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

Parent Directory | Revision Log

---

Revision 1669 - (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	1668	Name: Matthias Blume (blume (at) tti - c (dot) org)
17 :	mblume	1669	Date: 2004/10/27 17:50:00 CDT
18 :			Tag: blume-20041027-btrace-msg
19 :			Description:
20 :
21 :			BackTrace.monitor now also reports the source of the exception that
22 :			triggered the trace.
23 :
24 :			----------------------------------------------------------------------
25 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
26 :	mblume	1668	Date: 2004/10/27 17:20:00 CDT
27 :			Tag: blume-20041027-x86-c-calls
28 :			Description:
29 :
30 :			This is the HISTORY entry for two earlier commits, both concerning
31 :			the x86 c-calls code in MLRISC:
32 :
33 :			- added a missing LOAD in the code that deals with struct arguments
34 :			- made sure the caller does not add the wrong number of bytes to the
35 :			stack pointer after a call of a function returning a struct
36 :			(the callee already pops the implicit argument which points to
37 :			the space reserved for the result)
38 :
39 :			----------------------------------------------------------------------
40 :	allenleung	1662	Name: Allen Leung (leunga (at) reservoir (dot) com)
41 :			Date: 2004/10/24 14:00:00 EST
42 :			Tag: leunga-20041024-x86-gas-fucomip
43 :			Description:
44 :
45 :			John discovered a bug in the syntax of fucomip.
46 :			The opcodes FU?COMIP? have been changed to
47 :
48 :			fu?comip? %st(i), %st
49 :
50 :			----------------------------------------------------------------------
51 :	mblume	1622	Name: Matthias Blume (blume (at) tti - c (dot) org)
52 :	mblume	1656	Date: 2004/10/20 15:06:00 CDT
53 :			Tag: blume-20041020-standalone-backtrace
54 :			Description:
55 :
56 :			Added a mechanism for getting back-trace information from standalone
57 :			programs. Here is how it works:
58 :
59 :			1. The part of the program from which you want to get backtrace
60 :			information (usually the whole program) should be wrapped with
61 :			BackTrace.monitor. This is a (unit->'a)->'a function, and your
62 :			main program could be modified from something like
63 :
64 :			fun main (pgm, args) = ...
65 :
66 :			to
67 :
68 :			fun main (pgm, args) = BackTrace.monitor (fn () => ...)
69 :
70 :			2. To be able to access BackTrace.monitor, you have to add
71 :			library $smlnj-tdp/plugins.cm to the .cm file that contains your
72 :			main function.
73 :
74 :			3. Remove all compiled code (i.e., all the .cm/ subdirectories that
75 :			CM might have created in the past for your project).
76 :
77 :			4. Build the system using this command line:
78 :
79 :			ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
80 :			myprog.cm MyProg.main myprog
81 :
82 :			instead of the usual
83 :
84 :			ml-build myprog.cm MyProg.main myprog
85 :
86 :			I changed a library name:
87 :
88 :			$/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
89 :
90 :			New libraries:
91 :
92 :			$smlnj-tdp/back-trace.cm
93 :			-- when loaded causes the back-trace plugin to be installed
94 :			$smlnj-tdp/coverage.cm
95 :			-- when loaded causes the coverage plugin to be installed
96 :
97 :			----------------------------------------------------------------------
98 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
99 :	mblume	1655	Date: 2004/10/18 16:45:00 CDT
100 :			Tag: blume-20041018-groupowner
101 :			Description:
102 :
103 :			Added an "obsolete" warning for the "group owner" syntax to CM's parser.
104 :
105 :			Eliminated group owner specs from .cm files throughout the source tree.
106 :
107 :			----------------------------------------------------------------------
108 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
109 :	mblume	1653	Date: 2004/10/15 15:45:00 CDT
110 :			Tag: blume-20041015-coverage
111 :			Description:
112 :
113 :			* Test coverage tool added!
114 :			* Further reorganization of tracing-, debugging-, and profiling support:
115 :
116 :			- moved original BTImp -- now called BackTrace -- into a separate
117 :			library called $/trace-debug-profile.cm
118 :			- eliminated all mentions of BTrace from SMLofNJ.Internals
119 :			- only the instrumentation mechanism is now left in the compiler proper
120 :			- BackTrace module is a plugin which is NOT plugged in by default
121 :			- Coverage module is another such plugin
122 :
123 :			To get the benefits of any of these plugin modules, the code in
124 :			question must be compiled with tdp instrumentation turned on. This
125 :			can be done by setting SMLofNJ.Internals.TDP.mode to true.
126 :			(The ref cell is also controlled via the -Ctdp.instrument=... switch.)
127 :
128 :			Plugins are selected at link time. (Pre-compiled instrumented code
129 :			can be re-loaded with different plugins in effect.) When an
130 :			instrumented module is linked, whatever plugins are at that time
131 :			enabled will come into effect for that module.
132 :
133 :			To enable the back-trace plugin, load library $/trace-debug-profile.cm
134 :			and invoke BackTrace.install() (e.g., from the interactive prompt).
135 :			To enable the coverage plugin, load the same library and invoke
136 :			Coverage.install().
137 :
138 :			Back-traces are generated automatically on uncaught exceptions and
139 :			when the code in question explicitly invokes BackTrace.trigger().
140 :
141 :			Coverage (and execution frequency-) information must be queried
142 :			explicitly by calling Coverage.not_covered and Coverage.hot_spots.
143 :
144 :			----------------------------------------------------------------------
145 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
146 :	mblume	1652	Date: 2004/10/14 17:40:00 CDT
147 :			Tag: blume-20041014-tdp-core
148 :			Description:
149 :
150 :			Snapshot of a significant overhaul of how the trace/debug/profile support
151 :			is hooked into the system (specifically: Core and SMLofNJ.Internals).
152 :
153 :			----------------------------------------------------------------------
154 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
155 :	mblume	1650	Date: 2004/10/13 16:34:00 CDT
156 :			Tag: blume-20041013-tdp
157 :			Description:
158 :
159 :			Some rationalization of names:
160 :
161 :			structure BTrace -> structure TDPInstrument
162 :			etc.
163 :
164 :			This is is preparation of using the original back-trace
165 :			instrumentation for other purposes. "TDP" stands for
166 :			Trace/Debug/Profile.
167 :
168 :			The control flag controlling whether instrumentation is on or off is now
169 :			registered under a different name, so instead of running sml as
170 :
171 :			sml -Cinstrument.btrace-mode=true
172 :
173 :			one has to say
174 :
175 :			sml -Ctdp.instrument=true
176 :
177 :			----------------------------------------------------------------------
178 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
179 :	mblume	1645	Date: 2004/10/11 16:37:00 CDT
180 :			Tag: blume-20041011-regions
181 :			Description:
182 :
183 :			Made some minor modifications to elabcore.sml to have source regions
184 :			be propagated more tightly -- resulting in better (i.e., smaller)
185 :			regions being reported in error- and debug messages.
186 :
187 :			----------------------------------------------------------------------
188 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
189 :	mblume	1644	Date: 2004/10/08 22:50:00 CDT
190 :			Tag: blume-20041008-cmkw
191 :			Description:
192 :
193 :			Fixed handling of keywords in .cm files: After seeing "is" the lexer
194 :			treats subsequent occurrences of "group", "library", "source", "is",
195 :			"*", and "-" as ordinary identifiers rather than keywords.
196 :
197 :			Most seriously, this fixes a problem with CM's "shell" tool. The tool
198 :			is supposed to accept a tool argument called "source", but this did
199 :			not work because of the clash with the keyword.
200 :
201 :			----------------------------------------------------------------------
202 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
203 :	mblume	1642	Date: 2004/10/07 16:00:00 CDT
204 :			Tag: blume-20041007-cleanup
205 :			Description:
206 :
207 :			Assorted cleanup work:
208 :
209 :			- got rid of intstrmap in favor of using the library's
210 :			hash table implementation
211 :			- threw out most of the pathnames stuff, as it was not used anyway
212 :			- simplified tokentable implementation
213 :			- fixed some minor spelling errors
214 :
215 :			----------------------------------------------------------------------
216 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
217 :	mblume	1641	Date: 2004/10/06 15:15:15 CDT
218 :			Tag: blume-20041006-handler
219 :			Description:
220 :
221 :			Cleaned up the absyn to reflect the invariant that HANDLE always
222 :			carries a FNexp as part of the type definition. This eliminates some
223 :			superfluous sanity checks at runtime down the road.
224 :
225 :			Some minor cleanup of the btrace code.
226 :
227 :			----------------------------------------------------------------------
228 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
229 :	mblume	1640	Date: 2004/10/01 10:20:30 CDT
230 :			Tag: blume-20041001-slave
231 :			Description:
232 :
233 :			Added hack to make slave mode work in the presence of the version
234 :			tool. (Still, since the master does two passes over the code for
235 :			CMB.make, the release number gets bumped twice when slaves are
236 :			attached. I don't know if this is worth fixing...)
237 :
238 :			----------------------------------------------------------------------
239 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
240 :	mblume	1639	Date: 2004/09/30 10:55:00 CDT
241 :			Tag: blume-20040930-version
242 :			Description:
243 :
244 :			* Moved the "version" magic into its own little library under
245 :			src/system/smlnj/internal. This avoids expensive reconstruction
246 :			of a stable src/compiler/core.cm.
247 :
248 :			* At the same time, structure CompilerVersion is now known as structure
249 :			SMLNJVersion.
250 :
251 :			* Arranged for the version tool to NOT kick in when rebuilding the system
252 :			(makeml -rebuild, fixpt). Otherwise one would never reach a fixpoint.
253 :			Also, loading the versiontool does not work when rebuilding the system
254 :			because CM is not properly initialized at that time.
255 :
256 :			----------------------------------------------------------------------
257 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
258 :	mblume	1638	Date: 2004/09/29 14:00:00 CDT
259 :			Tag: blume-20040929-autoversion
260 :			Description:
261 :
262 :			Implemented some CM magic to have
263 :			file src/compiler/TopLevel/main/version.sml
264 :			generated automagically.
265 :			The version is taken from two files: config/version and config/release.
266 :			The first is expected to contain a two-part version number such as 110.49.
267 :			The second should contain a single number, but it may be missing.
268 :
269 :			If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
270 :			time the version tool is loaded (which is the first time you say CMB.make),
271 :			then the tool will increment the value stored in config/release every
272 :			time CMB.make is invoked.
273 :
274 :			The binfile format is now insensitive to anything beyond the first
275 :			two components of a version number, so bumping the release does not render
276 :			binfiles incompatible. Auto-bumping can be used to keep track of versions
277 :			during development without invalidating existing binfiles.
278 :
279 :			In any case, every CMB.make updates the date information in version.sml.
280 :			(This is the date that is printed in the banner.)
281 :
282 :			----------------------------------------------------------------------
283 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
284 :	mblume	1634	Date: 2004/09/28 10:53:00 CDT
285 :			Tag: blume-20040928-controls
286 :			Description:
287 :
288 :			Some cleanup of the controls code.
289 :
290 :			----------------------------------------------------------------------
291 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
292 :	mblume	1633	Date: 2004/09/27 22:08:00 CDT
293 :			Tag: blume-20040927-controls
294 :			Description:
295 :
296 :			Added two pieces of functionality to the Controls interface:
297 :
298 :			1. val save'restore: 'a control -> unit -> unit
299 :
300 :			grabs the current value of the control in stage 1 and restores it
301 :			in stage 2.
302 :
303 :			2. val set' : 'a control * 'a -> unit -> unit
304 :
305 :			stores the given value into the control in stage 2 (i.e., delayed)
306 :			but does all error checking in stage 1.
307 :			(This is for string controls that need to do parse their argument --
308 :			something that might fail. In some cases, notably in CM, one
309 :			already knows the intended argument but wants to delay the actual
310 :			assignment until a time when error recovery would be more difficult.)
311 :
312 :			Changed the handling of controls in tool arguments to classes "sml" and
313 :			"lazysml":
314 :			- use Controls.save'restore as a more robust way of restoring the
315 :			old value (in particular: without having to re-parse the string)
316 :			- use controls to handle the "overload" keyword in the init group
317 :			(I believe this change actually fixes a long-standing obscure bug.)
318 :
319 :			----------------------------------------------------------------------
320 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
321 :	mblume	1632	Date: 2004/09/27 17:00:00 CDT
322 :			Tag: blume-20040927-lazysml
323 :			Description:
324 :
325 :			Added a new tool class called "lazysml" to CM's tool chest. The only
326 :			difference to "sml" is that compilation is done with Control.lazysml
327 :			set to true. A source of class "lazysml" is automatically recognized
328 :			by a file name suffix of ".lml".
329 :
330 :			In addition to the above feature, the original class "sml" now also
331 :			supports a tool argument "lazy" which has the same effect. As a
332 :			result, the following three lines are equivalent:
333 :
334 :			foo.sml : lazysml
335 :			foo.sml : sml (lazy)
336 :			foo.sml (lazy)
337 :
338 :			The setting goes into effect both during parsing and during
339 :			compilation. The original setting is restored right after parsing and
340 :			after compilation, respectively.
341 :
342 :			In addition to all the above, there is also a general mechanism to set
343 :			ANY of the "controls" that are available at the command line via
344 :			"-C..." on a per-sml-file basis. The same rules that apply for "lazy"
345 :			apply as well. (In fact, "lazy" is implemented as a special case of
346 :			the general mechanism.)
347 :
348 :			The .cm file syntax uses a new keyword tool argument called "with".
349 :			There are several ways of indicating the desired settings:
350 :
351 :			foo.sml (with:parser.quotations=true)
352 :			foo.sml (with:(name:parser.quotations value:true))
353 :			foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
354 :			foo.sml (with:(name1=value1 name2=value2 ...))
355 :			foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
356 :
357 :			etc.
358 :
359 :			Another possible abbreviation is to leave out the =v or value:v part
360 :			if the name refers to a boolean control (in which case the value is
361 :			taken to be true). Thus, one could get lazy sml also by saying:
362 :
363 :			foo.sml (with:parser.lazy-keyword=true)
364 :			foo.sml (with:parser.lazy-keyword)
365 :			foo.sml (with:(name:parser.lazy-keyword value:true))
366 :			foo.sml (with:(name:parser.lazy-keyword))
367 :
368 :			----------------------------------------------------------------------
369 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
370 :	mblume	1631	Date: 2004/09/24 16:22:00 CDT
371 :			Tag: blume-20040924-ppc-long-branch
372 :			Description:
373 :
374 :			Turned message about "emiting long form of branch" off by default.
375 :			Added a control flag to turn it back on when desired.
376 :
377 :			----------------------------------------------------------------------
378 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
379 :	mblume	1630	Date: 2004/09/24 16:05:00 CDT
380 :			Tag: blume-20040924-rounding
381 :			Description:
382 :
383 :			Applied patch for setting rounding modes under Mac OS X. Thanks to
384 :			Melissa O'Neill for providing the code!
385 :
386 :			----------------------------------------------------------------------
387 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
388 :	mblume	1629	Date: 2004/09/23 17:30:00 CDT
389 :			Tag: blume-20040923-envvars
390 :			Description:
391 :
392 :			1. Changed definition of type ControlRegistry.registry_tree to
393 :			include control_info (i.e., the name of the controlling environment
394 :			variable).
395 :
396 :			2. Added command-line flags -e and -E to print the names of environment
397 :			variables that can be used to control internal settings.
398 :			(This uses the new API mentioned in 1.)
399 :
400 :			----------------------------------------------------------------------
401 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
402 :	mblume	1622	Date: 2004/09/13 16:50:00 CDT
403 :			Tag: Release_110_49
404 :			Description:
405 :
406 :			New working version (110.49). NEW BOOTFILES!
407 :
408 :			----------------------------------------------------------------------
409 :	mblume	1621	Name: Matthias Blume
410 :			Date: 2004/09/13 16:20:00 CDT
411 :			Tag: blume-20040913-config-mlrisc
412 :			Description:
413 :
414 :			Put target "mlrisc" back into the default list.
415 :			(There is no harm in having it, and some users have expressed their
416 :			wish to have "mlrisc" included by default.)
417 :
418 :			----------------------------------------------------------------------
419 :	jhr	1620	Name: John Reppy
420 :			Date: 2004/09/13
421 :			Tag: jhr-20040913-signals
422 :			Description:
423 :
424 :			Fixed the signal masking code to properly nest mask/unmask operations
425 :			on a per-signal basis.
426 :
427 :			----------------------------------------------------------------------
428 :	mblume	1619	Name: Matthias Blume (blume (at) tti - c (dot) org)
429 :			Date: 2004/09/08 13:20:00 CDT
430 :			Tag: blume-20040908-heap-magic
431 :			Description:
432 :
433 :			Bumped the heap macig number to 0x09082004 to account for the changed
434 :			layout of the ML frame under MacOS X.
435 :
436 :			----------------------------------------------------------------------
437 :	allenleung	1618	Name: Allen Leung (leunga (at) reservoir (dot) com)
438 :			Date: 2004/09/03 11:26:00 EST
439 :			Tag: leunga-20040903-cygwin-install
440 :			Description:
441 :
442 :			Added a patch to _arch-n-opsys to enable the Cygwin runtime.
443 :			The Cygwin runtime is turned on by setting the environment
444 :			variable SMLNJ_CYGWIN_RUNTIME to 1.
445 :
446 :			----------------------------------------------------------------------
447 :	mblume	1615	Name: Matthias Blume (blume (at) tti - c (dot) org)
448 :	mblume	1617	Date: 2004/08/31 17:15:00 CDT
449 :			Tag: blume-20040831-core
450 :			Description:
451 :
452 :			Added some exports to src/compiler/core.cm upon request by J. Joemann.
453 :
454 :			----------------------------------------------------------------------
455 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
456 :	mblume	1616	Date: 2004/08/30 17:55:00 CDT
457 :			Tag: blume-20040830-installer
458 :			Description:
459 :
460 :			Upon request by Johannes Joemann:
461 :
462 :			- improved ML code of installer to fall back to coping when renaming
463 :			fails (i.e., when source and target are on different file systems);
464 :			the code compiles but has yet to be tested in anger
465 :			- removed mlrisc from list of default targets (config/targets)
466 :
467 :			----------------------------------------------------------------------
468 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
469 :	mblume	1615	Date: 2004/08/27 17:20:00 CDT
470 :			Tag: blume-20040827-ptreql
471 :			Description:
472 :
473 :			Added ptreql primop to structure InlineT (upon request from Larry
474 :			Paulson).
475 :
476 :			----------------------------------------------------------------------
477 :	allenleung	1613	Name: Allen Leung (leunga (at) reservoir (dot) com)
478 :	allenleung	1614	Date: 2004/08/15 21:21:00 EST
479 :			Tag: leunga-110_48-udgraph
480 :			Description:
481 :
482 :			Another bug fix from Carl Hauser:
483 :
484 :			diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
485 :			> 48c48
486 :			> < | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
487 :			> e::L)
488 :			> ---
489 :			> > | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
490 :			> Without this, any deletion of an edge in an undirected graph does severe
491 :			> violence to the graph.
492 :
493 :			----------------------------------------------------------------------
494 :			Name: Allen Leung (leunga (at) reservoir (dot) com)
495 :	allenleung	1613	Date: 2004/08/10 23:35:00 EST
496 :			Tag: leunga-110_48-ppc
497 :			Description:
498 :
499 :			The IBM/MacOS syntax switch on PPC was incorrectly swapped.
500 :
501 :			----------------------------------------------------------------------
502 :	mblume	1602	Name: Matthias Blume (blume (at) tti - c (dot) org)
503 :	mblume	1615	Date: 2004/08/10 12:00:00 CDT
504 :	mblume	1602	Tag: Release_110_48
505 :			Description:
506 :
507 :			New working version (110.48). NEW BOOTFILES!
508 :
509 :			----------------------------------------------------------------------
510 :	allenleung	1599	Name: Allen Leung (leunga (at) reservoir (dot) com)
511 :			Date: 2004/08/09 12:21:00 EST
512 :			Tag: leunga-110_47-dijsktra
513 :			Description:
514 :
515 :			Bug fix from Carl Hauser:
516 :
517 :			single_source_shortest_paths in dijkstra.sml was observed to get wrong
518 :			answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
519 :
520 :			The problem is that following the expression A.update(dist,s,Num.zero)
521 :			it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
522 :
523 :			----------------------------------------------------------------------
524 :	mblume	1595	Name: Matthias Blume (blume (at) tti - c (dot) org)
525 :			Date: 2004/08/06 18:10:00 CDT
526 :			Tag: blume-20040806-cmdline
527 :			Description:
528 :
529 :			Fiddled with handling of command-line options:
530 :
531 :			* sml now quits after processing the command line
532 :			if -H, -S, -h<n>, or -s<n> appears as the last
533 :			command-line argument
534 :			* a new option -q terminates the session when encountered on
535 :			the command line; subsequent arguments will be ignored
536 :			* bug fixes: short (erroneous) arguments are no longer ignored
537 :			completely
538 :
539 :			----------------------------------------------------------------------
540 :	allenleung	1593	Name: Allen Leung (leunga (at) reservoir (dot) com)
541 :			Date: 2004/08/04 18:17:00 EST
542 :			Tag: leunga-110_47-ppc-ibm-asm
543 :			Description:
544 :
545 :			- Added minimal IBM assembly syntax support for PowerPC.
546 :
547 :			- Cygwin: manually changed the file cygwin.def. Some exported symbols have
548 :			been altered in the runtime. We need an automatic way to keep the file
549 :			in sync.
550 :
551 :			----------------------------------------------------------------------
552 :	mblume	1548	Name: Matthias Blume (blume (at) tti - c (dot) org)
553 :	mblume	1581	Date: 2004/08/04 14:00:00 CDT
554 :			Tag: Release_110_47
555 :			Description:
556 :
557 :			New working version (110.47). NEW BOOTFILES!
558 :
559 :			----------------------------------------------------------------------
560 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
561 :	mblume	1580	Date: 2004/08/03 14:25:00 CDT
562 :			Tag: blume-20040803-callingconv
563 :			Description:
564 :
565 :			Added low-level support for choosing C calling conventions by
566 :			twiddling the type of rawccall. (See
567 :			src/compiler/Semant/types/cproto.sml for details.)
568 :
569 :			----------------------------------------------------------------------
570 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
571 :	mblume	1579	Date: 2004/08/02 15:55:00 CDT
572 :			Tag: blume-20040802-backout
573 :			Description:
574 :
575 :			Backed out of change to win32-filesys.c. The earlier patch to
576 :			get_file_time caused CM to produce files with the wrong time stamp.
577 :
578 :			----------------------------------------------------------------------
579 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
580 :	mblume	1573	Date: 2004/08/02 14:45:00 CDT
581 :			Tag: blume-20040802-nlffi-win32
582 :			Description:
583 :
584 :			Added NLFFI support for Win32, adapted from a patch provided by David
585 :			Hansel. This is currently completely untested. Also, the issue
586 :			concerning stdcall vs. ccall is still unresolved.
587 :
588 :			----------------------------------------------------------------------
589 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
590 :	mblume	1572	Date: 2004/07/30 17:55:00 CDT
591 :			Tag: blume-20040730-various
592 :			Description:
593 :
594 :			Gearing up towards 110.47...
595 :
596 :			- various minor bugfixes to ml-nlffigen
597 :			- a beginning of a manual for nlffi
598 :
599 :			- eliminated 'export name=value' in config/install.sh as this does
600 :			not work with certain versions of /bin/sh
601 :			(Thanks to David King at Motorola for catching this.)
602 :
603 :			- several bugfixes provided or suggested by David Hansel at Reactive Systems:
604 :			- added a test for tm==NULL to gmtime.c and localtime.c
605 :			- applied patch for incorrect GetFileTime under win32
606 :			- toSeconds -> toMilliseconds in Win32/win32-process.sml
607 :
608 :			----------------------------------------------------------------------
609 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
610 :	mblume	1557	Date: 2004/07/21 18:20:00 CDT
611 :			Tag: blume-20040721-nlffigen
612 :			Description:
613 :
614 :			- Fixed minor issue in ml-nlffigen: Now generate structure T_foo
615 :			for a typedef to an incomplete type, but leave out the "typ" member.
616 :			(This is just for consistency.)
617 :			- Started to produce what is supposed to become better (i.e., comprehensive)
618 :			documentation of what ml-nlffigen does and produces.
619 :
620 :			----------------------------------------------------------------------
621 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
622 :	mblume	1548	Date: 2004/07/14 16:25:00 CDT
623 :			Tag: blume-20040714-union
624 :			Description:
625 :
626 :			Added C_UNION to c-calls/c-types.sml and updated the machinery
627 :			(ml-nlffigen, cproto.sml) that conveys C function interface
628 :			information to the code generator.
629 :
630 :			However, the actual architecture-specific implementation of function
631 :			arguments and results that are C unions is still not implemented.
632 :
633 :			----------------------------------------------------------------------
634 :	allenleung	1539	Name: Allen Leung (leunga (at) reservoir (dot) com)
635 :	allenleung	1546	Date: 2004/07/14 14:38:00 EST
636 :			Tag: leunga-110_46_1-ppc-lwzu
637 :			Description:
638 :
639 :			Added these instructions to the PowerPC architecture:
640 :			LBZU(X), LHZU(X), LWZU(X),
641 :			STWU(X), STFDU, STFSU
642 :
643 :			etc...
644 :
645 :			Note: I haven't added their instruction encoding into the description.
646 :
647 :			----------------------------------------------------------------------
648 :			Name: Allen Leung (leunga (at) reservoir (dot) com)
649 :	allenleung	1539	Date: 2004/07/13 15:04:00 EST
650 :			Tag: leunga-110_46_1-ppc-lwarx
651 :			Description:
652 :
653 :			Added the two instructions LWARX and STWCX to the PowerPC
654 :			instruction set.
655 :
656 :			A (untested) rewrite of loop-structure.sml. The old version
657 :			is completely broken.
658 :
659 :			----------------------------------------------------------------------
660 :	mblume	1479	Name: Matthias Blume (blume (at) tti - c (dot) org)
661 :	mblume	1537	Date: 2004/07/13 13:50:00 CDT
662 :			Tag: blume-20040713-nlffi
663 :			Description:
664 :
665 :			- use paramAlloc to report c-calls with too many arguments
666 :			(for PPC version where parameter area is pre-allocated)
667 :			- added ccall_maxargspace to machspec (to implement the above)
668 :			- made "make" commend in CM's "make" tool configurable
669 :			- added option (default: on) for passing the name of the SML/NJ's "bin"
670 :			directory to "make"; the call looks like this:
671 :
672 :			make <options> SMLNJ_BINDIR=<dir> <target>
673 :
674 :			This can be used by the Makefile to, e.g., pick the "right" version
675 :			of ml-nlffigen.
676 :			- minor code tweaks
677 :
678 :			----------------------------------------------------------------------
679 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
680 :	mblume	1532	Date: 2004/07/12 22:50:00 CDT
681 :			Tag: blume-110_46_1-macosx-nlffi
682 :			Description:
683 :
684 :			NLFFI under Mac OS X now working (sort of). This is largely untested,
685 :			though.
686 :
687 :			Note: 1. You have to make a new, clean build of the runtime system.
688 :			2. There are new BOOTFILES, you have to use them!
689 :			(Doing the bootstrap process yourself would be *very* painful!
690 :			If you absolutely have to do it, build the system under
691 :			a different architecture and then cross-compile.)
692 :
693 :			Version bumped to 110.46.1 to account for runtime data format changes.
694 :
695 :			----------------------------------------------------------------------
696 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
697 :	mblume	1515	Date: 2004/06/18 14:30:00 CDT
698 :			Tag: blume-20040618-unix
699 :			Description:
700 :
701 :			Changed the implementation of structure Unix so that the same stream
702 :			is returned every time one of the {text,bin}{In,Out}streamOf functions
703 :			is invoked on the same proc. This is not what the spec currently
704 :			says -- although IMO it arguably should. (See discussion below.)
705 :
706 :			----------------------------------------------------------------------
707 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
708 :	mblume	1507	Date: 2004/06/17 18:15:00 CDT
709 :			Tag: Release_110_46
710 :			Description:
711 :
712 :			New working version (110.46). NEW BOOTFILES!
713 :
714 :			----------------------------------------------------------------------
715 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
716 :	mblume	1506	Date: 2004/06/17 17:20:00 CDT
717 :			Tag: blume-20040617-timer-unix
718 :			Description:
719 :
720 :			Changed the interface of structures Timer and Unix to match the most
721 :			recent Basis spec.
722 :
723 :			In the case of Unix there still seems to be an open/weird issue:
724 :
725 :			The {text,bin}{In,Out}streamOf functions are supposed to create
726 :			fresh streams whenever they are called -- as opposed to have them
727 :			return the same stream every time. This design is supposed to
728 :			prevent space leaks caused by proc values hanging on to streams.
729 :
730 :			The reap function, on the other hand, is supposed to close the
731 :			streams. This cannot be done without having a handle on the
732 :			stream in proc after all...
733 :
734 :			I took the liberty to implement the following stopgap solution:
735 :
736 :			The proc value hangs on to the most recently created stream(s).
737 :			Reap closes those. If either or both of the two streams hadn't
738 :			been created at all yet, then reap will close the corresponding
739 :			file descriptors directly.
740 :
741 :			PS: I don't understand the original space leak argument anymore. If
742 :			a proc hangs on to the imperative stream, then I/O operations on those
743 :			will advance the state of the cached stream and avoid the space leak.
744 :
745 :			----------------------------------------------------------------------
746 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
747 :	mblume	1500	Date: 2004/05/28 16:45:00 CDT
748 :			Tag: blume-20040528-basis
749 :			Description:
750 :
751 :			Added signature PACK_REAL and exported functor PrimIO.
752 :
753 :			----------------------------------------------------------------------
754 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
755 :	mblume	1498	Date: 2004/05/25 16:00:00 CDT
756 :			Tag: blume-20040525-group-owner
757 :			Description:
758 :
759 :			CM now ignores (but still accepts) the "owner" information in group
760 :			descriptions. The owner of a group is its next enclosing
761 :			library. Each group must have a unique owner. (There is a virtual
762 :	mblume	1505	"toplevel" library that owns groups which are not nested within a real
763 :	mblume	1498	library.) Previously, each group had to explicitly declare its owner,
764 :			and CM would check that such a declaration is correct. The new scheme
765 :			is to have CM check that for each group there is precisely one owning
766 :			library.
767 :
768 :			The advantage of the new scheme is that the programmer no longer needs
769 :			to maintain the somewhat annoying owner information. The downside is
770 :			that CM cannot enforce the ownership rule across multiple runs of
771 :			CM.make. Fortunately, enclosing the same group in two different
772 :			libraries A and B which are not part of the same program does not
773 :			cause real problems.
774 :
775 :			----------------------------------------------------------------------
776 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
777 :	mblume	1496	Date: 2004/05/20 16:00:00 CDT
778 :			Tag: blume-20040520-win32
779 :			Description:
780 :
781 :			Made the win32 version work again. (Strangely, a misplaced comma had
782 :			slipped into win32-process.c which prevented the runtime from being
783 :			compiled correctly.)
784 :
785 :			Also, included a minor addition to ml-build.bat analogous to what was
786 :			done in blume-20040519-ml-build.
787 :
788 :			----------------------------------------------------------------------
789 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
790 :	mblume	1494	Date: 2004/05/19 22:10:00 CDT
791 :			Tag: blume-20040519-ml-build
792 :			Description:
793 :
794 :			Arranged for ml-build to clean up after itself a little bit better.
795 :			The script generates a temporary SML source file and compiles it using
796 :			CM, so CM generates metadata (GUID, SKEL, objectfile) for it. It now
797 :			gets rid of those at the end, so they don't accumulate under .cm.
798 :
799 :			This required a minor change to install.sh because the name of the
800 :			metadata directory (default: .cm) is actually configurable at
801 :			installation time.
802 :
803 :			----------------------------------------------------------------------
804 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
805 :	mblume	1491	Date: 2004/05/18 15:50:00 CDT
806 :			Tag: blume-20040518-mkreader
807 :			Description:
808 :
809 :			Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
810 :			implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
811 :
812 :			----------------------------------------------------------------------
813 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
814 :	mblume	1489	Date: 2004/05/11 14:35:00 CDT
815 :			Tag: blume-20040511-win32sock
816 :			Description:
817 :
818 :			Added previously missing support for many socket-related functions
819 :			under win32. Thanks to David Hansel <hansel@reactive-systems.com>
820 :			for the voluminous patch!
821 :
822 :			(I have not tested this patch under win32 yet.)
823 :
824 :	mblume	1490	Here is David's e-mail:
825 :
826 :			Hi,
827 :
828 :			Attached to this email you find a diff against sml/nj 110.45
829 :			that will enable socket support under Windows.
830 :
831 :			To apply the patch (using unix or cygwin)
832 :			1) gunzip runtime.diff.gz
833 :			2) "cd" into "src/runtime" in the source tree of a fresh
834 :			110.45 installation.
835 :			3) patch -p 1 < [your/path/to]runtime.diff
836 :
837 :			The code compiles fine but has NOT yet been extensively tested.
838 :			I only ran a few tests for basic socket client functionality
839 :			(which worked fine). Especially the functions that use ioctl
840 :			are not tested at all and might not work (see below).
841 :
842 :			I implemented this since we want to move to a newer version of sml/nj
843 :			but need socket support in order to use it. This is the first time I
844 :			even had a look at the sml/nj source, so please review my changes
845 :			before making this part of the distribution! Here are a few issues
846 :			that I think might be better for someone to solve who is more
847 :			familiar with the sml/nj source (and socket programming):
848 :
849 :			- getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
850 :			exception since I could not figure out what the windows equivalent
851 :			of these functions is
852 :
853 :			- In sockets-osdep.h there are a some #include statements that are
854 :			only used in a few files that include sockets-osdep.h
855 :
856 :			- In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
857 :			does not process its return value since I don't know how to
858 :			report an error upwards.
859 :
860 :			- It would probably be good to have a call to WSACleanup() when
861 :			the library is unloaded (if there is such a possibility).
862 :			Otherwise I think Windows will take care of this automatically
863 :			when the process finishes.
864 :
865 :			- I used ioctlsocket() as a replacement for ioctl() but I have
866 :			no idea if that is actually the proper replacement on Windows.
867 :
868 :			- All these issues are marked in the code by "FIXME" comments.
869 :
870 :			We use sml/nj extensively in our products and are quite happy
871 :			with it. I hope this contribution will help you.
872 :
873 :			Keep up the good work!
874 :
875 :			David
876 :
877 :	mblume	1489	----------------------------------------------------------------------
878 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
879 :	mblume	1488	Date: 2004/05/11 14:20:00 CDT
880 :			Tag: blume-20040511-installml
881 :			Description:
882 :
883 :			Fixed two bugs in installml script. (Thanks to Vesa A. Norrman for
884 :			the patch.)
885 :
886 :			----------------------------------------------------------------------
887 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
888 :	mblume	1487	Date: 2004/05/11 14:05:00 CDT
889 :			Tag: blume-20040511-nlffi-netbsd
890 :			Description:
891 :
892 :			Added support for nlffi under netbsd. (Thanks to Vesa A. Norrman for
893 :			the patch.)
894 :
895 :			----------------------------------------------------------------------
896 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
897 :	mblume	1486	Date: 2004/05/11 12:05:00 CDT
898 :			Tag: blume-20040511-exports
899 :			Description:
900 :
901 :			As per request by Adam Chlipala <adam@hcoop.net>, extended various
902 :			export lists in compiler-related .cm-files.
903 :
904 :			----------------------------------------------------------------------
905 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
906 :	mblume	1485	Date: 2004/05/11 11:35:00 CDT
907 :			Tag: blume-20040511-allsource
908 :			Description:
909 :
910 :			The installer now honors the "src-smlnj" target again, although its meaning
911 :			has changed from "all sources required for the compiler" to "all sources
912 :			the installer knows about". In other words, if you enable "src-smlnj"
913 :			in the "targets" file, then the installer will pull in sources for
914 :			everything. (Notice that this refers to source code only. Compiled
915 :			code is still only installed for modules that were requested explicitly
916 :			or which are required for other modules that were requested explicitly.)
917 :
918 :			----------------------------------------------------------------------
919 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
920 :	mblume	1482	Date: 2004/04/23 17:40:00 CDT
921 :			Tag: blume-20040423-ieee-scan
922 :			Description:
923 :
924 :			Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
925 :			in the exponent calculation we get INF or ZERO (depending on the mantissa
926 :			and the sign of the exponent).
927 :
928 :			----------------------------------------------------------------------
929 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
930 :	mblume	1480	Date: 2004/04/23 10:40:00 CDT
931 :			Tag: blume-20040423-ml-build
932 :			Description:
933 :
934 :			The ml-build script now terminates with a non-0 status when something
935 :			goes wrong.
936 :
937 :			----------------------------------------------------------------------
938 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
939 :	mblume	1479	Date: 2004/04/22 16:35:00 CDT
940 :			Tag: blume-20040422-Option
941 :			Description:
942 :
943 :			Made exception Option to be the same as exception Option.Option
944 :			(as it should be).
945 :
946 :			----------------------------------------------------------------------
947 :	allenleung	1468	Name: Allen Leung (leunga (at) reservoir (dot) com)
948 :			Date: 2004/03/19 14:40:00 EST
949 :			Tag: leunga-20040319-cygwin-nlffi
950 :			Description:
951 :
952 :			Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
953 :			of SML/NJ. The problem is that
954 :
955 :			lib = dlopen(NULL, ...)
956 :			f = dlsym(lib, "malloc");
957 :
958 :			does not work on Windows unless we explicitly export symbols
959 :			such as 'malloc' during linking. We fixed this by explicitly
960 :			exporting the required symbols with the magic gcc incantation:
961 :
962 :			-Wl,--export-all cygwin.def
963 :
964 :			where cygwin.def is a file containing all the symbols that
965 :			we wish to export.
966 :
967 :			I suspect this is a Windows problem and we'll have to
968 :			do the same (somehow with windows compilers) when
969 :			we build the native win32 version with the system
970 :			calls LoadLibrary/GetProcAddress.
971 :
972 :			----------------------------------------------------------------------
973 :	mblume	1384	Name: Matthias Blume (blume (at) tti - c (dot) org)
974 :	mblume	1465	Date: 2004/03/04 16:35:00 CST
975 :			Tag: blume-20040304-intinf-fmt
976 :			Description:
977 :
978 :			Fixed problem with IntInf.fmt (sign would show up on the right instead
979 :			of on the left for BIN, OCT, and HEX).
980 :
981 :			----------------------------------------------------------------------
982 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
983 :	mblume	1464	Date: 2004/03/04 11:25:00 CST
984 :			Tag: blume-20040304-symlinks
985 :			Description:
986 :
987 :			Fixed problem with installer script (unix only) where bin/ml-yacc and
988 :			friends pointed (via symlinks) to absolute locations instead of just
989 :			.run-sml. This was reported by Vesa A Norrman.
990 :
991 :			----------------------------------------------------------------------
992 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
993 :	mblume	1455	Date: 2004/02/13 14:50:00 CST
994 :			Tag: Release_110_45
995 :			Description:
996 :
997 :			New working version (110.45). New bootfiles.
998 :
999 :			----------------------------------------------------------------------
1000 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1001 :	mblume	1448	Date: 2004/01/26 15:15:15 CST
1002 :			Tag: blume-20040126-toplevel
1003 :			Description:
1004 :
1005 :			Improved handling of exceptions at the interactive toplevel.
1006 :
1007 :			----------------------------------------------------------------------
1008 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1009 :			Date: 2004/01/26 11:25:00
1010 :	mblume	1446	Tag: blume-20040126-app
1011 :			Description:
1012 :
1013 :			Type of top-level "app" corrected.
1014 :	mblume	1447	Added code for setting vp_limitPtrMask to Win32-specific runtime.
1015 :	mblume	1446
1016 :			----------------------------------------------------------------------
1017 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1018 :	mblume	1427	Date: 2003/11/18 17:10 CST
1019 :			Tag: blume-20031118-basis-fiddle
1020 :			Description:
1021 :
1022 :			- changed Timer interface to what might become the spec
1023 :			- POSIX_FLAGS -> BIT_FLAGS according to spec
1024 :			- some other minor discrepancies wrt. spec eliminated
1025 :
1026 :			----------------------------------------------------------------------
1027 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1028 :	mblume	1414	Date: 2003/11/06 12:00:00 CST
1029 :			Tag: Release_110_44
1030 :			Description:
1031 :
1032 :			New working version (110.44). New bootfiles.
1033 :
1034 :			----------------------------------------------------------------------
1035 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1036 :	mblume	1411	Date: 2003/11/04 11:50:00 CST
1037 :			Tag: blume-20031104-move-libraries
1038 :			Description:
1039 :
1040 :			Eliminated the "dont_move_libraries" directive in config/targets.
1041 :			(The mechanism was broken and could not be fixed easily. Moreover,
1042 :			there does not seem to be any reason not to move all libraries into
1043 :			lib during installation. I originally implemented this directive as a
1044 :			backward-compatibility feature when I first introduced the new CM.
1045 :			Now that things have been stable for a long time and going back to the
1046 :			old CM is not an option, there is no reason to keep it around.)
1047 :
1048 :			----------------------------------------------------------------------
1049 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1050 :	mblume	1409	Date: 2003/11/03 16:00:00 CST
1051 :			Tag: blume-20031103-installdir
1052 :			Description:
1053 :
1054 :			Made installer honor INSTALLDIR variable again. (Thanks to Chris
1055 :			Richards for pointing out the problem and providing the solution.)
1056 :
1057 :			----------------------------------------------------------------------
1058 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1059 :	mblume	1400	Date: 2003/10/01 17:05:00 CDT
1060 :			Tag: blume-20031001-lal-mlrisc
1061 :			Description:
1062 :
1063 :			MLRISC bug fix from Lal.
1064 :
1065 :			----------------------------------------------------------------------
1066 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1067 :	mblume	1399	Date: 2003/09/30 16:10:00 CDT
1068 :			Tag: blume-20030930-primio-bat
1069 :	mblume	1398	Description:
1070 :
1071 :			1. Added openVector, nullRd, and nullWr to PRIM_IO.
1072 :			2. Improved .bat files (for Win32 port) to make things work under Win95.
1073 :			(thanks to Aaron S. Hawley for this one)
1074 :
1075 :			----------------------------------------------------------------------
1076 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1077 :	mblume	1395	Date: 2003/09/26 16:05:00 CDT
1078 :			Tag: blume-20030926-wrappriv
1079 :			Description:
1080 :
1081 :			Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1082 :
1083 :			----------------------------------------------------------------------
1084 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1085 :	mblume	1394	Date: 2003/09/26 15:00:00 CDT
1086 :			Tag: blume-20030926-110_43_3
1087 :			Description:
1088 :
1089 :			- additional cleanup
1090 :			- version number bump, NEW BOOTFILES
1091 :
1092 :			----------------------------------------------------------------------
1093 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1094 :	mblume	1393	Date: 2003/09/26 12:00:00 CDT
1095 :			Tag: blume-20030926-ppautoload
1096 :			Description:
1097 :
1098 :			I modified the read-eval-print loop so that the autoloader gets
1099 :			invoked whenever the prettyprinter tries to look up a symbol that
1100 :			is not currently defined in the toplevel environment but which
1101 :			appears in CM's autoload registry. As a result, we see far fewer of
1102 :			those ?.Foo.Bar.xxx names in the prettyprinter's output.
1103 :
1104 :			In addition to this I tried to clean up some pieces of the Basis
1105 :			implementation (e.g., Socket, Word8Array) in order to prevent other
1106 :			instances of these ?.Foo.Bar.xxx names from being printed.
1107 :
1108 :			The mechanism that picks names for types still needs some work, though.
1109 :			(Right now it seems that if there is a type A.t which is defined to
1110 :	mblume	1394	be B.u, but B is unavailable at toplevel, then A.t gets printed as
1111 :			"?.B.u" although the perhaps more sensible solution would be to use
1112 :	mblume	1393	"A.t" in this case. In other words, the prettyprinter should follow
1113 :			a chain of DEFtycs not farther than there are corresponding toplevel
1114 :			names in the current environment.)
1115 :
1116 :			----------------------------------------------------------------------
1117 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1118 :	mblume	1391	Date: 2003/09/24 16:31:00 CDT
1119 :			Tag: blume-20030924-installer
1120 :			Description:
1121 :
1122 :			Another installer tweak: All the ML code for the installer is now
1123 :			compiled during CMB.make and put into a little library called
1124 :			$smlnj/installer.cm. The installation then simply invokes
1125 :
1126 :			sml -m $smlnj/installer.cm
1127 :
1128 :			and everything happens automagically.
1129 :
1130 :			Win32: ML code senses value of environment variable SMLNJ_HOME.
1131 :			Unix: ML code senses values of environment variables ROOT, CONFIGDIR,
1132 :			and BINDIR.
1133 :
1134 :			The new scheme guarantees that the ML code responsible for the installation
1135 :			is in sync with the APIs of the main system. Also, the installer is
1136 :			somewhat faster because the installer script is precompiled.
1137 :
1138 :			----------------------------------------------------------------------
1139 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1140 :	mblume	1390	Date: 2003/09/24 15:35:00 CDT
1141 :			Tag: blume-20030924-synsock
1142 :			Description:
1143 :
1144 :			Added a signature SYNCHRONOUS_SOCKET to basis.cm. This is like SOCKET
1145 :			but excludes all non-blocking operations. Defined SOCKET (in Basis)
1146 :			and CML_SOCKET in terms of SYNCHRONOUS_SOCKET. Removed superfluous
1147 :			implementations of non-blocking operations from CML's Socket
1148 :			structure.
1149 :
1150 :			----------------------------------------------------------------------
1151 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1152 :	mblume	1389	Date: 2003/09/24 15:10:05 CDT
1153 :			Tag: blume-20030924-sockets
1154 :			Description:
1155 :
1156 :			1. Fixed SOCKET API and implementation to match Basis spec.
1157 :			This required changing the internal representation of sockets to one
1158 :			that remembers (for each socket file descriptor) whether it is currently
1159 :			blocking or non-blocking. This state is maintained lazily (i.e., a system
1160 :			call is made only if the state actually needs to change).
1161 :
1162 :			2. OS-specific details of sockets were moved into separate files, thus
1163 :			making it possible to unify the bulk of the socket implementations
1164 :			between Unix and Win32.
1165 :
1166 :			3. CML's socket API changed accordingly.
1167 :			(Note that we need to remove non-blocking functions from this API
1168 :			since they are redundant in the case of CML!)
1169 :
1170 :			4. CML's socket implementation now makes use of non-blocking functions
1171 :			provided by Basis, thus removing all OS-dependent code from this part
1172 :			of CML.
1173 :
1174 :			5. Changed Real64.precision from 52 to 53. Minor cleanup in Real64 code.
1175 :
1176 :			----------------------------------------------------------------------
1177 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1178 :	mblume	1387	Date: 2003/09/22 12:10:00 CDT
1179 :	mblume	1388	Tag: blume-20030922-110_43_2
1180 :	mblume	1387	Description:
1181 :
1182 :			Made a new interim version and bootfiles for developer's bootstrapping
1183 :			convenience.
1184 :
1185 :			110.43.2 -- NEW BOOTFILES
1186 :
1187 :			----------------------------------------------------------------------
1188 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1189 :	mblume	1385	Date: 2003/09/19 15:55:00 CDT
1190 :			Tag: blume-20030919-cmdir
1191 :			Description:
1192 :
1193 :			1. new-install.sh -> install.sh
1194 :			2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1195 :			3. tweaked installer so that another name instead of .cm can be chosen
1196 :			at install time (by setting the CM_DIR_ARC environment variable
1197 :			during installation); once installation is complete, the name is
1198 :			fixed
1199 :
1200 :			----------------------------------------------------------------------
1201 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1202 :	mblume	1384	Date: 2003/09/18 16:00:00 CDT
1203 :			Tag: blume-20030918-110_43_1
1204 :			Description:
1205 :
1206 :			Made a new interim version and bootfiles for developer's bootstrapping
1207 :			convenience.
1208 :
1209 :			110.43.1 -- NEW BOOTFILES
1210 :
1211 :			----------------------------------------------------------------------
1212 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1213 :	mblume	1383	Date: 2003/09/18 15:20:00 CDT
1214 :			Tag: blume-20030918-misc
1215 :			Description:
1216 :
1217 :			1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1218 :			this seems to be controversial at the moment) and used those in
1219 :			Posix.ProcEnv.times.
1220 :
1221 :			2. Added Time.{from,to}Nanoseconds to Time.
1222 :
1223 :			3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1224 :			For example, fromLargeInt never needs to look at more than 3 "big
1225 :			digits" to get its 53 bits of precision.
1226 :
1227 :			----------------------------------------------------------------------
1228 :	mblume	1384	Name: Matthias Blume (blume (at) tti - c (dot) org)
1229 :	mblume	1381	Date: 2003/09/17 16:30:00 CDT
1230 :			Tag: blume-20030917-real32-slices
1231 :			Description:
1232 :
1233 :			Added an entry to the primitive environment
1234 :			(compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1235 :			added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1236 :
1237 :			Removed some of the "magic" constants in real64.sml and replaced them
1238 :			with code that generates these values from their corresponding
1239 :			integer counterparts.
1240 :
1241 :			Made all(?) the slice-related changes to the Basis and made everything
1242 :			compile again...
1243 :
1244 :			----------------------------------------------------------------------
1245 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1246 :	mblume	1375	Date: 2003/09/15 17:45:00 CDT
1247 :			Tag: blume-20030915-rbase
1248 :			Description:
1249 :
1250 :			Fixed bug in Real.fromLargeInt.
1251 :
1252 :			----------------------------------------------------------------------
1253 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1254 :	mblume	1369	Date: 2003/09/13 18:11:00 CDT
1255 :			Tag: blume-20030913-libinstall
1256 :			Description:
1257 :
1258 :			Minor bugfix in config/libinstall (set anchor with path to
1259 :			standalone tool after installing it, otherwise libraries that
1260 :			need ml-lex or ml-yacc won't compile the first time the installer
1261 :			runs).
1262 :
1263 :			----------------------------------------------------------------------
1264 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1265 :	mblume	1368	Date: 2003/09/12 11:45:00 CDT
1266 :			Tag: blume-20030912-various
1267 :			Description:
1268 :
1269 :			- fixed bug in Real.toLargeInt
1270 :			- fixed bug in Posix.ProcEnv.times
1271 :			- changed inputLine functions to return an option
1272 :			- minor installer improvements / bugfixes
1273 :			- changed default @SMLalloc parameter for x86/celeron to 64k
1274 :
1275 :			----------------------------------------------------------------------
1276 :			Name: Matthias Blume (blume (at) tti - c (dot) org)
1277 :	mblume	1356	Date: 2003/09/09 22:00:00 CDT
1278 :			Tag: Release_110_43
1279 :			Description:
1280 :
1281 :			New working release 110.43. New bootfiles.
1282 :
1283 :			----------------------------------------------------------------------
1284 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1285 :	mblume	1354	Date: 2003/09/09 19:20:00 CDT
1286 :			Tag: blume-20030909-installer
1287 :			Description:
1288 :
1289 :			Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1290 :			Modified config/install.bat to take advantage of it. Also modified
1291 :			config/install.sh (and called it config/new-install.sh) to take advantage
1292 :			of it on Unix systems. (The SML code is (supposed to be) platform-
1293 :			independent.)
1294 :
1295 :			The installer can now install everything under Win32
1296 :			as well as under *nix as long as it compiles.
1297 :
1298 :			Other changes:
1299 :
1300 :			- made CML compile again under Win32
1301 :			- made eXene compile under Win32 (by providing a fake structure UnixSock
1302 :			and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1303 :			- fixed a bug in nowhere: it assumed that type OS.Process.status is the
1304 :			same as type int; under Win32 it isn't
1305 :			- fixed some slice-related problems in the win32-specific parts of CML
1306 :			- added a functor argument "sameVol" to os-path-fn.sml in the Basis
1307 :			(under Win32, the volume name is case-insensitive, and the
1308 :			OS.Path code compares volume names for equality)
1309 :
1310 :			----------------------------------------------------------------------
1311 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1312 :	mblume	1353	Date: 2003/09/08 11:55:00 CDT
1313 :			Tag: blume-20030908-fullpath
1314 :			Description:
1315 :
1316 :			Made Win32 version of OS.FileSys.fullPath return current directory
1317 :			when given an empty string. This is what the spec says, and incidentally,
1318 :			CM depends on it. (CM otherwise goes into an infinite loop in certain
1319 :			cases when presented with the name of a non-existing .cm file.)
1320 :
1321 :			----------------------------------------------------------------------
1322 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1323 :	mblume	1350	Date: 2003/09/04 16:30:00 CDT
1324 :	mblume	1352	Tag: blume-20030905-slices-etc
1325 :	mblume	1350	Description:
1326 :
1327 :			1. Changed interface to vectors and arrays in Basis to match
1328 :			(draft) Basis spec.
1329 :			2. Added signatures and implementations of slices according to
1330 :			Basis spec.
1331 :			3. Edited source code throughout the system to make it compile again
1332 :			under 1. and 2. (In some cases code had to be added to have it
1333 :			match the new signatures.)
1334 :			4. MLRISC should be backward-compatible: the copies of the originals
1335 :			of files that needed to change under 3. were retained, the .cm files
1336 :			check the compiler version number and use old versions when
1337 :			appropriate.
1338 :			5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1339 :			dirstream -> string option (in accordance with Basis spec).
1340 :			6. When generating code that counts lines, ml-lex used function
1341 :			CharVector.foldli, taking advantage of its old interface.
1342 :			This has been replaced with the corresponding code from
1343 :			CharVectorSlice. (html-lex must be re-lexed!)
1344 :			7. BitArray in smlnj-lib/Util has been extended/modified to match the
1345 :			new MONO_ARRAY signature. (Do we need BitArraySlice?)
1346 :			8. Removed temporary additions (fromInternal, toInternal) from the
1347 :			(now obsolete) IntInf in smlnj-lib/Util.
1348 :			9. Cleaned up structure Byte.
1349 :			10. Added localOffset, scan, and fromString to Date (according to spec).
1350 :			Cleaned/corrected implementation of Date.
1351 :			(Still need to check for correctness; implement better canonicalizeDate.)
1352 :			11. Added "scan" to signature IEEE_REAL.
1353 :			12. Some improvements to IntInf [in particular: efficiency-hack for
1354 :			mod and rem when second operand is 2 (for parity checks).]
1355 :			13. Changed representation of type Time.time, using a single IntInf.int
1356 :			value counting microseconds. This considerably simplified the
1357 :			implementation of structure Time. We now support negative time
1358 :			values; scan and fromString handle signs.
1359 :			14. Functor PrimIO now takes two additional arguments (VectorSlice and
1360 :			ArraySlice).
1361 :
1362 :			----------------------------------------------------------------------
1363 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1364 :	mblume	1347	Date: 2003/08/28 17:00:00 CDT
1365 :			Tag: blume-20030828-intinf
1366 :			Description:
1367 :
1368 :			This is a major update which comes with a version number bump
1369 :			(110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1370 :			and an implementation of IntInf in the Basis.
1371 :
1372 :			There are a fairly large number of related changes and updates throughout
1373 :			the system:
1374 :
1375 :			Basis:
1376 :			- Implemented IntInf.
1377 :			- Made LargeInt a projection of IntInf (by filtering through INTEGER).
1378 :	mblume	1348	- Added some missing Real64 operations, most notably Real.toLargeInt.
1379 :	mblume	1347	- Added FixedInt as a synonym for Int32.
1380 :
1381 :			compiler:
1382 :			* Added support for a built-in intinf type.
1383 :			- literals
1384 :			- pattern matching
1385 :			- conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1386 :			- overloading on literals and operations
1387 :
1388 :			This required adding a primitive type intinf, some additional
1389 :			primops, and implementations for several non-trivial intinf
1390 :			operations in Core. (The intinf type is completely abstract
1391 :			to the compiler; all operations get delegated back to the Core.)
1392 :
1393 :			* Intinf equality is handled by polyequal. However, the compiler
1394 :			does not print its usual warning in this case (since polyequal
1395 :			is the right thing to do there).
1396 :
1397 :			* Improved the organization of structure InlineT.
1398 :
1399 :			* A word about conversion primops:
1400 :			If conversions involving intinf do not cancel out during
1401 :	mblume	1348	CPS contract, then the compiler must insert calls to Core functions.
1402 :	mblume	1347	Since all core access must be resolved already during the FLINT
1403 :			translate phase, it would be too late a the time of CPS contract
1404 :	mblume	1348	to add new Core calls. For this reason, conversion primops
1405 :	mblume	1347	for intinf carry two arguments: 1. the numeric argument that
1406 :			they are supposed to convert, and 2. the Core function that
1407 :			can help with this conversion if necessary. If CPS contract
1408 :	mblume	1348	eliminates a primop, then the associated Core function becomes
1409 :			dead and goes away. Intinf conversion primops that do not get
1410 :			eliminated by CPS contract get rewritten into calls of their
1411 :			core functions by a separate, new phase.
1412 :	mblume	1347
1413 :			interactive system:
1414 :			- Control.Print.intinfDepth controls max length of intinf constants
1415 :			being printed. (Analogous to Control.Print.stringDepth.)
1416 :			- Cleanup in printutil and pputil: got rid of unused stuff and
1417 :			duplicates; replaced some of the code with code that makes better
1418 :			use of library functionality.
1419 :
1420 :			CM:
1421 :			Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1422 :			are no longer silent.
1423 :
1424 :			CKIT:
1425 :			Fixed mismatched uses of Int32 and LargeInt. I always decided
1426 :			in favor of LargeInt -- which is now the same as IntInf.
1427 :			CKIT-knowledgable people should check whether this is what's
1428 :			intended and otherwise change things back to using Int32 or
1429 :			FixedInt.
1430 :
1431 :			Throughout the code:
1432 :	mblume	1348	Started using IntInf.int literals and built-in operations
1433 :	mblume	1347	(e.g., comparison with 0) where this seems appropriate.
1434 :
1435 :
1436 :			----------------------------------------------------------------------
1437 :	macqueen	1343	Name: Dave MacQueen (dbm@cs.uchicago.edu)
1438 :			Date: 2003/08/13 11:36:00 CDT
1439 :			Tag: dbm-20030813-mcz-merge1
1440 :			Description:
1441 :
1442 :			Merging changes from the mcz-branch development branch into trunk.
1443 :			These changes involve replacement of the emulated old prettyprinter
1444 :			interface with direct use of the SML/NJ Lib PP library, and fixing
1445 :			of a couple of bugs (895, 1186) relating to error messages. A new
1446 :			prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1447 :			has been added.
1448 :
1449 :			----------------------------------------------------------------------
1450 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1451 :	mblume	1342	Date: 2003/08/11 15:45:00 CDT
1452 :			Tag: blume-20030811-windows
1453 :			Description:
1454 :
1455 :			Version number bumped to 110.42.9. NEW BOOTFILES!!!
1456 :
1457 :			http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1458 :
1459 :			This patch restores SML/NJ's ability to run under win32. There are a
1460 :			number of changes, including fixes for several bugs that had gone
1461 :			unnoticed until now:
1462 :
1463 :			- uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1464 :			trying to run the win32 version from within cygwin.)
1465 :
1466 :			- There are a number of simple .bat scripts that substitute for their
1467 :			corresponding Unix shell-scripts. (See below.)
1468 :
1469 :			- The internals of ml-build have been modified slightly. The main
1470 :			difference is that instead of calling ".link-sml" (or link-sml.bat)
1471 :			using OS.Process.system, the ML process delegates this task back
1472 :			to the script. Otherwise problems arise in mixed environments such
1473 :			as Cygwin where scripts look and work like Unix scripts, but
1474 :			where OS.Process.system cannot run them.
1475 :
1476 :			- In CM, the srcpath pickler used native pathname syntax -- which
1477 :			is incorrect in the case of cross-compilation. The new pickle format
1478 :			is independent of platform-specific naming conventions.
1479 :
1480 :			- Path configuration files (such as lib/pathconfig) can now choose
1481 :			between native and standard syntax. Placing a line of the form
1482 :
1483 :			standard!
1484 :
1485 :			into the file causes all subsequent paths to be interpreted using
1486 :			CM standard pathname syntax (= Unix conventions); a line
1487 :
1488 :			native!
1489 :
1490 :			switches back to native style. This was needed so that
1491 :			path config files can be written portably, see src/system/pathconfig.
1492 :
1493 :			- Runtime system:
1494 :
1495 :			- win32-filesys.c: get_file_time and set_file_time now
1496 :			access modification time, not creation time.
1497 :
1498 :			- I/O code made aware of new array representation.
1499 :
1500 :			- Bug fixes in X86.prim.masm.
1501 :
1502 :			- src/system/makeml made aware of win32. (For use under cygwin
1503 :			and other Unix-environments for windows.)
1504 :
1505 :			- In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1506 :			which caused BinIO.inputAll to fail consistently.
1507 :
1508 :			.bat scripts:
1509 :
1510 :			Windows .bat scripts assume that SMLNJ_HOME is defined.
1511 :
1512 :			- sml.bat, ml-yacc.bat, ml-lex.bat: Driver scripts for standalone
1513 :			applications (sml, ml-yacc, ml-lex).
1514 :			- ml-build.bat: analogous to ml-build.
1515 :			- config\install.bat: Analogous to config/install.sh. This requires
1516 :			that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1517 :			(nmake etc. must be on the path, and vcvars32 must have been run.)
1518 :			Moreover, sources for ml-lex and ml-yacc need to exist under src,
1519 :			and the bootfile hierarchy must have been unpacked under
1520 :			sml.boot.x86-win32.
1521 :			The script is very primitive and does a poor job at error checking.
1522 :			It only installs the base system, ml-lex, and ml-yacc. No other
1523 :			libraries are being installed (i.e., you get only those that
1524 :			are part of the compiler.)
1525 :			- link-sml.bat: analogous to .link-sml, but not currently used
1526 :
1527 :			Unrelated bug fixes:
1528 :
1529 :			- ml-nlffigen now exports structures ST_* corresponding to incomplete
1530 :			types.
1531 :			- Added getDevice to PP/src/pp-debug-fn.sml. (Would not compile
1532 :			otherwise.)
1533 :
1534 :			----------------------------------------------------------------------
1535 :	macqueen	1337	Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1536 :			Date: 2003/06/17
1537 :			Tag: macqueen-20030617-bug895
1538 :			Description:
1539 :
1540 :			Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1541 :			Tag will be used for new development branch (mcz-branch) for
1542 :			use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1543 :
1544 :			----------------------------------------------------------------------
1545 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1546 :	mblume	1335	Date: 2003/05/27 16:55:00 CDT
1547 :			Tag: blume-20030527-polyeq
1548 :			Description:
1549 :
1550 :			Tried to eliminated most cases of polymorphic equality.
1551 :
1552 :			----------------------------------------------------------------------
1553 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1554 :	mblume	1334	Date: 2003/05/21 17:45:00 CDT
1555 :			Tag: blume-20030517-complete
1556 :			Description:
1557 :
1558 :			Two changes:
1559 :
1560 :			1. Added a flag for controlling whether non-exhaustive bindings will
1561 :			be treated as errors (default is false).
1562 :			2. Cleaned up the *entire* source tree so that CMB.make goes through
1563 :			without a single non-exhaustive match- or bind warning.
1564 :
1565 :			----------------------------------------------------------------------
1566 :	mblume	1368	Name: Matthias Blume (blume (at) tti - c (dot) org)
1567 :	mblume	1332	Date: 2003/05/17 10:20:00 CDT
1568 :			Tag: blume-20030517-absyn
1569 :			Description:
1570 :
1571 :			1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1572 :
1573 :			This mainly affects the quality of error messages. However, some
1574 :			of the code is now more straightforward than before. (Treatment of
1575 :			the above four constructs in translate.sml is much simpler than
1576 :			the "macro-expansion" that was going on before. Plus, the mach-
1577 :			compiler no longer gets invoked just to be able to compile an
1578 :			if-expression.)
1579 :
1580 :			2. The ErrorMsg.Error exception is now caught and absorbed by the
1581 :			interactive loop.
1582 :
1583 :			----------------------------------------------------------------------
1584 :	allenleung	1327	Name: Allen Leung
1585 :			Date: 2003/05/16 13:05:00 CDT
1586 :			Tag: leunga-20030516-cygwin-runtime
1587 :			Description:
1588 :
1589 :			Ported the runtime system to cygwin, which uses the unix
1590 :			x86-unix bin files. Missing/buggy features:
1591 :
1592 :			o getnetbyname, getnetbyaddr: these functions seem to be missing in
1593 :			the Cygwin library.
1594 :			o Ctrl-C handling may be flaky.
1595 :			o Windows system calls and Windows I/O are not supported.
1596 :
1597 :			A new set of binfiles is located at:
1598 :
1599 :			http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1600 :
1601 :			This is only needed for bootstrapping the cygwin version of smlnj.
1602 :			Other x86 versions can use the existing binfiles.
1603 :
1604 :			----------------------------------------------------------------------
1605 :	mblume	1317	Name: Matthias Blume
1606 :			Date: 2003/04/08 15:42:00 CDT
1607 :			Tag: blume-20030408-listpair
1608 :			Description:
1609 :
1610 :			1. Added a target 'mlrisc' to installer.
1611 :
1612 :			2. Added missing elements to structure ListPair.
1613 :
1614 :			----------------------------------------------------------------------
1615 :	leunga	1296	Name: Allen Leung
1616 :			Date: 2003/01/07 10:40:00 EST
1617 :			Tag: leunga-20030107-int-rem
1618 :			Description:
1619 :
1620 :			Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1621 :			The arguments to the SUBL instruction were swapped.
1622 :
1623 :			----------------------------------------------------------------------
1624 :	blume	1229	Name: Matthias Blume
1625 :	blume	1294	Date: 2002/12/12 16:25:00 EST
1626 :			Tag: blume-20021212-risc-ra
1627 :			Description:
1628 :
1629 :			Fixed a serious bug in the rewrite code for FP spilling/reloading that
1630 :			sent the RA into an infinite loop when floating point registers get
1631 :			spilled. (Because of this bug, e.g., nucleic stopped compiling between
1632 :			110.37 and 110.38.)
1633 :			There was another set of potential problems related to the handling of
1634 :			MLRISC annotations (but those did not yet cause real problems, apparently).
1635 :
1636 :			----------------------------------------------------------------------
1637 :			Name: Matthias Blume
1638 :	blume	1291	Date: 2002/12/06 22:40:00 EST
1639 :			Tag: blume-20021206-cm-fileid
1640 :			Description:
1641 :
1642 :			Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1643 :			This fixes the problem of CM getting confused by files that suddenly
1644 :			change their identity (e.g., by getting unlinked and recreated by some
1645 :			text editor such as vi). There might be a better/cheaper/cleaner way
1646 :			of doing this, but for now this will have to do.
1647 :
1648 :			----------------------------------------------------------------------
1649 :			Name: Matthias Blume
1650 :	blume	1288	Date: 2002/10/28 09:50:00 EST
1651 :			Tag: blume-20021028-typecheck
1652 :			Description:
1653 :
1654 :			Exported structure Typecheck from $smlnj/viscomp/core.cm.
1655 :
1656 :			----------------------------------------------------------------------
1657 :			Name: Matthias Blume
1658 :	blume	1282	Date: 2002/10/17 09:10:00 EDT
1659 :	blume	1279	Tag: Release_110_42
1660 :			Description:
1661 :
1662 :	blume	1282	In good old tradition, there has been a slight hiccup so that we have
1663 :			to patch 110.42 after the fact. The old release tag has been replaced
1664 :			(see below).
1665 :
1666 :			The change solves a problem with two competing approaches the
1667 :			configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1668 :			each other's way.
1669 :
1670 :			This change only affects the runtime system code and the installer script.
1671 :			(No new bootfiles.)
1672 :
1673 :			----------------------------------------------------------------------
1674 :			Name: Matthias Blume
1675 :			Date: 2002/10/16 12:00:00 EDT
1676 :			Tag: Release_110_42_removed
1677 :			Description:
1678 :
1679 :	blume	1279	New working release. New bootfiles.
1680 :
1681 :			----------------------------------------------------------------------
1682 :			Name: Matthias Blume
1683 :	blume	1273	Date: 2002/10/10 13:10:00 EDT
1684 :			Tag: blume-20021010-ppc-divs
1685 :			Description:
1686 :
1687 :			The mltree operator DIVS must be implemented with an overflow check on
1688 :			the PPC because the hardware indicates divide-by-zero using "overflow" as
1689 :			well.
1690 :
1691 :			----------------------------------------------------------------------
1692 :			Name: Matthias Blume
1693 :	blume	1261	Date: 2002/07/23 11:45:00
1694 :			Tag: blume-20020723-smlnj-home
1695 :			Description:
1696 :
1697 :			Sml now senses the SMLNJ_HOME environment variable. If this is set,
1698 :			then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1699 :			CM_PATHCONFIG is also set), the path configuration file is assumed
1700 :			to be in $SMLNJ_HOME/lib/pathconfig. This way one can easily move
1701 :			the entire tree to some other place and everything will "just work".
1702 :
1703 :			(Companion commands such as ml-build and ml-makedepend also sense this
1704 :			variable.)
1705 :
1706 :			----------------------------------------------------------------------
1707 :			Name: Matthias Blume
1708 :	blume	1259	Date: 2002/07/12 21:19:00 EDT
1709 :			Tag: blume-20020712-liveness
1710 :			Description:
1711 :
1712 :			Exported two useful "step" functions from liveness module (MLRISC).
1713 :
1714 :			----------------------------------------------------------------------
1715 :			Name: Matthias Blume
1716 :	blume	1253	Date: 2002/07/05 16:00 EDT
1717 :			Tag: Release_110_41
1718 :			Description:
1719 :
1720 :			New working release. New bootfiles.
1721 :
1722 :			----------------------------------------------------------------------
1723 :			Name: Matthias Blume
1724 :	blume	1250	Date: 2002/07/05 10:25:00 EDT
1725 :			Tag: blume-20020705-btimp
1726 :			Description:
1727 :
1728 :			Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1729 :			other clients can set up backtracing support.
1730 :
1731 :			----------------------------------------------------------------------
1732 :			Name: Matthias Blume
1733 :	blume	1249	Date: 2002/06/25 17:23:00 EDT
1734 :			Tag: blume-20020625-fpmax
1735 :			Description:
1736 :
1737 :			Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1738 :			case. (The sense of the isNaN test was reversed -- which made min and
1739 :			max always return their first argument.)
1740 :
1741 :			----------------------------------------------------------------------
1742 :			Name: Matthias Blume
1743 :	blume	1240	Date: 2002/06/11
1744 :			Tag: blume-20020611-unixpath
1745 :			Description:
1746 :
1747 :			Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1748 :
1749 :			----------------------------------------------------------------------
1750 :			Name: Matthias Blume
1751 :	blume	1238	Date: 2002/06/10 16:35:00 EDT
1752 :			Tag: blume-20020610-ieeereal
1753 :			Description:
1754 :
1755 :			I back-ported my implementation of IEEEReal.fromString from the
1756 :			idlbasis-devel branch so that we can test it.
1757 :
1758 :			Another small change is that ppDec tries to give more information
1759 :			than just "<sig>" in the case of functors. However, this code is
1760 :			broken in some mysterious way if the functor's body's signature
1761 :			has not been declared by ascription but gets inferred from the
1762 :			implementation. This needs fixing...
1763 :
1764 :			----------------------------------------------------------------------
1765 :			Name: Matthias Blume
1766 :	blume	1229	Date: 2002/05/31
1767 :			Tag: blume-20020531-btrace-mode
1768 :			Description:
1769 :
1770 :			Resurrected SMLofNJ.Internals.BTrace.mode. (It accidentally fell by
1771 :			the wayside when I switched over to using Controls everywhere.)
1772 :
1773 :			----------------------------------------------------------------------
1774 :	george	1222	Name: Lal George
1775 :			Date: 2002/05/23 12:21:40 EDT
1776 :			Tag: george-20020523-visual-labels
1777 :			Description:
1778 :
1779 :			Labels are now displayed in the graphical output to make
1780 :			the fall-through and target blocks obvious.
1781 :
1782 :			----------------------------------------------------------------------
1783 :	blume	1204	Name: Matthias Blume
1784 :	blume	1220	Date: 2002/05/22 11:03:00 EDT
1785 :			Tag: blume-20020522-shrink
1786 :			Description:
1787 :
1788 :			John tweaked yesterday's fix for 1131 to handle an out-of-memory
1789 :			situation that comes up when allocating huge arrays.
1790 :
1791 :			----------------------------------------------------------------------
1792 :			Name: Matthias Blume
1793 :	blume	1215	Date: 2002/05/21 16:00:00 EDT
1794 :			Tag: Release_110_40
1795 :			Description:
1796 :
1797 :			New working release (110.40). New bootfiles.
1798 :
1799 :			[Also: John Reppy fixed GC bug 1131.]
1800 :
1801 :			----------------------------------------------------------------------
1802 :			Name: Matthias Blume
1803 :	blume	1212	Date: 2002/05/21 12:35:00 EDT
1804 :			Tag: blume-20020521-cmdoc
1805 :			Description:
1806 :
1807 :			CM documentation update.
1808 :
1809 :			----------------------------------------------------------------------
1810 :			Name: Matthias Blume
1811 :	blume	1211	Date: 2002/05/21 10:55:00 EDT
1812 :			Tag: blume-20020521-misc
1813 :			Description:
1814 :
1815 :			- John tweaked runtime to be silent on heap export (except when
1816 :			GC messages are on).
1817 :			- I added a few more things (cross-compiling versions of CMB) to
1818 :			config/preloads (as suggestions).
1819 :
1820 :			----------------------------------------------------------------------
1821 :			Name: Matthias Blume
1822 :	blume	1208	Date: 2002/05/20 22:25:00 EDT
1823 :			Tag: blume-20020520-controls
1824 :			Description:
1825 :
1826 :			- Added ControlUtil structure to control-lib.cm.
1827 :			- Use it throughout.
1828 :			- Used Controls facility to define MLRISC controls (as opposed to
1829 :			registering MLRISC control ref cells with Controls after the fact)
1830 :			- Fixed messed-up controls priorities.
1831 :
1832 :			* Removed again all the stuff from config/preloads that one wouldn't
1833 :			be able to preload at the time the initial heap image is built.
1834 :			(Many libraries, e.g., CML, do not exist yet at this time. The
1835 :			only libraries that can be preloaded via config/preloads are those
1836 :			that come bundled with the bootfiles.)
1837 :
1838 :			----------------------------------------------------------------------
1839 :			Name: Matthias Blume
1840 :	blume	1204	Date: 2002/05/20 10:59:00 EDT
1841 :			Tag: blume-20020520-preloads
1842 :			Description:
1843 :
1844 :			Added a lot of commented-out suggestions for things to be included
1845 :			in config/preloads.
1846 :
1847 :			----------------------------------------------------------------------
1848 :	leunga	1203	Name: Allen Leung
1849 :			Date: 2002/05/18 14:20:00 EDT
1850 :			Tag: leunga-20020518-mdl
1851 :			Description:
1852 :
1853 :			o Made the mdl tool stuff compile and run again.
1854 :			o I've disabled all the stuff that depends on RTL specifications; they
1855 :			are all badly broken anyway.
1856 :
1857 :			----------------------------------------------------------------------
1858 :	blume	1201	Name: Matthias Blume
1859 :			Date: 2002/05/17 16:49:00 EDT
1860 :			Tag: blume-20020517-controls
1861 :			Description:
1862 :
1863 :			0. John Reppy made several modifications to the SML/NJ library.
1864 :			In particular, there is a shiny new controls-lib.cm.
1865 :
1866 :			1. Pushed new controls interface through compiler so that everything
1867 :			compiles again.
1868 :
1869 :			2. Added FormatComb and FORMAT_COMB to the CML version of the
1870 :			SML/NJ library (so that CML compiles again).
1871 :
1872 :			3. Modified init scripts because XXX_DEFAULT environment variables
1873 :			are no longer with us. (Boot-time initialization is now done
1874 :			using the same environment variables that are also used for
1875 :			startup-time initialization of controls.)
1876 :
1877 :			----------------------------------------------------------------------
1878 :	george	1192	Name: Lal George
1879 :			Date: 2002/05/15 09:20:10 EDT
1880 :			Tag: george-20020515-pseudo-op-decls
1881 :			Description:
1882 :
1883 :			All pseudo-ops emitted before the first segment declaration
1884 :			such as TEXT, DATA, and BSS directives are assumed to be global
1885 :			declarations and are emitted first in the assembly file. This is
1886 :			useful in a number of situations where one has pseudo-ops that are not
1887 :			specific to any segment, and also works around the constraint that one
1888 :			cannot have client pseudo-ops in the TEXT segment.
1889 :
1890 :			Because no segment is associated with these declarations it is
1891 :			an error to allocate any space or objects before the first segment
1892 :			directive and an exception will be raised. However, we cannot make
1893 :			this check for client pseudo-ops.
1894 :
1895 :			These top level declarations are a field in the CFG graph_info.
1896 :			In theory you can continue to add to this field after the CFG has been
1897 :			built -- provided you know what you are doing;-)
1898 :
1899 :			----------------------------------------------------------------------
1900 :	blume	1175	Name: Matthias Blume
1901 :	blume	1190	Date: 2002/05/13 16:40:00 EDT
1902 :			Tag: blume-20020513-pp-etc
1903 :			Description:
1904 :
1905 :			A few minor bugfixes:
1906 :
1907 :			- Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1908 :			(Bogus printouts for redefined bindings still occur. Compiler
1909 :			bug should no longer occur now. We need to redo the prettyprinter
1910 :			from scratch.)
1911 :
1912 :			- CM pathname printer now also adds escape sequences for ( and )
1913 :
1914 :			- commend and docu fixes for ml-nlffi
1915 :
1916 :			----------------------------------------------------------------------
1917 :			Name: Matthias Blume
1918 :	blume	1189	Date: 2002/05/10 16:40:00 EDT
1919 :			Tag: blume-20020510-erg-textio
1920 :			Description:
1921 :
1922 :			Applied the following bugfix provided by Emden Gansner:
1923 :
1924 :			Output is corrupted when outputSubstr is used rather than output.
1925 :
1926 :			The problem occurs when a substring
1927 :
1928 :			ss = (s, dataStart, dataLen)
1929 :
1930 :			where dataStart > 0, fills a stream buffer with avail bytes left.
1931 :			avail bytes of s, starting at index dataStart, are copied into the
1932 :			buffer, the buffer is flushed, and then the remaining dataLen-avail
1933 :			bytes of ss are copied into the beginning of the buffer. Instead of
1934 :			starting this copy at index dataStart+avail in s, the current code
1935 :			starts the copy at index avail.
1936 :
1937 :			Fix:
1938 :			In text-io-fn.sml, change line 695 from
1939 :			val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1940 :			to
1941 :			val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1942 :
1943 :			----------------------------------------------------------------------
1944 :			Name: Matthias Blume
1945 :	blume	1186	Date: 2002/04/12 13:55:00 EDT
1946 :			Tag: blume-20020412-assyntax
1947 :			Description:
1948 :
1949 :			1. Grabbed newer assyntax.h from the XFree86 project.
1950 :			2. Fiddled with how to compile X86.prim.asm without warnings.
1951 :			3. (Very) Minor cleanup in CM.
1952 :
1953 :			----------------------------------------------------------------------
1954 :			Name: Matthias Blume
1955 :	blume	1185	Date: 2002/04/01 (no joke!) 17:07:00 EST
1956 :			Tag: blume-20020401-x86div
1957 :			Description:
1958 :
1959 :			Added full support for div/mod/rem/quot on the x86, using the machine
1960 :			instruction's two results (without clumsily recomputing the remainder)
1961 :			directly where appropriate.
1962 :
1963 :			Some more extensive power-of-two support was added to the x86 instruction
1964 :			selector (avoiding expensive divs, mods, and muls where they can be
1965 :			replaced with cheaper shifts and masks). However, this sort of thing
1966 :			ought to be done earlier, e.g., within the CPS optimizer so that
1967 :			all architectures benefit from it.
1968 :
1969 :			The compiler compiles to a fixed point, but changes might be somewhat
1970 :			fragile nevertheless. Please, report any strange things that you might
1971 :			see wrt. div/mod/quot/rem...
1972 :
1973 :			----------------------------------------------------------------------
1974 :			Name: Matthias Blume
1975 :	blume	1184	Date: 2002/03/29 17:22:00
1976 :			Tag: blume-20020329-div
1977 :			Description:
1978 :
1979 :			Fixed my broken div/mod logic. Unfortunately, this means that the
1980 :			inline code for div/mod now has one more comparison than before.
1981 :			Fast paths (quotient > 0 or remainder = 0) are not affected, though.
1982 :			The problem was with quotient = 0, because that alone does not tell
1983 :			us which way the rounding went. One then has to look at whether
1984 :			remainder and divisor have the same sign... :(
1985 :
1986 :			Anyway, I replaced the bootfiles with fresh ones...
1987 :
1988 :			----------------------------------------------------------------------
1989 :			Name: Matthias Blume
1990 :	blume	1183	Date: 2002/03/29 14:10:00 EST
1991 :			Tag: blume-20020329-inlprims
1992 :			Description:
1993 :
1994 :			NEW BOOTFILES!!! Version number bumped to 110.39.3.
1995 :
1996 :			Primops have changed. This means that the bin/boot-file formats have
1997 :			changed as well.
1998 :
1999 :			To make sure that there is no confusion, I made a new version.
2000 :
2001 :
2002 :			CHANGES:
2003 :
2004 :			* removed REMT from mltree (remainder should never overflow).
2005 :
2006 :			* added primops to deal with divisions of all flavors to the frontend
2007 :
2008 :			* handled these primops all the way through so they map to their respective
2009 :			MLRISC support
2010 :
2011 :			* used these primops in the implementation of Int, Int32, Word, Word32
2012 :
2013 :			* removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2014 :
2015 :			* parameterized INLMIN, INLMAX, and INLABS by a numkind
2016 :
2017 :			* translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2018 :			floating point
2019 :
2020 :			* used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2021 :			and Real (but Real.abs maps to a separate floating-point-only primop)
2022 :
2023 :
2024 :			TODO items:
2025 :
2026 :			* Hacked Alpha32 instruction selection, disabling the selection of REMx
2027 :			instructions because the machine instruction encoder cannot handle
2028 :			them. (Hppa, PPC, and Sparc instruction selection did not handle
2029 :			REM in the first place, and REM is supported by the x86 machine coder.)
2030 :
2031 :			* Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2032 :			selection phase. (The two can be streamlined because the hardware
2033 :			delivers both quotient and remainder at the same time anyway.)
2034 :
2035 :			* Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2036 :			(Currently the behavior is inconsistent both across architectures and
2037 :			wrt. the draft Basis spec.)
2038 :
2039 :			* Word8 should eventually be handled natively, too.
2040 :
2041 :			* There seems to be one serious bug in mltree-gen.sml. It appears, though,
2042 :			as if there currently is no execution path that could trigger it in
2043 :			SML/NJ. (The assumptions underlying functions arith and promotable do not
2044 :			hold for things like multiplication and division.)
2045 :
2046 :			----------------------------------------------------------------------
2047 :			Name: Matthias Blume
2048 :	blume	1181	Date: 2002/03/27 16:27:00 EST
2049 :			Tag: blume-20020327-mlrisc-divisions
2050 :			Description:
2051 :
2052 :			Added support for all four division operations (ML's div, mod, quot,
2053 :			and rem) to MLRISC. In the course of doing so, I also rationalized
2054 :			the naming (no more annoying switch-around of DIV and QUOT), by
2055 :			parameterizing the operation by div_rounding_mode (which can be either
2056 :			DIV_TO_ZERO or DIV_TO_NEGINF).
2057 :
2058 :			The generic MLTreeGen functor takes care of compiling all four
2059 :			operations down to only round-to-zero div.
2060 :
2061 :			Missing pieces:
2062 :
2063 :			* Doing something smarter than relying on MLTreeGen on architectures
2064 :			like, e.g., the x86 where hardware division delivers both quotient and
2065 :			remainder at the same time. With this, the implementation of the
2066 :			round-to-neginf operations could be further streamlined.
2067 :
2068 :			* Remove inlining support for div/mod/rem from the frontend and replace it
2069 :			with primops that get carried through to the backend. Do this for all
2070 :			int and word types.
2071 :
2072 :			----------------------------------------------------------------------
2073 :			Name: Matthias Blume
2074 :	blume	1180	Date: 2002/03/25 17:25:00 EST
2075 :			Tag: blume-20020325-divmod
2076 :			Description:
2077 :
2078 :			I improved (hopefully without breaking them) the implementation of Int.div,
2079 :			Int.mod, and Int.rem. For this, the code in translate.sml now takes
2080 :			advantage of the following observations:
2081 :
2082 :			Let q = x quot y r = x rem y
2083 :			d = x div y m = x mod y
2084 :
2085 :			where "quot" is the round-to-zero version of integer division that
2086 :			hardware usually provides. Then we have:
2087 :
2088 :			r = x - q * y where neither the * nor the - will overflow
2089 :			d = if q >= 0 orelse x = q * y then q else q - 1
2090 :			where neither the * nor the - will overflow
2091 :			m = if q >= 0 orelse r = 0 then r else r + y
2092 :			where the + will not overflow
2093 :
2094 :			This results in substantial simplification of the generated code.
2095 :			The following table shows the number of CFG nodes and edges generated
2096 :			for
2097 :			fun f (x, y) = x OPER y
2098 :			(* with OPER \in div, mod, quot, rem *)
2099 :
2100 :
2101 :			OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2102 :			--------------------------------------------------------
2103 :			div | 24 | 39 | 12 | 16
2104 :			mod | 41 | 71 | 12 | 16
2105 :			quot | 8 | 10 | 8 | 10
2106 :			rem | 10 | 14 | 8 | 10
2107 :
2108 :
2109 :			----------------------------------------------------------------------
2110 :			Name: Matthias Blume
2111 :	blume	1179	Date: 2002/03/25 22:06:00 EST
2112 :			Tag: blume-20020325-cprotobug
2113 :			Description:
2114 :
2115 :			Fixed a bug in cproto (c prototype decoder).
2116 :
2117 :			----------------------------------------------------------------------
2118 :			Name: Matthias Blume
2119 :	blume	1178	Date: 2002/03/25 16:00:00 EST
2120 :			Tag: blume-20020325-raw-primops
2121 :			Description:
2122 :
2123 :			I did some cleanup to Allen's new primop code and
2124 :			replaced yesterday's bootfiles with new ones.
2125 :			(But they are stored in the same place.)
2126 :
2127 :			----------------------------------------------------------------------
2128 :			Name: Matthias Blume
2129 :	blume	1175	Date: 2002/03/24 22:40:00 EST
2130 :			Tag: blume-20020324-bootfiles
2131 :			Description:
2132 :
2133 :			Made the bootfiles that Allen asked for.
2134 :
2135 :			----------------------------------------------------------------------
2136 :	leunga	1174	Name: Allen Leung
2137 :			Date: 2002/03/23 15:50:00 EST
2138 :			Tag: leunga-20020323-flint-cps-rcc-primops
2139 :			Description:
2140 :
2141 :			1. Changes to FLINT primops:
2142 :
2143 :			(* make a call to a C-function;
2144 :			* The primop carries C function prototype information and specifies
2145 :			* which of its (ML-) arguments are floating point. C prototype
2146 :			* information is for use by the backend, ML information is for
2147 :			* use by the CPS converter. *)
2148 :			| RAW_CCALL of { c_proto: CTypes.c_proto,
2149 :			ml_args: ccall_type list,
2150 :			ml_res_opt: ccall_type option,
2151 :			reentrant : bool
2152 :			} option
2153 :			(* Allocate uninitialized storage on the heap.
2154 :			* The record is meant to hold short-lived C objects, i.e., they
2155 :			* are not ML pointers. With the tag, the representation is
2156 :			* the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2157 :			*)
2158 :			| RAW_RECORD of {tag:bool,sz:int}
2159 :			and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2160 :
2161 :			2. These CPS primops are now overloaded:
2162 :
2163 :			rawload of {kind:numkind}
2164 :			rawstore of {kind:numkind}
2165 :
2166 :			The one argument form is:
2167 :
2168 :			rawload {kind} address
2169 :
2170 :			The two argument form is:
2171 :
2172 :			rawload {kind} [ml object, byte-offset]
2173 :
2174 :			3. RAW_CCALL/RCC now takes two extra arguments:
2175 :
2176 :			a. The first is whether the C call is reentrant, i.e., whether
2177 :			ML state should be saved and restored.
2178 :			b. The second argument is a string argument specifying the name of
2179 :			library and the C function.
2180 :
2181 :			These things are currently not handled in the code generator, yet.
2182 :
2183 :			4. In CProto,
2184 :
2185 :			An encoding type of "bool" means "ml object" and is mapped into
2186 :			C prototype of PTR. Note that "bool" is different than "string",
2187 :			even though "string" is also mapped into PTR, because "bool"
2188 :			is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2189 :
2190 :			5. Pickler/unpicker
2191 :
2192 :			Changed to handle RAW_RECORD and newest RAW_CCALL
2193 :
2194 :			6. MLRiscGen,
2195 :
2196 :			1. Changed to handle the new rawload/rawstore/rawrecord operators.
2197 :			2. Code for handling C Calls has been moved to a new module CPSCCalls,
2198 :			in the file CodeGen/cpscompile/cps-c-calls.sml
2199 :
2200 :			7. Added the conditional move operator
2201 :
2202 :			condmove of branch
2203 :
2204 :			to cps. Generation of this is still buggy so it is currently
2205 :			disabled.
2206 :
2207 :			----------------------------------------------------------------------
2208 :	george	1168	Name: Lal George
2209 :			Date: 2002/03/22 14:18:25 EST
2210 :	george	1171	Tag: george-20020322-cps-branch-prob
2211 :	george	1168	Description:
2212 :
2213 :	george	1171	Implemented the Ball-Larus branch prediction-heuristics, and
2214 :	george	1168	incorporated graphical viewers for control flow graphs.
2215 :
2216 :	george	1171	Ball-Larus Heuristics:
2217 :	george	1168	---------------------
2218 :			See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2219 :
2220 :			By design it uses the Dempster-Shafer theory for combining
2221 :			probabilities. For example, in the function:
2222 :
2223 :			fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2224 :
2225 :			the ball-larus heuristics predicts that the n=0 is unlikely
2226 :			(OH-heuristic), and the 'then' branch is unlikely because of the
2227 :			RH-heuristic -- giving the 'then' branch an even lower combined
2228 :	george	1171	probability using the Dempster-Shafer theory.
2229 :	george	1168
2230 :			Finally, John Reppy's loop analysis in MLRISC, further lowers the
2231 :			probability of the 'then' branch because of the loop in the else
2232 :			branch.
2233 :
2234 :
2235 :			Graphical Viewing:
2236 :			------------------
2237 :			I merely plugged in Allen's graphical viewers into the compiler. The
2238 :			additional code is not much. At the top level, saying:
2239 :
2240 :			Control.MLRISC.getFlag "cfg-graphical-view" := true;
2241 :
2242 :			will display the graphical view of the control flow graph just before
2243 :			back-patching. daVinci must be in your path for this to work. If
2244 :			daVinci is not available, then the default viewer can be changed
2245 :			using:
2246 :
2247 :			Control.MLRISC.getString "viewer"
2248 :
2249 :			which can be set to "dot" or "vcg" for the corresponding viewers. Of
2250 :			course, these viewers must be in your path.
2251 :
2252 :			The above will display the compilation unit at the level of clusters,
2253 :			many of which are small, boring, and un-interesting. Also setting:
2254 :
2255 :			Control.MLRISC.getInt "cfg-graphical-view_size"
2256 :
2257 :			will display clusters that are larger than the value set by the above.
2258 :
2259 :
2260 :			----------------------------------------------------------------------
2261 :	blume	1159	Name: Matthias Blume
2262 :			Date: 2002/03/21 22:20:00 EST
2263 :			Tag: blume-20020321-kmp-bugfix
2264 :			Description:
2265 :
2266 :			Changed the interface to the KMP routine in PreString and fixed
2267 :			a minor bug in one place where it was used.
2268 :
2269 :			----------------------------------------------------------------------
2270 :	leunga	1156	Name: Allen Leung
2271 :	leunga	1158	Date: 2002/03/21 20:30:00 EST
2272 :			Tag: leunga-20020321-cfg
2273 :			Description:
2274 :
2275 :			Fixed a potential problem in cfg edge splitting.
2276 :
2277 :			----------------------------------------------------------------------
2278 :			Name: Allen Leung
2279 :	leunga	1156	Date: 2002/03/21 17:15:00 EST
2280 :			Tag: leunga-20020321-x86-fp-cfg
2281 :			Description:
2282 :
2283 :			1. Recoded the buggy parts of x86-fp.
2284 :
2285 :			a. All the block reordering code has been removed.
2286 :			We now depend on the block placement phases to do this work.
2287 :
2288 :			b. Critical edge splitting code has been simplified and moved into the
2289 :			CFG modules, as where they belong.
2290 :
2291 :			Both of these were quite buggy and complex. The code is now much, much
2292 :			simpler.
2293 :
2294 :			2. X86 backend.
2295 :
2296 :			a. Added instructions for 64-bit support. Instruction selection for
2297 :			64-bit has not been committed, however, since that
2298 :			requires changes to MLTREE which haven't been approved by
2299 :			Lal and John.
2300 :
2301 :			b. Added support for FUCOMI and FUCOMIP when generating code for
2302 :			PentiumPro and above. We only generate these instructions in
2303 :			the fast-fp mode.
2304 :
2305 :			c. Added cases for JP and JNP in X86FreqProps.
2306 :
2307 :			3. CFG
2308 :
2309 :			CFG now has a bunch of methods for edge splitting and merging.
2310 :
2311 :			4. Machine description.
2312 :
2313 :			John's simplification of MLTREE_BASIS.fcond broke a few machine
2314 :			description things:
2315 :
2316 :			rtl-build.{sig,sml} and hppa.mdl fixed.
2317 :
2318 :			NOTE: the machine description stuff in the repository is still broken.
2319 :			Again, I can't put my fixes in because that involves
2320 :			changes to MLTREE.
2321 :
2322 :			----------------------------------------------------------------------
2323 :	blume	1145	Name: Matthias Blume
2324 :	blume	1155	Date: 2002/03/20 15:55:00 EST
2325 :			Tag: blume-20020320-kmp
2326 :			Description:
2327 :
2328 :			Implemented Knuth-Morris-Pratt string matching in PreString and used
2329 :			it for String.isSubstring, Substring.isSubstring, and
2330 :			Substring.position.
2331 :
2332 :			(Might need some stress-testing. Simple examples worked fine.)
2333 :
2334 :			----------------------------------------------------------------------
2335 :			Name: Matthias Blume
2336 :	blume	1152	Date: 2002/03/19 16:37:00 EST
2337 :			Tag: blume-20020319-witnesses
2338 :			Description:
2339 :
2340 :			Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2341 :
2342 :			This implements a generic mechanism for changing constness qualifiers
2343 :			anywhere within big C types without resorting to outright "casts".
2344 :			(So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2345 :			modify the constness at the outermost level.)
2346 :			The implementation of "convert" is based on the idea of "witness"
2347 :			values -- values that are not used by the operation but whose types
2348 :			"testify" to their applicability. On the implementation side, "convert"
2349 :			is simply a projection (returning its second curried argument). With
2350 :			cross-module inlining, it should not result in any machine code being
2351 :			generated.
2352 :
2353 :			----------------------------------------------------------------------
2354 :			Name: Matthias Blume
2355 :	blume	1148	Date: 2002/03/15 16:40:00 EST
2356 :			Tag: blume-20020315-basis
2357 :			Description:
2358 :
2359 :			Provided (preliminary?) implementations for
2360 :
2361 :			{String,Substring}.{concatWith,isSuffix,isSubstring}
2362 :
2363 :			and
2364 :
2365 :			Substring.full
2366 :
2367 :			Those are in the Basis spec but they were missing in SML/NJ.
2368 :
2369 :			----------------------------------------------------------------------
2370 :			Name: Matthias Blume
2371 :	blume	1145	Date: 2002/03/14 21:30:00 EST
2372 :			Tag: blume-20020314-controls
2373 :			Description:
2374 :
2375 :			Controls:
2376 :			---------
2377 :
2378 :			1. Factored out the recently-added Controls : CONTROLS stuff and put
2379 :			it into its own library $/controls-lib.cm. The source tree for
2380 :			this is under src/smlnj-lib/Controls.
2381 :
2382 :			2. Changed the names of types and functions in this interface, so they
2383 :			make a bit more "sense":
2384 :
2385 :			module -> registry
2386 :			'a registry -> 'a group
2387 :
2388 :			3. The interface now deals in ref cells only. The getter/setter interface
2389 :			is (mostly) gone.
2390 :
2391 :			4. Added a function that lets one register an already-existing ref cell.
2392 :
2393 :			5. Made the corresponding modifications to the rest of the code so that
2394 :			everything compiles again.
2395 :
2396 :			6. Changed the implementation of Controls.MLRISC back to something closer
2397 :			to the original. In particular, this module (and therefore MLRISC)
2398 :			does not depend on Controls. There now is some link-time code in
2399 :			int-sys.sml that registers the MLRISC controls with the Controls
2400 :			module.
2401 :
2402 :			CM:
2403 :			---
2404 :
2405 :			* One can now specify the lambda-split aggressiveness in init.cmi.
2406 :
2407 :			----------------------------------------------------------------------
2408 :	leunga	1142	Name: Allen Leung
2409 :			Date: 2002/03/13 17:30:00 EST
2410 :			Tag: leunga-20020313-x86-fp-unary
2411 :			Description:
2412 :
2413 :			Bug fix for:
2414 :
2415 :			> leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2416 :			> Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2417 :			> - fun f(x,(y,z)) = Real.~ y;
2418 :			> [autoloading]
2419 :			> [autoloading done]
2420 :			> fchsl (%eax), 184(%esp)
2421 :			> Error: MLRisc bug: X86MCEmitter.emitInstr
2422 :			>
2423 :			> uncaught exception Error
2424 :			> raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2425 :
2426 :			The problem was that the code generator did not generate any fp registers
2427 :			in this case, and the ra didn't know that it needed to run the X86FP phase to
2428 :			translate the pseudo fp instruction. This only happened with unary fp
2429 :			operators in certain situations.
2430 :
2431 :			----------------------------------------------------------------------
2432 :	blume	1137	Name: Matthias Blume
2433 :	blume	1140	Date: 2002/03/13 14:00:00 EST
2434 :			Tag: blume-20020313-overload-etc
2435 :			Description:
2436 :
2437 :			1. Added _overload as a synonym for overload for backward compatibility.
2438 :			(Control.overloadKW must be true for either version to be accepted.)
2439 :
2440 :			2. Fixed bug in install script that caused more things to be installed
2441 :			than what was requested in config/targets.
2442 :
2443 :			3. Made CM aware of the (_)overload construct so that autoloading
2444 :			works.
2445 :
2446 :			----------------------------------------------------------------------
2447 :			Name: Matthias Blume
2448 :	blume	1138	Date: 2002/03/12 22:03:00 EST
2449 :			Tag: blume-20020312-url
2450 :			Description:
2451 :
2452 :			Forgot to update BOOT and srcarchiveurl.
2453 :
2454 :			----------------------------------------------------------------------
2455 :			Name: Matthias Blume
2456 :	blume	1137	Date: 2002/03/12 17:30:00 EST
2457 :			Tag: blume-20020312-version110392
2458 :			Description:
2459 :
2460 :			Yet another version number bump (because of small changes to the
2461 :			binfile format). Version number is now 110.39.2. NEW BOOTFILES!
2462 :
2463 :			Changes:
2464 :
2465 :			The new pid generation scheme described a few weeks ago was overly
2466 :			complicated. I implemented a new mechanism that is simpler and
2467 :			provides a bit more "stability": Once CM has seen a compilation
2468 :			unit, it keeps its identity constant (as long as you do not delete
2469 :			those crucial CM/GUID/* files). This means that when you change
2470 :	blume	1145	an interface, compile, then go back to the old interface, and
2471 :	blume	1137	compile again, you arrive at the original pid.
2472 :
2473 :			There now also is a mechanism that instructs CM to use the plain
2474 :			environment hash as a module's pid (effectively making its GUID
2475 :			the empty string). For this, "noguid" must be specified as an
2476 :			option to the .sml file in question within its .cm file.
2477 :			This is most useful for code that is being generated by tools such
2478 :			as ml-nlffigen (because during development programmers tend to
2479 :			erase the tool's entire output directory tree including CM's cached
2480 :			GUIDs). "noguid" is somewhat dangerous (since it can be used to locally
2481 :			revert to the old, broken behavior of SML/NJ, but in specific cases
2482 :			where there is no danger of interface confusion, its use is ok
2483 :			(I think).
2484 :
2485 :			ml-nlffigen by default generates "noguid" annotations. They can be
2486 :			turned off by specifying -guid in its command line.
2487 :
2488 :			----------------------------------------------------------------------
2489 :	george	1136	Name: Lal George
2490 :			Date: 2002/03/12 12 14:42:36 EST
2491 :			Tag: george-20020312-frequency-computation
2492 :			Description:
2493 :
2494 :			Integrated jump chaining and static block frequency into the
2495 :			compiler. More details and numbers later.
2496 :	leunga	1025
2497 :	george	1003	----------------------------------------------------------------------
2498 :	george	1133	Name: Lal George
2499 :			Date: 2002/03/11 11 22:38:53 EST
2500 :			Tag: george-20020311-jump-chain-elim
2501 :			Description:
2502 :
2503 :			Tested the jump chain elimination on all architectures (except the
2504 :			hppa). This is on by default right now and is profitable for the
2505 :			alpha and x86, however, it may not be profitable for the sparc and ppc
2506 :			when compiling the compiler.
2507 :
2508 :			The gc test will typically jump to a label at the end of the cluster,
2509 :			where there is another jump to an external cluster containing the actual
2510 :			code to invoke gc. This is to allow factoring of common gc invocation
2511 :			sequences. That is to say, we generate:
2512 :
2513 :			f:
2514 :			testgc
2515 :			ja L1 % jump if above to L1
2516 :
2517 :			L1:
2518 :			jmp L2
2519 :
2520 :
2521 :			After jump chain elimination the 'ja L1' instructions is converted to
2522 :			'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2523 :			up being implemented in their long form (if L2 is far away) using:
2524 :
2525 :			jbe L3 % jump if below or equal to L3
2526 :			jmp L2
2527 :			L3:
2528 :			...
2529 :
2530 :
2531 :			For large compilation units L2 may be far away.
2532 :
2533 :
2534 :			----------------------------------------------------------------------
2535 :	blume	1132	Name: Matthias Blume
2536 :			Date: 2002/03/11 13:30:00 EST
2537 :			Tag: blume-20020311-mltreeeval
2538 :			Description:
2539 :
2540 :			A functor parameter was missing.
2541 :
2542 :			----------------------------------------------------------------------
2543 :	leunga	1130	Name: Allen Leung
2544 :	leunga	1131	Date: 2002/03/11 10:30:00 EST
2545 :	leunga	1142	Tag: leunga-20020311-runtime-string0
2546 :	leunga	1131	Description:
2547 :
2548 :			The representation of the empty string now points to a
2549 :			legal null terminated C string instead of unit. It is now possible
2550 :			to convert an ML string into C string with InlineT.CharVector.getData.
2551 :			This compiles into one single machine instruction.
2552 :
2553 :			----------------------------------------------------------------------
2554 :			Name: Allen Leung
2555 :	leunga	1130	Date: 2002/03/10 23:55:00 EST
2556 :			Tag: leunga-20020310-x86-call
2557 :			Description:
2558 :
2559 :			Added machine generation for CALL instruction (relative displacement mode)
2560 :
2561 :			----------------------------------------------------------------------
2562 :	blume	1128	Name: Matthias Blume
2563 :			Date: 2002/03/08 16:05:00
2564 :			Tag: blume-20020308-entrypoints
2565 :			Description:
2566 :
2567 :			Version number bumped to 110.39.1. NEW BOOTFILES!
2568 :
2569 :			Entrypoints: non-zero offset into a code object where execution should begin.
2570 :
2571 :			- Added the notion of an entrypoint to CodeObj.
2572 :			- Added reading/writing of entrypoint info to Binfile.
2573 :			- Made runtime system bootloader aware of entrypoints.
2574 :			- Use the address of the label of the first function given to mlriscGen
2575 :			as the entrypoint. This address is currently always 0, but it will
2576 :			not be 0 once we turn on block placement.
2577 :			- Removed the linkage cluster code (which was The Other Way(tm) of dealing
2578 :			with entry points) from mlriscGen.
2579 :
2580 :			----------------------------------------------------------------------
2581 :	leunga	1127	Name: Allen Leung
2582 :			Date: 2002/03/07 20:45:00 EST
2583 :			Tag: leunga-20020307-x86-cmov
2584 :			Description:
2585 :
2586 :			Bug fixes for CMOVcc on x86.
2587 :
2588 :			1. Added machine code generation for CMOVcc
2589 :			2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2590 :			3. CMOVcc cannot have an immediate operand as argument.
2591 :
2592 :			----------------------------------------------------------------------
2593 :	blume	1126	Name: Matthias Blume
2594 :			Date: 2002/03/07 16:15:00 EST
2595 :			Tag: blume-20020307-controls
2596 :			Description:
2597 :
2598 :			This is a very large but mostly boring patch which makes (almost)
2599 :			every tuneable compiler knob (i.e., pretty much everything under
2600 :			Control.* plus a few other things) configurable via both the command
2601 :			line and environment variables in the style CM did its configuration
2602 :			until now.
2603 :
2604 :			Try starting sml with '-h' (or, if you are brave, '-H')
2605 :
2606 :			To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2607 :			implements the underlying generic mechanism.
2608 :
2609 :			The interface to some of the existing such facilities has changed somewhat.
2610 :			For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2611 :			(The getFoo interface is still there for backward-compatibility, but its
2612 :			use is deprecated.)
2613 :
2614 :			The ml-build script passes -Cxxx=yyy command-line arguments through so
2615 :			that one can now twiddle the compiler settings when using this "batch"
2616 :			compiler.
2617 :
2618 :			TODO items:
2619 :
2620 :			We should go through and throw out all controls that are no longer
2621 :			connected to anything. Moreover, we should go through and provide
2622 :			meaningful (and correct!) documentation strings for those controls
2623 :			that still are connected.
2624 :
2625 :			Currently, multiple calls to Controls.new are accepted (only the first
2626 :			has any effect). Eventually we should make sure that every control
2627 :			is being made (via Controls.new) exactly once. Future access can then
2628 :			be done using Controls.acc.
2629 :
2630 :			Finally, it would probably be a good idea to use the getter-setter
2631 :			interface to controls rather than ref cells. For the time being, both
2632 :			styles are provided by the Controls module, but getter-setter pairs are
2633 :			better if thread-safety is of any concern because they can be wrapped.
2634 :
2635 :			*****************************************
2636 :
2637 :			One bug fix: The function blockPlacement in three of the MLRISC
2638 :			backpatch files used to be hard-wired to one of two possibilities at
2639 :			link time (according to the value of the placementFlag). But (I
2640 :			think) it should rather sense the flag every time.
2641 :
2642 :			*****************************************
2643 :
2644 :			Other assorted changes (by other people who did not supply a HISTORY entry):
2645 :
2646 :			1. the cross-module inliner now works much better (Monnier)
2647 :			2. representation of weights, frequencies, and probabilities in MLRISC
2648 :			changed in preparation of using those for weighted block placement
2649 :			(Reppy, George)
2650 :
2651 :			----------------------------------------------------------------------
2652 :	george	1116	Name: Lal George
2653 :	george	1124	Date: 2002/03/07 14:44:24 EST 2002
2654 :			Tag: george-20020307-weighted-block-placement
2655 :
2656 :			Tested the weighted block placement optimization on all architectures
2657 :			(except the hppa) using AMPL to generate the block and edge frequencies.
2658 :			Changes were required in the machine properties to correctly
2659 :			categorize trap instructions. There is an MLRISC flag
2660 :			"weighted-block-placement" that can be used to enable weighted block
2661 :			placement, but this will be ineffective without block/edge
2662 :			frequencies (coming soon).
2663 :
2664 :
2665 :			----------------------------------------------------------------------
2666 :			Name: Lal George
2667 :	george	1116	Date: 2002/03/05 17:24:48 EST
2668 :			Tag: george-20020305-linkage-cluster
2669 :
2670 :	george	1124	In order to support the block placement optimization, a new cluster
2671 :			is generated as the very first cluster (called the linkage cluster).
2672 :			It contains a single jump to the 'real' entry point for the compilation
2673 :			unit. Block placement has no effect on the linkage cluster itself, but
2674 :			all the other clusters have full freedom in the manner in which they
2675 :			reorder blocks or functions.
2676 :	george	1116
2677 :			On the x86 the typical linkage code that is generated is:
2678 :			----------------------
2679 :			.align 2
2680 :			L0:
2681 :			addl $L1-L0, 72(%esp)
2682 :	george	1124	jmp L1
2683 :	george	1116
2684 :
2685 :			.align 2
2686 :			L1:
2687 :			----------------------
2688 :
2689 :			72(%esp) is the memory location for the stdlink register. This
2690 :			must contain the address of the CPS function being called. In the
2691 :	george	1124	above example, it contains the address of L0; before
2692 :	george	1116	calling L1 (the real entry point for the compilation unit), it
2693 :			must contain the address for L1, and hence
2694 :
2695 :			addl $L1-L0, 72(%esp)
2696 :
2697 :	george	1124	I have tested this on all architectures except the hppa.The increase
2698 :			in code size is of course negligible
2699 :	george	1116
2700 :			----------------------------------------------------------------------
2701 :	leunga	1115	Name: Allen Leung
2702 :			Date: 2002/03/03 13:20:00 EST
2703 :			Tag: leunga-20020303-mlrisc-tools
2704 :
2705 :			Added #[ ... ] expressions to mlrisc tools
2706 :
2707 :			----------------------------------------------------------------------
2708 :	blume	1096	Name: Matthias Blume
2709 :	blume	1098	Date: 2002/02/27 12:29:00 EST
2710 :			Tag: blume-20020227-cdebug
2711 :			Description:
2712 :
2713 :			- made types in structure C and C_Debug to be equal
2714 :			- got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2715 :			- there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2716 :
2717 :			----------------------------------------------------------------------
2718 :			Name: Matthias Blume
2719 :	blume	1096	Date: 2002/02/26 12:00:00 EST
2720 :			Tag: blume-20020226-ffi
2721 :			Description:
2722 :
2723 :			1. Fixed a minor bug in CM's "noweb" tool:
2724 :			If numbering is turned off, then truly don't number (i.e., do not
2725 :			supply the -L option to noweb). The previous behavior was to supply
2726 :			-L'' -- which caused noweb to use the "default" line numbering scheme.
2727 :			Thanks to Chris Richards for pointing this out (and supplying the fix).
2728 :
2729 :			2. Once again, I reworked some aspects of the FFI:
2730 :
2731 :			A. The incomplete/complete type business:
2732 :
2733 :			- Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2734 :			gone!
2735 :			- ML types representing an incomplete type are now *equal* to
2736 :			ML types representing their corresponding complete types (just like
2737 :			in C). This is still safe because ml-nlffigen will not generate
2738 :			RTTI for incomplete types, nor will it generate functions that
2739 :			require access to such RTTI. But when ML code generated from both
2740 :			incomplete and complete versions of the C type meet, the ML types
2741 :			are trivially interoperable.
2742 :
2743 :			NOTE: These changes restore the full generality of the translation
2744 :			(which was previously lost when I eliminated functorization)!
2745 :
2746 :			B. Enum types:
2747 :
2748 :			- Structure C now has a type constructor "enum" that is similar to
2749 :			how the "su" constructor works. However, "enum" is not a phantom
2750 :			type because each "T enum" has values (and is isomorphic to
2751 :			MLRep.Signed.int).
2752 :			- There are generic access operations for enum objects (using
2753 :			MLRep.Signed.int).
2754 :			- ml-nlffigen will generate a structure E_foo for each "enum foo".
2755 :			* The structure contains the definition of type "mlrep" (the ML-side
2756 :			representation type of the enum). Normally, mlrep is the same
2757 :			as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2758 :			then mlrep will be defined as a datatype -- thus facilitating
2759 :			pattern matching on mlrep values.
2760 :			("-ec" will be suppressed if there are duplicate values in an
2761 :			enumeration.)
2762 :			* Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2763 :			will be generated for each C enum constant xxx.
2764 :			* Conversion functions m2i and i2m convert between mlrep and
2765 :			MLRep.Signed.int. (Without "-ec", these functions are identities.)
2766 :			* Coversion functions c and ml convert between mlrep and "tag enum".
2767 :			* Access functions (get/set) fetch and store mlrep values.
2768 :			- By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2769 :			enumerations are merged into one single enumeration represented by
2770 :			structure E_'.
2771 :
2772 :			----------------------------------------------------------------------
2773 :	leunga	1080	Name: Allen Leung
2774 :	leunga	1094	Date: 2002/02/25 04:45:00 EST
2775 :			Tag: leunga-20020225-cps-spill
2776 :
2777 :			This is a new implementation of the CPS spill phase.
2778 :			The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2779 :			In case of problems, replace it with the old file spill.sml
2780 :
2781 :			The current compiler runs into some serious performance problems when
2782 :			constructing a large record. This can happen when we try to compile a
2783 :			structure with many items. Even a very simple structure like the following
2784 :			makes the compiler slow down.
2785 :
2786 :			structure Foo = struct
2787 :			val x_1 = 0w1 : Word32.int
2788 :			val x_2 = 0w2 : Word32.int
2789 :			val x_3 = 0w3 : Word32.int
2790 :			...
2791 :			val x_N = 0wN : Word32.int
2792 :			end
2793 :
2794 :			The following table shows the compile time, from N=1000 to N=4000,
2795 :			with the old compiler:
2796 :
2797 :			N
2798 :			1000 CPS 100 spill 0.04u 0.00s 0.00g
2799 :			MLRISC ra 0.06u 0.00s 0.05g
2800 :			(spills = 0 reloads = 0)
2801 :			TOTAL 0.63u 0.07s 0.21g
2802 :
2803 :			1100 CPS 100 spill 8.25u 0.32s 0.64g
2804 :			MLRISC ra 5.68u 0.59s 3.93g
2805 :			(spills = 0 reloads = 0)
2806 :			TOTAL 14.71u 0.99s 4.81g
2807 :
2808 :			1500 CPS 100 spill 58.55u 2.34s 1.74g
2809 :			MLRISC ra 5.54u 0.65s 3.91g
2810 :			(spills = 543 reloads = 1082)
2811 :			TOTAL 65.40u 3.13s 6.00g
2812 :
2813 :			2000 CPS 100 spill 126.69u 4.84s 3.08g
2814 :			MLRISC ra 0.80u 0.10s 0.55g
2815 :			(spills = 42 reloads = 84)
2816 :			TOTAL 129.42u 5.10s 4.13g
2817 :
2818 :			3000 CPS 100 spill 675.59u 19.03s 11.64g
2819 :			MLRISC ra 2.69u 0.27s 1.38g
2820 :			(spills = 62 reloads = 124)
2821 :			TOTAL 682.48u 19.61s 13.99g
2822 :
2823 :			4000 CPS 100 spill 2362.82u 56.28s 43.60g
2824 :			MLRISC ra 4.96u 0.27s 2.72g
2825 :			(spills = 85 reloads = 170)
2826 :			TOTAL 2375.26u 57.21s 48.00g
2827 :
2828 :			As you can see the old cps spill module suffers from some serious
2829 :	leunga	1115	performance problem. But since I cannot decipher the old code fully,
2830 :			instead of patching the problems up, I'm reimplementing it
2831 :	leunga	1094	with a different algorithm. The new code is more modular,
2832 :			smaller when compiled, and substantially faster
2833 :			(O(n log n) time and O(n) space). Timing of the new spill module:
2834 :
2835 :			4000 CPS 100 spill 0.02u 0.00s 0.00g
2836 :			MLRISC ra 0.25u 0.02s 0.15g
2837 :			(spills=1 reloads=3)
2838 :			TOTAL 7.74u 0.34s 1.62g
2839 :
2840 :			Implementation details:
2841 :
2842 :			As far as I can tell, the purpose of the CPS spill module is to make sure the
2843 :			number of live variables at any program point (the bandwidth)
2844 :			does not exceed a certain limit, which is determined by the
2845 :			size of the spill area.
2846 :
2847 :			When the bandwidth is too large, we decrease the register pressure by
2848 :			packing live variables into spill records. How we achieve this is
2849 :			completely different than what we did in the old code.
2850 :
2851 :			First, there is something about the MLRiscGen code generator
2852 :			that we should be aware of:
2853 :
2854 :			o MLRiscGen performs code motion!
2855 :
2856 :			In particular, it will move floating point computations and
2857 :			address computations involving only the heap pointer to
2858 :			their use sites (if there is only a single use).
2859 :			What this means is that if we have a CPS record construction
2860 :			statement
2861 :
2862 :			RECORD(k,vl,w,e)
2863 :
2864 :			we should never count the new record address w as live if w
2865 :			has only one use (which is often the case).
2866 :
2867 :			We should do something similar to floating point, but the transformation
2868 :			there is much more complex, so I won't deal with that.
2869 :
2870 :			Secondly, there are now two new cps primops at our disposal:
2871 :
2872 :			1. rawrecord of record_kind option
2873 :			This pure operator allocates some uninitialized storage from the heap.
2874 :			There are two forms:
2875 :
2876 :			rawrecord NONE [INT n] allocates a tagless record of length n
2877 :			rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2878 :			and initializes the tag.
2879 :
2880 :			2. rawupdate of cty
2881 :			rawupdate cty (v,i,x)
2882 :			Assigns to x to the ith component of record v.
2883 :			The storelist is not updated.
2884 :
2885 :			We use these new primops for both spilling and increment record construction.
2886 :
2887 :			1. Spilling.
2888 :
2889 :			This is implemented with a linear scan algorithm (but generalized
2890 :			to trees). The algorithm will create a single spill record at the
2891 :			beginning of the cps function and use rawupdate to spill to it,
2892 :			and SELECT or SELp to reload from it. So both spills and reloads
2893 :			are fine-grain operations. In contrast, in the old algorithm
2894 :			"spills" have to be bundled together in records.
2895 :
2896 :			Ideally, we should sink the spill record construction to where
2897 :			it is needed. We can even split the spill record into multiple ones
2898 :			at the places where they are needed. But CPS is not a good
2899 :			representation for global code motion, so I'll keep it simple and
2900 :			am not attempting this.
2901 :
2902 :			2. Incremental record construction (aka record splitting).
2903 :
2904 :			Long records with many component values which are simulatenously live
2905 :			(recall that single use record addresses are not considered to
2906 :			be live) are constructed with rawrecord and rawupdate.
2907 :			We allocate space on the heap with rawrecord first, then gradually
2908 :			fill it in with rawupdate. This is the technique suggested to me
2909 :			by Matthias.
2910 :
2911 :			Some restrictions on when this is applicable:
2912 :			1. It is not a VECTOR record. The code generator currently does not handle
2913 :			this case. VECTOR record uses double indirection like arrays.
2914 :			2. All the record component values are defined in the same "basic block"
2915 :			as the record constructor. This is to prevent speculative
2916 :			record construction.
2917 :
2918 :			----------------------------------------------------------------------
2919 :			Name: Allen Leung
2920 :	leunga	1086	Date: 2002/02/22 01:02:00 EST
2921 :			Tag: leunga-20020222-mlrisc-tools
2922 :
2923 :			Minor bug fixes in the parser and rewriter
2924 :
2925 :			----------------------------------------------------------------------
2926 :			Name: Allen Leung
2927 :	leunga	1085	Date: 2002/02/21 20:20:00 EST
2928 :			Tag: leunga-20020221-peephole
2929 :
2930 :			Regenerated the peephole files. Some contained typos in the specification
2931 :			and some didn't compile because of pretty printing bugs in the old version
2932 :			of 'nowhere'.
2933 :
2934 :			----------------------------------------------------------------------
2935 :			Name: Allen Leung
2936 :	leunga	1079	Date: 2002/02/19 20:20:00 EST
2937 :			Tag: leunga-20020219-mlrisc-tools
2938 :			Description:
2939 :
2940 :			Minor bug fixes to the mlrisc-tools library:
2941 :
2942 :			1. Fixed up parsing colon suffixed keywords
2943 :			2. Added the ability to shut the error messages up
2944 :			3. Reimplemented the pretty printer and fixed up/improved
2945 :			the pretty printing of handle and -> types.
2946 :			4. Fixed up generation of literal symbols in the nowhere tool.
2947 :			5. Added some SML keywords to to sml.sty
2948 :
2949 :			----------------------------------------------------------------------
2950 :			Name: Matthias Blume
2951 :	blume	1078	Date: 2002/02/19 16:20:00 EST
2952 :			Tag: blume-20020219-cmffi
2953 :			Description:
2954 :
2955 :			A wild mix of changes, some minor, some major:
2956 :
2957 :			* All C FFI-related libraries are now anchored under $c:
2958 :			$/c.cm --> $c/c.cm
2959 :			$/c-int.cm --> $c/internals/c-int.cm
2960 :			$/memory.cm --> $c/memory/memory.cm
2961 :
2962 :			* "make" tool (in CM) now treats its argument pathname slightly
2963 :			differently:
2964 :			1. If the native expansion is an absolute name, then before invoking
2965 :			the "make" command on it, CM will apply OS.Path.mkRelative
2966 :			(with relativeTo = OS.FileSys.getDir()) to it.
2967 :			2. The argument will be passed through to subsequent phases of CM
2968 :			processing without "going native". In particular, if the argument
2969 :			was an anchored path, then "make" will not lose track of that anchor.
2970 :
2971 :			* Compiler backends now "know" their respective C calling conventions
2972 :			instead of having to be told about it by ml-nlffigen. This relieves
2973 :			ml-nlffigen from one of its burdens.
2974 :
2975 :			* The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
2976 :
2977 :			* Export C_DEBUG and C_Debug from $c/c.cm.
2978 :
2979 :			* C type encoding in ml-nlffi-lib has been improved to model the conceptual
2980 :			subtyping relationship between incomplete pointers and their complete
2981 :			counterparts. For this, ('t, 'c) ptr has been changed to 'o ptr --
2982 :			with the convention of instantiating 'o with ('t, 'c) obj whenever
2983 :			the pointer target type is complete. In the incomplete case, 'o
2984 :			will be instantiated with some "'c iobj" -- a type obtained by
2985 :			using one of the functors PointerToIncompleteType or PointerToCompleteType.
2986 :
2987 :			Operations that work on both incomplete and complete pointer types are
2988 :			typed as taking an 'o ptr while operations that require the target to
2989 :			be known are typed as taking some ('t, 'c) obj ptr.
2990 :
2991 :			voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
2992 :			where void is an eqtype without any values. This makes it possible
2993 :			to work on voidptr values using functions meant to operate on light
2994 :			incomplete pointers.
2995 :
2996 :			* As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
2997 :			been vastly simplified.
2998 :
2999 :			----------------------------------------------------------------------
3000 :			Name: Matthias Blume
3001 :	blume	1076	Date: 2002/02/19 10:48:00 EST
3002 :			Tag: blume-20020219-pqfix
3003 :			Description:
3004 :
3005 :			Applied Chris Okasaki's bug fix for priority queues.
3006 :
3007 :			----------------------------------------------------------------------
3008 :			Name: Matthias Blume
3009 :	blume	1073	Date: 2002/02/15 17:05:00
3010 :	blume	1069	Tag: Release_110_39
3011 :			Description:
3012 :
3013 :	blume	1073	Last-minute retagging is becoming a tradition... :-(
3014 :
3015 :			This is the working release 110.39.
3016 :
3017 :			----------------------------------------------------------------------
3018 :			Name: Matthias Blume
3019 :			Date: 2002/02/15 16:00:00 EST
3020 :			Tag: Release_110_39-orig
3021 :			Description:
3022 :
3023 :	blume	1069	Working release 110.39. New bootfiles.
3024 :
3025 :	blume	1073	(Update: There was a small bug in the installer so it wouldn't work
3026 :			with all shells. So I retagged. -Matthias)
3027 :
3028 :	blume	1069	----------------------------------------------------------------------
3029 :			Name: Matthias Blume
3030 :	blume	1068	Date: 2002/02/15 14:17:00 EST
3031 :			Tag: blume-20020215-showbindings
3032 :			Description:
3033 :
3034 :			Added EnvRef.listBoundSymbols and CM.State.showBindings. Especially
3035 :			the latter can be useful for exploring what bindings are available at
3036 :			the interactive prompt. (The first function returns only the list
3037 :			of symbols that are really bound, the second prints those but also the
3038 :			ones that CM's autoloading mechanism knows about.)
3039 :
3040 :			----------------------------------------------------------------------
3041 :			Name: Matthias Blume
3042 :	blume	1067	Date: 2002/02/15 12:08:00 EST
3043 :			Tag: blume-20020215-iptrs
3044 :			Description:
3045 :
3046 :			Two improvements to ml-nlffigen:
3047 :
3048 :			1. Write files only if they do not exist or if their current contents
3049 :			do not coincide with what's being written. (That is, avoid messing
3050 :			with the time stamps unless absolutely necessary.)
3051 :
3052 :			2. Implement a "repository" mechanism for generated files related
3053 :			to "incomplete pointer types". See the README file for details.
3054 :
3055 :			----------------------------------------------------------------------
3056 :			Name: Matthias Blume
3057 :	blume	1066	Date: 2002/02/14 11:50:00 EST
3058 :			Tag: blume-20020214-quote
3059 :			Description:
3060 :
3061 :			Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm). This is required
3062 :			because of the new and improved tag generation scheme. (Thanks to Allen
3063 :			Leung for pointing it out.)
3064 :
3065 :			----------------------------------------------------------------------
3066 :	george	1065	Name: Lal George
3067 :			Date: 2002/02/14 09:55:27 EST 2002
3068 :			Tag: george-20020214-isabelle-bug
3069 :			Description:
3070 :
3071 :			Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3072 :			of Isabelle on the x86.
3073 :
3074 :			From Allen:
3075 :			-----------
3076 :			I've found the problem:
3077 :
3078 :			in ra-core.sml, I use the counter "blocked" to keep track of the
3079 :			true number of elements in the freeze queue. When the counter goes
3080 :			to zero, I skip examining the queue. But I've messed up the
3081 :			bookkeeping in combine():
3082 :
3083 :			else ();
3084 :			case !ucol of
3085 :			PSEUDO => (if !cntv > 0 then
3086 :			(if !cntu > 0 then blocked := !blocked - 1 else ();
3087 :			^^^^^^^^^^^^^^^^^^^^^^^
3088 :			moveu := mergeMoveList(!movev, !moveu)
3089 :			)
3090 :			else ();
3091 :
3092 :			combine() is called to coalesce two nodes u and v.
3093 :			I think I was thinking that if the move counts of u and v are both
3094 :			greater than zero then after they are coalesced then one node is
3095 :			removed from the freeze queue. Apparently I was thinking that
3096 :			both u and v are of low degree, but that's clearly not necessarily true.
3097 :
3098 :
3099 :			02/12/2002:
3100 :			Here's the patch. HOL now compiles.
3101 :
3102 :			I don't know how this impact on performance (compile
3103 :			time or runtime). This bug caused the RA (especially on the x86)
3104 :			to go thru the potential spill phase when there are still nodes on the
3105 :			freeze queue.
3106 :
3107 :
3108 :
3109 :
3110 :			----------------------------------------------------------------------
3111 :	blume	1058	Name: Matthias Blume
3112 :	blume	1064	Date: 2002/02/13 22:40:00 EST
3113 :			Tag: blume-20020213-fptr-rtti
3114 :			Description:
3115 :
3116 :			Fixed a bug in ml-nlffigen that was introduced with one of the previous
3117 :			updates.
3118 :
3119 :			----------------------------------------------------------------------
3120 :			Name: Matthias Blume
3121 :	blume	1063	Date: 2002/02/13 16:41:00 EST
3122 :			Tag: blume-20020213-cmlpq
3123 :			Description:
3124 :
3125 :			Added new priority queue export symbols (which have just been added to
3126 :			smlnj-lib.cm) to CML's version of smlnj-lib.cm. (Otherwise CML would
3127 :			not compile and the installer would choke.)
3128 :
3129 :			----------------------------------------------------------------------
3130 :			Name: Matthias Blume
3131 :	blume	1062	Date: 2002/02/13 16:15:00 EST
3132 :			Tag: blume-20020213-various
3133 :			Description:
3134 :
3135 :			1. More tweaks to ml-nlffigen:
3136 :
3137 :			- better internal datastructures (resulting in slight speedup)
3138 :			- "-match" option requires exact match
3139 :			- "localized" gensym counters (untagged structs/unions nested within
3140 :			other structs/unions or within typedefs get a fresh counter; their
3141 :			tag will be prefixed by a concatenation of their parents' tags)
3142 :			- bug fixes (related to calculation of transitive closure of types
3143 :			to be included in the output)
3144 :
3145 :			2. Minor Basis updates:
3146 :
3147 :			- added implementations for List.collate and Option.app
3148 :
3149 :			----------------------------------------------------------------------
3150 :			Name: Matthias Blume
3151 :	blume	1060	Date: 2002/02/11 15:55:00 EST
3152 :			Tag: blume-20020211-gensym
3153 :			Description:
3154 :
3155 :			Added a "-gensym" option to command line of ml-nlffigen. This can be
3156 :			used to specify a "stem" -- a string that is inserted in all "gensym'd"
3157 :			names (ML structure names that correspond to unnamed C structs, unions,
3158 :			and enums), so that separate runs of ml-nlffigen do not clash.
3159 :
3160 :			----------------------------------------------------------------------
3161 :			Name: Matthias Blume
3162 :	blume	1059	Date: 2002/02/11 12:05:00 EST
3163 :			Tag: blume-20020211-gensml
3164 :			Description:
3165 :
3166 :			A quick fix for a problem with GenSML (in the pgraph-util library):
3167 :			Make generation of toplevel "local" optional. (Strictly speaking,
3168 :			signature definitions within "local" are not legal SML.)
3169 :
3170 :			Other than that: updates to INSTALL and cm/TODO.
3171 :
3172 :			----------------------------------------------------------------------
3173 :			Name: Matthias Blume
3174 :	blume	1058	Date: 2002/02/08 15:00:00 EST
3175 :			Tag: blume-20020208-uniquepid
3176 :			Description:
3177 :
3178 :			0. Version number has been bumped to 110.38.1. NEW BOOTFILES!!!
3179 :
3180 :			1. The installer (config/install.sh) has gotten smarter:
3181 :
3182 :			- Configuration options are a bit easier to specify now
3183 :			(in config/targets).
3184 :			- Bug in recognizing .tar.bz2 files fixed.
3185 :			- Installer automatically resolves dependencies between
3186 :			configuration options (e.g., if you ask for eXene, you will
3187 :			also get cml -- regardless whether you asked for it or not).
3188 :			- Installer can run in "quieter mode" by setting the environment
3189 :			variable INSTALL_QUIETLY to "true". "Quieter" does not mean
3190 :			"completely silent", though.
3191 :			- Build HashCons library as part of smlnj-lib.
3192 :
3193 :			2. A new scheme for assigning persistent identifiers to compilation
3194 :			units (and, by extension, to types etc.) has been put into place.
3195 :			This fixes a long-standing bug where types and even dynamic values
3196 :			can get internally confused, thereby compromising type safety
3197 :			(abstraction) and dynamic correctness. See
3198 :
3199 :			http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3200 :
3201 :			for an example of how things could go wrong until now.
3202 :
3203 :			The downside of the new scheme is that pids are not quite as
3204 :			persistent as they used to be: CM will generate a fresh pid
3205 :			for every compilation unit that it thinks it sees for the first
3206 :			time. That means that if you compile starting from a clean, fresh
3207 :			source tree at two different times, you end up with different
3208 :			binaries.
3209 :
3210 :			Cutoff recompilation, however, has not been compromised because
3211 :			CM keeps pid information in special caches between runs.
3212 :
3213 :			----------------------------------------------------------------------
3214 :	george	1053	Name: Lal George
3215 :	george	1055	Date: 2002/02/07 15:34:13 EST 2002
3216 :			Tag: <none>
3217 :			Description:
3218 :
3219 :			Compilers that generate assembly code may produce global labels
3220 :			whose value is resolved at link time. The various peephole optimization
3221 :			modules did not take this in account.
3222 :
3223 :			TODO. The Labels.addrOf function should really return an option
3224 :			type so that clients are forced to deal with this issue, rather
3225 :			than an exception being raised.
3226 :
3227 :			----------------------------------------------------------------------
3228 :			Name: Lal George
3229 :	george	1053	Date: 2002/02/06 13:55:02 EST
3230 :			Tag: george-20020206-ra-breakup
3231 :			Description:
3232 :
3233 :			1. A bug fix from Allen.
3234 :
3235 :			A typo causes extra fstp %st(0)'s to be generated at compensation
3236 :			edges, which might cause stack underflow traps at runtime. This
3237 :			occurs in fft where there are extraneous fstps right before the 'into'
3238 :			trap instruction (in this case they are harmless since none of the
3239 :			integers overflow.)
3240 :
3241 :	george	1054	2. Pulled out various utility modules that were embedded in the modules
3242 :	george	1053	of the register allocator. I need these modules for other purposes, but
3243 :	george	1054	they are not complete enough to put into a library (just yet).
3244 :	george	1053	----------------------------------------------------------------------
3245 :	blume	1049	Name: Matthias Blume
3246 :			Date: 2002/01/31 16:05:00 EST
3247 :			Tag: blume-20020131-sparc-ccalls
3248 :			Description:
3249 :
3250 :			1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3251 :			of extra stack space by mistake. Fixed.
3252 :
3253 :			2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3254 :
3255 :			----------------------------------------------------------------------
3256 :	leunga	1048	Name: Allen Leung
3257 :			Date: 2002/01/30
3258 :			Tag: leunga-20020130-nowhere-bug-fix
3259 :			Description:
3260 :
3261 :			MLRISC bug fixes:
3262 :			1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3263 :			2. MachineInt.fromString was negating its value.
3264 :
3265 :			----------------------------------------------------------------------
3266 :	blume	1039	Name: Matthias Blume
3267 :	blume	1047	Date: 2002/01/29
3268 :			Tag: blume-20020129-INSTALL
3269 :			Description:
3270 :
3271 :			- Added somewhat detailed installation instructions (file INSTALL).
3272 :			- Fixed curl-detection bug in config/install.sh.
3273 :			- It is now possible to select the URL getter using the URLGETTER
3274 :			environment variable:
3275 :
3276 :			not set / "unknown" --> automatic detection (script tries wget,
3277 :			curl, and lynx)
3278 :			"wget" / "curl" / "lynx" --> use the specified program (script "knows"
3279 :			how to properly invoke them)
3280 :			other --> use $URLGETTER directly, it must take
3281 :			precisely two command-line arguments
3282 :			(source URL and destination file name)
3283 :
3284 :			----------------------------------------------------------------------
3285 :			Name: Matthias Blume
3286 :	blume	1044	Date: 2002/01/28
3287 :	blume	1045	Tag: blume-20020128-sparc-ccalls
3288 :			Description:
3289 :
3290 :			- Fixed problem with calculation of "used" registers in sparc-c-calls.
3291 :			- Make use of the allocParam argument in sparc-c-calls.
3292 :
3293 :			----------------------------------------------------------------------
3294 :			Name: Matthias Blume
3295 :			Date: 2002/01/28
3296 :	blume	1044	Tag: blume-20020128-allocParam
3297 :			Description:
3298 :
3299 :			John Reppy: Changes c-calls API to accept client-callback for
3300 :			allocating extra stack space.
3301 :			me: Corresponding changes to mlriscGen (using a dummy argument that
3302 :			does not change the current behavior).
3303 :
3304 :			----------------------------------------------------------------------
3305 :			Name: Matthias Blume
3306 :	blume	1040	Date: 2002/01/28 12:00:00
3307 :			Tag: Release_110_38
3308 :			Description:
3309 :
3310 :			This time for real!!!
3311 :
3312 :			----------------------------------------------------------------------
3313 :			Name: Matthias Blume
3314 :	blume	1039	Date: 2002/01/28 10:56:00 EST
3315 :			Tag: blume-20020128-retraction
3316 :			Description:
3317 :
3318 :			0. Retracted earlier 110.38. (The Release_110_38 tag has been replaced
3319 :			with blume-Release_110_38-retracted.)
3320 :
3321 :			1. Fixed a problem with incorrect rounding modes in real64.sml.
3322 :			(Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3323 :
3324 :			2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3325 :			and enums fixed. The naming of corresponding ML identifiers should
3326 :			now be consistent again.
3327 :
3328 :			----------------------------------------------------------------------
3329 :	leunga	1038	Name: Allen Leung
3330 :			Date: 2002/01/27
3331 :			Tag: leunga-20020127-nowhere
3332 :			Description:
3333 :
3334 :			Added a target called nowhere in the configuration scripts.
3335 :			Enabling this will build the MLRISC 'nowhere' tool (for translating
3336 :			programs with where-clauses into legal SML code) during installation.
3337 :
3338 :			----------------------------------------------------------------------
3339 :	blume	1036	Name: Matthias Blume
3340 :	leunga	1038	Date: 2002/01/25 21:27:00 EST
3341 :	blume	1039	Tag: blume-Release_110_38-retracted
3342 :	blume	1037	Description:
3343 :
3344 :			Call it a (working) release! Version is 110.38. Bootfiles are ready.
3345 :
3346 :			README will be added later.
3347 :
3348 :	blume	1039	!!! NOTE: Re-tagged as blume-Release_110_38-retracted. Original tag
3349 :			(Release_110_38) removed. Reason: Last-minute bug fixes.
3350 :
3351 :	blume	1037	----------------------------------------------------------------------
3352 :			Name: Matthias Blume
3353 :	blume	1036	Date: 2002/01/25
3354 :			Tag: blume-20020125-ffi
3355 :			Description:
3356 :
3357 :			A large number of tweaks and improvements to ml-nlffi-lib and
3358 :			ml-nlffigen:
3359 :
3360 :			- ML represenation types have been streamlined
3361 :			- getter and setter functions work with concrete values, not abstract
3362 :			ones where possible
3363 :			- ml-nlffigen command line more flexible (see README file there)
3364 :			- some bugs have been fixed (hopefully)
3365 :
3366 :			----------------------------------------------------------------------
3367 :	george	1034	Name: Lal George
3368 :			Date: 2002/01/24
3369 :			Tag: george-20020124-risc-ra-interface
3370 :			Description:
3371 :
3372 :			There is a dramatic simplification in the interface to the
3373 :			register allocator for RISC architectures as a result of making
3374 :			parallel copy instructions explicit.
3375 :
3376 :			----------------------------------------------------------------------
3377 :	blume	1028	Name: Matthias Blume
3378 :	blume	1032	Date: 2002/01/22
3379 :			Tag: blume-20020122-x86-ccalls
3380 :			Description:
3381 :
3382 :			Bug fix for c-calls on x86 (having to do with how char- and
3383 :			short-arguments are being handled).
3384 :
3385 :			----------------------------------------------------------------------
3386 :			Name: Matthias Blume
3387 :	blume	1031	Date: 2002/01/21
3388 :			Tag: blume-20020121-ff
3389 :			Description:
3390 :
3391 :			Another day of fiddling with the FFI...
3392 :
3393 :			1. Bug fix/workaround: CKIT does not complain about negative array
3394 :			dimensions, so ml-nlffigen has to guard itself against this possibility.
3395 :			(Otherwise a negative dimension would send it into an infinite loop.)
3396 :
3397 :			2. Some of the abstract types (light objects, light pointers, most "base"
3398 :			types) in structure C are now eqtypes.
3399 :
3400 :			3. Added constructors and test functions for NULL function pointers.
3401 :
3402 :			----------------------------------------------------------------------
3403 :			Name: Matthias Blume
3404 :	blume	1028	Date: 2002/01/18
3405 :	blume	1030	Tag: blume-20020118-ready-for-new-release
3406 :			Description:
3407 :
3408 :			Made config/srcarchiveurl point to a new place. (Will provide boot
3409 :			files shortly.)
3410 :
3411 :			Maybe we christen this to be 110.38?
3412 :
3413 :			----------------------------------------------------------------------
3414 :			Name: Matthias Blume
3415 :			Date: 2002/01/18
3416 :	blume	1029	Tag: blume-20020118-more-ffifiddle
3417 :			Description:
3418 :
3419 :			Today's FFI fiddling:
3420 :
3421 :			- Provided a structure CGetSet with "convenient" versions of C.Get.* and
3422 :			C.Set.* that use concrete (MLRep.*) arguments and results instead
3423 :			of abstract ones.
3424 :
3425 :			- Provided word-style bit operations etc. for "int" representation
3426 :			types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3427 :			and Long.
3428 :
3429 :			----------------------------------------------------------------------
3430 :			Name: Matthias Blume
3431 :			Date: 2002/01/18
3432 :	blume	1028	Tag: blume-20020118-use-x86-fp
3433 :			Description:
3434 :
3435 :			Now that x86-fast-fp seems to be working, I turned it back on again
3436 :			by default. (Seems to work fine now, even with the FFI.)
3437 :
3438 :			Other than that, I added some documentation about the FFI to
3439 :			src/ml-nlffigen/README and updated the FFI test examples in
3440 :			src/ml-nlffi-lib/Tests/*.
3441 :
3442 :			----------------------------------------------------------------------
3443 :	leunga	1025	Name: Allen Leung
3444 :			Date: 2002/01/17
3445 :			Tag: leunga-20020117-x86-fast-fp-call
3446 :			Description:
3447 :
3448 :			1. Fixed a problem with handling return fp values when x86's fast fp
3449 :			mode is turned on.
3450 :
3451 :			2. Minor pretty printing fix for cellset. Print %st(0) as %st(0) instead
3452 :			of %f32.
3453 :
3454 :			3. Added a constructor INT32lit to the ast of MLRISC tools.
3455 :
3456 :			----------------------------------------------------------------------
3457 :	blume	1021	Name: Matthias Blume
3458 :			Date: 2002/01/16
3459 :			Tag: blume-20020116-ffifiddle
3460 :			Description:
3461 :
3462 :			More fiddling with the FFI interface:
3463 :
3464 :			- Make constness 'c instead of rw wherever possible. This eliminates
3465 :			the need for certain explicit coercions. (However, due to ML's
3466 :			value polymorphism, there will still be many cases where explicit
3467 :			coercions are necessary. Phantom types are not the whole answer
3468 :			to modeling a subtyping relationship in ML.)
3469 :
3470 :			- ro/rw coersions for pointers added. (Avoids the detour through */&.)
3471 :
3472 :			- "printf" test example added to src/ml-nlffi-lib/Tests. (Demonstrates
3473 :			clumsy workaround for varargs problem.)
3474 :
3475 :			----------------------------------------------------------------------
3476 :	george	1016	Name: Lal George
3477 :			Date: 2002/01/15
3478 :			Tag: <none>
3479 :			Description:
3480 :
3481 :			1. Since COPY instructions are no longer native to the architecture,
3482 :			a generic functor can be used to implement the expandCopies function.
3483 :
3484 :			2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3485 :			TEXT segment.
3486 :
3487 :			----------------------------------------------------------------------
3488 :	blume	1011	Name: Matthias Blume
3489 :	blume	1015	Date: 2002/01/15
3490 :			Tag: blume-20020115-ffiupdates
3491 :			Description:
3492 :
3493 :			1. Fix for bug resulting in single-precision float values being returned
3494 :			incorrectly from FFI calls.
3495 :
3496 :			2. Small modifications to C FFI API:
3497 :
3498 :			- memory-allocation routines return straight objects (no options)
3499 :			and raise an exception in out-of-memory situations
3500 :			- unsafe extensions to cast between function pointers and pointers
3501 :			from/to ints
3502 :			- added structure C_Debug as an alternative to structure C where
3503 :			pointer-dereferencing (|*| and |*!) always check for null-pointers
3504 :			- added open_lib' to DynLinkage; open_lib' works like open_lib
3505 :			but also takes a (possibly empty) list of existing library handles
3506 :			that the current library depends on
3507 :
3508 :			----------------------------------------------------------------------
3509 :			Name: Matthias Blume
3510 :	blume	1011	Date: 2002/01/10
3511 :			Tag: blume-20020110-newffigen
3512 :			Description:
3513 :
3514 :			1. Updates to portable graph code.
3515 :
3516 :			2. Major update to ml-nlffigen and ml-nlffi-lib. Things are much
3517 :			more scalable now so that even huge interfaces such as the one
3518 :			for GTK compile in finite time and space. :-)
3519 :			See src/ml-nlffigen/README for details on what's new.
3520 :
3521 :			----------------------------------------------------------------------
3522 :	george	1003	Name: Lal George
3523 :	george	1009	Date: 2001/01/09 14:31:35 EST 2002
3524 :			Tag: george-20011206-rm-native-copy
3525 :			Description:
3526 :
3527 :			Removed the native COPY and FCOPY instructions
3528 :			from all the architectures and replaced it with the
3529 :			explicit COPY instruction from the previous commit.
3530 :
3531 :			It is now possible to simplify many of the optimizations
3532 :			modules that manipulate copies. This has not been
3533 :			done in this change.
3534 :
3535 :			----------------------------------------------------------------------
3536 :			Name: Lal George
3537 :	george	1003	Date: 2001/12/06 16:50:13 EST 2001
3538 :			Tag: george-20011206-mlrisc-instruction
3539 :			Description:
3540 :	george	824
3541 :	george	1003	Changed the representation of instructions from being fully abstract
3542 :			to being partially concrete. That is to say:
3543 :
3544 :			from
3545 :			type instruction
3546 :
3547 :			to
3548 :			type instr (* machine instruction *)
3549 :
3550 :			datatype instruction =
3551 :			LIVE of {regs: C.cellset, spilled: C.cellset}
3552 :			| KILL of {regs: C.cellset, spilled: C.cellset}
3553 :			| COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3554 :			| ANNOTATION of {i: instruction, a: Annotations.annotation}
3555 :			| INSTR of instr
3556 :
3557 :			This makes the handling of certain special instructions that appear on
3558 :			all architectures easier and uniform.
3559 :
3560 :			LIVE and KILL say that a list of registers are live or killed at the
3561 :			program point where they appear. No spill code is generated when an
3562 :			element of the 'regs' field is spilled, but the register is moved to
3563 :			the 'spilled' (which is present, more for debugging than anything else).
3564 :
3565 :			LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3566 :			We used to generate:
3567 :
3568 :			DEFFREG f1
3569 :			f1 := f2 + f3
3570 :			trapb
3571 :
3572 :			but now generate:
3573 :
3574 :			f1 := f2 + f3
3575 :			trapb
3576 :			LIVE {regs=[f1,f2,f3], spilled=[]}
3577 :
3578 :			Furthermore, the DEFFREG (hack) required that all floating point instruction
3579 :			use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3580 :			defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3581 :			in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3582 :			this architecture).
3583 :
3584 :			COPYXXX is intended to replace the parallel COPY and FCOPY available on
3585 :			all the architectures. This will result in further simplification of the
3586 :			register allocator that must be aware of them for coalescing purposes, and
3587 :			will also simplify certain aspects of the machine description that provides
3588 :			callbacks related to parallel copies.
3589 :
3590 :			ANNOTATION should be obvious, and now INSTR represents the honest to God
3591 :			machine instruction set!
3592 :
3593 :			The <arch>/instructions/<arch>Instr.sml files define certain utility
3594 :			functions for making porting easier -- essentially converting upper case
3595 :			to lower case. All machine instructions (of type instr) are in upper case,
3596 :			and the lower case form generates an MLRISC instruction. For example on
3597 :			the alpha we have:
3598 :
3599 :			datatype instr =
3600 :			LDA of {r:cell, b:cell, d:operand}
3601 :			| ...
3602 :
3603 :			val lda : {r:cell, b:cell, d:operand} -> instruction
3604 :			...
3605 :
3606 :			where lda is just (INSTR o LDA), etc.
3607 :
3608 :	leunga	815	----------------------------------------------------------------------
3609 :	blume	986	Name: Matthias Blume
3610 :	blume	993	Date: 2001/11/22 21:40:00 EST
3611 :			Tag: Release_110_37
3612 :			Description:
3613 :
3614 :			Release 110.37. This time for real.
3615 :
3616 :			----------------------------------------------------------------------
3617 :			Name: Matthias Blume
3618 :	blume	988	Date: 2001/11/21 16:35:00 EST
3619 :			Tag: blume-20011121-foot-in-mouth
3620 :			Description:
3621 :
3622 :			Removed the "Release_110_37" tag because of a serious bug.
3623 :			This will be re-tagged once the bug is fixed.
3624 :
3625 :			----------------------------------------------------------------------
3626 :			Name: Matthias Blume
3627 :	blume	987	Date: 2001/11/21 16:14:00 EST
3628 :			Tag: blume-20011121-forgottenfile
3629 :			Description:
3630 :
3631 :			Forgot to add a file. (Just a .tex-file -- part of
3632 :			the CM manual source.)
3633 :
3634 :			----------------------------------------------------------------------
3635 :			Name: Matthias Blume
3636 :	blume	986	Date: 2001/11/21 16:10:00 EST
3637 :	blume	988	Tag: blume-20011121-invalid_110_37
3638 :	blume	986	Description:
3639 :
3640 :	blume	988	Note: I removed the original tag "Release_110_37" from this commit
3641 :			because we found a serious bug in all non-x86 backends.
3642 :			- Matthias
3643 :
3644 :	blume	986	1. Modifications to the SML/NJ code generator and to the runtime system
3645 :			so that code object name strings are directly inserted into code
3646 :			objects at code generation time. The only business the runtime system
3647 :			has with this is now to read the name strings on occasions.
3648 :			(The encoding of the name string has also changed somewhat.)
3649 :
3650 :			2. CM now implements a simple "set calculus" for specifying export lists.
3651 :			In particular, it is now possible to refer to the export lists of
3652 :			other libraries/groups/sources and form unions as well as differences.
3653 :			See the latest CM manual for details.
3654 :
3655 :			3. An separate notion of "proxy" libraries has again be eliminated from
3656 :			CM's model. (Proxy libraries are now simply a special case of using
3657 :			the export list calculus.)
3658 :
3659 :			4. Some of the existing libraries now take advantage of the new set
3660 :			calculus.
3661 :			(Notice that not all libraries have been converted because some
3662 :			of the existing .cm-files are supposed to be backward compatible
3663 :			with 110.0.x.)
3664 :
3665 :			5. Some cleanup in stand-alone programs. (Don't use "exnMessage" -- use
3666 :			"General.exnMessage"! The former relies on a certain hook to be
3667 :			initialized, and that often does not happen in the stand-alone case.)
3668 :
3669 :			----------------------------------------------------------------------
3670 :	george	984	Name: Lal George
3671 :			Date: 2001/11/21 13:56:18 EST
3672 :			Tag: george-2001121-pseudo-ops
3673 :			Description:
3674 :
3675 :			Implemented a complete redesign of MLRISC pseudo-ops. Now there
3676 :			ought to never be any question of incompatabilities with
3677 :			pseudo-op syntax expected by host assemblers.
3678 :
3679 :			For now, only modules supporting GAS syntax are implemented
3680 :			but more should follow, such as MASM, and vendor assembler
3681 :			syntax, e.g. IBM as, Sun as, etc.
3682 :
3683 :			----------------------------------------------------------------------
3684 :	blume	977	Name: Matthias Blume
3685 :			Date: 2001/11/14 11:52:00 EST
3686 :			Tag: blume-20011114-srcname
3687 :			Description:
3688 :
3689 :			1. Routed the name of the current source file to mlriscgen where it
3690 :			should be directly emitted into the code object. (This last part
3691 :			is yet to be done.)
3692 :
3693 :			2. Some cleanup of the pgraph code to make it match the proposal that
3694 :			I put out the other day. (The proposal notwithstanding, things are
3695 :			still in flux here.)
3696 :
3697 :			----------------------------------------------------------------------
3698 :	george	976	Name: Lal George
3699 :			Date: 2001/11/14 09:44:04 EST
3700 :			Tag:
3701 :			Description:
3702 :
3703 :			Fix for a backpatching bug reported by Allen.
3704 :
3705 :			Because the boundary between short and long span-dependent
3706 :			instructions is +/- 128, there are an astounding number of
3707 :			span-dependent instructions whose size is over estimated.
3708 :
3709 :			Allen came up with the idea of letting the size of span
3710 :			dependent instructions be non-monotonic, for a maxIter
3711 :			number of times, after which the size must be monotonically
3712 :			increasing.
3713 :
3714 :			This table shows the number of span-dependent instructions
3715 :			whose size was over-estimated as a function of maxIter, for the
3716 :			file Parse/parse/ml.grm.sml:
3717 :
3718 :			maxIter # of instructions:
3719 :			10 687
3720 :			20 438
3721 :			30 198
3722 :			40 0
3723 :
3724 :			In compiling the compiler, there is no significant difference in
3725 :			compilation speed between maxIter=10 and maxIter=40. Actually,
3726 :			my measurements showed that maxIter=40 was a tad faster than
3727 :			maxIter=10! Also 96% of the files in the compiler reach a fix
3728 :			point within 13 iterations, so fixing maxIter at 40, while high,
3729 :			is okay.
3730 :
3731 :			----------------------------------------------------------------------
3732 :	blume	975	Name: Matthias Blume
3733 :			Date: 2001/10/31 15:25:00 EST
3734 :			Tag: blume-20011031-pgraph
3735 :			Description:
3736 :
3737 :			CKIT:
3738 :			* Changed the "Function" constructor of type Ast.ctype to carry optional
3739 :			argument identifiers.
3740 :			* Changed the return type of TypeUtil.getFunction accordingly.
3741 :			* Type equality ignores the argument names.
3742 :			* TypeUtil.composite tries to preserve argument names but gives up quickly
3743 :			if there is a mismatch.
3744 :
3745 :			installation script:
3746 :			* attempts to use "curl" if available (unless "wget" is available as well)
3747 :
3748 :			CM:
3749 :			* has an experimental implementation of "portable graphs" which I will
3750 :			soon propose as an implementation-independent library format
3751 :			* there are also new libraries $/pgraph.cm and $/pgraph-util.cm
3752 :
3753 :			NLFFI-LIB:
3754 :			* some cleanup (all cosmetic)
3755 :
3756 :			NLFFIGEN:
3757 :			* temporarily disabled the mechanism that suppresses ML output for
3758 :			C definitions whose identifiers start with an underscore character
3759 :			* generate val bindings for enum constants
3760 :			* user can request that only one style (light or heavy) is being used;
3761 :			default is to use both (command-line arguments: -heavy and -light)
3762 :			* fixed bug in handling of function types involving incomplete pointers
3763 :			* generate ML entry points that take record arguments (i.e., using
3764 :			named arguments) for C functions that have a prototype with named
3765 :			arguments
3766 :			(see changes to CKIT)
3767 :
3768 :			----------------------------------------------------------------------
3769 :	leunga	972	Name: Allen Leung
3770 :			Date: 2001/10/27 20:34:00 EDT
3771 :			Tag: leunga-20011027-x86-fast-fp-call
3772 :			Description:
3773 :
3774 :			Fixed the bug described in blume-20010920-slowfp.
3775 :
3776 :			The fix involves
3777 :			1. generating FCOPYs in FSTP in ia32-svid
3778 :			2. marking a CALL with the appropriate annotation
3779 :
3780 :			----------------------------------------------------------------------
3781 :	blume	961	Name: Matthias Blume
3782 :			Date: 2001/10/16 11:32:00 EDT
3783 :			Tag: blume-20011016-netbsd
3784 :			Description:
3785 :
3786 :			Underscore patch from Chris Richards (fixing problem with compiling
3787 :			runtime system under recent NetBSD).
3788 :
3789 :			----------------------------------------------------------------------
3790 :	george	959	Name: Allen Leung
3791 :			Date: 2001/10/12 17:18:32 EDT 2001
3792 :			Tag: leung-20011012-x86-printflowgraph
3793 :			Description:
3794 :
3795 :			X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
3796 :
3797 :			----------------------------------------------------------------------
3798 :	george	952	Name: Lal George
3799 :	george	958	Date: 2001/10/11 23:51:34 EDT
3800 :			Tag: george-20011011-too-many-instrs
3801 :			Description:
3802 :
3803 :			The representation of a program point never expected to see more
3804 :			than 65536 instructions in a basic block!
3805 :
3806 :			----------------------------------------------------------------------
3807 :			Name: Lal George
3808 :	george	952	Date: 2001/10/09 09:41:37 EDT
3809 :			Tag: george-20011008-mlrisc-labels
3810 :			Description:
3811 :
3812 :			Changed the machine description files to support printing of
3813 :			local and global labels in assembly code, based on host assembler
3814 :			conventions.
3815 :
3816 :			----------------------------------------------------------------------
3817 :	blume	916	Name: Matthias Blume
3818 :	blume	939	Date: 2001/09/25 15:25:00 EDT
3819 :			Tag: blume-20010925-exninfo
3820 :			Description:
3821 :
3822 :			I provided a non-hook implementation of exnName (at the toplevel) and
3823 :			made the "dummy" implementation of exnMessage (at the toplevel) more
3824 :			useful: if nothing gets "hooked in", then at least you are going to
3825 :			see the exception name and a message indicating why you don't see more.
3826 :
3827 :			[For the time being, programs that need exnMessage and want to use
3828 :			ml-build should either use General.exnMessage (strongly recommended) or
3829 :			refer to structure General at some other point so that CM sees a
3830 :			static dependency.]
3831 :
3832 :			[Similar remarks go for "print" and "use": If you want to use their
3833 :			functionality in stand-alone programs generated by ml-build, then use
3834 :			TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
3835 :
3836 :			----------------------------------------------------------------------
3837 :			Name: Matthias Blume
3838 :	blume	938	Date: 2001/09/20 17:28:00 EDT
3839 :			Tag: blume-20010920-slowfp
3840 :			Description:
3841 :
3842 :			Allen says that x86-fast-fp is not safe yet, so I turned it off again...
3843 :
3844 :			----------------------------------------------------------------------
3845 :			Name: Matthias Blume
3846 :	blume	937	Date: 2001/09/20 17:20:00 EDT
3847 :			Tag: blume-20010920-canonicalpaths
3848 :			Description:
3849 :
3850 :			0. Updated the BOOT file (something that I forgot to do earlier).
3851 :
3852 :			1. Small internal change to CM so that it avoids "/../" in filenames
3853 :			as much as possible (but only where it is safe).
3854 :
3855 :			2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
3856 :			that arguments that contain delimiters are passed through correctly.
3857 :			This change also means that all "special" arguments of the form
3858 :			@SMLxxx... must come first.
3859 :
3860 :			3. Changed install script to put relative anchor names for tool commands
3861 :			into pathconfig.
3862 :
3863 :			----------------------------------------------------------------------
3864 :			Name: Matthias Blume
3865 :	blume	932	Date: 2001/09/18 15:35:00 EDT
3866 :			Tag: blume-20010918-readme11036
3867 :			Description:
3868 :
3869 :			Added README files.
3870 :
3871 :			----------------------------------------------------------------------
3872 :			Name: Matthias Blume
3873 :	blume	929	Date: 2001/09/18 11:45:00 EDT
3874 :			Tag: Release_110_36 (retag)
3875 :			Description:
3876 :
3877 :			Fixed mistake in config/preloads. Retagged as 110.36.
3878 :
3879 :			----------------------------------------------------------------------
3880 :			Name: Matthias Blume
3881 :	blume	928	Date: 2001/09/18 09:40:00 EDT
3882 :	blume	929	Tag: Release_110_36_orig (tag changed)
3883 :	blume	928	Description:
3884 :
3885 :			New version (110.36). New bootfiles.
3886 :
3887 :			----------------------------------------------------------------------
3888 :			Name: Matthias Blume
3889 :	blume	927	Date: 2001/09/14 16:15:00 EDT
3890 :			Tag: blume-20010914-x86fastfp
3891 :			Description:
3892 :
3893 :			John committed some changes that Allen made, in particular a (hopefully)
3894 :			correctly working version of the x86-fp module.
3895 :
3896 :			I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
3897 :			flag to "true". Everything seems to compile to a fixpoint ok, and
3898 :			"mandelbrot" speeds up by about 15%.
3899 :
3900 :			----------------------------------------------------------------------
3901 :			Name: Matthias Blume
3902 :	blume	923	Date: 2001/09/13 11:20:00 EDT
3903 :			Tag: blume-20010913-minimal
3904 :			Description:
3905 :
3906 :			1. Stefan Monnier's patch to fix a miscompilation problem that
3907 :			was brought to light by John Reppy's work on Moby.
3908 :
3909 :			2. Implemented a minimal "structure Compiler" that contains just
3910 :			"version" and "architecture". The minimal version will be
3911 :			available when the full version is not. This is for backward-
3912 :			compatibility with code that wants to test Compiler.version.
3913 :
3914 :			----------------------------------------------------------------------
3915 :			Name: Matthias Blume
3916 :	blume	920	Date: 2001/08/28 14:03:00 EDT
3917 :			Tag: blume-20010828-ml-lex
3918 :			Description:
3919 :
3920 :			Fix for bug 1581, received from Neophytos Michael.
3921 :
3922 :			----------------------------------------------------------------------
3923 :			Name: Matthias Blume
3924 :	blume	918	Date: 2001/08/27 11:20:00 EDT
3925 :			Tag: blume-20010827-readme11035
3926 :			Description:
3927 :
3928 :			Fleshed out the README file for 110.35.
3929 :
3930 :			----------------------------------------------------------------------
3931 :			Name: Matthias Blume
3932 :	blume	916	Date: 2001/08/24 17:10:00 EDT
3933 :			Tag: Release_110_35
3934 :			Description:
3935 :
3936 :			New version number (110.35). New bootfiles.
3937 :
3938 :			----------------------------------------------------------------------
3939 :	george	915	Name: Lal George
3940 :	george	909	Date: 2001/08/24 13:47:18 EDT 2001
3941 :			Tag: george-20010824-MLRISC-graphs
3942 :			Description:
3943 :
3944 :			removed clusters from MLRISC completely and replaced with graphs.
3945 :
3946 :			----------------------------------------------------------------------
3947 :			Name: Matthias Blume
3948 :	blume	905	Date: 2001/08/23 17:50:00 EDT
3949 :			Tag: blume-20010823-toplevel
3950 :			Description:
3951 :
3952 :			- some reorganization of the code that implements various kinds of
3953 :			environments in the compiler (static, dynamic, symbolic, combined)
3954 :			- re-implemented the EnvRef module so that evalStream works properly
3955 :			(if the stream contains references to "use", "CM.make", etc.)
3956 :			- cleaned up evalloop.sml and interact.sml (but they need more cleaning)
3957 :
3958 :			----------------------------------------------------------------------
3959 :			Name: Matthias Blume
3960 :	blume	904	Date: 2001/08/20 15:50 EDT
3961 :			Tag: blume20010820-slipup
3962 :			Description:
3963 :
3964 :			I forgot to commit a few files. Here they are...
3965 :
3966 :			----------------------------------------------------------------------
3967 :			Name: Matthias Blume
3968 :	blume	903	Date: 2001/08/20 15:35:00 EDT
3969 :			Tag: blume-20010820-debugprof
3970 :			Description:
3971 :
3972 :			!!!! NEW BOOTFILES !!!!
3973 :
3974 :			This is another round of reorganizing the compiler sources. This
3975 :			time the main goal was to factor out all the "instrumentation"
3976 :			passes (for profiling and backtracing) into their own library.
3977 :			The difficulty was to do it in such a way that it does not depend
3978 :			on elaborate.cm but only on elabdata.cm.
3979 :
3980 :			Therefore there have been further changes to both elaborate.cm and
3981 :			elabdata.cm -- more "generic" things have been moved from the former
3982 :			to the latter. As a result, I was forced to split the assignment
3983 :			of numbers indicating "primtyc"s into two portions: SML-generic and
3984 :			SML/NJ-specific. Since it would have been awkward to maintain,
3985 :			I bit the bullet and actually _changed_ the mapping between these
3986 :			numbers and primtycs. The bottom line of this is that you need
3987 :			a new set of bin- and bootfiles.
3988 :
3989 :			I have built new bootfiles for all architectures, so doing a fresh
3990 :			checkout and config/install.sh should be all you need.
3991 :
3992 :			The newly created library's name is
3993 :
3994 :			$smlnj/viscomp/debugprof.cm
3995 :
3996 :			and its sources live under
3997 :
3998 :			src/compiler/DebugProf
3999 :
4000 :			----------------------------------------------------------------------
4001 :			Name: Matthias Blume
4002 :	blume	902	Date: 2001/08/15 17:15:00 EDT
4003 :			Tag: blume-20010815-compreorg
4004 :			Description:
4005 :
4006 :			This is a first cut at reorganizing the CM libraries that make up the
4007 :			core of the compiler. The idea is to separate out pieces that could
4008 :			be used independently by tools, e.g., the parser, the typechecker, etc.
4009 :
4010 :			The current status is a step in this direction, but it is not quite
4011 :			satisfactory yet. Expect more changes in the future.
4012 :
4013 :			Here is the current (new) organization...
4014 :
4015 :			What used to be $smlnj/viscomp/core.cm is now divided into
4016 :			six CM libraries:
4017 :
4018 :			$smlnj/viscomp/basics.cm
4019 :			/parser.cm
4020 :			/elabdata.cm
4021 :			/elaborate.cm
4022 :			/execute.cm
4023 :			/core.cm
4024 :
4025 :			The CM files for these libraries live under src/system/smlnj/viscomp.
4026 :			All these libraries are proxy libraries that contain precisely
4027 :			one CM library component. Here are the locations of the components
4028 :			(all within the src/compiler tree):
4029 :
4030 :			Basics/basics.cm
4031 :			Parse/parser.cm
4032 :			ElabData/elabdata.cm
4033 :			Elaborator/elaborate.cm
4034 :			Execution/execute.cm
4035 :			core.cm
4036 :
4037 :			[This organization is the same that has been used already
4038 :			for a while for the architecture-specific parts of the visible
4039 :			compiler and for the old version of core.cm.]
4040 :
4041 :			As you will notice, many source files have been moved from their
4042 :			respective original locations to a new home in one of the above
4043 :			subtrees.
4044 :
4045 :			The division of labor between the new libraries is the following:
4046 :
4047 :			basics.cm:
4048 :			- Simple, basic definitions that pertain to many (or all) of
4049 :			the other libraries.
4050 :			parser.cm:
4051 :			- The SML parser, producing output of type Ast.dec.
4052 :			- The type family for Ast is also defined and exported here.
4053 :			elabdata.cm:
4054 :			- The datatypes that describe input and output of the elaborator.
4055 :			This includes types, absyn, and static environments.
4056 :			elaborator.cm:
4057 :			- The SML/NJ type checker and elaborator.
4058 :			This maps an Ast.dec (with a given static environment) to
4059 :			an Absyn.dec (with a new static environment).
4060 :			- This libraries implements certain modules that used to be
4061 :			structures as functors (to remove dependencies on FLINT).
4062 :			execute.cm:
4063 :			- Everything having to do with executing binary code objects.
4064 :			- Dynamic environments.
4065 :			core.cm:
4066 :			- SML/NJ-specific instantiations of the elaborator and MLRISC.
4067 :			- Top-level modules.
4068 :			- FLINT (this should eventually become its own library)
4069 :
4070 :			Notes:
4071 :
4072 :			I am not 100% happy with the way I separated the elaborator (and its
4073 :			data structures) from FLINT. Two instances of the same problem:
4074 :
4075 :			1. Data structures contain certain fields that carry FLINT-specific
4076 :			information. I hacked around this using exn and the property list
4077 :			module from smlnj-lib. But the fact that there are middle-end
4078 :			specific fields around at all is a bit annoying.
4079 :
4080 :			2. The elaborator calculates certain FLINT-related information. I tried
4081 :			to make this as abstract as I could using functorization, but, again,
4082 :			the fact that the elaborator has to perform calculations on behalf
4083 :			of the middle-end at all is not nice.
4084 :
4085 :			3. Having to used exn and property lists is unfortunate because it
4086 :			weakens type checking. The other alternative (parameterizing
4087 :			nearly *everything*) is not appealing, though.
4088 :
4089 :			I removed the "rebinding =" warning hack because due to the new organization
4090 :			it was awkward to maintain it. As a result, the compiler now issues some of
4091 :			these warnings when compiling init.cmi during bootstrap compilation. On
4092 :			the plus side, you also get a warning when you do, for example:
4093 :			val op = = Int32.+
4094 :			which was not the case up to now.
4095 :
4096 :			I placed "assign" and "deref" into the _Core structure so that the
4097 :			code that deals with the "lazy" keyword can find them there. This
4098 :			removes the need for having access to the primitive environment
4099 :			during elaboration.
4100 :
4101 :			----------------------------------------------------------------------
4102 :			Name: Matthias Blume
4103 :	blume	898	Date: 2001/08/13
4104 :			Tag: blume-20010813-closures
4105 :			Description:
4106 :
4107 :			This fix was sent to us by Zhong Shao. It is supposed to improve the
4108 :			performance of certain loops by avoiding needless closure allocation.
4109 :
4110 :			----------------------------------------------------------------------
4111 :	george	890	Name: Lal George
4112 :	george	896	Date: 2001/07/31 10:03:23 EDT 2001
4113 :			Tag: george-20010731-x86-fmalloc
4114 :			Description: Fixed bug in x86 calls
4115 :
4116 :			There was a bug where call instructions would mysteriously
4117 :			vanish. The call instruction had to be one that returned
4118 :			a floating point value.
4119 :
4120 :			----------------------------------------------------------------------
4121 :			Name: Lal George
4122 :	george	890	Date: 2001/07/19 16:36:29 EDT 2001
4123 :			Tag: george-20010719-simple-cells
4124 :			Description:
4125 :
4126 :			I have dramatically simplified the interface for CELLS in MLRISC.
4127 :
4128 :			In summary, the cells interface is broken up into three parts:
4129 :
4130 :			1. CellsBasis : CELLS_BASIS
4131 :
4132 :			CellsBasis is a top level structure and common for all
4133 :			architectures. it contains the definitions of basic datatypes
4134 :			and utility functions over these types.
4135 :
4136 :			2. functor Cells() : CELLS
4137 :
4138 :			Cells generates an interface for CELLS that incorporates the
4139 :			specific resources on the target architecture, such as the
4140 :			presence of special register classes, their number and size,
4141 :			and various useful substructures.
4142 :
4143 :			3. <ARCH>CELLS
4144 :
4145 :			e.g. SparcCells: SPARCCELLS
4146 :
4147 :			<ARCH>CELLS usually contains additional bindings for special
4148 :			registers on the architecture, such as:
4149 :
4150 :			val r0 : cell (* register zero *)
4151 :			val y : cell (* Y register *)
4152 :			val psr : cell (* processor status register *)
4153 :			...
4154 :
4155 :			The structure returned by applying the Cells functor is opened
4156 :			in this interface.
4157 :
4158 :			The main implication of all this is that the datatypes for cells is
4159 :			split between CellsBasis and CELLS -- a fairly simple change for user
4160 :			code.
4161 :
4162 :			In the old scheme the CELLS interface had a definitional binding of
4163 :			the form:
4164 :
4165 :			signature CELLS = sig
4166 :
4167 :			structure CellsBasis = CellsBasis
4168 :
4169 :			...
4170 :
4171 :			end
4172 :
4173 :			With all the sharing constraints that goes on in MLRISC, this old
4174 :			design quickly leads to errors such as:
4175 :
4176 :			"structure definition spec inside of sharing ... "
4177 :
4178 :
4179 :			and appears to require an unacceptable amount of sharing and where
4180 :			constraint hackery.
4181 :
4182 :			I think this error message (the interaction of definitional specs and
4183 :			sharing) requires more explanation on our web page.
4184 :
4185 :			----------------------------------------------------------------------
4186 :	blume	826	Name: Matthias Blume
4187 :	blume	879	Date: 2001/07/19 15:00:00 EDT
4188 :			Tag: blume-20010719-libreorg
4189 :			Description:
4190 :
4191 :			This update puts together a fairly extensive but straightforward change
4192 :			to the way the libraries that implement the interactive system are
4193 :			organized:
4194 :
4195 :			The biggest change is the elimination of structure Compiler. As a
4196 :			replacement for this structure, there is now a CM library
4197 :			(known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
4198 :			that exports all the substructures of the original structure Compiler
4199 :			directly. So instead of saying Compiler.Foo.bar one now simply
4200 :			says Foo.bar. (The CM libraries actually export a collection of
4201 :			structures that is richer than the collection of substructures of
4202 :			structure Compiler.)
4203 :
4204 :			To make the transition smooth, there is a separate library called
4205 :			$smlnj/compiler/compiler.cm which puts together and exports the
4206 :			original structure Compiler (or at least something very close to it).
4207 :
4208 :			There are five members of the original structure Compiler
4209 :			that are not exported directly but which instead became members
4210 :			of a new structure Backend (described by signature BACKEND). These are:
4211 :			structure Profile (: PROFILE), structure Compile (: COMPILE), structure
4212 :			Interact (: INTERACT), structure Machine (: MACHINE), and val
4213 :			architecture (: string).
4214 :
4215 :			Structure Compiler.Version has become structure CompilerVersion.
4216 :
4217 :			Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
4218 :			by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
4219 :			or x86, respectively.
4220 :			Each of these exports the same frontend structures that
4221 :			$smlnj/compiler.cm exports. But they do not have a structure Backend
4222 :			and instead export some structure <Arch>Backend where <Arch> is Alpha32,
4223 :			Hppa, PPC, Sparc, or X86, respectively.
4224 :
4225 :			Library $smlnj/compiler/all.cm exports the union of the exports of
4226 :			$smlnj/compiler/<arch>.cm
4227 :
4228 :			There are no structures <Arch>Compiler anymore, use
4229 :			$smlnj/compiler/<arch>.cm instead.
4230 :
4231 :			Library host-compiler-0.cm is gone. Instead, the internal library
4232 :			that instantiates CM is now called cm0.cm. Selection of the host
4233 :			compiler (backend) is no longer done here but. (Responsibility for it
4234 :			now lies with $smlnj/compiler/current.cm. This seems to be more
4235 :			logical.)
4236 :
4237 :			Many individual files have been moved or renamed. Some files have
4238 :			been split into multiple files, and some "dead" files have been deleted.
4239 :
4240 :			Aside from these changes to library organization, there are also changes
4241 :			to the way the code itself is organized:
4242 :
4243 :			Structure Binfile has been re-implemented in such a way that it no
4244 :			longer needs any knowledge of the compiler. It exclusively deals
4245 :			with the details of binfile layout. It no longer invokes the
4246 :			compiler (for the purpose of creating new prospective binfile
4247 :			content), and it no longer has any knowledge of how to interpret
4248 :			pickles.
4249 :
4250 :			Structure Compile (: COMPILE) has been stripped down to the bare
4251 :			essentials of compilation. It no longer deals with linking/execution.
4252 :			The interface has been cleaned up considerably.
4253 :
4254 :			Utility routines for dealing with linking and execution have been
4255 :			moved into their own substructures.
4256 :
4257 :			(The ultimate goal of these changes is to provide a light-weight
4258 :			binfile loader/linker (at least for, e.g., stable libraries) that
4259 :			does not require CM or the compiler to be present.)
4260 :
4261 :			CM documentation has been updated to reflect the changes to library
4262 :			organization.
4263 :
4264 :			----------------------------------------------------------------------
4265 :			Name: Matthias Blume
4266 :	blume	875	Date: 2001/07/10 17:30:00 EDT
4267 :			Tag: Release_110_34
4268 :			Description:
4269 :
4270 :	blume	876	Minor tweak to 110.34 (re-tagged):
4271 :
4272 :			- README.html file added to CVS repository
4273 :			- runtime compiles properly under FreeBSD 3.X and 4.X
4274 :
4275 :			----------------------------------------------------------------------
4276 :			Name: Matthias Blume
4277 :			Date: 2001/07/10 17:30:00 EDT
4278 :			Tag: Release_110_34
4279 :			Description:
4280 :
4281 :	blume	875	New version number (110.34). New bootfiles.
4282 :
4283 :			----------------------------------------------------------------------
4284 :			Name: Matthias Blume
4285 :	blume	874	Date: 2001/07/09 16:00:00 EDT
4286 :			Tag: blume-20010709-more-varargs
4287 :			Description:
4288 :
4289 :			I changed the handling of varargs in ml-nlffigen again:
4290 :			The ellipsis ... will now simply be ignored (with an accompanying warning).
4291 :
4292 :			The immediate effect is that you can actually call a varargs function
4293 :			from ML -- but you can't actually supply any arguments beyond the ones
4294 :			specified explicitly. (For example, you can call printf with its format
4295 :			string, but you cannot pass additional arguments.)
4296 :
4297 :			This behavior is only marginally more useful than the one before, but
4298 :			it has the advantage that a function or, more importantly, a function
4299 :			type never gets dropped on the floor, thus avoiding follow-up problems with
4300 :			other types that refer to the offending one.
4301 :
4302 :			----------------------------------------------------------------------
4303 :			Name: Matthias Blume
4304 :	blume	873	Date: 2001/07/09 11:25:00 EDT
4305 :			Tag: blume-20010709-varargs
4306 :			Description:
4307 :
4308 :			1. ckit-lib.cm now exports structure Error
4309 :			2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
4310 :			with a warning accompanied by a source location. Moreover, it
4311 :			merely skips the offending function or type and proceeds with the