404 :			Making Word64 official would mean that LargeWord becomes Word64. But
405 :			this requires extreme care because most word-word conversions have to
406 :			go through LargeWord, so making a mistake means loss of efficiency or
407 :			worse. Eventually there will be a solution similar to (but actually
408 :			simpler than) what I did with IntInf.
409 :
410 :			----------------------------------------------------------------------
411 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
412 :	mblume	1683	Date: 2004/11/10 18:12:00 CST
413 :			Tag: blume-20041110-64bit
414 :			Description:
415 :
416 :			More 64-bit hacking (but still not even half-way there yet).
417 :			Also, some assorted improvements to the handling of 8-bit words.
418 :
419 :			----------------------------------------------------------------------
420 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
421 :	mblume	1682	Date: 2004/11/09 17:50:00 CST
422 :	mblume	1683	Tag: <none>
423 :	mblume	1682	Description:
424 :
425 :			Started adding some infrastructure for supporting 64-bit int- and
426 :			word-types. (Still in its very early stages.)
427 :
428 :			----------------------------------------------------------------------
429 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
430 :	mblume	1672	Date: 2004/10/28 10:45:00 CDT
431 :			Tag: Release_110_50
432 :			Description:
433 :
434 :			New working version (110.50). NEW BOOTFILES!
435 :
436 :			=====================
437 :
438 :			Also:
439 :
440 :			- Changend config/srcarchiveurl from a file just
441 :			containing the URL string into a file containing
442 :			shell script code. The code has access to the $VERSION variable.
443 :			- Made corresponding changes to config/install.sh and config/unpack.
444 :			- Default contents of config/srcarchiveurl uses $VERSION and
445 :			normally does not have to be edited to reflect a version change.
446 :
447 :			(As a result, a version change can be done by just editing
448 :			config/version, the rest is now automatic.)
449 :
450 :			----------------------------------------------------------------------
451 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
452 :	mblume	1669	Date: 2004/10/27 17:50:00 CDT
453 :			Tag: blume-20041027-btrace-msg
454 :			Description:
455 :
456 :			BackTrace.monitor now also reports the source of the exception that
457 :			triggered the trace.
458 :
459 :			----------------------------------------------------------------------
460 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
461 :	mblume	1668	Date: 2004/10/27 17:20:00 CDT
462 :			Tag: blume-20041027-x86-c-calls
463 :			Description:
464 :
465 :			This is the HISTORY entry for two earlier commits, both concerning
466 :			the x86 c-calls code in MLRISC:
467 :
468 :			- added a missing LOAD in the code that deals with struct arguments
469 :			- made sure the caller does not add the wrong number of bytes to the
470 :			stack pointer after a call of a function returning a struct
471 :			(the callee already pops the implicit argument which points to
472 :			the space reserved for the result)
473 :
474 :			----------------------------------------------------------------------
475 :	allenleung	1662	Name: Allen Leung (leunga (at) reservoir (dot) com)
476 :			Date: 2004/10/24 14:00:00 EST
477 :			Tag: leunga-20041024-x86-gas-fucomip
478 :			Description:
479 :
480 :			John discovered a bug in the syntax of fucomip.
481 :			The opcodes FU?COMIP? have been changed to
482 :
483 :			fu?comip? %st(i), %st
484 :
485 :			----------------------------------------------------------------------
486 :	mblume	1622	Name: Matthias Blume (blume (at) tti - c (dot) org)
487 :	mblume	1656	Date: 2004/10/20 15:06:00 CDT
488 :			Tag: blume-20041020-standalone-backtrace
489 :			Description:
490 :
491 :			Added a mechanism for getting back-trace information from standalone
492 :			programs. Here is how it works:
493 :
494 :			1. The part of the program from which you want to get backtrace
495 :			information (usually the whole program) should be wrapped with
496 :			BackTrace.monitor. This is a (unit->'a)->'a function, and your
497 :			main program could be modified from something like
498 :
499 :			fun main (pgm, args) = ...
500 :
501 :			to
502 :
503 :			fun main (pgm, args) = BackTrace.monitor (fn () => ...)
504 :
505 :			2. To be able to access BackTrace.monitor, you have to add
506 :			library $smlnj-tdp/plugins.cm to the .cm file that contains your
507 :			main function.
508 :
509 :			3. Remove all compiled code (i.e., all the .cm/ subdirectories that
510 :			CM might have created in the past for your project).
511 :
512 :			4. Build the system using this command line:
513 :
514 :			ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
515 :			myprog.cm MyProg.main myprog
516 :
517 :			instead of the usual
518 :
519 :			ml-build myprog.cm MyProg.main myprog
520 :
521 :			I changed a library name:
522 :
523 :			$/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
524 :
525 :			New libraries:
526 :
527 :			$smlnj-tdp/back-trace.cm
528 :			-- when loaded causes the back-trace plugin to be installed
529 :			$smlnj-tdp/coverage.cm
530 :			-- when loaded causes the coverage plugin to be installed
531 :
532 :			----------------------------------------------------------------------
533 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
534 :	mblume	1655	Date: 2004/10/18 16:45:00 CDT
535 :			Tag: blume-20041018-groupowner
536 :			Description:
537 :
538 :			Added an "obsolete" warning for the "group owner" syntax to CM's parser.
539 :
540 :			Eliminated group owner specs from .cm files throughout the source tree.
541 :
542 :			----------------------------------------------------------------------
543 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
544 :	mblume	1653	Date: 2004/10/15 15:45:00 CDT
545 :			Tag: blume-20041015-coverage
546 :			Description:
547 :
548 :			* Test coverage tool added!
549 :			* Further reorganization of tracing-, debugging-, and profiling support:
550 :
551 :			- moved original BTImp -- now called BackTrace -- into a separate
552 :			library called $/trace-debug-profile.cm
553 :			- eliminated all mentions of BTrace from SMLofNJ.Internals
554 :			- only the instrumentation mechanism is now left in the compiler proper
555 :			- BackTrace module is a plugin which is NOT plugged in by default
556 :			- Coverage module is another such plugin
557 :
558 :			To get the benefits of any of these plugin modules, the code in
559 :			question must be compiled with tdp instrumentation turned on. This
560 :			can be done by setting SMLofNJ.Internals.TDP.mode to true.
561 :			(The ref cell is also controlled via the -Ctdp.instrument=... switch.)
562 :
563 :			Plugins are selected at link time. (Pre-compiled instrumented code
564 :			can be re-loaded with different plugins in effect.) When an
565 :			instrumented module is linked, whatever plugins are at that time
566 :			enabled will come into effect for that module.
567 :
568 :			To enable the back-trace plugin, load library $/trace-debug-profile.cm
569 :			and invoke BackTrace.install() (e.g., from the interactive prompt).
570 :			To enable the coverage plugin, load the same library and invoke
571 :			Coverage.install().
572 :
573 :			Back-traces are generated automatically on uncaught exceptions and
574 :			when the code in question explicitly invokes BackTrace.trigger().
575 :
576 :			Coverage (and execution frequency-) information must be queried
577 :			explicitly by calling Coverage.not_covered and Coverage.hot_spots.
578 :
579 :			----------------------------------------------------------------------
580 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
581 :	mblume	1652	Date: 2004/10/14 17:40:00 CDT
582 :			Tag: blume-20041014-tdp-core
583 :			Description:
584 :
585 :			Snapshot of a significant overhaul of how the trace/debug/profile support
586 :			is hooked into the system (specifically: Core and SMLofNJ.Internals).
587 :
588 :			----------------------------------------------------------------------
589 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
590 :	mblume	1650	Date: 2004/10/13 16:34:00 CDT
591 :			Tag: blume-20041013-tdp
592 :			Description:
593 :
594 :			Some rationalization of names:
595 :
596 :			structure BTrace -> structure TDPInstrument
597 :			etc.
598 :
599 :			This is is preparation of using the original back-trace
600 :			instrumentation for other purposes. "TDP" stands for
601 :			Trace/Debug/Profile.
602 :
603 :			The control flag controlling whether instrumentation is on or off is now
604 :			registered under a different name, so instead of running sml as
605 :
606 :			sml -Cinstrument.btrace-mode=true
607 :
608 :			one has to say
609 :
610 :			sml -Ctdp.instrument=true
611 :
612 :			----------------------------------------------------------------------
613 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
614 :	mblume	1645	Date: 2004/10/11 16:37:00 CDT
615 :			Tag: blume-20041011-regions
616 :			Description:
617 :
618 :			Made some minor modifications to elabcore.sml to have source regions
619 :			be propagated more tightly -- resulting in better (i.e., smaller)
620 :			regions being reported in error- and debug messages.
621 :
622 :			----------------------------------------------------------------------
623 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
624 :	mblume	1644	Date: 2004/10/08 22:50:00 CDT
625 :			Tag: blume-20041008-cmkw
626 :			Description:
627 :
628 :			Fixed handling of keywords in .cm files: After seeing "is" the lexer
629 :			treats subsequent occurrences of "group", "library", "source", "is",
630 :			"*", and "-" as ordinary identifiers rather than keywords.
631 :
632 :			Most seriously, this fixes a problem with CM's "shell" tool. The tool
633 :			is supposed to accept a tool argument called "source", but this did
634 :			not work because of the clash with the keyword.
635 :
636 :			----------------------------------------------------------------------
637 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
638 :	mblume	1642	Date: 2004/10/07 16:00:00 CDT
639 :			Tag: blume-20041007-cleanup
640 :			Description:
641 :
642 :			Assorted cleanup work:
643 :
644 :			- got rid of intstrmap in favor of using the library's
645 :			hash table implementation
646 :			- threw out most of the pathnames stuff, as it was not used anyway
647 :			- simplified tokentable implementation
648 :			- fixed some minor spelling errors
649 :
650 :			----------------------------------------------------------------------
651 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
652 :	mblume	1641	Date: 2004/10/06 15:15:15 CDT
653 :			Tag: blume-20041006-handler
654 :			Description:
655 :
656 :			Cleaned up the absyn to reflect the invariant that HANDLE always
657 :			carries a FNexp as part of the type definition. This eliminates some
658 :			superfluous sanity checks at runtime down the road.
659 :
660 :			Some minor cleanup of the btrace code.
661 :
662 :			----------------------------------------------------------------------
663 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
664 :	mblume	1640	Date: 2004/10/01 10:20:30 CDT
665 :			Tag: blume-20041001-slave
666 :			Description:
667 :
668 :			Added hack to make slave mode work in the presence of the version
669 :			tool. (Still, since the master does two passes over the code for
670 :			CMB.make, the release number gets bumped twice when slaves are
671 :			attached. I don't know if this is worth fixing...)
672 :
673 :			----------------------------------------------------------------------
674 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
675 :	mblume	1639	Date: 2004/09/30 10:55:00 CDT
676 :			Tag: blume-20040930-version
677 :			Description:
678 :
679 :			* Moved the "version" magic into its own little library under
680 :	mblume	1672	src/system/smlnj/internal. This avoids expensive reconstruction of
681 :			a stable src/compiler/core.cm.
682 :	mblume	1639
683 :	mblume	1672	* At the same time, structure CompilerVersion is now known as
684 :			structure SMLNJVersion.
685 :	mblume	1639
686 :			* Arranged for the version tool to NOT kick in when rebuilding the system
687 :			(makeml -rebuild, fixpt). Otherwise one would never reach a fixpoint.
688 :			Also, loading the versiontool does not work when rebuilding the system
689 :			because CM is not properly initialized at that time.
690 :
691 :			----------------------------------------------------------------------
692 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
693 :	mblume	1638	Date: 2004/09/29 14:00:00 CDT
694 :			Tag: blume-20040929-autoversion
695 :			Description:
696 :
697 :			Implemented some CM magic to have
698 :			file src/compiler/TopLevel/main/version.sml
699 :			generated automagically.
700 :			The version is taken from two files: config/version and config/release.
701 :			The first is expected to contain a two-part version number such as 110.49.
702 :			The second should contain a single number, but it may be missing.
703 :
704 :			If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
705 :			time the version tool is loaded (which is the first time you say CMB.make),
706 :			then the tool will increment the value stored in config/release every
707 :			time CMB.make is invoked.
708 :
709 :			The binfile format is now insensitive to anything beyond the first
710 :			two components of a version number, so bumping the release does not render
711 :			binfiles incompatible. Auto-bumping can be used to keep track of versions
712 :			during development without invalidating existing binfiles.
713 :
714 :			In any case, every CMB.make updates the date information in version.sml.
715 :			(This is the date that is printed in the banner.)
716 :
717 :			----------------------------------------------------------------------
718 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
719 :	mblume	1634	Date: 2004/09/28 10:53:00 CDT
720 :			Tag: blume-20040928-controls
721 :			Description:
722 :
723 :			Some cleanup of the controls code.
724 :
725 :			----------------------------------------------------------------------
726 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
727 :	mblume	1633	Date: 2004/09/27 22:08:00 CDT
728 :			Tag: blume-20040927-controls
729 :			Description:
730 :
731 :			Added two pieces of functionality to the Controls interface:
732 :
733 :			1. val save'restore: 'a control -> unit -> unit
734 :
735 :			grabs the current value of the control in stage 1 and restores it
736 :			in stage 2.
737 :
738 :			2. val set' : 'a control * 'a -> unit -> unit
739 :
740 :			stores the given value into the control in stage 2 (i.e., delayed)
741 :			but does all error checking in stage 1.
742 :			(This is for string controls that need to do parse their argument --
743 :			something that might fail. In some cases, notably in CM, one
744 :			already knows the intended argument but wants to delay the actual
745 :			assignment until a time when error recovery would be more difficult.)
746 :
747 :			Changed the handling of controls in tool arguments to classes "sml" and
748 :			"lazysml":
749 :			- use Controls.save'restore as a more robust way of restoring the
750 :			old value (in particular: without having to re-parse the string)
751 :			- use controls to handle the "overload" keyword in the init group
752 :			(I believe this change actually fixes a long-standing obscure bug.)
753 :
754 :			----------------------------------------------------------------------
755 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
756 :	mblume	1632	Date: 2004/09/27 17:00:00 CDT
757 :			Tag: blume-20040927-lazysml
758 :			Description:
759 :
760 :			Added a new tool class called "lazysml" to CM's tool chest. The only
761 :			difference to "sml" is that compilation is done with Control.lazysml
762 :			set to true. A source of class "lazysml" is automatically recognized
763 :			by a file name suffix of ".lml".
764 :
765 :			In addition to the above feature, the original class "sml" now also
766 :			supports a tool argument "lazy" which has the same effect. As a
767 :			result, the following three lines are equivalent:
768 :
769 :			foo.sml : lazysml
770 :			foo.sml : sml (lazy)
771 :			foo.sml (lazy)
772 :
773 :			The setting goes into effect both during parsing and during
774 :			compilation. The original setting is restored right after parsing and
775 :			after compilation, respectively.
776 :
777 :			In addition to all the above, there is also a general mechanism to set
778 :			ANY of the "controls" that are available at the command line via
779 :			"-C..." on a per-sml-file basis. The same rules that apply for "lazy"
780 :			apply as well. (In fact, "lazy" is implemented as a special case of
781 :			the general mechanism.)
782 :
783 :			The .cm file syntax uses a new keyword tool argument called "with".
784 :			There are several ways of indicating the desired settings:
785 :
786 :			foo.sml (with:parser.quotations=true)
787 :			foo.sml (with:(name:parser.quotations value:true))
788 :			foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
789 :			foo.sml (with:(name1=value1 name2=value2 ...))
790 :			foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
791 :
792 :			etc.
793 :
794 :			Another possible abbreviation is to leave out the =v or value:v part
795 :			if the name refers to a boolean control (in which case the value is
796 :			taken to be true). Thus, one could get lazy sml also by saying:
797 :
798 :			foo.sml (with:parser.lazy-keyword=true)
799 :			foo.sml (with:parser.lazy-keyword)
800 :			foo.sml (with:(name:parser.lazy-keyword value:true))
801 :			foo.sml (with:(name:parser.lazy-keyword))
802 :
803 :			----------------------------------------------------------------------
804 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
805 :	mblume	1631	Date: 2004/09/24 16:22:00 CDT
806 :			Tag: blume-20040924-ppc-long-branch
807 :			Description:
808 :
809 :			Turned message about "emiting long form of branch" off by default.
810 :			Added a control flag to turn it back on when desired.
811 :
812 :			----------------------------------------------------------------------
813 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
814 :	mblume	1630	Date: 2004/09/24 16:05:00 CDT
815 :			Tag: blume-20040924-rounding
816 :			Description:
817 :
818 :			Applied patch for setting rounding modes under Mac OS X. Thanks to
819 :			Melissa O'Neill for providing the code!
820 :
821 :			----------------------------------------------------------------------
822 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
823 :	mblume	1629	Date: 2004/09/23 17:30:00 CDT
824 :			Tag: blume-20040923-envvars
825 :			Description:
826 :
827 :			1. Changed definition of type ControlRegistry.registry_tree to
828 :			include control_info (i.e., the name of the controlling environment
829 :			variable).
830 :
831 :			2. Added command-line flags -e and -E to print the names of environment
832 :			variables that can be used to control internal settings.
833 :			(This uses the new API mentioned in 1.)
834 :
835 :			----------------------------------------------------------------------
836 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
837 :	mblume	1622	Date: 2004/09/13 16:50:00 CDT
838 :			Tag: Release_110_49
839 :			Description:
840 :
841 :			New working version (110.49). NEW BOOTFILES!
842 :
843 :			----------------------------------------------------------------------
844 :	mblume	1621	Name: Matthias Blume
845 :			Date: 2004/09/13 16:20:00 CDT
846 :			Tag: blume-20040913-config-mlrisc
847 :			Description:
848 :
849 :			Put target "mlrisc" back into the default list.
850 :			(There is no harm in having it, and some users have expressed their
851 :			wish to have "mlrisc" included by default.)
852 :
853 :			----------------------------------------------------------------------
854 :	jhr	1620	Name: John Reppy
855 :			Date: 2004/09/13
856 :			Tag: jhr-20040913-signals
857 :			Description:
858 :
859 :			Fixed the signal masking code to properly nest mask/unmask operations
860 :			on a per-signal basis.
861 :
862 :			----------------------------------------------------------------------
863 :	mblume	1619	Name: Matthias Blume (blume (at) tti - c (dot) org)
864 :			Date: 2004/09/08 13:20:00 CDT
865 :			Tag: blume-20040908-heap-magic
866 :			Description:
867 :
868 :			Bumped the heap macig number to 0x09082004 to account for the changed
869 :			layout of the ML frame under MacOS X.
870 :
871 :			----------------------------------------------------------------------
872 :	allenleung	1618	Name: Allen Leung (leunga (at) reservoir (dot) com)
873 :			Date: 2004/09/03 11:26:00 EST
874 :			Tag: leunga-20040903-cygwin-install
875 :			Description:
876 :
877 :			Added a patch to _arch-n-opsys to enable the Cygwin runtime.
878 :			The Cygwin runtime is turned on by setting the environment
879 :			variable SMLNJ_CYGWIN_RUNTIME to 1.
880 :
881 :			----------------------------------------------------------------------
882 :	mblume	1615	Name: Matthias Blume (blume (at) tti - c (dot) org)
883 :	mblume	1617	Date: 2004/08/31 17:15:00 CDT
884 :			Tag: blume-20040831-core
885 :			Description:
886 :
887 :			Added some exports to src/compiler/core.cm upon request by J. Joemann.
888 :
889 :			----------------------------------------------------------------------
890 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
891 :	mblume	1616	Date: 2004/08/30 17:55:00 CDT
892 :			Tag: blume-20040830-installer
893 :			Description:
894 :
895 :			Upon request by Johannes Joemann:
896 :
897 :			- improved ML code of installer to fall back to coping when renaming
898 :			fails (i.e., when source and target are on different file systems);
899 :			the code compiles but has yet to be tested in anger
900 :			- removed mlrisc from list of default targets (config/targets)
901 :
902 :			----------------------------------------------------------------------
903 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
904 :	mblume	1615	Date: 2004/08/27 17:20:00 CDT
905 :			Tag: blume-20040827-ptreql
906 :			Description:
907 :
908 :			Added ptreql primop to structure InlineT (upon request from Larry
909 :			Paulson).
910 :
911 :			----------------------------------------------------------------------
912 :	allenleung	1613	Name: Allen Leung (leunga (at) reservoir (dot) com)
913 :	allenleung	1614	Date: 2004/08/15 21:21:00 EST
914 :			Tag: leunga-110_48-udgraph
915 :			Description:
916 :
917 :			Another bug fix from Carl Hauser:
918 :
919 :			diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
920 :			> 48c48
921 :			> < | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
922 :			> e::L)
923 :			> ---
924 :			> > | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
925 :			> Without this, any deletion of an edge in an undirected graph does severe
926 :			> violence to the graph.
927 :
928 :			----------------------------------------------------------------------
929 :			Name: Allen Leung (leunga (at) reservoir (dot) com)
930 :	allenleung	1613	Date: 2004/08/10 23:35:00 EST
931 :			Tag: leunga-110_48-ppc
932 :			Description:
933 :
934 :			The IBM/MacOS syntax switch on PPC was incorrectly swapped.
935 :
936 :			----------------------------------------------------------------------
937 :	mblume	1602	Name: Matthias Blume (blume (at) tti - c (dot) org)
938 :	mblume	1615	Date: 2004/08/10 12:00:00 CDT
939 :	mblume	1602	Tag: Release_110_48
940 :			Description:
941 :
942 :			New working version (110.48). NEW BOOTFILES!
943 :
944 :			----------------------------------------------------------------------
945 :	allenleung	1599	Name: Allen Leung (leunga (at) reservoir (dot) com)
946 :			Date: 2004/08/09 12:21:00 EST
947 :			Tag: leunga-110_47-dijsktra
948 :			Description:
949 :
950 :			Bug fix from Carl Hauser:
951 :
952 :			single_source_shortest_paths in dijkstra.sml was observed to get wrong
953 :			answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
954 :
955 :			The problem is that following the expression A.update(dist,s,Num.zero)
956 :			it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
957 :
958 :			----------------------------------------------------------------------
959 :	mblume	1595	Name: Matthias Blume (blume (at) tti - c (dot) org)
960 :			Date: 2004/08/06 18:10:00 CDT
961 :			Tag: blume-20040806-cmdline
962 :			Description:
963 :
964 :			Fiddled with handling of command-line options:
965 :
966 :			* sml now quits after processing the command line
967 :			if -H, -S, -h<n>, or -s<n> appears as the last
968 :			command-line argument
969 :			* a new option -q terminates the session when encountered on
970 :			the command line; subsequent arguments will be ignored
971 :			* bug fixes: short (erroneous) arguments are no longer ignored
972 :			completely
973 :
974 :			----------------------------------------------------------------------
975 :	allenleung	1593	Name: Allen Leung (leunga (at) reservoir (dot) com)
976 :			Date: 2004/08/04 18:17:00 EST
977 :			Tag: leunga-110_47-ppc-ibm-asm
978 :			Description:
979 :
980 :			- Added minimal IBM assembly syntax support for PowerPC.
981 :
982 :			- Cygwin: manually changed the file cygwin.def. Some exported symbols have
983 :			been altered in the runtime. We need an automatic way to keep the file
984 :			in sync.
985 :
986 :			----------------------------------------------------------------------
987 :	mblume	1548	Name: Matthias Blume (blume (at) tti - c (dot) org)
988 :	mblume	1581	Date: 2004/08/04 14:00:00 CDT
989 :			Tag: Release_110_47
990 :			Description:
991 :
992 :			New working version (110.47). NEW BOOTFILES!
993 :
994 :			----------------------------------------------------------------------
995 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
996 :	mblume	1580	Date: 2004/08/03 14:25:00 CDT
997 :			Tag: blume-20040803-callingconv
998 :			Description:
999 :
1000 :			Added low-level support for choosing C calling conventions by
1001 :			twiddling the type of rawccall. (See
1002 :			src/compiler/Semant/types/cproto.sml for details.)
1003 :
1004 :			----------------------------------------------------------------------
1005 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1006 :	mblume	1579	Date: 2004/08/02 15:55:00 CDT
1007 :			Tag: blume-20040802-backout
1008 :			Description:
1009 :
1010 :			Backed out of change to win32-filesys.c. The earlier patch to
1011 :			get_file_time caused CM to produce files with the wrong time stamp.
1012 :
1013 :			----------------------------------------------------------------------
1014 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1015 :	mblume	1573	Date: 2004/08/02 14:45:00 CDT
1016 :			Tag: blume-20040802-nlffi-win32
1017 :			Description:
1018 :
1019 :			Added NLFFI support for Win32, adapted from a patch provided by David
1020 :			Hansel. This is currently completely untested. Also, the issue
1021 :			concerning stdcall vs. ccall is still unresolved.
1022 :
1023 :			----------------------------------------------------------------------
1024 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1025 :	mblume	1572	Date: 2004/07/30 17:55:00 CDT
1026 :			Tag: blume-20040730-various
1027 :			Description:
1028 :
1029 :			Gearing up towards 110.47...
1030 :
1031 :			- various minor bugfixes to ml-nlffigen
1032 :			- a beginning of a manual for nlffi
1033 :
1034 :			- eliminated 'export name=value' in config/install.sh as this does
1035 :			not work with certain versions of /bin/sh
1036 :			(Thanks to David King at Motorola for catching this.)
1037 :
1038 :			- several bugfixes provided or suggested by David Hansel at Reactive Systems:
1039 :			- added a test for tm==NULL to gmtime.c and localtime.c
1040 :			- applied patch for incorrect GetFileTime under win32
1041 :			- toSeconds -> toMilliseconds in Win32/win32-process.sml
1042 :
1043 :			----------------------------------------------------------------------
1044 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1045 :	mblume	1557	Date: 2004/07/21 18:20:00 CDT
1046 :			Tag: blume-20040721-nlffigen
1047 :			Description:
1048 :
1049 :			- Fixed minor issue in ml-nlffigen: Now generate structure T_foo
1050 :			for a typedef to an incomplete type, but leave out the "typ" member.
1051 :			(This is just for consistency.)
1052 :			- Started to produce what is supposed to become better (i.e., comprehensive)
1053 :			documentation of what ml-nlffigen does and produces.
1054 :
1055 :			----------------------------------------------------------------------
1056 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1057 :	mblume	1548	Date: 2004/07/14 16:25:00 CDT
1058 :			Tag: blume-20040714-union
1059 :			Description:
1060 :
1061 :			Added C_UNION to c-calls/c-types.sml and updated the machinery
1062 :			(ml-nlffigen, cproto.sml) that conveys C function interface
1063 :			information to the code generator.
1064 :
1065 :			However, the actual architecture-specific implementation of function
1066 :			arguments and results that are C unions is still not implemented.
1067 :
1068 :			----------------------------------------------------------------------
1069 :	allenleung	1539	Name: Allen Leung (leunga (at) reservoir (dot) com)
1070 :	allenleung	1546	Date: 2004/07/14 14:38:00 EST
1071 :			Tag: leunga-110_46_1-ppc-lwzu
1072 :			Description:
1073 :
1074 :			Added these instructions to the PowerPC architecture:
1075 :			LBZU(X), LHZU(X), LWZU(X),
1076 :			STWU(X), STFDU, STFSU
1077 :
1078 :			etc...
1079 :
1080 :			Note: I haven't added their instruction encoding into the description.
1081 :
1082 :			----------------------------------------------------------------------
1083 :			Name: Allen Leung (leunga (at) reservoir (dot) com)
1084 :	allenleung	1539	Date: 2004/07/13 15:04:00 EST
1085 :			Tag: leunga-110_46_1-ppc-lwarx
1086 :			Description:
1087 :
1088 :			Added the two instructions LWARX and STWCX to the PowerPC
1089 :			instruction set.
1090 :
1091 :			A (untested) rewrite of loop-structure.sml. The old version
1092 :			is completely broken.
1093 :
1094 :			----------------------------------------------------------------------
1095 :	mblume	1479	Name: Matthias Blume (blume (at) tti - c (dot) org)
1096 :	mblume	1537	Date: 2004/07/13 13:50:00 CDT
1097 :			Tag: blume-20040713-nlffi
1098 :			Description:
1099 :
1100 :			- use paramAlloc to report c-calls with too many arguments
1101 :			(for PPC version where parameter area is pre-allocated)
1102 :			- added ccall_maxargspace to machspec (to implement the above)
1103 :			- made "make" commend in CM's "make" tool configurable
1104 :			- added option (default: on) for passing the name of the SML/NJ's "bin"
1105 :			directory to "make"; the call looks like this:
1106 :
1107 :			make <options> SMLNJ_BINDIR=<dir> <target>
1108 :
1109 :			This can be used by the Makefile to, e.g., pick the "right" version
1110 :			of ml-nlffigen.
1111 :			- minor code tweaks
1112 :
1113 :			----------------------------------------------------------------------
1114 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1115 :	mblume	1532	Date: 2004/07/12 22:50:00 CDT
1116 :			Tag: blume-110_46_1-macosx-nlffi
1117 :			Description:
1118 :
1119 :			NLFFI under Mac OS X now working (sort of). This is largely untested,
1120 :			though.
1121 :
1122 :			Note: 1. You have to make a new, clean build of the runtime system.
1123 :			2. There are new BOOTFILES, you have to use them!
1124 :			(Doing the bootstrap process yourself would be *very* painful!
1125 :			If you absolutely have to do it, build the system under
1126 :			a different architecture and then cross-compile.)
1127 :
1128 :			Version bumped to 110.46.1 to account for runtime data format changes.
1129 :
1130 :			----------------------------------------------------------------------
1131 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1132 :	mblume	1515	Date: 2004/06/18 14:30:00 CDT
1133 :			Tag: blume-20040618-unix
1134 :			Description:
1135 :
1136 :			Changed the implementation of structure Unix so that the same stream
1137 :			is returned every time one of the {text,bin}{In,Out}streamOf functions
1138 :			is invoked on the same proc. This is not what the spec currently
1139 :			says -- although IMO it arguably should. (See discussion below.)
1140 :
1141 :			----------------------------------------------------------------------
1142 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1143 :	mblume	1507	Date: 2004/06/17 18:15:00 CDT
1144 :			Tag: Release_110_46
1145 :			Description:
1146 :
1147 :			New working version (110.46). NEW BOOTFILES!
1148 :
1149 :			----------------------------------------------------------------------
1150 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1151 :	mblume	1506	Date: 2004/06/17 17:20:00 CDT
1152 :			Tag: blume-20040617-timer-unix
1153 :			Description:
1154 :
1155 :			Changed the interface of structures Timer and Unix to match the most
1156 :			recent Basis spec.
1157 :
1158 :			In the case of Unix there still seems to be an open/weird issue:
1159 :
1160 :			The {text,bin}{In,Out}streamOf functions are supposed to create
1161 :			fresh streams whenever they are called -- as opposed to have them
1162 :			return the same stream every time. This design is supposed to
1163 :			prevent space leaks caused by proc values hanging on to streams.
1164 :
1165 :			The reap function, on the other hand, is supposed to close the
1166 :			streams. This cannot be done without having a handle on the
1167 :			stream in proc after all...
1168 :
1169 :			I took the liberty to implement the following stopgap solution:
1170 :
1171 :			The proc value hangs on to the most recently created stream(s).
1172 :			Reap closes those. If either or both of the two streams hadn't
1173 :			been created at all yet, then reap will close the corresponding
1174 :			file descriptors directly.
1175 :
1176 :			PS: I don't understand the original space leak argument anymore. If
1177 :			a proc hangs on to the imperative stream, then I/O operations on those
1178 :			will advance the state of the cached stream and avoid the space leak.
1179 :
1180 :			----------------------------------------------------------------------
1181 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1182 :	mblume	1500	Date: 2004/05/28 16:45:00 CDT
1183 :			Tag: blume-20040528-basis
1184 :			Description:
1185 :
1186 :			Added signature PACK_REAL and exported functor PrimIO.
1187 :
1188 :			----------------------------------------------------------------------
1189 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1190 :	mblume	1498	Date: 2004/05/25 16:00:00 CDT
1191 :			Tag: blume-20040525-group-owner
1192 :			Description:
1193 :
1194 :			CM now ignores (but still accepts) the "owner" information in group
1195 :			descriptions. The owner of a group is its next enclosing
1196 :			library. Each group must have a unique owner. (There is a virtual
1197 :	mblume	1505	"toplevel" library that owns groups which are not nested within a real
1198 :	mblume	1498	library.) Previously, each group had to explicitly declare its owner,
1199 :			and CM would check that such a declaration is correct. The new scheme
1200 :			is to have CM check that for each group there is precisely one owning
1201 :			library.
1202 :
1203 :			The advantage of the new scheme is that the programmer no longer needs
1204 :			to maintain the somewhat annoying owner information. The downside is
1205 :			that CM cannot enforce the ownership rule across multiple runs of
1206 :			CM.make. Fortunately, enclosing the same group in two different
1207 :			libraries A and B which are not part of the same program does not
1208 :			cause real problems.
1209 :
1210 :			----------------------------------------------------------------------
1211 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1212 :	mblume	1496	Date: 2004/05/20 16:00:00 CDT
1213 :			Tag: blume-20040520-win32
1214 :			Description:
1215 :
1216 :			Made the win32 version work again. (Strangely, a misplaced comma had
1217 :			slipped into win32-process.c which prevented the runtime from being
1218 :			compiled correctly.)
1219 :
1220 :			Also, included a minor addition to ml-build.bat analogous to what was
1221 :			done in blume-20040519-ml-build.
1222 :
1223 :			----------------------------------------------------------------------
1224 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1225 :	mblume	1494	Date: 2004/05/19 22:10:00 CDT
1226 :			Tag: blume-20040519-ml-build
1227 :			Description:
1228 :
1229 :			Arranged for ml-build to clean up after itself a little bit better.
1230 :			The script generates a temporary SML source file and compiles it using
1231 :			CM, so CM generates metadata (GUID, SKEL, objectfile) for it. It now
1232 :			gets rid of those at the end, so they don't accumulate under .cm.
1233 :
1234 :			This required a minor change to install.sh because the name of the
1235 :			metadata directory (default: .cm) is actually configurable at
1236 :			installation time.
1237 :
1238 :			----------------------------------------------------------------------
1239 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1240 :	mblume	1491	Date: 2004/05/18 15:50:00 CDT
1241 :			Tag: blume-20040518-mkreader
1242 :			Description:
1243 :
1244 :			Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1245 :			implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1246 :
1247 :			----------------------------------------------------------------------
1248 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1249 :	mblume	1489	Date: 2004/05/11 14:35:00 CDT
1250 :			Tag: blume-20040511-win32sock
1251 :			Description:
1252 :
1253 :			Added previously missing support for many socket-related functions
1254 :			under win32. Thanks to David Hansel <hansel@reactive-systems.com>
1255 :			for the voluminous patch!
1256 :
1257 :			(I have not tested this patch under win32 yet.)
1258 :
1259 :	mblume	1490	Here is David's e-mail:
1260 :
1261 :			Hi,
1262 :
1263 :			Attached to this email you find a diff against sml/nj 110.45
1264 :			that will enable socket support under Windows.
1265 :
1266 :			To apply the patch (using unix or cygwin)
1267 :			1) gunzip runtime.diff.gz
1268 :			2) "cd" into "src/runtime" in the source tree of a fresh
1269 :			110.45 installation.
1270 :			3) patch -p 1 < [your/path/to]runtime.diff
1271 :
1272 :			The code compiles fine but has NOT yet been extensively tested.
1273 :			I only ran a few tests for basic socket client functionality
1274 :			(which worked fine). Especially the functions that use ioctl
1275 :			are not tested at all and might not work (see below).
1276 :
1277 :			I implemented this since we want to move to a newer version of sml/nj
1278 :			but need socket support in order to use it. This is the first time I
1279 :			even had a look at the sml/nj source, so please review my changes
1280 :			before making this part of the distribution! Here are a few issues
1281 :			that I think might be better for someone to solve who is more
1282 :			familiar with the sml/nj source (and socket programming):
1283 :
1284 :			- getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1285 :			exception since I could not figure out what the windows equivalent
1286 :			of these functions is
1287 :
1288 :			- In sockets-osdep.h there are a some #include statements that are
1289 :			only used in a few files that include sockets-osdep.h
1290 :
1291 :			- In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1292 :			does not process its return value since I don't know how to
1293 :			report an error upwards.
1294 :
1295 :			- It would probably be good to have a call to WSACleanup() when
1296 :			the library is unloaded (if there is such a possibility).
1297 :			Otherwise I think Windows will take care of this automatically
1298 :			when the process finishes.
1299 :
1300 :			- I used ioctlsocket() as a replacement for ioctl() but I have
1301 :			no idea if that is actually the proper replacement on Windows.
1302 :
1303 :			- All these issues are marked in the code by "FIXME" comments.
1304 :
1305 :			We use sml/nj extensively in our products and are quite happy
1306 :			with it. I hope this contribution will help you.
1307 :
1308 :			Keep up the good work!
1309 :
1310 :			David
1311 :
1312 :	mblume	1489	----------------------------------------------------------------------
1313 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1314 :	mblume	1488	Date: 2004/05/11 14:20:00 CDT
1315 :			Tag: blume-20040511-installml
1316 :			Description:
1317 :
1318 :			Fixed two bugs in installml script. (Thanks to Vesa A. Norrman for
1319 :			the patch.)
1320 :
1321 :			----------------------------------------------------------------------
1322 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1323 :	mblume	1487	Date: 2004/05/11 14:05:00 CDT
1324 :			Tag: blume-20040511-nlffi-netbsd
1325 :			Description:
1326 :
1327 :			Added support for nlffi under netbsd. (Thanks to Vesa A. Norrman for
1328 :			the patch.)
1329 :
1330 :			----------------------------------------------------------------------
1331 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1332 :	mblume	1486	Date: 2004/05/11 12:05:00 CDT
1333 :			Tag: blume-20040511-exports
1334 :			Description:
1335 :
1336 :			As per request by Adam Chlipala <adam@hcoop.net>, extended various
1337 :			export lists in compiler-related .cm-files.
1338 :
1339 :			----------------------------------------------------------------------
1340 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1341 :	mblume	1485	Date: 2004/05/11 11:35:00 CDT
1342 :			Tag: blume-20040511-allsource
1343 :			Description:
1344 :
1345 :			The installer now honors the "src-smlnj" target again, although its meaning
1346 :			has changed from "all sources required for the compiler" to "all sources
1347 :			the installer knows about". In other words, if you enable "src-smlnj"
1348 :			in the "targets" file, then the installer will pull in sources for
1349 :			everything. (Notice that this refers to source code only. Compiled
1350 :			code is still only installed for modules that were requested explicitly
1351 :			or which are required for other modules that were requested explicitly.)
1352 :
1353 :			----------------------------------------------------------------------
1354 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1355 :	mblume	1482	Date: 2004/04/23 17:40:00 CDT
1356 :			Tag: blume-20040423-ieee-scan
1357 :			Description:
1358 :
1359 :			Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1360 :			in the exponent calculation we get INF or ZERO (depending on the mantissa
1361 :			and the sign of the exponent).
1362 :
1363 :			----------------------------------------------------------------------
1364 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1365 :	mblume	1480	Date: 2004/04/23 10:40:00 CDT
1366 :			Tag: blume-20040423-ml-build
1367 :			Description:
1368 :
1369 :			The ml-build script now terminates with a non-0 status when something
1370 :			goes wrong.
1371 :
1372 :			----------------------------------------------------------------------
1373 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1374 :	mblume	1479	Date: 2004/04/22 16:35:00 CDT
1375 :			Tag: blume-20040422-Option
1376 :			Description:
1377 :
1378 :			Made exception Option to be the same as exception Option.Option
1379 :			(as it should be).
1380 :
1381 :			----------------------------------------------------------------------
1382 :	allenleung	1468	Name: Allen Leung (leunga (at) reservoir (dot) com)
1383 :			Date: 2004/03/19 14:40:00 EST
1384 :			Tag: leunga-20040319-cygwin-nlffi
1385 :			Description:
1386 :
1387 :			Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1388 :			of SML/NJ. The problem is that
1389 :
1390 :			lib = dlopen(NULL, ...)
1391 :			f = dlsym(lib, "malloc");
1392 :
1393 :			does not work on Windows unless we explicitly export symbols
1394 :			such as 'malloc' during linking. We fixed this by explicitly
1395 :			exporting the required symbols with the magic gcc incantation:
1396 :
1397 :			-Wl,--export-all cygwin.def
1398 :
1399 :			where cygwin.def is a file containing all the symbols that
1400 :			we wish to export.
1401 :
1402 :			I suspect this is a Windows problem and we'll have to
1403 :			do the same (somehow with windows compilers) when
1404 :			we build the native win32 version with the system
1405 :			calls LoadLibrary/GetProcAddress.
1406 :
1407 :			----------------------------------------------------------------------
1408 :	mblume	1384	Name: Matthias Blume (blume (at) tti - c (dot) org)
1409 :	mblume	1465	Date: 2004/03/04 16:35:00 CST
1410 :			Tag: blume-20040304-intinf-fmt
1411 :			Description:
1412 :
1413 :			Fixed problem with IntInf.fmt (sign would show up on the right instead
1414 :			of on the left for BIN, OCT, and HEX).
1415 :
1416 :			----------------------------------------------------------------------
1417 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1418 :	mblume	1464	Date: 2004/03/04 11:25:00 CST
1419 :			Tag: blume-20040304-symlinks
1420 :			Description:
1421 :
1422 :			Fixed problem with installer script (unix only) where bin/ml-yacc and
1423 :			friends pointed (via symlinks) to absolute locations instead of just
1424 :			.run-sml. This was reported by Vesa A Norrman.
1425 :
1426 :			----------------------------------------------------------------------
1427 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1428 :	mblume	1455	Date: 2004/02/13 14:50:00 CST
1429 :			Tag: Release_110_45
1430 :			Description:
1431 :
1432 :			New working version (110.45). New bootfiles.
1433 :
1434 :			----------------------------------------------------------------------
1435 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1436 :	mblume	1448	Date: 2004/01/26 15:15:15 CST
1437 :			Tag: blume-20040126-toplevel
1438 :			Description:
1439 :
1440 :			Improved handling of exceptions at the interactive toplevel.
1441 :
1442 :			----------------------------------------------------------------------
1443 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1444 :			Date: 2004/01/26 11:25:00
1445 :	mblume	1446	Tag: blume-20040126-app
1446 :			Description:
1447 :
1448 :			Type of top-level "app" corrected.
1449 :	mblume	1447	Added code for setting vp_limitPtrMask to Win32-specific runtime.
1450 :	mblume	1446
1451 :			----------------------------------------------------------------------
1452 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1453 :	mblume	1427	Date: 2003/11/18 17:10 CST
1454 :			Tag: blume-20031118-basis-fiddle
1455 :			Description:
1456 :
1457 :			- changed Timer interface to what might become the spec
1458 :			- POSIX_FLAGS -> BIT_FLAGS according to spec
1459 :			- some other minor discrepancies wrt. spec eliminated
1460 :
1461 :			----------------------------------------------------------------------
1462 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1463 :	mblume	1414	Date: 2003/11/06 12:00:00 CST
1464 :			Tag: Release_110_44
1465 :			Description:
1466 :
1467 :			New working version (110.44). New bootfiles.
1468 :
1469 :			----------------------------------------------------------------------
1470 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1471 :	mblume	1411	Date: 2003/11/04 11:50:00 CST
1472 :			Tag: blume-20031104-move-libraries
1473 :			Description:
1474 :
1475 :			Eliminated the "dont_move_libraries" directive in config/targets.
1476 :			(The mechanism was broken and could not be fixed easily. Moreover,
1477 :			there does not seem to be any reason not to move all libraries into
1478 :			lib during installation. I originally implemented this directive as a
1479 :			backward-compatibility feature when I first introduced the new CM.
1480 :			Now that things have been stable for a long time and going back to the
1481 :			old CM is not an option, there is no reason to keep it around.)
1482 :
1483 :			----------------------------------------------------------------------
1484 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1485 :	mblume	1409	Date: 2003/11/03 16:00:00 CST
1486 :			Tag: blume-20031103-installdir
1487 :			Description:
1488 :
1489 :			Made installer honor INSTALLDIR variable again. (Thanks to Chris
1490 :			Richards for pointing out the problem and providing the solution.)
1491 :
1492 :			----------------------------------------------------------------------
1493 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1494 :	mblume	1400	Date: 2003/10/01 17:05:00 CDT
1495 :			Tag: blume-20031001-lal-mlrisc
1496 :			Description:
1497 :
1498 :			MLRISC bug fix from Lal.
1499 :
1500 :			----------------------------------------------------------------------
1501 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1502 :	mblume	1399	Date: 2003/09/30 16:10:00 CDT
1503 :			Tag: blume-20030930-primio-bat
1504 :	mblume	1398	Description:
1505 :
1506 :			1. Added openVector, nullRd, and nullWr to PRIM_IO.
1507 :			2. Improved .bat files (for Win32 port) to make things work under Win95.
1508 :			(thanks to Aaron S. Hawley for this one)
1509 :
1510 :			----------------------------------------------------------------------
1511 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1512 :	mblume	1395	Date: 2003/09/26 16:05:00 CDT
1513 :			Tag: blume-20030926-wrappriv
1514 :			Description:
1515 :
1516 :			Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1517 :
1518 :			----------------------------------------------------------------------
1519 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1520 :	mblume	1394	Date: 2003/09/26 15:00:00 CDT
1521 :			Tag: blume-20030926-110_43_3
1522 :			Description:
1523 :
1524 :			- additional cleanup
1525 :			- version number bump, NEW BOOTFILES
1526 :
1527 :			----------------------------------------------------------------------
1528 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1529 :	mblume	1393	Date: 2003/09/26 12:00:00 CDT
1530 :			Tag: blume-20030926-ppautoload
1531 :			Description:
1532 :
1533 :			I modified the read-eval-print loop so that the autoloader gets
1534 :			invoked whenever the prettyprinter tries to look up a symbol that
1535 :			is not currently defined in the toplevel environment but which
1536 :			appears in CM's autoload registry. As a result, we see far fewer of
1537 :			those ?.Foo.Bar.xxx names in the prettyprinter's output.
1538 :
1539 :			In addition to this I tried to clean up some pieces of the Basis
1540 :			implementation (e.g., Socket, Word8Array) in order to prevent other
1541 :			instances of these ?.Foo.Bar.xxx names from being printed.
1542 :
1543 :			The mechanism that picks names for types still needs some work, though.
1544 :			(Right now it seems that if there is a type A.t which is defined to
1545 :	mblume	1394	be B.u, but B is unavailable at toplevel, then A.t gets printed as
1546 :			"?.B.u" although the perhaps more sensible solution would be to use
1547 :	mblume	1393	"A.t" in this case. In other words, the prettyprinter should follow
1548 :			a chain of DEFtycs not farther than there are corresponding toplevel
1549 :			names in the current environment.)
1550 :
1551 :			----------------------------------------------------------------------
1552 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1553 :	mblume	1391	Date: 2003/09/24 16:31:00 CDT
1554 :			Tag: blume-20030924-installer
1555 :			Description:
1556 :
1557 :			Another installer tweak: All the ML code for the installer is now
1558 :			compiled during CMB.make and put into a little library called
1559 :			$smlnj/installer.cm. The installation then simply invokes
1560 :
1561 :			sml -m $smlnj/installer.cm
1562 :
1563 :			and everything happens automagically.
1564 :
1565 :			Win32: ML code senses value of environment variable SMLNJ_HOME.
1566 :			Unix: ML code senses values of environment variables ROOT, CONFIGDIR,
1567 :			and BINDIR.
1568 :
1569 :			The new scheme guarantees that the ML code responsible for the installation
1570 :			is in sync with the APIs of the main system. Also, the installer is
1571 :			somewhat faster because the installer script is precompiled.
1572 :
1573 :			----------------------------------------------------------------------
1574 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1575 :	mblume	1390	Date: 2003/09/24 15:35:00 CDT
1576 :			Tag: blume-20030924-synsock
1577 :			Description:
1578 :
1579 :			Added a signature SYNCHRONOUS_SOCKET to basis.cm. This is like SOCKET
1580 :			but excludes all non-blocking operations. Defined SOCKET (in Basis)
1581 :			and CML_SOCKET in terms of SYNCHRONOUS_SOCKET. Removed superfluous
1582 :			implementations of non-blocking operations from CML's Socket
1583 :			structure.
1584 :
1585 :			----------------------------------------------------------------------
1586 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1587 :	mblume	1389	Date: 2003/09/24 15:10:05 CDT
1588 :			Tag: blume-20030924-sockets
1589 :			Description:
1590 :
1591 :			1. Fixed SOCKET API and implementation to match Basis spec.
1592 :			This required changing the internal representation of sockets to one
1593 :			that remembers (for each socket file descriptor) whether it is currently
1594 :			blocking or non-blocking. This state is maintained lazily (i.e., a system
1595 :			call is made only if the state actually needs to change).
1596 :
1597 :			2. OS-specific details of sockets were moved into separate files, thus
1598 :			making it possible to unify the bulk of the socket implementations
1599 :			between Unix and Win32.
1600 :
1601 :			3. CML's socket API changed accordingly.
1602 :			(Note that we need to remove non-blocking functions from this API
1603 :			since they are redundant in the case of CML!)
1604 :
1605 :			4. CML's socket implementation now makes use of non-blocking functions
1606 :			provided by Basis, thus removing all OS-dependent code from this part
1607 :			of CML.
1608 :
1609 :			5. Changed Real64.precision from 52 to 53. Minor cleanup in Real64 code.
1610 :
1611 :			----------------------------------------------------------------------
1612 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1613 :	mblume	1387	Date: 2003/09/22 12:10:00 CDT
1614 :	mblume	1388	Tag: blume-20030922-110_43_2
1615 :	mblume	1387	Description:
1616 :
1617 :			Made a new interim version and bootfiles for developer's bootstrapping
1618 :			convenience.
1619 :
1620 :			110.43.2 -- NEW BOOTFILES
1621 :
1622 :			----------------------------------------------------------------------
1623 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1624 :	mblume	1385	Date: 2003/09/19 15:55:00 CDT
1625 :			Tag: blume-20030919-cmdir
1626 :			Description:
1627 :
1628 :			1. new-install.sh -> install.sh
1629 :			2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1630 :			3. tweaked installer so that another name instead of .cm can be chosen
1631 :			at install time (by setting the CM_DIR_ARC environment variable
1632 :			during installation); once installation is complete, the name is
1633 :			fixed
1634 :
1635 :			----------------------------------------------------------------------
1636 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1637 :	mblume	1384	Date: 2003/09/18 16:00:00 CDT
1638 :			Tag: blume-20030918-110_43_1
1639 :			Description:
1640 :
1641 :			Made a new interim version and bootfiles for developer's bootstrapping
1642 :			convenience.
1643 :
1644 :			110.43.1 -- NEW BOOTFILES
1645 :
1646 :			----------------------------------------------------------------------
1647 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1648 :	mblume	1383	Date: 2003/09/18 15:20:00 CDT
1649 :			Tag: blume-20030918-misc
1650 :			Description:
1651 :
1652 :			1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1653 :			this seems to be controversial at the moment) and used those in
1654 :			Posix.ProcEnv.times.
1655 :
1656 :			2. Added Time.{from,to}Nanoseconds to Time.
1657 :
1658 :			3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1659 :			For example, fromLargeInt never needs to look at more than 3 "big
1660 :			digits" to get its 53 bits of precision.
1661 :
1662 :			----------------------------------------------------------------------
1663 :	mblume	1384	Name: Matthias Blume (blume (at) tti - c (dot) org)
1664 :	mblume	1381	Date: 2003/09/17 16:30:00 CDT
1665 :			Tag: blume-20030917-real32-slices
1666 :			Description:
1667 :
1668 :			Added an entry to the primitive environment
1669 :			(compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1670 :			added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1671 :
1672 :			Removed some of the "magic" constants in real64.sml and replaced them
1673 :			with code that generates these values from their corresponding
1674 :			integer counterparts.
1675 :
1676 :			Made all(?) the slice-related changes to the Basis and made everything
1677 :			compile again...
1678 :
1679 :			----------------------------------------------------------------------
1680 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1681 :	mblume	1375	Date: 2003/09/15 17:45:00 CDT
1682 :			Tag: blume-20030915-rbase
1683 :			Description:
1684 :
1685 :			Fixed bug in Real.fromLargeInt.
1686 :
1687 :			----------------------------------------------------------------------
1688 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1689 :	mblume	1369	Date: 2003/09/13 18:11:00 CDT
1690 :			Tag: blume-20030913-libinstall
1691 :			Description:
1692 :
1693 :			Minor bugfix in config/libinstall (set anchor with path to
1694 :			standalone tool after installing it, otherwise libraries that
1695 :			need ml-lex or ml-yacc won't compile the first time the installer
1696 :			runs).
1697 :
1698 :			----------------------------------------------------------------------
1699 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1700 :	mblume	1368	Date: 2003/09/12 11:45:00 CDT
1701 :			Tag: blume-20030912-various
1702 :			Description:
1703 :
1704 :			- fixed bug in Real.toLargeInt
1705 :			- fixed bug in Posix.ProcEnv.times
1706 :			- changed inputLine functions to return an option
1707 :			- minor installer improvements / bugfixes
1708 :			- changed default @SMLalloc parameter for x86/celeron to 64k
1709 :
1710 :			----------------------------------------------------------------------
1711 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1712 :	mblume	1356	Date: 2003/09/09 22:00:00 CDT
1713 :			Tag: Release_110_43
1714 :			Description:
1715 :
1716 :			New working release 110.43. New bootfiles.
1717 :
1718 :			----------------------------------------------------------------------
1719 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1720 :	mblume	1354	Date: 2003/09/09 19:20:00 CDT
1721 :			Tag: blume-20030909-installer
1722 :			Description:
1723 :
1724 :			Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1725 :			Modified config/install.bat to take advantage of it. Also modified
1726 :			config/install.sh (and called it config/new-install.sh) to take advantage
1727 :			of it on Unix systems. (The SML code is (supposed to be) platform-
1728 :			independent.)
1729 :
1730 :			The installer can now install everything under Win32
1731 :			as well as under *nix as long as it compiles.
1732 :
1733 :			Other changes:
1734 :
1735 :			- made CML compile again under Win32
1736 :			- made eXene compile under Win32 (by providing a fake structure UnixSock
1737 :			and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1738 :			- fixed a bug in nowhere: it assumed that type OS.Process.status is the
1739 :			same as type int; under Win32 it isn't
1740 :			- fixed some slice-related problems in the win32-specific parts of CML
1741 :			- added a functor argument "sameVol" to os-path-fn.sml in the Basis
1742 :			(under Win32, the volume name is case-insensitive, and the
1743 :			OS.Path code compares volume names for equality)
1744 :
1745 :			----------------------------------------------------------------------
1746 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1747 :	mblume	1353	Date: 2003/09/08 11:55:00 CDT
1748 :			Tag: blume-20030908-fullpath
1749 :			Description:
1750 :
1751 :			Made Win32 version of OS.FileSys.fullPath return current directory
1752 :			when given an empty string. This is what the spec says, and incidentally,
1753 :			CM depends on it. (CM otherwise goes into an infinite loop in certain
1754 :			cases when presented with the name of a non-existing .cm file.)
1755 :
1756 :			----------------------------------------------------------------------
1757 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1758 :	mblume	1350	Date: 2003/09/04 16:30:00 CDT
1759 :	mblume	1352	Tag: blume-20030905-slices-etc
1760 :	mblume	1350	Description:
1761 :
1762 :			1. Changed interface to vectors and arrays in Basis to match
1763 :			(draft) Basis spec.
1764 :			2. Added signatures and implementations of slices according to
1765 :			Basis spec.
1766 :			3. Edited source code throughout the system to make it compile again
1767 :			under 1. and 2. (In some cases code had to be added to have it
1768 :			match the new signatures.)
1769 :			4. MLRISC should be backward-compatible: the copies of the originals
1770 :			of files that needed to change under 3. were retained, the .cm files
1771 :			check the compiler version number and use old versions when
1772 :			appropriate.
1773 :			5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1774 :			dirstream -> string option (in accordance with Basis spec).
1775 :			6. When generating code that counts lines, ml-lex used function
1776 :			CharVector.foldli, taking advantage of its old interface.
1777 :			This has been replaced with the corresponding code from
1778 :			CharVectorSlice. (html-lex must be re-lexed!)
1779 :			7. BitArray in smlnj-lib/Util has been extended/modified to match the
1780 :			new MONO_ARRAY signature. (Do we need BitArraySlice?)
1781 :			8. Removed temporary additions (fromInternal, toInternal) from the
1782 :			(now obsolete) IntInf in smlnj-lib/Util.
1783 :			9. Cleaned up structure Byte.
1784 :			10. Added localOffset, scan, and fromString to Date (according to spec).
1785 :			Cleaned/corrected implementation of Date.
1786 :			(Still need to check for correctness; implement better canonicalizeDate.)
1787 :			11. Added "scan" to signature IEEE_REAL.
1788 :			12. Some improvements to IntInf [in particular: efficiency-hack for
1789 :			mod and rem when second operand is 2 (for parity checks).]
1790 :			13. Changed representation of type Time.time, using a single IntInf.int
1791 :			value counting microseconds. This considerably simplified the
1792 :			implementation of structure Time. We now support negative time
1793 :			values; scan and fromString handle signs.
1794 :			14. Functor PrimIO now takes two additional arguments (VectorSlice and
1795 :			ArraySlice).
1796 :
1797 :			----------------------------------------------------------------------
1798 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1799 :	mblume	1347	Date: 2003/08/28 17:00:00 CDT
1800 :			Tag: blume-20030828-intinf
1801 :			Description:
1802 :
1803 :			This is a major update which comes with a version number bump
1804 :			(110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1805 :			and an implementation of IntInf in the Basis.
1806 :
1807 :			There are a fairly large number of related changes and updates throughout
1808 :			the system:
1809 :
1810 :			Basis:
1811 :			- Implemented IntInf.
1812 :			- Made LargeInt a projection of IntInf (by filtering through INTEGER).
1813 :	mblume	1348	- Added some missing Real64 operations, most notably Real.toLargeInt.
1814 :	mblume	1347	- Added FixedInt as a synonym for Int32.
1815 :
1816 :			compiler:
1817 :			* Added support for a built-in intinf type.
1818 :			- literals
1819 :			- pattern matching
1820 :			- conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1821 :			- overloading on literals and operations
1822 :
1823 :			This required adding a primitive type intinf, some additional
1824 :			primops, and implementations for several non-trivial intinf
1825 :			operations in Core. (The intinf type is completely abstract
1826 :			to the compiler; all operations get delegated back to the Core.)
1827 :
1828 :			* Intinf equality is handled by polyequal. However, the compiler
1829 :			does not print its usual warning in this case (since polyequal
1830 :			is the right thing to do there).
1831 :
1832 :			* Improved the organization of structure InlineT.
1833 :
1834 :			* A word about conversion primops:
1835 :			If conversions involving intinf do not cancel out during
1836 :	mblume	1348	CPS contract, then the compiler must insert calls to Core functions.
1837 :	mblume	1347	Since all core access must be resolved already during the FLINT
1838 :			translate phase, it would be too late a the time of CPS contract
1839 :	mblume	1348	to add new Core calls. For this reason, conversion primops
1840 :	mblume	1347	for intinf carry two arguments: 1. the numeric argument that
1841 :			they are supposed to convert, and 2. the Core function that
1842 :			can help with this conversion if necessary. If CPS contract
1843 :	mblume	1348	eliminates a primop, then the associated Core function becomes
1844 :			dead and goes away. Intinf conversion primops that do not get
1845 :			eliminated by CPS contract get rewritten into calls of their
1846 :			core functions by a separate, new phase.
1847 :	mblume	1347
1848 :			interactive system:
1849 :			- Control.Print.intinfDepth controls max length of intinf constants
1850 :			being printed. (Analogous to Control.Print.stringDepth.)
1851 :			- Cleanup in printutil and pputil: got rid of unused stuff and
1852 :			duplicates; replaced some of the code with code that makes better
1853 :			use of library functionality.
1854 :
1855 :			CM:
1856 :			Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1857 :			are no longer silent.
1858 :
1859 :			CKIT:
1860 :			Fixed mismatched uses of Int32 and LargeInt. I always decided
1861 :			in favor of LargeInt -- which is now the same as IntInf.
1862 :			CKIT-knowledgable people should check whether this is what's
1863 :			intended and otherwise change things back to using Int32 or
1864 :			FixedInt.
1865 :
1866 :			Throughout the code:
1867 :	mblume	1348	Started using IntInf.int literals and built-in operations
1868 :	mblume	1347	(e.g., comparison with 0) where this seems appropriate.
1869 :
1870 :
1871 :			----------------------------------------------------------------------
1872 :	macqueen	1343	Name: Dave MacQueen (dbm@cs.uchicago.edu)
1873 :			Date: 2003/08/13 11:36:00 CDT
1874 :			Tag: dbm-20030813-mcz-merge1
1875 :			Description:
1876 :
1877 :			Merging changes from the mcz-branch development branch into trunk.
1878 :			These changes involve replacement of the emulated old prettyprinter
1879 :			interface with direct use of the SML/NJ Lib PP library, and fixing
1880 :			of a couple of bugs (895, 1186) relating to error messages. A new
1881 :			prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1882 :			has been added.
1883 :
1884 :			----------------------------------------------------------------------
1885 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1886 :	mblume	1342	Date: 2003/08/11 15:45:00 CDT
1887 :			Tag: blume-20030811-windows
1888 :			Description:
1889 :
1890 :			Version number bumped to 110.42.9. NEW BOOTFILES!!!
1891 :
1892 :			http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1893 :
1894 :			This patch restores SML/NJ's ability to run under win32. There are a
1895 :			number of changes, including fixes for several bugs that had gone
1896 :			unnoticed until now:
1897 :
1898 :			- uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1899 :			trying to run the win32 version from within cygwin.)
1900 :
1901 :			- There are a number of simple .bat scripts that substitute for their
1902 :			corresponding Unix shell-scripts. (See below.)
1903 :
1904 :			- The internals of ml-build have been modified slightly. The main
1905 :			difference is that instead of calling ".link-sml" (or link-sml.bat)
1906 :			using OS.Process.system, the ML process delegates this task back
1907 :			to the script. Otherwise problems arise in mixed environments such
1908 :			as Cygwin where scripts look and work like Unix scripts, but
1909 :			where OS.Process.system cannot run them.
1910 :
1911 :			- In CM, the srcpath pickler used native pathname syntax -- which
1912 :			is incorrect in the case of cross-compilation. The new pickle format
1913 :			is independent of platform-specific naming conventions.
1914 :
1915 :			- Path configuration files (such as lib/pathconfig) can now choose
1916 :			between native and standard syntax. Placing a line of the form
1917 :
1918 :			standard!
1919 :
1920 :			into the file causes all subsequent paths to be interpreted using
1921 :			CM standard pathname syntax (= Unix conventions); a line
1922 :
1923 :			native!
1924 :
1925 :			switches back to native style. This was needed so that
1926 :			path config files can be written portably, see src/system/pathconfig.
1927 :
1928 :			- Runtime system:
1929 :
1930 :			- win32-filesys.c: get_file_time and set_file_time now
1931 :			access modification time, not creation time.
1932 :
1933 :			- I/O code made aware of new array representation.
1934 :
1935 :			- Bug fixes in X86.prim.masm.
1936 :
1937 :			- src/system/makeml made aware of win32. (For use under cygwin
1938 :			and other Unix-environments for windows.)
1939 :
1940 :			- In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1941 :			which caused BinIO.inputAll to fail consistently.
1942 :
1943 :			.bat scripts:
1944 :
1945 :			Windows .bat scripts assume that SMLNJ_HOME is defined.
1946 :
1947 :			- sml.bat, ml-yacc.bat, ml-lex.bat: Driver scripts for standalone
1948 :			applications (sml, ml-yacc, ml-lex).
1949 :			- ml-build.bat: analogous to ml-build.
1950 :			- config\install.bat: Analogous to config/install.sh. This requires
1951 :			that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1952 :			(nmake etc. must be on the path, and vcvars32 must have been run.)
1953 :			Moreover, sources for ml-lex and ml-yacc need to exist under src,
1954 :			and the bootfile hierarchy must have been unpacked under
1955 :			sml.boot.x86-win32.
1956 :			The script is very primitive and does a poor job at error checking.
1957 :			It only installs the base system, ml-lex, and ml-yacc. No other
1958 :			libraries are being installed (i.e., you get only those that
1959 :			are part of the compiler.)
1960 :			- link-sml.bat: analogous to .link-sml, but not currently used
1961 :
1962 :			Unrelated bug fixes:
1963 :
1964 :			- ml-nlffigen now exports structures ST_* corresponding to incomplete
1965 :			types.
1966 :			- Added getDevice to PP/src/pp-debug-fn.sml. (Would not compile
1967 :			otherwise.)
1968 :
1969 :			----------------------------------------------------------------------
1970 :	macqueen	1337	Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1971 :			Date: 2003/06/17
1972 :			Tag: macqueen-20030617-bug895
1973 :			Description:
1974 :
1975 :			Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1976 :			Tag will be used for new development branch (mcz-branch) for
1977 :			use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1978 :
1979 :			----------------------------------------------------------------------
1980 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1981 :	mblume	1335	Date: 2003/05/27 16:55:00 CDT
1982 :			Tag: blume-20030527-polyeq
1983 :			Description:
1984 :
1985 :			Tried to eliminated most cases of polymorphic equality.
1986 :
1987 :			----------------------------------------------------------------------
1988 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1989 :	mblume	1334	Date: 2003/05/21 17:45:00 CDT
1990 :			Tag: blume-20030517-complete
1991 :			Description:
1992 :
1993 :			Two changes:
1994 :
1995 :			1. Added a flag for controlling whether non-exhaustive bindings will
1996 :			be treated as errors (default is false).
1997 :			2. Cleaned up the *entire* source tree so that CMB.make goes through
1998 :			without a single non-exhaustive match- or bind warning.
1999 :
2000 :			----------------------------------------------------------------------
2001 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
2002 :	mblume	1332	Date: 2003/05/17 10:20:00 CDT
2003 :			Tag: blume-20030517-absyn
2004 :			Description:
2005 :
2006 :			1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
2007 :
2008 :			This mainly affects the quality of error messages. However, some
2009 :			of the code is now more straightforward than before. (Treatment of
2010 :			the above four constructs in translate.sml is much simpler than
2011 :			the "macro-expansion" that was going on before. Plus, the mach-
2012 :			compiler no longer gets invoked just to be able to compile an
2013 :			if-expression.)
2014 :
2015 :			2. The ErrorMsg.Error exception is now caught and absorbed by the
2016 :			interactive loop.
2017 :
2018 :			----------------------------------------------------------------------
2019 :	allenleung	1327	Name: Allen Leung
2020 :			Date: 2003/05/16 13:05:00 CDT
2021 :			Tag: leunga-20030516-cygwin-runtime
2022 :			Description:
2023 :
2024 :			Ported the runtime system to cygwin, which uses the unix
2025 :			x86-unix bin files. Missing/buggy features:
2026 :
2027 :			o getnetbyname, getnetbyaddr: these functions seem to be missing in
2028 :			the Cygwin library.
2029 :			o Ctrl-C handling may be flaky.
2030 :			o Windows system calls and Windows I/O are not supported.
2031 :
2032 :			A new set of binfiles is located at:
2033 :
2034 :			http://www.dorsai.org/~leunga/boot.x86-unix.tgz
2035 :
2036 :			This is only needed for bootstrapping the cygwin version of smlnj.
2037 :			Other x86 versions can use the existing binfiles.
2038 :
2039 :			----------------------------------------------------------------------
2040 :	mblume	1317	Name: Matthias Blume
2041 :			Date: 2003/04/08 15:42:00 CDT
2042 :			Tag: blume-20030408-listpair
2043 :			Description:
2044 :
2045 :			1. Added a target 'mlrisc' to installer.
2046 :
2047 :			2. Added missing elements to structure ListPair.
2048 :
2049 :			----------------------------------------------------------------------
2050 :	leunga	1296	Name: Allen Leung
2051 :			Date: 2003/01/07 10:40:00 EST
2052 :			Tag: leunga-20030107-int-rem
2053 :			Description:
2054 :
2055 :			Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
2056 :			The arguments to the SUBL instruction were swapped.
2057 :
2058 :			----------------------------------------------------------------------
2059 :	blume	1229	Name: Matthias Blume
2060 :	blume	1294	Date: 2002/12/12 16:25:00 EST
2061 :			Tag: blume-20021212-risc-ra
2062 :			Description:
2063 :
2064 :			Fixed a serious bug in the rewrite code for FP spilling/reloading that
2065 :			sent the RA into an infinite loop when floating point registers get
2066 :			spilled. (Because of this bug, e.g., nucleic stopped compiling between
2067 :			110.37 and 110.38.)
2068 :			There was another set of potential problems related to the handling of
2069 :			MLRISC annotations (but those did not yet cause real problems, apparently).
2070 :
2071 :			----------------------------------------------------------------------
2072 :			Name: Matthias Blume
2073 :	blume	1291	Date: 2002/12/06 22:40:00 EST
2074 :			Tag: blume-20021206-cm-fileid
2075 :			Description:
2076 :
2077 :			Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
2078 :			This fixes the problem of CM getting confused by files that suddenly
2079 :			change their identity (e.g., by getting unlinked and recreated by some
2080 :			text editor such as vi). There might be a better/cheaper/cleaner way
2081 :			of doing this, but for now this will have to do.
2082 :
2083 :			----------------------------------------------------------------------
2084 :			Name: Matthias Blume
2085 :	blume	1288	Date: 2002/10/28 09:50:00 EST
2086 :			Tag: blume-20021028-typecheck
2087 :			Description:
2088 :
2089 :			Exported structure Typecheck from $smlnj/viscomp/core.cm.
2090 :
2091 :			----------------------------------------------------------------------
2092 :			Name: Matthias Blume
2093 :	blume	1282	Date: 2002/10/17 09:10:00 EDT
2094 :	blume	1279	Tag: Release_110_42
2095 :			Description:
2096 :
2097 :	blume	1282	In good old tradition, there has been a slight hiccup so that we have
2098 :			to patch 110.42 after the fact. The old release tag has been replaced
2099 :			(see below).
2100 :
2101 :			The change solves a problem with two competing approaches the
2102 :			configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
2103 :			each other's way.
2104 :
2105 :			This change only affects the runtime system code and the installer script.
2106 :			(No new bootfiles.)
2107 :
2108 :			----------------------------------------------------------------------
2109 :			Name: Matthias Blume
2110 :			Date: 2002/10/16 12:00:00 EDT
2111 :			Tag: Release_110_42_removed
2112 :			Description:
2113 :
2114 :	blume	1279	New working release. New bootfiles.
2115 :
2116 :			----------------------------------------------------------------------
2117 :			Name: Matthias Blume
2118 :	blume	1273	Date: 2002/10/10 13:10:00 EDT
2119 :			Tag: blume-20021010-ppc-divs
2120 :			Description:
2121 :
2122 :			The mltree operator DIVS must be implemented with an overflow check on
2123 :			the PPC because the hardware indicates divide-by-zero using "overflow" as
2124 :			well.
2125 :
2126 :			----------------------------------------------------------------------
2127 :			Name: Matthias Blume
2128 :	blume	1261	Date: 2002/07/23 11:45:00
2129 :			Tag: blume-20020723-smlnj-home
2130 :			Description:
2131 :
2132 :			Sml now senses the SMLNJ_HOME environment variable. If this is set,
2133 :			then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
2134 :			CM_PATHCONFIG is also set), the path configuration file is assumed
2135 :			to be in $SMLNJ_HOME/lib/pathconfig. This way one can easily move
2136 :			the entire tree to some other place and everything will "just work".
2137 :
2138 :			(Companion commands such as ml-build and ml-makedepend also sense this
2139 :			variable.)
2140 :
2141 :			----------------------------------------------------------------------
2142 :			Name: Matthias Blume
2143 :	blume	1259	Date: 2002/07/12 21:19:00 EDT
2144 :			Tag: blume-20020712-liveness
2145 :			Description:
2146 :
2147 :			Exported two useful "step" functions from liveness module (MLRISC).
2148 :
2149 :			----------------------------------------------------------------------
2150 :			Name: Matthias Blume
2151 :	blume	1253	Date: 2002/07/05 16:00 EDT
2152 :			Tag: Release_110_41
2153 :			Description:
2154 :
2155 :			New working release. New bootfiles.
2156 :
2157 :			----------------------------------------------------------------------
2158 :			Name: Matthias Blume
2159 :	blume	1250	Date: 2002/07/05 10:25:00 EDT
2160 :			Tag: blume-20020705-btimp
2161 :			Description:
2162 :
2163 :			Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
2164 :			other clients can set up backtracing support.
2165 :
2166 :			----------------------------------------------------------------------
2167 :			Name: Matthias Blume
2168 :	blume	1249	Date: 2002/06/25 17:23:00 EDT
2169 :			Tag: blume-20020625-fpmax
2170 :			Description:
2171 :
2172 :			Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
2173 :			case. (The sense of the isNaN test was reversed -- which made min and
2174 :			max always return their first argument.)
2175 :
2176 :			----------------------------------------------------------------------
2177 :			Name: Matthias Blume
2178 :	blume	1240	Date: 2002/06/11
2179 :			Tag: blume-20020611-unixpath
2180 :			Description:
2181 :
2182 :			Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
2183 :
2184 :			----------------------------------------------------------------------
2185 :			Name: Matthias Blume
2186 :	blume	1238	Date: 2002/06/10 16:35:00 EDT
2187 :			Tag: blume-20020610-ieeereal
2188 :			Description:
2189 :
2190 :			I back-ported my implementation of IEEEReal.fromString from the
2191 :			idlbasis-devel branch so that we can test it.
2192 :
2193 :			Another small change is that ppDec tries to give more information
2194 :			than just "<sig>" in the case of functors. However, this code is
2195 :			broken in some mysterious way if the functor's body's signature
2196 :			has not been declared by ascription but gets inferred from the
2197 :			implementation. This needs fixing...
2198 :
2199 :			----------------------------------------------------------------------
2200 :			Name: Matthias Blume
2201 :	blume	1229	Date: 2002/05/31
2202 :			Tag: blume-20020531-btrace-mode
2203 :			Description:
2204 :
2205 :			Resurrected SMLofNJ.Internals.BTrace.mode. (It accidentally fell by
2206 :			the wayside when I switched over to using Controls everywhere.)
2207 :
2208 :			----------------------------------------------------------------------
2209 :	george	1222	Name: Lal George
2210 :			Date: 2002/05/23 12:21:40 EDT
2211 :			Tag: george-20020523-visual-labels
2212 :			Description:
2213 :
2214 :			Labels are now displayed in the graphical output to make
2215 :			the fall-through and target blocks obvious.
2216 :
2217 :			----------------------------------------------------------------------
2218 :	blume	1204	Name: Matthias Blume
2219 :	blume	1220	Date: 2002/05/22 11:03:00 EDT
2220 :			Tag: blume-20020522-shrink
2221 :			Description:
2222 :
2223 :			John tweaked yesterday's fix for 1131 to handle an out-of-memory
2224 :			situation that comes up when allocating huge arrays.
2225 :
2226 :			----------------------------------------------------------------------
2227 :			Name: Matthias Blume
2228 :	blume	1215	Date: 2002/05/21 16:00:00 EDT
2229 :			Tag: Release_110_40
2230 :			Description:
2231 :
2232 :			New working release (110.40). New bootfiles.
2233 :
2234 :			[Also: John Reppy fixed GC bug 1131.]
2235 :
2236 :			----------------------------------------------------------------------
2237 :			Name: Matthias Blume
2238 :	blume	1212	Date: 2002/05/21 12:35:00 EDT
2239 :			Tag: blume-20020521-cmdoc
2240 :			Description:
2241 :
2242 :			CM documentation update.
2243 :
2244 :			----------------------------------------------------------------------
2245 :			Name: Matthias Blume
2246 :	blume	1211	Date: 2002/05/21 10:55:00 EDT
2247 :			Tag: blume-20020521-misc
2248 :			Description:
2249 :
2250 :			- John tweaked runtime to be silent on heap export (except when
2251 :			GC messages are on).
2252 :			- I added a few more things (cross-compiling versions of CMB) to
2253 :			config/preloads (as suggestions).
2254 :
2255 :			----------------------------------------------------------------------
2256 :			Name: Matthias Blume
2257 :	blume	1208	Date: 2002/05/20 22:25:00 EDT
2258 :			Tag: blume-20020520-controls
2259 :			Description:
2260 :
2261 :			- Added ControlUtil structure to control-lib.cm.
2262 :			- Use it throughout.
2263 :			- Used Controls facility to define MLRISC controls (as opposed to
2264 :			registering MLRISC control ref cells with Controls after the fact)
2265 :			- Fixed messed-up controls priorities.
2266 :
2267 :			* Removed again all the stuff from config/preloads that one wouldn't
2268 :			be able to preload at the time the initial heap image is built.
2269 :			(Many libraries, e.g., CML, do not exist yet at this time. The
2270 :			only libraries that can be preloaded via config/preloads are those
2271 :			that come bundled with the bootfiles.)
2272 :
2273 :			----------------------------------------------------------------------
2274 :			Name: Matthias Blume
2275 :	blume	1204	Date: 2002/05/20 10:59:00 EDT
2276 :			Tag: blume-20020520-preloads
2277 :			Description:
2278 :
2279 :			Added a lot of commented-out suggestions for things to be included
2280 :			in config/preloads.
2281 :
2282 :			----------------------------------------------------------------------
2283 :	leunga	1203	Name: Allen Leung
2284 :			Date: 2002/05/18 14:20:00 EDT
2285 :			Tag: leunga-20020518-mdl
2286 :			Description:
2287 :
2288 :			o Made the mdl tool stuff compile and run again.
2289 :			o I've disabled all the stuff that depends on RTL specifications; they
2290 :			are all badly broken anyway.
2291 :
2292 :			----------------------------------------------------------------------
2293 :	blume	1201	Name: Matthias Blume
2294 :			Date: 2002/05/17 16:49:00 EDT
2295 :			Tag: blume-20020517-controls
2296 :			Description:
2297 :
2298 :			0. John Reppy made several modifications to the SML/NJ library.
2299 :			In particular, there is a shiny new controls-lib.cm.
2300 :
2301 :			1. Pushed new controls interface through compiler so that everything
2302 :			compiles again.
2303 :
2304 :			2. Added FormatComb and FORMAT_COMB to the CML version of the
2305 :			SML/NJ library (so that CML compiles again).
2306 :
2307 :			3. Modified init scripts because XXX_DEFAULT environment variables
2308 :			are no longer with us. (Boot-time initialization is now done
2309 :			using the same environment variables that are also used for
2310 :			startup-time initialization of controls.)
2311 :
2312 :			----------------------------------------------------------------------
2313 :	george	1192	Name: Lal George
2314 :			Date: 2002/05/15 09:20:10 EDT
2315 :			Tag: george-20020515-pseudo-op-decls
2316 :			Description:
2317 :
2318 :			All pseudo-ops emitted before the first segment declaration
2319 :			such as TEXT, DATA, and BSS directives are assumed to be global
2320 :			declarations and are emitted first in the assembly file. This is
2321 :			useful in a number of situations where one has pseudo-ops that are not
2322 :			specific to any segment, and also works around the constraint that one
2323 :			cannot have client pseudo-ops in the TEXT segment.
2324 :
2325 :			Because no segment is associated with these declarations it is
2326 :			an error to allocate any space or objects before the first segment
2327 :			directive and an exception will be raised. However, we cannot make
2328 :			this check for client pseudo-ops.
2329 :
2330 :			These top level declarations are a field in the CFG graph_info.
2331 :			In theory you can continue to add to this field after the CFG has been
2332 :			built -- provided you know what you are doing;-)
2333 :
2334 :			----------------------------------------------------------------------
2335 :	blume	1175	Name: Matthias Blume
2336 :	blume	1190	Date: 2002/05/13 16:40:00 EDT
2337 :			Tag: blume-20020513-pp-etc
2338 :			Description:
2339 :
2340 :			A few minor bugfixes:
2341 :
2342 :			- Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2343 :			(Bogus printouts for redefined bindings still occur. Compiler
2344 :			bug should no longer occur now. We need to redo the prettyprinter
2345 :			from scratch.)
2346 :
2347 :			- CM pathname printer now also adds escape sequences for ( and )
2348 :
2349 :			- commend and docu fixes for ml-nlffi
2350 :
2351 :			----------------------------------------------------------------------
2352 :			Name: Matthias Blume
2353 :	blume	1189	Date: 2002/05/10 16:40:00 EDT
2354 :			Tag: blume-20020510-erg-textio
2355 :			Description:
2356 :
2357 :			Applied the following bugfix provided by Emden Gansner:
2358 :
2359 :			Output is corrupted when outputSubstr is used rather than output.
2360 :
2361 :			The problem occurs when a substring
2362 :
2363 :			ss = (s, dataStart, dataLen)
2364 :
2365 :			where dataStart > 0, fills a stream buffer with avail bytes left.
2366 :			avail bytes of s, starting at index dataStart, are copied into the
2367 :			buffer, the buffer is flushed, and then the remaining dataLen-avail
2368 :			bytes of ss are copied into the beginning of the buffer. Instead of
2369 :			starting this copy at index dataStart+avail in s, the current code
2370 :			starts the copy at index avail.
2371 :
2372 :			Fix:
2373 :			In text-io-fn.sml, change line 695 from
2374 :			val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2375 :			to
2376 :			val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2377 :
2378 :			----------------------------------------------------------------------
2379 :			Name: Matthias Blume
2380 :	blume	1186	Date: 2002/04/12 13:55:00 EDT
2381 :			Tag: blume-20020412-assyntax
2382 :			Description:
2383 :
2384 :			1. Grabbed newer assyntax.h from the XFree86 project.
2385 :			2. Fiddled with how to compile X86.prim.asm without warnings.
2386 :			3. (Very) Minor cleanup in CM.
2387 :
2388 :			----------------------------------------------------------------------
2389 :			Name: Matthias Blume
2390 :	blume	1185	Date: 2002/04/01 (no joke!) 17:07:00 EST
2391 :			Tag: blume-20020401-x86div
2392 :			Description:
2393 :
2394 :			Added full support for div/mod/rem/quot on the x86, using the machine
2395 :			instruction's two results (without clumsily recomputing the remainder)
2396 :			directly where appropriate.
2397 :
2398 :			Some more extensive power-of-two support was added to the x86 instruction
2399 :			selector (avoiding expensive divs, mods, and muls where they can be
2400 :			replaced with cheaper shifts and masks). However, this sort of thing
2401 :			ought to be done earlier, e.g., within the CPS optimizer so that
2402 :			all architectures benefit from it.
2403 :
2404 :			The compiler compiles to a fixed point, but changes might be somewhat
2405 :			fragile nevertheless. Please, report any strange things that you might
2406 :			see wrt. div/mod/quot/rem...
2407 :
2408 :			----------------------------------------------------------------------
2409 :			Name: Matthias Blume
2410 :	blume	1184	Date: 2002/03/29 17:22:00
2411 :			Tag: blume-20020329-div
2412 :			Description:
2413 :
2414 :			Fixed my broken div/mod logic. Unfortunately, this means that the
2415 :			inline code for div/mod now has one more comparison than before.
2416 :			Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2417 :			The problem was with quotient = 0, because that alone does not tell
2418 :			us which way the rounding went. One then has to look at whether
2419 :			remainder and divisor have the same sign... :(
2420 :
2421 :			Anyway, I replaced the bootfiles with fresh ones...
2422 :
2423 :			----------------------------------------------------------------------
2424 :			Name: Matthias Blume
2425 :	blume	1183	Date: 2002/03/29 14:10:00 EST
2426 :			Tag: blume-20020329-inlprims
2427 :			Description:
2428 :
2429 :			NEW BOOTFILES!!! Version number bumped to 110.39.3.
2430 :
2431 :			Primops have changed. This means that the bin/boot-file formats have
2432 :			changed as well.
2433 :
2434 :			To make sure that there is no confusion, I made a new version.
2435 :
2436 :
2437 :			CHANGES:
2438 :
2439 :			* removed REMT from mltree (remainder should never overflow).
2440 :
2441 :			* added primops to deal with divisions of all flavors to the frontend
2442 :
2443 :			* handled these primops all the way through so they map to their respective
2444 :			MLRISC support
2445 :
2446 :			* used these primops in the implementation of Int, Int32, Word, Word32
2447 :
2448 :			* removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2449 :
2450 :			* parameterized INLMIN, INLMAX, and INLABS by a numkind
2451 :
2452 :			* translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2453 :			floating point
2454 :
2455 :			* used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2456 :			and Real (but Real.abs maps to a separate floating-point-only primop)
2457 :
2458 :
2459 :			TODO items:
2460 :
2461 :			* Hacked Alpha32 instruction selection, disabling the selection of REMx
2462 :			instructions because the machine instruction encoder cannot handle
2463 :			them. (Hppa, PPC, and Sparc instruction selection did not handle
2464 :			REM in the first place, and REM is supported by the x86 machine coder.)
2465 :
2466 :			* Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2467 :			selection phase. (The two can be streamlined because the hardware
2468 :			delivers both quotient and remainder at the same time anyway.)
2469 :
2470 :			* Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2471 :			(Currently the behavior is inconsistent both across architectures and
2472 :			wrt. the draft Basis spec.)
2473 :
2474 :			* Word8 should eventually be handled natively, too.
2475 :
2476 :			* There seems to be one serious bug in mltree-gen.sml. It appears, though,
2477 :			as if there currently is no execution path that could trigger it in
2478 :			SML/NJ. (The assumptions underlying functions arith and promotable do not
2479 :			hold for things like multiplication and division.)
2480 :
2481 :			----------------------------------------------------------------------
2482 :			Name: Matthias Blume
2483 :	blume	1181	Date: 2002/03/27 16:27:00 EST
2484 :			Tag: blume-20020327-mlrisc-divisions
2485 :			Description:
2486 :
2487 :			Added support for all four division operations (ML's div, mod, quot,
2488 :			and rem) to MLRISC. In the course of doing so, I also rationalized
2489 :			the naming (no more annoying switch-around of DIV and QUOT), by
2490 :			parameterizing the operation by div_rounding_mode (which can be either
2491 :			DIV_TO_ZERO or DIV_TO_NEGINF).
2492 :
2493 :			The generic MLTreeGen functor takes care of compiling all four
2494 :			operations down to only round-to-zero div.
2495 :
2496 :			Missing pieces:
2497 :
2498 :			* Doing something smarter than relying on MLTreeGen on architectures
2499 :			like, e.g., the x86 where hardware division delivers both quotient and
2500 :			remainder at the same time. With this, the implementation of the
2501 :			round-to-neginf operations could be further streamlined.
2502 :
2503 :			* Remove inlining support for div/mod/rem from the frontend and replace it
2504 :			with primops that get carried through to the backend. Do this for all
2505 :			int and word types.
2506 :
2507 :			----------------------------------------------------------------------
2508 :			Name: Matthias Blume
2509 :	blume	1180	Date: 2002/03/25 17:25:00 EST
2510 :			Tag: blume-20020325-divmod
2511 :			Description:
2512 :
2513 :			I improved (hopefully without breaking them) the implementation of Int.div,
2514 :			Int.mod, and Int.rem. For this, the code in translate.sml now takes
2515 :			advantage of the following observations:
2516 :
2517 :			Let q = x quot y r = x rem y
2518 :			d = x div y m = x mod y
2519 :
2520 :			where "quot" is the round-to-zero version of integer division that
2521 :			hardware usually provides. Then we have:
2522 :
2523 :			r = x - q * y where neither the * nor the - will overflow
2524 :			d = if q >= 0 orelse x = q * y then q else q - 1
2525 :			where neither the * nor the - will overflow
2526 :			m = if q >= 0 orelse r = 0 then r else r + y
2527 :			where the + will not overflow
2528 :
2529 :			This results in substantial simplification of the generated code.
2530 :			The following table shows the number of CFG nodes and edges generated
2531 :			for
2532 :			fun f (x, y) = x OPER y
2533 :			(* with OPER \in div, mod, quot, rem *)
2534 :
2535 :
2536 :			OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2537 :			--------------------------------------------------------
2538 :			div | 24 | 39 | 12 | 16
2539 :			mod | 41 | 71 | 12 | 16
2540 :			quot | 8 | 10 | 8 | 10
2541 :			rem | 10 | 14 | 8 | 10
2542 :
2543 :
2544 :			----------------------------------------------------------------------
2545 :			Name: Matthias Blume
2546 :	blume	1179	Date: 2002/03/25 22:06:00 EST
2547 :			Tag: blume-20020325-cprotobug
2548 :			Description:
2549 :
2550 :			Fixed a bug in cproto (c prototype decoder).
2551 :
2552 :			----------------------------------------------------------------------
2553 :			Name: Matthias Blume
2554 :	blume	1178	Date: 2002/03/25 16:00:00 EST
2555 :			Tag: blume-20020325-raw-primops
2556 :			Description:
2557 :
2558 :			I did some cleanup to Allen's new primop code and
2559 :			replaced yesterday's bootfiles with new ones.
2560 :			(But they are stored in the same place.)
2561 :
2562 :			----------------------------------------------------------------------
2563 :			Name: Matthias Blume
2564 :	blume	1175	Date: 2002/03/24 22:40:00 EST
2565 :			Tag: blume-20020324-bootfiles
2566 :			Description:
2567 :
2568 :			Made the bootfiles that Allen asked for.
2569 :
2570 :			----------------------------------------------------------------------
2571 :	leunga	1174	Name: Allen Leung
2572 :			Date: 2002/03/23 15:50:00 EST
2573 :			Tag: leunga-20020323-flint-cps-rcc-primops
2574 :			Description:
2575 :
2576 :			1. Changes to FLINT primops:
2577 :
2578 :			(* make a call to a C-function;
2579 :			* The primop carries C function prototype information and specifies
2580 :			* which of its (ML-) arguments are floating point. C prototype
2581 :			* information is for use by the backend, ML information is for
2582 :			* use by the CPS converter. *)
2583 :			| RAW_CCALL of { c_proto: CTypes.c_proto,
2584 :			ml_args: ccall_type list,
2585 :			ml_res_opt: ccall_type option,
2586 :			reentrant : bool
2587 :			} option
2588 :			(* Allocate uninitialized storage on the heap.
2589 :			* The record is meant to hold short-lived C objects, i.e., they
2590 :			* are not ML pointers. With the tag, the representation is
2591 :			* the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2592 :			*)
2593 :			| RAW_RECORD of {tag:bool,sz:int}
2594 :			and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2595 :
2596 :			2. These CPS primops are now overloaded:
2597 :
2598 :			rawload of {kind:numkind}
2599 :			rawstore of {kind:numkind}
2600 :
2601 :			The one argument form is:
2602 :
2603 :			rawload {kind} address
2604 :
2605 :			The two argument form is:
2606 :
2607 :			rawload {kind} [ml object, byte-offset]
2608 :
2609 :			3. RAW_CCALL/RCC now takes two extra arguments:
2610 :
2611 :			a. The first is whether the C call is reentrant, i.e., whether
2612 :			ML state should be saved and restored.
2613 :			b. The second argument is a string argument specifying the name of
2614 :			library and the C function.
2615 :
2616 :			These things are currently not handled in the code generator, yet.
2617 :
2618 :			4. In CProto,
2619 :
2620 :			An encoding type of "bool" means "ml object" and is mapped into
2621 :			C prototype of PTR. Note that "bool" is different than "string",
2622 :			even though "string" is also mapped into PTR, because "bool"
2623 :			is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2624 :
2625 :			5. Pickler/unpicker
2626 :
2627 :			Changed to handle RAW_RECORD and newest RAW_CCALL
2628 :
2629 :			6. MLRiscGen,
2630 :
2631 :			1. Changed to handle the new rawload/rawstore/rawrecord operators.
2632 :			2. Code for handling C Calls has been moved to a new module CPSCCalls,
2633 :			in the file CodeGen/cpscompile/cps-c-calls.sml
2634 :
2635 :			7. Added the conditional move operator
2636 :
2637 :			condmove of branch
2638 :
2639 :			to cps. Generation of this is still buggy so it is currently
2640 :			disabled.
2641 :
2642 :			----------------------------------------------------------------------
2643 :	george	1168	Name: Lal George
2644 :			Date: 2002/03/22 14:18:25 EST
2645 :	george	1171	Tag: george-20020322-cps-branch-prob
2646 :	george	1168	Description:
2647 :
2648 :	george	1171	Implemented the Ball-Larus branch prediction-heuristics, and
2649 :	george	1168	incorporated graphical viewers for control flow graphs.
2650 :
2651 :	george	1171	Ball-Larus Heuristics:
2652 :	george	1168	---------------------
2653 :			See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2654 :
2655 :			By design it uses the Dempster-Shafer theory for combining
2656 :			probabilities. For example, in the function:
2657 :
2658 :			fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2659 :
2660 :			the ball-larus heuristics predicts that the n=0 is unlikely
2661 :			(OH-heuristic), and the 'then' branch is unlikely because of the
2662 :			RH-heuristic -- giving the 'then' branch an even lower combined
2663 :	george	1171	probability using the Dempster-Shafer theory.
2664 :	george	1168
2665 :			Finally, John Reppy's loop analysis in MLRISC, further lowers the
2666 :			probability of the 'then' branch because of the loop in the else
2667 :			branch.
2668 :
2669 :
2670 :			Graphical Viewing:
2671 :			------------------
2672 :			I merely plugged in Allen's graphical viewers into the compiler. The
2673 :			additional code is not much. At the top level, saying:
2674 :
2675 :			Control.MLRISC.getFlag "cfg-graphical-view" := true;
2676 :
2677 :			will display the graphical view of the control flow graph just before
2678 :			back-patching. daVinci must be in your path for this to work. If
2679 :			daVinci is not available, then the default viewer can be changed
2680 :			using:
2681 :
2682 :			Control.MLRISC.getString "viewer"
2683 :
2684 :			which can be set to "dot" or "vcg" for the corresponding viewers. Of
2685 :			course, these viewers must be in your path.
2686 :
2687 :			The above will display the compilation unit at the level of clusters,
2688 :			many of which are small, boring, and un-interesting. Also setting:
2689 :
2690 :			Control.MLRISC.getInt "cfg-graphical-view_size"
2691 :
2692 :			will display clusters that are larger than the value set by the above.
2693 :
2694 :
2695 :			----------------------------------------------------------------------
2696 :	blume	1159	Name: Matthias Blume
2697 :			Date: 2002/03/21 22:20:00 EST
2698 :			Tag: blume-20020321-kmp-bugfix
2699 :			Description:
2700 :
2701 :			Changed the interface to the KMP routine in PreString and fixed
2702 :			a minor bug in one place where it was used.
2703 :
2704 :			----------------------------------------------------------------------
2705 :	leunga	1156	Name: Allen Leung
2706 :	leunga	1158	Date: 2002/03/21 20:30:00 EST
2707 :			Tag: leunga-20020321-cfg
2708 :			Description:
2709 :
2710 :			Fixed a potential problem in cfg edge splitting.
2711 :
2712 :			----------------------------------------------------------------------
2713 :			Name: Allen Leung
2714 :	leunga	1156	Date: 2002/03/21 17:15:00 EST
2715 :			Tag: leunga-20020321-x86-fp-cfg
2716 :			Description:
2717 :
2718 :			1. Recoded the buggy parts of x86-fp.
2719 :
2720 :			a. All the block reordering code has been removed.
2721 :			We now depend on the block placement phases to do this work.
2722 :
2723 :			b. Critical edge splitting code has been simplified and moved into the
2724 :			CFG modules, as where they belong.
2725 :
2726 :			Both of these were quite buggy and complex. The code is now much, much
2727 :			simpler.
2728 :
2729 :			2. X86 backend.
2730 :
2731 :			a. Added instructions for 64-bit support. Instruction selection for
2732 :			64-bit has not been committed, however, since that
2733 :			requires changes to MLTREE which haven't been approved by
2734 :			Lal and John.
2735 :
2736 :			b. Added support for FUCOMI and FUCOMIP when generating code for
2737 :			PentiumPro and above. We only generate these instructions in
2738 :			the fast-fp mode.
2739 :
2740 :			c. Added cases for JP and JNP in X86FreqProps.
2741 :
2742 :			3. CFG
2743 :
2744 :			CFG now has a bunch of methods for edge splitting and merging.
2745 :
2746 :			4. Machine description.
2747 :
2748 :			John's simplification of MLTREE_BASIS.fcond broke a few machine
2749 :			description things:
2750 :
2751 :			rtl-build.{sig,sml} and hppa.mdl fixed.
2752 :
2753 :			NOTE: the machine description stuff in the repository is still broken.
2754 :			Again, I can't put my fixes in because that involves
2755 :			changes to MLTREE.
2756 :
2757 :			----------------------------------------------------------------------
2758 :	blume	1145	Name: Matthias Blume
2759 :	blume	1155	Date: 2002/03/20 15:55:00 EST
2760 :			Tag: blume-20020320-kmp
2761 :			Description:
2762 :
2763 :			Implemented Knuth-Morris-Pratt string matching in PreString and used
2764 :			it for String.isSubstring, Substring.isSubstring, and
2765 :			Substring.position.
2766 :
2767 :			(Might need some stress-testing. Simple examples worked fine.)
2768 :
2769 :			----------------------------------------------------------------------
2770 :			Name: Matthias Blume
2771 :	blume	1152	Date: 2002/03/19 16:37:00 EST
2772 :			Tag: blume-20020319-witnesses
2773 :			Description:
2774 :
2775 :			Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2776 :
2777 :			This implements a generic mechanism for changing constness qualifiers
2778 :			anywhere within big C types without resorting to outright "casts".
2779 :			(So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2780 :			modify the constness at the outermost level.)
2781 :			The implementation of "convert" is based on the idea of "witness"
2782 :			values -- values that are not used by the operation but whose types
2783 :			"testify" to their applicability. On the implementation side, "convert"
2784 :			is simply a projection (returning its second curried argument). With
2785 :			cross-module inlining, it should not result in any machine code being
2786 :			generated.
2787 :
2788 :			----------------------------------------------------------------------
2789 :			Name: Matthias Blume
2790 :	blume	1148	Date: 2002/03/15 16:40:00 EST
2791 :			Tag: blume-20020315-basis
2792 :			Description:
2793 :
2794 :			Provided (preliminary?) implementations for
2795 :
2796 :			{String,Substring}.{concatWith,isSuffix,isSubstring}
2797 :
2798 :			and
2799 :
2800 :			Substring.full
2801 :
2802 :			Those are in the Basis spec but they were missing in SML/NJ.
2803 :
2804 :			----------------------------------------------------------------------
2805 :			Name: Matthias Blume
2806 :	blume	1145	Date: 2002/03/14 21:30:00 EST
2807 :			Tag: blume-20020314-controls
2808 :			Description:
2809 :
2810 :			Controls:
2811 :			---------
2812 :
2813 :			1. Factored out the recently-added Controls : CONTROLS stuff and put
2814 :			it into its own library $/controls-lib.cm. The source tree for
2815 :			this is under src/smlnj-lib/Controls.
2816 :
2817 :			2. Changed the names of types and functions in this interface, so they
2818 :			make a bit more "sense":
2819 :
2820 :			module -> registry
2821 :			'a registry -> 'a group
2822 :
2823 :			3. The interface now deals in ref cells only. The getter/setter interface
2824 :			is (mostly) gone.
2825 :
2826 :			4. Added a function that lets one register an already-existing ref cell.
2827 :
2828 :			5. Made the corresponding modifications to the rest of the code so that
2829 :			everything compiles again.
2830 :
2831 :			6. Changed the implementation of Controls.MLRISC back to something closer
2832 :			to the original. In particular, this module (and therefore MLRISC)
2833 :			does not depend on Controls. There now is some link-time code in
2834 :			int-sys.sml that registers the MLRISC controls with the Controls
2835 :			module.
2836 :
2837 :			CM:
2838 :			---
2839 :
2840 :			* One can now specify the lambda-split aggressiveness in init.cmi.
2841 :
2842 :			----------------------------------------------------------------------
2843 :	leunga	1142	Name: Allen Leung
2844 :			Date: 2002/03/13 17:30:00 EST
2845 :			Tag: leunga-20020313-x86-fp-unary
2846 :			Description:
2847 :
2848 :			Bug fix for:
2849 :
2850 :			> leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2851 :			> Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2852 :			> - fun f(x,(y,z)) = Real.~ y;
2853 :			> [autoloading]
2854 :			> [autoloading done]
2855 :			> fchsl (%eax), 184(%esp)
2856 :			> Error: MLRisc bug: X86MCEmitter.emitInstr
2857 :			>
2858 :			> uncaught exception Error
2859 :			> raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2860 :
2861 :			The problem was that the code generator did not generate any fp registers
2862 :			in this case, and the ra didn't know that it needed to run the X86FP phase to
2863 :			translate the pseudo fp instruction. This only happened with unary fp
2864 :			operators in certain situations.
2865 :
2866 :			----------------------------------------------------------------------
2867 :	blume	1137	Name: Matthias Blume
2868 :	blume	1140	Date: 2002/03/13 14:00:00 EST
2869 :			Tag: blume-20020313-overload-etc
2870 :			Description:
2871 :
2872 :			1. Added _overload as a synonym for overload for backward compatibility.
2873 :			(Control.overloadKW must be true for either version to be accepted.)
2874 :
2875 :			2. Fixed bug in install script that caused more things to be installed
2876 :			than what was requested in config/targets.
2877 :
2878 :			3. Made CM aware of the (_)overload construct so that autoloading
2879 :			works.
2880 :
2881 :			----------------------------------------------------------------------
2882 :			Name: Matthias Blume
2883 :	blume	1138	Date: 2002/03/12 22:03:00 EST
2884 :			Tag: blume-20020312-url
2885 :			Description:
2886 :
2887 :			Forgot to update BOOT and srcarchiveurl.
2888 :
2889 :			----------------------------------------------------------------------
2890 :			Name: Matthias Blume
2891 :	blume	1137	Date: 2002/03/12 17:30:00 EST
2892 :			Tag: blume-20020312-version110392
2893 :			Description:
2894 :
2895 :			Yet another version number bump (because of small changes to the
2896 :			binfile format). Version number is now 110.39.2. NEW BOOTFILES!
2897 :
2898 :			Changes:
2899 :
2900 :			The new pid generation scheme described a few weeks ago was overly
2901 :			complicated. I implemented a new mechanism that is simpler and
2902 :			provides a bit more "stability": Once CM has seen a compilation
2903 :			unit, it keeps its identity constant (as long as you do not delete
2904 :			those crucial CM/GUID/* files). This means that when you change
2905 :	blume	1145	an interface, compile, then go back to the old interface, and
2906 :	blume	1137	compile again, you arrive at the original pid.
2907 :
2908 :			There now also is a mechanism that instructs CM to use the plain
2909 :			environment hash as a module's pid (effectively making its GUID
2910 :			the empty string). For this, "noguid" must be specified as an
2911 :			option to the .sml file in question within its .cm file.
2912 :			This is most useful for code that is being generated by tools such
2913 :			as ml-nlffigen (because during development programmers tend to
2914 :			erase the tool's entire output directory tree including CM's cached
2915 :			GUIDs). "noguid" is somewhat dangerous (since it can be used to locally
2916 :			revert to the old, broken behavior of SML/NJ, but in specific cases
2917 :			where there is no danger of interface confusion, its use is ok
2918 :			(I think).
2919 :
2920 :			ml-nlffigen by default generates "noguid" annotations. They can be
2921 :			turned off by specifying -guid in its command line.
2922 :
2923 :			----------------------------------------------------------------------
2924 :	george	1136	Name: Lal George
2925 :			Date: 2002/03/12 12 14:42:36 EST
2926 :			Tag: george-20020312-frequency-computation
2927 :			Description:
2928 :
2929 :			Integrated jump chaining and static block frequency into the
2930 :			compiler. More details and numbers later.
2931 :	leunga	1025
2932 :	george	1003	----------------------------------------------------------------------
2933 :	george	1133	Name: Lal George
2934 :			Date: 2002/03/11 11 22:38:53 EST
2935 :			Tag: george-20020311-jump-chain-elim
2936 :			Description:
2937 :
2938 :			Tested the jump chain elimination on all architectures (except the
2939 :			hppa). This is on by default right now and is profitable for the
2940 :			alpha and x86, however, it may not be profitable for the sparc and ppc
2941 :			when compiling the compiler.
2942 :
2943 :			The gc test will typically jump to a label at the end of the cluster,
2944 :			where there is another jump to an external cluster containing the actual
2945 :			code to invoke gc. This is to allow factoring of common gc invocation
2946 :			sequences. That is to say, we generate:
2947 :
2948 :			f:
2949 :			testgc
2950 :			ja L1 % jump if above to L1
2951 :
2952 :			L1:
2953 :			jmp L2
2954 :
2955 :
2956 :			After jump chain elimination the 'ja L1' instructions is converted to
2957 :			'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2958 :			up being implemented in their long form (if L2 is far away) using:
2959 :
2960 :			jbe L3 % jump if below or equal to L3
2961 :			jmp L2
2962 :			L3:
2963 :			...
2964 :
2965 :
2966 :			For large compilation units L2 may be far away.
2967 :
2968 :
2969 :			----------------------------------------------------------------------
2970 :	blume	1132	Name: Matthias Blume
2971 :			Date: 2002/03/11 13:30:00 EST
2972 :			Tag: blume-20020311-mltreeeval
2973 :			Description:
2974 :
2975 :			A functor parameter was missing.
2976 :
2977 :			----------------------------------------------------------------------
2978 :	leunga	1130	Name: Allen Leung
2979 :	leunga	1131	Date: 2002/03/11 10:30:00 EST
2980 :	leunga	1142	Tag: leunga-20020311-runtime-string0
2981 :	leunga	1131	Description:
2982 :
2983 :			The representation of the empty string now points to a
2984 :			legal null terminated C string instead of unit. It is now possible
2985 :			to convert an ML string into C string with InlineT.CharVector.getData.
2986 :			This compiles into one single machine instruction.
2987 :
2988 :			----------------------------------------------------------------------
2989 :			Name: Allen Leung
2990 :	leunga	1130	Date: 2002/03/10 23:55:00 EST
2991 :			Tag: leunga-20020310-x86-call
2992 :			Description:
2993 :
2994 :			Added machine generation for CALL instruction (relative displacement mode)
2995 :
2996 :			----------------------------------------------------------------------
2997 :	blume	1128	Name: Matthias Blume
2998 :			Date: 2002/03/08 16:05:00
2999 :			Tag: blume-20020308-entrypoints
3000 :			Description:
3001 :
3002 :			Version number bumped to 110.39.1. NEW BOOTFILES!
3003 :
3004 :			Entrypoints: non-zero offset into a code object where execution should begin.
3005 :
3006 :			- Added the notion of an entrypoint to CodeObj.
3007 :			- Added reading/writing of entrypoint info to Binfile.
3008 :			- Made runtime system bootloader aware of entrypoints.
3009 :			- Use the address of the label of the first function given to mlriscGen
3010 :			as the entrypoint. This address is currently always 0, but it will
3011 :			not be 0 once we turn on block placement.
3012 :			- Removed the linkage cluster code (which was The Other Way(tm) of dealing
3013 :			with entry points) from mlriscGen.
3014 :
3015 :			----------------------------------------------------------------------
3016 :	leunga	1127	Name: Allen Leung
3017 :			Date: 2002/03/07 20:45:00 EST
3018 :			Tag: leunga-20020307-x86-cmov
3019 :			Description:
3020 :
3021 :			Bug fixes for CMOVcc on x86.
3022 :
3023 :			1. Added machine code generation for CMOVcc
3024 :			2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
3025 :			3. CMOVcc cannot have an immediate operand as argument.
3026 :
3027 :			----------------------------------------------------------------------
3028 :	blume	1126	Name: Matthias Blume
3029 :			Date: 2002/03/07 16:15:00 EST
3030 :			Tag: blume-20020307-controls
3031 :			Description:
3032 :
3033 :			This is a very large but mostly boring patch which makes (almost)
3034 :			every tuneable compiler knob (i.e., pretty much everything under
3035 :			Control.* plus a few other things) configurable via both the command
3036 :			line and environment variables in the style CM did its configuration
3037 :			until now.
3038 :
3039 :			Try starting sml with '-h' (or, if you are brave, '-H')
3040 :
3041 :			To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
3042 :			implements the underlying generic mechanism.
3043 :
3044 :			The interface to some of the existing such facilities has changed somewhat.
3045 :			For example, the MLRiscControl module now provides mkFoo instead of getFoo.
3046 :			(The getFoo interface is still there for backward-compatibility, but its
3047 :			use is deprecated.)
3048 :
3049 :			The ml-build script passes -Cxxx=yyy command-line arguments through so
3050 :			that one can now twiddle the compiler settings when using this "batch"
3051 :			compiler.
3052 :
3053 :			TODO items:
3054 :
3055 :			We should go through and throw out all controls that are no longer
3056 :			connected to anything. Moreover, we should go through and provide
3057 :			meaningful (and correct!) documentation strings for those controls
3058 :			that still are connected.
3059 :
3060 :			Currently, multiple calls to Controls.new are accepted (only the first
3061 :			has any effect). Eventually we should make sure that every control
3062 :			is being made (via Controls.new) exactly once. Future access can then
3063 :			be done using Controls.acc.
3064 :
3065 :			Finally, it would probably be a good idea to use the getter-setter
3066 :			interface to controls rather than ref cells. For the time being, both
3067 :			styles are provided by the Controls module, but getter-setter pairs are
3068 :			better if thread-safety is of any concern because they can be wrapped.
3069 :
3070 :			*****************************************
3071 :
3072 :			One bug fix: The function blockPlacement in three of the MLRISC
3073 :			backpatch files used to be hard-wired to one of two possibilities at
3074 :			link time (according to the value of the placementFlag). But (I
3075 :			think) it should rather sense the flag every time.
3076 :
3077 :			*****************************************
3078 :
3079 :			Other assorted changes (by other people who did not supply a HISTORY entry):
3080 :
3081 :			1. the cross-module inliner now works much better (Monnier)
3082 :			2. representation of weights, frequencies, and probabilities in MLRISC
3083 :			changed in preparation of using those for weighted block placement
3084 :			(Reppy, George)
3085 :
3086 :			----------------------------------------------------------------------
3087 :	george	1116	Name: Lal George
3088 :	george	1124	Date: 2002/03/07 14:44:24 EST 2002
3089 :			Tag: george-20020307-weighted-block-placement
3090 :
3091 :			Tested the weighted block placement optimization on all architectures
3092 :			(except the hppa) using AMPL to generate the block and edge frequencies.
3093 :			Changes were required in the machine properties to correctly
3094 :			categorize trap instructions. There is an MLRISC flag
3095 :			"weighted-block-placement" that can be used to enable weighted block
3096 :			placement, but this will be ineffective without block/edge
3097 :			frequencies (coming soon).
3098 :
3099 :
3100 :			----------------------------------------------------------------------
3101 :			Name: Lal George
3102 :	george	1116	Date: 2002/03/05 17:24:48 EST
3103 :			Tag: george-20020305-linkage-cluster
3104 :
3105 :	george	1124	In order to support the block placement optimization, a new cluster
3106 :			is generated as the very first cluster (called the linkage cluster).
3107 :			It contains a single jump to the 'real' entry point for the compilation
3108 :			unit. Block placement has no effect on the linkage cluster itself, but
3109 :			all the other clusters have full freedom in the manner in which they
3110 :			reorder blocks or functions.
3111 :	george	1116
3112 :			On the x86 the typical linkage code that is generated is:
3113 :			----------------------
3114 :			.align 2
3115 :			L0:
3116 :			addl $L1-L0, 72(%esp)
3117 :	george	1124	jmp L1
3118 :	george	1116
3119 :
3120 :			.align 2
3121 :			L1:
3122 :			----------------------
3123 :
3124 :			72(%esp) is the memory location for the stdlink register. This
3125 :			must contain the address of the CPS function being called. In the
3126 :	george	1124	above example, it contains the address of L0; before
3127 :	george	1116	calling L1 (the real entry point for the compilation unit), it
3128 :			must contain the address for L1, and hence
3129 :
3130 :			addl $L1-L0, 72(%esp)
3131 :
3132 :	george	1124	I have tested this on all architectures except the hppa.The increase
3133 :			in code size is of course negligible
3134 :	george	1116
3135 :			----------------------------------------------------------------------
3136 :	leunga	1115	Name: Allen Leung
3137 :			Date: 2002/03/03 13:20:00 EST
3138 :			Tag: leunga-20020303-mlrisc-tools
3139 :
3140 :			Added #[ ... ] expressions to mlrisc tools
3141 :
3142 :			----------------------------------------------------------------------
3143 :	blume	1096	Name: Matthias Blume
3144 :	blume	1098	Date: 2002/02/27 12:29:00 EST
3145 :			Tag: blume-20020227-cdebug
3146 :			Description:
3147 :
3148 :			- made types in structure C and C_Debug to be equal
3149 :			- got rid of code duplication (c-int.sml vs. c-int-debug.sml)
3150 :			- there no longer is a C_Int_Debug (C_Debug is directly derived from C)
3151 :
3152 :			----------------------------------------------------------------------
3153 :			Name: Matthias Blume
3154 :	blume	1096	Date: 2002/02/26 12:00:00 EST
3155 :			Tag: blume-20020226-ffi
3156 :			Description:
3157 :
3158 :			1. Fixed a minor bug in CM's "noweb" tool:
3159 :			If numbering is turned off, then truly don't number (i.e., do not
3160 :			supply the -L option to noweb). The previous behavior was to supply
3161 :			-L'' -- which caused noweb to use the "default" line numbering scheme.
3162 :			Thanks to Chris Richards for pointing this out (and supplying the fix).
3163 :
3164 :			2. Once again, I reworked some aspects of the FFI:
3165 :
3166 :			A. The incomplete/complete type business:
3167 :
3168 :			- Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
3169 :			gone!
3170 :			- ML types representing an incomplete type are now *equal* to
3171 :			ML types representing their corresponding complete types (just like
3172 :			in C). This is still safe because ml-nlffigen will not generate
3173 :			RTTI for incomplete types, nor will it generate functions that
3174 :			require access to such RTTI. But when ML code generated from both
3175 :			incomplete and complete versions of the C type meet, the ML types
3176 :			are trivially interoperable.
3177 :
3178 :			NOTE: These changes restore the full generality of the translation
3179 :			(which was previously lost when I eliminated functorization)!
3180 :
3181 :			B. Enum types:
3182 :
3183 :			- Structure C now has a type constructor "enum" that is similar to
3184 :			how the "su" constructor works. However, "enum" is not a phantom
3185 :			type because each "T enum" has values (and is isomorphic to
3186 :			MLRep.Signed.int).
3187 :			- There are generic access operations for enum objects (using
3188 :			MLRep.Signed.int).
3189 :			- ml-nlffigen will generate a structure E_foo for each "enum foo".
3190 :			* The structure contains the definition of type "mlrep" (the ML-side
3191 :			representation type of the enum). Normally, mlrep is the same
3192 :			as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3193 :			then mlrep will be defined as a datatype -- thus facilitating
3194 :			pattern matching on mlrep values.
3195 :			("-ec" will be suppressed if there are duplicate values in an
3196 :			enumeration.)
3197 :			* Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3198 :			will be generated for each C enum constant xxx.
3199 :			* Conversion functions m2i and i2m convert between mlrep and
3200 :			MLRep.Signed.int. (Without "-ec", these functions are identities.)
3201 :			* Coversion functions c and ml convert between mlrep and "tag enum".
3202 :			* Access functions (get/set) fetch and store mlrep values.
3203 :			- By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3204 :			enumerations are merged into one single enumeration represented by
3205 :			structure E_'.
3206 :
3207 :			----------------------------------------------------------------------
3208 :	leunga	1080	Name: Allen Leung
3209 :	leunga	1094	Date: 2002/02/25 04:45:00 EST
3210 :			Tag: leunga-20020225-cps-spill
3211 :
3212 :			This is a new implementation of the CPS spill phase.
3213 :			The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3214 :			In case of problems, replace it with the old file spill.sml
3215 :
3216 :			The current compiler runs into some serious performance problems when
3217 :			constructing a large record. This can happen when we try to compile a
3218 :			structure with many items. Even a very simple structure like the following
3219 :			makes the compiler slow down.
3220 :
3221 :			structure Foo = struct
3222 :			val x_1 = 0w1 : Word32.int
3223 :			val x_2 = 0w2 : Word32.int
3224 :			val x_3 = 0w3 : Word32.int
3225 :			...
3226 :			val x_N = 0wN : Word32.int
3227 :			end
3228 :
3229 :			The following table shows the compile time, from N=1000 to N=4000,
3230 :			with the old compiler:
3231 :
3232 :			N
3233 :			1000 CPS 100 spill 0.04u 0.00s 0.00g
3234 :			MLRISC ra 0.06u 0.00s 0.05g
3235 :			(spills = 0 reloads = 0)
3236 :			TOTAL 0.63u 0.07s 0.21g
3237 :
3238 :			1100 CPS 100 spill 8.25u 0.32s 0.64g
3239 :			MLRISC ra 5.68u 0.59s 3.93g
3240 :			(spills = 0 reloads = 0)
3241 :			TOTAL 14.71u 0.99s 4.81g
3242 :
3243 :			1500 CPS 100 spill 58.55u 2.34s 1.74g
3244 :			MLRISC ra 5.54u 0.65s 3.91g
3245 :			(spills = 543 reloads = 1082)
3246 :			TOTAL 65.40u 3.13s 6.00g
3247 :
3248 :			2000 CPS 100 spill 126.69u 4.84s 3.08g
3249 :			MLRISC ra 0.80u 0.10s 0.55g
3250 :			(spills = 42 reloads = 84)
3251 :			TOTAL 129.42u 5.10s 4.13g
3252 :
3253 :			3000 CPS 100 spill 675.59u 19.03s 11.64g
3254 :			MLRISC ra 2.69u 0.27s 1.38g
3255 :			(spills = 62 reloads = 124)
3256 :			TOTAL 682.48u 19.61s 13.99g
3257 :
3258 :			4000 CPS 100 spill 2362.82u 56.28s 43.60g
3259 :			MLRISC ra 4.96u 0.27s 2.72g
3260 :			(spills = 85 reloads = 170)
3261 :			TOTAL 2375.26u 57.21s 48.00g
3262 :
3263 :			As you can see the old cps spill module suffers from some serious
3264 :	leunga	1115	performance problem. But since I cannot decipher the old code fully,
3265 :			instead of patching the problems up, I'm reimplementing it
3266 :	leunga	1094	with a different algorithm. The new code is more modular,
3267 :			smaller when compiled, and substantially faster
3268 :			(O(n log n) time and O(n) space). Timing of the new spill module:
3269 :
3270 :			4000 CPS 100 spill 0.02u 0.00s 0.00g
3271 :			MLRISC ra 0.25u 0.02s 0.15g
3272 :			(spills=1 reloads=3)
3273 :			TOTAL 7.74u 0.34s 1.62g
3274 :
3275 :			Implementation details:
3276 :
3277 :			As far as I can tell, the purpose of the CPS spill module is to make sure the
3278 :			number of live variables at any program point (the bandwidth)
3279 :			does not exceed a certain limit, which is determined by the
3280 :			size of the spill area.
3281 :
3282 :			When the bandwidth is too large, we decrease the register pressure by
3283 :			packing live variables into spill records. How we achieve this is
3284 :			completely different than what we did in the old code.
3285 :
3286 :			First, there is something about the MLRiscGen code generator
3287 :			that we should be aware of:
3288 :
3289 :			o MLRiscGen performs code motion!
3290 :
3291 :			In particular, it will move floating point computations and
3292 :			address computations involving only the heap pointer to
3293 :			their use sites (if there is only a single use).
3294 :			What this means is that if we have a CPS record construction
3295 :			statement
3296 :
3297 :			RECORD(k,vl,w,e)
3298 :
3299 :			we should never count the new record address w as live if w
3300 :			has only one use (which is often the case).
3301 :
3302 :			We should do something similar to floating point, but the transformation
3303 :			there is much more complex, so I won't deal with that.
3304 :
3305 :			Secondly, there are now two new cps primops at our disposal:
3306 :
3307 :			1. rawrecord of record_kind option
3308 :			This pure operator allocates some uninitialized storage from the heap.
3309 :			There are two forms:
3310 :
3311 :			rawrecord NONE [INT n] allocates a tagless record of length n
3312 :			rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3313 :			and initializes the tag.
3314 :
3315 :			2. rawupdate of cty
3316 :			rawupdate cty (v,i,x)
3317 :			Assigns to x to the ith component of record v.
3318 :			The storelist is not updated.
3319 :
3320 :			We use these new primops for both spilling and increment record construction.
3321 :
3322 :			1. Spilling.
3323 :
3324 :			This is implemented with a linear scan algorithm (but generalized
3325 :			to trees). The algorithm will create a single spill record at the
3326 :			beginning of the cps function and use rawupdate to spill to it,
3327 :			and SELECT or SELp to reload from it. So both spills and reloads
3328 :			are fine-grain operations. In contrast, in the old algorithm
3329 :			"spills" have to be bundled together in records.
3330 :
3331 :			Ideally, we should sink the spill record construction to where
3332 :			it is needed. We can even split the spill record into multiple ones
3333 :			at the places where they are needed. But CPS is not a good
3334 :			representation for global code motion, so I'll keep it simple and
3335 :			am not attempting this.
3336 :
3337 :			2. Incremental record construction (aka record splitting).
3338 :
3339 :			Long records with many component values which are simulatenously live
3340 :			(recall that single use record addresses are not considered to
3341 :			be live) are constructed with rawrecord and rawupdate.
3342 :			We allocate space on the heap with rawrecord first, then gradually
3343 :			fill it in with rawupdate. This is the technique suggested to me
3344 :			by Matthias.
3345 :
3346 :			Some restrictions on when this is applicable:
3347 :			1. It is not a VECTOR record. The code generator currently does not handle
3348 :			this case. VECTOR record uses double indirection like arrays.
3349 :			2. All the record component values are defined in the same "basic block"
3350 :			as the record constructor. This is to prevent speculative
3351 :			record construction.
3352 :
3353 :			----------------------------------------------------------------------
3354 :			Name: Allen Leung
3355 :	leunga	1086	Date: 2002/02/22 01:02:00 EST
3356 :			Tag: leunga-20020222-mlrisc-tools
3357 :
3358 :			Minor bug fixes in the parser and rewriter
3359 :
3360 :			----------------------------------------------------------------------
3361 :			Name: Allen Leung
3362 :	leunga	1085	Date: 2002/02/21 20:20:00 EST
3363 :			Tag: leunga-20020221-peephole
3364 :
3365 :			Regenerated the peephole files. Some contained typos in the specification
3366 :			and some didn't compile because of pretty printing bugs in the old version
3367 :			of 'nowhere'.
3368 :
3369 :			----------------------------------------------------------------------
3370 :			Name: Allen Leung
3371 :	leunga	1079	Date: 2002/02/19 20:20:00 EST
3372 :			Tag: leunga-20020219-mlrisc-tools
3373 :			Description:
3374 :
3375 :			Minor bug fixes to the mlrisc-tools library:
3376 :
3377 :			1. Fixed up parsing colon suffixed keywords
3378 :			2. Added the ability to shut the error messages up
3379 :			3. Reimplemented the pretty printer and fixed up/improved
3380 :			the pretty printing of handle and -> types.
3381 :			4. Fixed up generation of literal symbols in the nowhere tool.
3382 :			5. Added some SML keywords to to sml.sty
3383 :
3384 :			----------------------------------------------------------------------
3385 :			Name: Matthias Blume
3386 :	blume	1078	Date: 2002/02/19 16:20:00 EST
3387 :			Tag: blume-20020219-cmffi
3388 :			Description:
3389 :
3390 :			A wild mix of changes, some minor, some major:
3391 :
3392 :			* All C FFI-related libraries are now anchored under $c:
3393 :			$/c.cm --> $c/c.cm
3394 :			$/c-int.cm --> $c/internals/c-int.cm
3395 :			$/memory.cm --> $c/memory/memory.cm
3396 :
3397 :			* "make" tool (in CM) now treats its argument pathname slightly
3398 :			differently:
3399 :			1. If the native expansion is an absolute name, then before invoking
3400 :			the "make" command on it, CM will apply OS.Path.mkRelative
3401 :			(with relativeTo = OS.FileSys.getDir()) to it.
3402 :			2. The argument will be passed through to subsequent phases of CM
3403 :			processing without "going native". In particular, if the argument
3404 :			was an anchored path, then "make" will not lose track of that anchor.
3405 :
3406 :			* Compiler backends now "know" their respective C calling conventions
3407 :			instead of having to be told about it by ml-nlffigen. This relieves
3408 :			ml-nlffigen from one of its burdens.
3409 :
3410 :			* The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3411 :
3412 :			* Export C_DEBUG and C_Debug from $c/c.cm.
3413 :
3414 :			* C type encoding in ml-nlffi-lib has been improved to model the conceptual
3415 :			subtyping relationship between incomplete pointers and their complete
3416 :			counterparts. For this, ('t, 'c) ptr has been changed to 'o ptr --
3417 :			with the convention of instantiating 'o with ('t, 'c) obj whenever
3418 :			the pointer target type is complete. In the incomplete case, 'o
3419 :			will be instantiated with some "'c iobj" -- a type obtained by
3420 :			using one of the functors PointerToIncompleteType or PointerToCompleteType.
3421 :
3422 :			Operations that work on both incomplete and complete pointer types are
3423 :			typed as taking an 'o ptr while operations that require the target to
3424 :			be known are typed as taking some ('t, 'c) obj ptr.
3425 :
3426 :			voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3427 :			where void is an eqtype without any values. This makes it possible
3428 :			to work on voidptr values using functions meant to operate on light
3429 :			incomplete pointers.
3430 :
3431 :			* As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3432 :			been vastly simplified.
3433 :
3434 :			----------------------------------------------------------------------
3435 :			Name: Matthias Blume
3436 :	blume	1076	Date: 2002/02/19 10:48:00 EST
3437 :			Tag: blume-20020219-pqfix
3438 :			Description:
3439 :
3440 :			Applied Chris Okasaki's bug fix for priority queues.
3441 :
3442 :			----------------------------------------------------------------------
3443 :			Name: Matthias Blume
3444 :	blume	1073	Date: 2002/02/15 17:05:00
3445 :	blume	1069	Tag: Release_110_39
3446 :			Description:
3447 :
3448 :	blume	1073	Last-minute retagging is becoming a tradition... :-(
3449 :
3450 :			This is the working release 110.39.
3451 :
3452 :			----------------------------------------------------------------------
3453 :			Name: Matthias Blume
3454 :			Date: 2002/02/15 16:00:00 EST
3455 :			Tag: Release_110_39-orig
3456 :			Description:
3457 :
3458 :	blume	1069	Working release 110.39. New bootfiles.
3459 :
3460 :	blume	1073	(Update: There was a small bug in the installer so it wouldn't work
3461 :			with all shells. So I retagged. -Matthias)
3462 :
3463 :	blume	1069	----------------------------------------------------------------------
3464 :			Name: Matthias Blume
3465 :	blume	1068	Date: 2002/02/15 14:17:00 EST
3466 :			Tag: blume-20020215-showbindings
3467 :			Description:
3468 :
3469 :			Added EnvRef.listBoundSymbols and CM.State.showBindings. Especially
3470 :			the latter can be useful for exploring what bindings are available at
3471 :			the interactive prompt. (The first function returns only the list
3472 :			of symbols that are really bound, the second prints those but also the
3473 :			ones that CM's autoloading mechanism knows about.)
3474 :
3475 :			----------------------------------------------------------------------
3476 :			Name: Matthias Blume
3477 :	blume	1067	Date: 2002/02/15 12:08:00 EST
3478 :			Tag: blume-20020215-iptrs
3479 :			Description:
3480 :
3481 :			Two improvements to ml-nlffigen:
3482 :
3483 :			1. Write files only if they do not exist or if their current contents
3484 :			do not coincide with what's being written. (That is, avoid messing
3485 :			with the time stamps unless absolutely necessary.)
3486 :
3487 :			2. Implement a "repository" mechanism for generated files related
3488 :			to "incomplete pointer types". See the README file for details.
3489 :
3490 :			----------------------------------------------------------------------
3491 :			Name: Matthias Blume
3492 :	blume	1066	Date: 2002/02/14 11:50:00 EST
3493 :			Tag: blume-20020214-quote
3494 :			Description:
3495 :
3496 :			Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm). This is required
3497 :			because of the new and improved tag generation scheme. (Thanks to Allen
3498 :			Leung for pointing it out.)
3499 :
3500 :			----------------------------------------------------------------------
3501 :	george	1065	Name: Lal George
3502 :			Date: 2002/02/14 09:55:27 EST 2002
3503 :			Tag: george-20020214-isabelle-bug
3504 :			Description:
3505 :
3506 :			Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3507 :			of Isabelle on the x86.
3508 :
3509 :			From Allen:
3510 :			-----------
3511 :			I've found the problem:
3512 :
3513 :			in ra-core.sml, I use the counter "blocked" to keep track of the
3514 :			true number of elements in the freeze queue. When the counter goes
3515 :			to zero, I skip examining the queue. But I've messed up the
3516 :			bookkeeping in combine():
3517 :
3518 :			else ();
3519 :			case !ucol of
3520 :			PSEUDO => (if !cntv > 0 then
3521 :			(if !cntu > 0 then blocked := !blocked - 1 else ();
3522 :			^^^^^^^^^^^^^^^^^^^^^^^
3523 :			moveu := mergeMoveList(!movev, !moveu)
3524 :			)
3525 :			else ();
3526 :
3527 :			combine() is called to coalesce two nodes u and v.
3528 :			I think I was thinking that if the move counts of u and v are both
3529 :			greater than zero then after they are coalesced then one node is
3530 :			removed from the freeze queue. Apparently I was thinking that
3531 :			both u and v are of low degree, but that's clearly not necessarily true.
3532 :
3533 :
3534 :			02/12/2002:
3535 :			Here's the patch. HOL now compiles.
3536 :
3537 :			I don't know how this impact on performance (compile
3538 :			time or runtime). This bug caused the RA (especially on the x86)
3539 :			to go thru the potential spill phase when there are still nodes on the
3540 :			freeze queue.
3541 :
3542 :
3543 :
3544 :
3545 :			----------------------------------------------------------------------
3546 :	blume	1058	Name: Matthias Blume
3547 :	blume	1064	Date: 2002/02/13 22:40:00 EST
3548 :			Tag: blume-20020213-fptr-rtti
3549 :			Description:
3550 :
3551 :			Fixed a bug in ml-nlffigen that was introduced with one of the previous
3552 :			updates.
3553 :
3554 :			----------------------------------------------------------------------
3555 :			Name: Matthias Blume
3556 :	blume	1063	Date: 2002/02/13 16:41:00 EST
3557 :			Tag: blume-20020213-cmlpq
3558 :			Description:
3559 :
3560 :			Added new priority queue export symbols (which have just been added to
3561 :			smlnj-lib.cm) to CML's version of smlnj-lib.cm. (Otherwise CML would
3562 :			not compile and the installer would choke.)
3563 :
3564 :			----------------------------------------------------------------------
3565 :			Name: Matthias Blume
3566 :	blume	1062	Date: 2002/02/13 16:15:00 EST
3567 :			Tag: blume-20020213-various
3568 :			Description:
3569 :
3570 :			1. More tweaks to ml-nlffigen:
3571 :
3572 :			- better internal datastructures (resulting in slight speedup)
3573 :			- "-match" option requires exact match
3574 :			- "localized" gensym counters (untagged structs/unions nested within
3575 :			other structs/unions or within typedefs get a fresh counter; their
3576 :			tag will be prefixed by a concatenation of their parents' tags)
3577 :			- bug fixes (related to calculation of transitive closure of types
3578 :			to be included in the output)
3579 :
3580 :			2. Minor Basis updates:
3581 :
3582 :			- added implementations for List.collate and Option.app
3583 :
3584 :			----------------------------------------------------------------------
3585 :			Name: Matthias Blume
3586 :	blume	1060	Date: 2002/02/11 15:55:00 EST
3587 :			Tag: blume-20020211-gensym
3588 :			Description:
3589 :
3590 :			Added a "-gensym" option to command line of ml-nlffigen. This can be
3591 :			used to specify a "stem" -- a string that is inserted in all "gensym'd"
3592 :			names (ML structure names that correspond to unnamed C structs, unions,
3593 :			and enums), so that separate runs of ml-nlffigen do not clash.
3594 :
3595 :			----------------------------------------------------------------------
3596 :			Name: Matthias Blume
3597 :	blume	1059	Date: 2002/02/11 12:05:00 EST
3598 :			Tag: blume-20020211-gensml
3599 :			Description:
3600 :
3601 :			A quick fix for a problem with GenSML (in the pgraph-util library):
3602 :			Make generation of toplevel "local" optional. (Strictly speaking,
3603 :			signature definitions within "local" are not legal SML.)
3604 :
3605 :			Other than that: updates to INSTALL and cm/TODO.
3606 :
3607 :			----------------------------------------------------------------------
3608 :			Name: Matthias Blume
3609 :	blume	1058	Date: 2002/02/08 15:00:00 EST
3610 :			Tag: blume-20020208-uniquepid
3611 :			Description:
3612 :
3613 :			0. Version number has been bumped to 110.38.1. NEW BOOTFILES!!!
3614 :
3615 :			1. The installer (config/install.sh) has gotten smarter:
3616 :
3617 :			- Configuration options are a bit easier to specify now
3618 :			(in config/targets).
3619 :			- Bug in recognizing .tar.bz2 files fixed.
3620 :			- Installer automatically resolves dependencies between
3621 :			configuration options (e.g., if you ask for eXene, you will
3622 :			also get cml -- regardless whether you asked for it or not).
3623 :			- Installer can run in "quieter mode" by setting the environment
3624 :			variable INSTALL_QUIETLY to "true". "Quieter" does not mean
3625 :			"completely silent", though.
3626 :			- Build HashCons library as part of smlnj-lib.
3627 :
3628 :			2. A new scheme for assigning persistent identifiers to compilation
3629 :			units (and, by extension, to types etc.) has been put into place.
3630 :			This fixes a long-standing bug where types and even dynamic values
3631 :			can get internally confused, thereby compromising type safety
3632 :			(abstraction) and dynamic correctness. See
3633 :
3634 :			http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3635 :
3636 :			for an example of how things could go wrong until now.
3637 :
3638 :			The downside of the new scheme is that pids are not quite as
3639 :			persistent as they used to be: CM will generate a fresh pid
3640 :			for every compilation unit that it thinks it sees for the first
3641 :			time. That means that if you compile starting from a clean, fresh
3642 :			source tree at two different times, you end up with different
3643 :			binaries.
3644 :
3645 :			Cutoff recompilation, however, has not been compromised because
3646 :			CM keeps pid information in special caches between runs.
3647 :
3648 :			----------------------------------------------------------------------
3649 :	george	1053	Name: Lal George
3650 :	george	1055	Date: 2002/02/07 15:34:13 EST 2002
3651 :			Tag: <none>
3652 :			Description:
3653 :
3654 :			Compilers that generate assembly code may produce global labels
3655 :			whose value is resolved at link time. The various peephole optimization
3656 :			modules did not take this in account.
3657 :
3658 :			TODO. The Labels.addrOf function should really return an option
3659 :			type so that clients are forced to deal with this issue, rather
3660 :			than an exception being raised.
3661 :
3662 :			----------------------------------------------------------------------
3663 :			Name: Lal George
3664 :	george	1053	Date: 2002/02/06 13:55:02 EST
3665 :			Tag: george-20020206-ra-breakup
3666 :			Description:
3667 :
3668 :			1. A bug fix from Allen.
3669 :
3670 :			A typo causes extra fstp %st(0)'s to be generated at compensation
3671 :			edges, which might cause stack underflow traps at runtime. This
3672 :			occurs in fft where there are extraneous fstps right before the 'into'
3673 :			trap instruction (in this case they are harmless since none of the
3674 :			integers overflow.)
3675 :
3676 :	george	1054	2. Pulled out various utility modules that were embedded in the modules
3677 :	george	1053	of the register allocator. I need these modules for other purposes, but
3678 :	george	1054	they are not complete enough to put into a library (just yet).
3679 :	george	1053	----------------------------------------------------------------------
3680 :	blume	1049	Name: Matthias Blume
3681 :			Date: 2002/01/31 16:05:00 EST
3682 :			Tag: blume-20020131-sparc-ccalls
3683 :			Description:
3684 :
3685 :			1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3686 :			of extra stack space by mistake. Fixed.
3687 :
3688 :			2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3689 :
3690 :			----------------------------------------------------------------------
3691 :	leunga	1048	Name: Allen Leung
3692 :			Date: 2002/01/30
3693 :			Tag: leunga-20020130-nowhere-bug-fix
3694 :			Description:
3695 :
3696 :			MLRISC bug fixes:
3697 :			1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3698 :			2. MachineInt.fromString was negating its value.
3699 :
3700 :			----------------------------------------------------------------------
3701 :	blume	1039	Name: Matthias Blume
3702 :	blume	1047	Date: 2002/01/29
3703 :			Tag: blume-20020129-INSTALL
3704 :			Description:
3705 :
3706 :			- Added somewhat detailed installation instructions (file INSTALL).
3707 :			- Fixed curl-detection bug in config/install.sh.
3708 :			- It is now possible to select the URL getter using the URLGETTER
3709 :			environment variable:
3710 :
3711 :			not set / "unknown" --> automatic detection (script tries wget,
3712 :			curl, and lynx)
3713 :			"wget" / "curl" / "lynx" --> use the specified program (script "knows"
3714 :			how to properly invoke them)
3715 :			other --> use $URLGETTER directly, it must take
3716 :			precisely two command-line arguments
3717 :			(source URL and destination file name)
3718 :
3719 :			----------------------------------------------------------------------
3720 :			Name: Matthias Blume
3721 :	blume	1044	Date: 2002/01/28
3722 :	blume	1045	Tag: blume-20020128-sparc-ccalls
3723 :			Description:
3724 :
3725 :			- Fixed problem with calculation of "used" registers in sparc-c-calls.
3726 :			- Make use of the allocParam argument in sparc-c-calls.
3727 :
3728 :			----------------------------------------------------------------------
3729 :			Name: Matthias Blume
3730 :			Date: 2002/01/28
3731 :	blume	1044	Tag: blume-20020128-allocParam
3732 :			Description:
3733 :
3734 :			John Reppy: Changes c-calls API to accept client-callback for
3735 :			allocating extra stack space.
3736 :			me: Corresponding changes to mlriscGen (using a dummy argument that
3737 :			does not change the current behavior).
3738 :
3739 :			----------------------------------------------------------------------
3740 :			Name: Matthias Blume
3741 :	blume	1040	Date: 2002/01/28 12:00:00
3742 :			Tag: Release_110_38
3743 :			Description:
3744 :
3745 :			This time for real!!!
3746 :
3747 :			----------------------------------------------------------------------
3748 :			Name: Matthias Blume
3749 :	blume	1039	Date: 2002/01/28 10:56:00 EST
3750 :			Tag: blume-20020128-retraction
3751 :			Description:
3752 :
3753 :			0. Retracted earlier 110.38. (The Release_110_38 tag has been replaced
3754 :			with blume-Release_110_38-retracted.)
3755 :
3756 :			1. Fixed a problem with incorrect rounding modes in real64.sml.
3757 :			(Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3758 :
3759 :			2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3760 :			and enums fixed. The naming of corresponding ML identifiers should
3761 :			now be consistent again.
3762 :
3763 :			----------------------------------------------------------------------
3764 :	leunga	1038	Name: Allen Leung
3765 :			Date: 2002/01/27
3766 :			Tag: leunga-20020127-nowhere
3767 :			Description:
3768 :
3769 :			Added a target called nowhere in the configuration scripts.
3770 :			Enabling this will build the MLRISC 'nowhere' tool (for translating
3771 :			programs with where-clauses into legal SML code) during installation.
3772 :
3773 :			----------------------------------------------------------------------
3774 :	blume	1036	Name: Matthias Blume
3775 :	leunga	1038	Date: 2002/01/25 21:27:00 EST
3776 :	blume	1039	Tag: blume-Release_110_38-retracted
3777 :	blume	1037	Description:
3778 :
3779 :			Call it a (working) release! Version is 110.38. Bootfiles are ready.
3780 :
3781 :			README will be added later.
3782 :
3783 :	blume	1039	!!! NOTE: Re-tagged as blume-Release_110_38-retracted. Original tag
3784 :			(Release_110_38) removed. Reason: Last-minute bug fixes.
3785 :
3786 :	blume	1037	----------------------------------------------------------------------
3787 :			Name: Matthias Blume
3788 :	blume	1036	Date: 2002/01/25
3789 :			Tag: blume-20020125-ffi
3790 :			Description:
3791 :
3792 :			A large number of tweaks and improvements to ml-nlffi-lib and
3793 :			ml-nlffigen:
3794 :
3795 :			- ML represenation types have been streamlined
3796 :			- getter and setter functions work with concrete values, not abstract
3797 :			ones where possible
3798 :			- ml-nlffigen command line more flexible (see README file there)
3799 :			- some bugs have been fixed (hopefully)
3800 :
3801 :			----------------------------------------------------------------------
3802 :	george	1034	Name: Lal George
3803 :			Date: 2002/01/24
3804 :			Tag: george-20020124-risc-ra-interface
3805 :			Description:
3806 :
3807 :			There is a dramatic simplification in the interface to the
3808 :			register allocator for RISC architectures as a result of making
3809 :			parallel copy instructions explicit.
3810 :
3811 :			----------------------------------------------------------------------
3812 :	blume	1028	Name: Matthias Blume
3813 :	blume	1032	Date: 2002/01/22
3814 :			Tag: blume-20020122-x86-ccalls
3815 :			Description:
3816 :
3817 :			Bug fix for c-calls on x86 (having to do with how char- and
3818 :			short-arguments are being handled).
3819 :
3820 :			----------------------------------------------------------------------
3821 :			Name: Matthias Blume
3822 :	blume	1031	Date: 2002/01/21
3823 :			Tag: blume-20020121-ff
3824 :			Description:
3825 :
3826 :			Another day of fiddling with the FFI...
3827 :
3828 :			1. Bug fix/workaround: CKIT does not complain about negative array
3829 :			dimensions, so ml-nlffigen has to guard itself against this possibility.
3830 :			(Otherwise a negative dimension would send it into an infinite loop.)
3831 :
3832 :			2. Some of the abstract types (light objects, light pointers, most "base"
3833 :			types) in structure C are now eqtypes.
3834 :
3835 :			3. Added constructors and test functions for NULL function pointers.
3836 :
3837 :			----------------------------------------------------------------------
3838 :			Name: Matthias Blume
3839 :	blume	1028	Date: 2002/01/18
3840 :	blume	1030	Tag: blume-20020118-ready-for-new-release
3841 :			Description:
3842 :
3843 :			Made config/srcarchiveurl point to a new place. (Will provide boot
3844 :			files shortly.)
3845 :
3846 :			Maybe we christen this to be 110.38?
3847 :
3848 :			----------------------------------------------------------------------
3849 :			Name: Matthias Blume
3850 :			Date: 2002/01/18
3851 :	blume	1029	Tag: blume-20020118-more-ffifiddle
3852 :			Description:
3853 :
3854 :			Today's FFI fiddling:
3855 :
3856 :			- Provided a structure CGetSet with "convenient" versions of C.Get.* and
3857 :			C.Set.* that use concrete (MLRep.*) arguments and results instead
3858 :			of abstract ones.
3859 :
3860 :			- Provided word-style bit operations etc. for "int" representation
3861 :			types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3862 :			and Long.
3863 :
3864 :			----------------------------------------------------------------------
3865 :			Name: Matthias Blume
3866 :			Date: 2002/01/18
3867 :	blume	1028	Tag: blume-20020118-use-x86-fp
3868 :			Description:
3869 :
3870 :			Now that x86-fast-fp seems to be working, I turned it back on again
3871 :			by default. (Seems to work fine now, even with the FFI.)
3872 :
3873 :			Other than that, I added some documentation about the FFI to
3874 :			src/ml-nlffigen/README and updated the FFI test examples in
3875 :			src/ml-nlffi-lib/Tests/*.
3876 :
3877 :			----------------------------------------------------------------------
3878 :	leunga	1025	Name: Allen Leung
3879 :			Date: 2002/01/17
3880 :			Tag: leunga-20020117-x86-fast-fp-call
3881 :			Description:
3882 :
3883 :			1. Fixed a problem with handling return fp values when x86's fast fp
3884 :			mode is turned on.
3885 :
3886 :			2. Minor pretty printing fix for cellset. Print %st(0) as %st(0) instead
3887 :			of %f32.
3888 :
3889 :			3. Added a constructor INT32lit to the ast of MLRISC tools.
3890 :
3891 :			----------------------------------------------------------------------
3892 :	blume	1021	Name: Matthias Blume
3893 :			Date: 2002/01/16
3894 :			Tag: blume-20020116-ffifiddle
3895 :			Description:
3896 :
3897 :			More fiddling with the FFI interface:
3898 :
3899 :			- Make constness 'c instead of rw wherever possible. This eliminates
3900 :			the need for certain explicit coercions. (However, due to ML's
3901 :			value polymorphism, there will still be many cases where explicit
3902 :			coercions are necessary. Phantom types are not the whole answer
3903 :			to modeling a subtyping relationship in ML.)
3904 :
3905 :			- ro/rw coersions for pointers added. (Avoids the detour through */&.)
3906 :
3907 :			- "printf" test example added to src/ml-nlffi-lib/Tests. (Demonstrates
3908 :			clumsy workaround for varargs problem.)
3909 :
3910 :			----------------------------------------------------------------------
3911 :	george	1016	Name: Lal George
3912 :			Date: 2002/01/15
3913 :			Tag: <none>
3914 :			Description:
3915 :
3916 :			1. Since COPY instructions are no longer native to the architecture,
3917 :			a generic functor can be used to implement the expandCopies function.
3918 :
3919 :			2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3920 :			TEXT segment.
3921 :
3922 :			----------------------------------------------------------------------
3923 :	blume	1011	Name: Matthias Blume
3924 :	blume	1015	Date: 2002/01/15
3925 :			Tag: blume-20020115-ffiupdates
3926 :			Description:
3927 :
3928 :			1. Fix for bug resulting in single-precision float values being returned
3929 :			incorrectly from FFI calls.
3930 :
3931 :			2. Small modifications to C FFI API:
3932 :
3933 :			- memory-allocation routines return straight objects (no options)
3934 :			and raise an exception in out-of-memory situations
3935 :			- unsafe extensions to cast between function pointers and pointers
3936 :			from/to ints
3937 :			- added structure C_Debug as an alternative to structure C where
3938 :			pointer-dereferencing (|*| and |*!) always check for null-pointers
3939 :			- added open_lib' to DynLinkage; open_lib' works like open_lib
3940 :			but also takes a (possibly empty) list of existing library handles
3941 :			that the current library depends on
3942 :
3943 :			----------------------------------------------------------------------
3944 :			Name: Matthias Blume
3945 :	blume	1011	Date: 2002/01/10
3946 :			Tag: blume-20020110-newffigen
3947 :			Description:
3948 :
3949 :			1. Updates to portable graph code.
3950 :
3951 :			2. Major update to ml-nlffigen and ml-nlffi-lib. Things are much
3952 :			more scalable now so that even huge interfaces such as the one
3953 :			for GTK compile in finite time and space. :-)
3954 :			See src/ml-nlffigen/README for details on what's new.
3955 :
3956 :			----------------------------------------------------------------------
3957 :	george	1003	Name: Lal George
3958 :	george	1009	Date: 2001/01/09 14:31:35 EST 2002
3959 :			Tag: george-20011206-rm-native-copy
3960 :			Description:
3961 :
3962 :			Removed the native COPY and FCOPY instructions
3963 :			from all the architectures and replaced it with the
3964 :			explicit COPY instruction from the previous commit.
3965 :
3966 :			It is now possible to simplify many of the optimizations
3967 :			modules that manipulate copies. This has not been
3968 :			done in this change.
3969 :
3970 :			----------------------------------------------------------------------
3971 :			Name: Lal George
3972 :	george	1003	Date: 2001/12/06 16:50:13 EST 2001
3973 :			Tag: george-20011206-mlrisc-instruction
3974 :			Description:
3975 :	george	824
3976 :	george	1003	Changed the representation of instructions from being fully abstract
3977 :			to being partially concrete. That is to say:
3978 :
3979 :			from
3980 :			type instruction
3981 :
3982 :			to
3983 :			type instr (* machine instruction *)
3984 :
3985 :			datatype instruction =
3986 :			LIVE of {regs: C.cellset, spilled: C.cellset}
3987 :			| KILL of {regs: C.cellset, spilled: C.cellset}
3988 :			| COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3989 :			| ANNOTATION of {i: instruction, a: Annotations.annotation}
3990 :			| INSTR of instr
3991 :
3992 :			This makes the handling of certain special instructions that appear on
3993 :			all architectures easier and uniform.
3994 :
3995 :			LIVE and KILL say that a list of registers are live or killed at the
3996 :			program point where they appear. No spill code is generated when an
3997 :			element of the 'regs' field is spilled, but the register is moved to
3998 :			the 'spilled' (which is present, more for debugging than anything else).
3999 :
4000 :			LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
4001 :			We used to generate:
4002 :
4003 :			DEFFREG f1
4004 :			f1 := f2 + f3
4005 :			trapb
4006 :
4007 :			but now generate:
4008 :
4009 :			f1 := f2 + f3
4010 :			trapb
4011 :			LIVE {regs=[f1,f2,f3], spilled=[]}
4012 :
4013 :			Furthermore, the DEFFREG (hack) required that all floating point instruction
4014 :			use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
4015 :			defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
4016 :			in a cleaner alpha implementation. (Hopefully, intel will not get rid of
4017 :			this architecture).
4018 :
4019 :			COPYXXX is intended to replace the parallel COPY and FCOPY available on
4020 :			all the architectures. This will result in further simplification of the
4021 :			register allocator that must be aware of them for coalescing purposes, and
4022 :			will also simplify certain aspects of the machine description that provides
4023 :			callbacks related to parallel copies.
4024 :
4025 :			ANNOTATION should be obvious, and now INSTR represents the honest to God
4026 :			machine instruction set!
4027 :
4028 :			The <arch>/instructions/<arch>Instr.sml files define certain utility
4029 :			functions for making porting easier -- essentially converting upper case
4030 :			to lower case. All machine instructions (of type instr) are in upper case,
4031 :			and the lower case form generates an MLRISC instruction. For example on
4032 :			the alpha we have:
4033 :
4034 :			datatype instr =
4035 :			LDA of {r:cell, b:cell, d:operand}
4036 :			| ...
4037 :
4038 :			val lda : {r:cell, b:cell, d:operand} -> instruction
4039 :			...
4040 :
4041 :			where lda is just (INSTR o LDA), etc.
4042 :
4043 :	leunga	815	----------------------------------------------------------------------
4044 :	blume	986	Name: Matthias Blume
4045 :	blume	993	Date: 2001/11/22 21:40:00 EST
4046 :			Tag: Release_110_37
4047 :			Description:
4048 :
4049 :			Release 110.37. This time for real.
4050 :
4051 :			----------------------------------------------------------------------
4052 :			Name: Matthias Blume
4053 :	blume	988	Date: 2001/11/21 16:35:00 EST
4054 :			Tag: blume-20011121-foot-in-mouth
4055 :			Description:
4056 :
4057 :			Removed the "Release_110_37" tag because of a serious bug.
4058 :			This will be re-tagged once the bug is fixed.
4059 :
4060 :			----------------------------------------------------------------------
4061 :			Name: Matthias Blume
4062 :	blume	987	Date: 2001/11/21 16:14:00 EST
4063 :			Tag: blume-20011121-forgottenfile
4064 :			Description:
4065 :
4066 :			Forgot to add a file. (Just a .tex-file -- part of
4067 :			the CM manual source.)
4068 :
4069 :			----------------------------------------------------------------------
4070 :			Name: Matthias Blume
4071 :	blume	986	Date: 2001/11/21 16:10:00 EST
4072 :	blume	988	Tag: blume-20011121-invalid_110_37
4073 :	blume	986	Description:
4074 :
4075 :	blume	988	Note: I removed the original tag "Release_110_37" from this commit
4076 :			because we found a serious bug in all non-x86 backends.
4077 :			- Matthias
4078 :
4079 :	blume	986	1. Modifications to the SML/NJ code generator and to the runtime system
4080 :			so that code object name strings are directly inserted into code
4081 :			objects at code generation time. The only business the runtime system
4082 :			has with this is now to read the name strings on occasions.
4083 :			(The encoding of the name string has also changed somewhat.)
4084 :
4085 :			2. CM now implements a simple "set calculus" for specifying export lists.
4086 :			In particular, it is now possible to refer to the export lists of
4087 :			other libraries/groups/sources and form unions as well as differences.
4088 :			See the latest CM manual for details.
4089 :
4090 :			3. An separate notion of "proxy" libraries has again be eliminated from
4091 :			CM's model. (Proxy libraries are now simply a special case of using
4092 :			the export list calculus.)
4093 :
4094 :			4. Some of the existing libraries now take advantage of the new set
4095 :			calculus.
4096 :			(Notice that not all libraries have been converted because some
4097 :			of the existing .cm-files are supposed to be backward compatible
4098 :			with 110.0.x.)
4099 :
4100 :			5. Some cleanup in stand-alone programs. (Don't use "exnMessage" -- use
4101 :			"General.exnMessage"! The former relies on a certain hook to be
4102 :			initialized, and that often does not happen in the stand-alone case.)
4103 :
4104 :			----------------------------------------------------------------------
4105 :	george	984	Name: Lal George
4106 :			Date: 2001/11/21 13:56:18 EST
4107 :			Tag: george-2001121-pseudo-ops
4108 :			Description:
4109 :
4110 :			Implemented a complete redesign of MLRISC pseudo-ops. Now there
4111 :			ought to never be any question of incompatabilities with
4112 :			pseudo-op syntax expected by host assemblers.
4113 :
4114 :			For now, only modules supporting GAS syntax are implemented
4115 :			but more should follow, such as MASM, and vendor assembler
4116 :			syntax, e.g. IBM as, Sun as, etc.
4117 :
4118 :			----------------------------------------------------------------------
4119 :	blume	977	Name: Matthias Blume
4120 :			Date: 2001/11/14 11:52:00 EST
4121 :			Tag: blume-20011114-srcname
4122 :			Description:
4123 :
4124 :			1. Routed the name of the current source file to mlriscgen where it
4125 :			should be directly emitted into the code object. (This last part
4126 :			is yet to be done.)
4127 :
4128 :			2. Some cleanup of the pgraph code to make it match the proposal that
4129 :			I put out the other day. (The proposal notwithstanding, things are
4130 :			still in flux here.)
4131 :
4132 :			----------------------------------------------------------------------
4133 :	george	976	Name: Lal George
4134 :			Date: 2001/11/14 09:44:04 EST
4135 :			Tag:
4136 :			Description:
4137 :
4138 :			Fix for a backpatching bug reported by Allen.
4139 :
4140 :			Because the boundary between short and long span-dependent
4141 :			instructions is +/- 128, there are an astounding number of
4142 :			span-dependent instructions whose size is over estimated.
4143 :
4144 :			Allen came up with the idea of letting the size of span
4145 :			dependent instructions be non-monotonic, for a maxIter
4146 :			number of times, after which the size must be monotonically
4147 :			increasing.
4148 :
4149 :			This table shows the number of span-dependent instructions
4150 :			whose size was over-estimated as a function of maxIter, for the
4151 :			file Parse/parse/ml.grm.sml:
4152 :
4153 :			maxIter # of instructions:
4154 :			10 687
4155 :			20 438
4156 :			30 198
4157 :			40 0
4158 :
4159 :			In compiling the compiler, there is no significant difference in
4160 :			compilation speed between maxIter=10 and maxIter=40. Actually,
4161 :			my measurements showed that maxIter=40 was a tad faster than
4162 :			maxIter=10! Also 96% of the files in the compiler reach a fix
4163 :			point within 13 iterations, so fixing maxIter at 40, while high,
4164 :			is okay.
4165 :
4166 :			----------------------------------------------------------------------
4167 :	blume	975	Name: Matthias Blume
4168 :			Date: 2001/10/31 15:25:00 EST
4169 :			Tag: blume-20011031-pgraph
4170 :			Description:
4171 :
4172 :			CKIT:
4173 :			* Changed the "Function" constructor of type Ast.ctype to carry optional
4174 :			argument identifiers.
4175 :			* Changed the return type of TypeUtil.getFunction accordingly.
4176 :			* Type equality ignores the argument names.
4177 :			* TypeUtil.composite tries to preserve argument names but gives up quickly
4178 :			if there is a mismatch.
4179 :
4180 :			installation script:
4181 :			* attempts to use "curl" if available (unless "wget" is available as well)
4182 :
4183 :			CM:
4184 :			* has an experimental implementation of "portable graphs" which I will
4185 :			soon propose as an implementation-independent library format
4186 :			* there are also new libraries $/pgraph.cm and $/pgraph-util.cm
4187 :
4188 :			NLFFI-LIB:
4189 :			* some cleanup (all cosmetic)
4190 :
4191 :			NLFFIGEN:
4192 :			* temporarily disabled the mechanism that suppresses ML output for
4193 :			C definitions whose identifiers start with an underscore character
4194 :			* generate val bindings for enum constants
4195 :			* user can request that only one style (light or heavy) is being used;
4196 :			default is to use both (command-line arguments: -heavy and -light)
4197 :			* fixed bug in handling of function types involving incomplete pointers
4198 :			* generate ML entry points that take record arguments (i.e., using
4199 :			named arguments) for C functions that have a prototype with named
4200 :			arguments
4201 :			(see changes to CKIT)
4202 :
4203 :			----------------------------------------------------------------------
4204 :	leunga	972	Name: Allen Leung
4205 :			Date: 2001/10/27 20:34:00 EDT
4206 :			Tag: leunga-20011027-x86-fast-fp-call
4207 :			Description:
4208 :
4209 :			Fixed the bug described in blume-20010920-slowfp.
4210 :
4211 :			The fix involves
4212 :			1. generating FCOPYs in FSTP in ia32-svid
4213 :			2. marking a CALL with the appropriate annotation
4214 :
4215 :			----------------------------------------------------------------------
4216 :	blume	961	Name: Matthias Blume
4217 :			Date: 2001/10/16 11:32:00 EDT
4218 :			Tag: blume-20011016-netbsd
4219 :			Description:
4220 :
4221 :			Underscore patch from Chris Richards (fixing problem with compiling
4222 :			runtime system under recent NetBSD).
4223 :
4224 :			----------------------------------------------------------------------
4225 :	george	959	Name: Allen Leung
4226 :			Date: 2001/10/12 17:18:32 EDT 2001
4227 :			Tag: leung-20011012-x86-printflowgraph
4228 :			Description:
4229 :
4230 :			X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
4231 :
4232 :			----------------------------------------------------------------------
4233 :	george	952	Name: Lal George
4234 :	george	958	Date: 2001/10/11 23:51:34 EDT
4235 :			Tag: george-20011011-too-many-instrs
4236 :			Description:
4237 :
4238 :			The representation of a program point never expected to see more
4239 :			than 65536 instructions in a basic block!
4240 :
4241 :			----------------------------------------------------------------------
4242 :			Name: Lal George
4243 :	george	952	Date: 2001/10/09 09:41:37 EDT
4244 :			Tag: george-20011008-mlrisc-labels
4245 :			Description:
4246 :
4247 :			Changed the machine description files to support printing of
4248 :			local and global labels in assembly code, based on host assembler
4249 :			conventions.
4250 :
4251 :			----------------------------------------------------------------------
4252 :	blume	916	Name: Matthias Blume
4253 :	blume	939	Date: 2001/09/25 15:25:00 EDT
4254 :			Tag: blume-20010925-exninfo
4255 :			Description:
4256 :
4257 :			I provided a non-hook implementation of exnName (at the toplevel) and
4258 :			made the "dummy" implementation of exnMessage (at the toplevel) more
4259 :			useful: if nothing gets "hooked in", then at least you are going to
4260 :			see the exception name and a message indicating why you don't see more.
4261 :
4262 :			[For the time being, programs that need exnMessage and want to use
4263 :			ml-build should either use General.exnMessage (strongly recommended) or
4264 :			refer to structure General at some other point so that CM sees a
4265 :			static dependency.]
4266 :
4267 :			[Similar remarks go for "print" and "use": If you want to use their
4268 :			functionality in stand-alone programs generated by ml-build, then use
4269 :			TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
4270 :
4271 :			----------------------------------------------------------------------
4272 :			Name: Matthias Blume
4273 :	blume	938	Date: 2001/09/20 17:28:00 EDT
4274 :			Tag: blume-20010920-slowfp
4275 :			Description:
4276 :
4277 :			Allen says that x86-fast-fp is not safe yet, so I turned it off again...
4278 :
4279 :			----------------------------------------------------------------------
4280 :			Name: Matthias Blume
4281 :	blume	937	Date: 2001/09/20 17:20:00 EDT
4282 :			Tag: blume-20010920-canonicalpaths
4283 :			Description:
4284 :
4285 :			0. Updated the BOOT file (something that I forgot to do earlier).
4286 :
4287 :			1. Small internal change to CM so that it avoids "/../" in filenames
4288 :			as much as possible (but only where it is safe).
4289 :
4290 :			2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
4291 :			that arguments that contain delimiters are passed through correctly.
4292 :			This change also means that all "special" arguments of the form
4293 :			@SMLxxx... must come first.
4294 :
4295 :			3. Changed install script to put relative anchor names for tool commands
4296 :			into pathconfig.
4297 :
4298 :			----------------------------------------------------------------------
4299 :			Name: Matthias Blume