Annotation of /sml/releases/release-110.36/HISTORY

Parent Directory | Revision Log

---

Revision 930 - (view) (download)

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 :			Name:
11 :			Date:
12 :			Tag: <post-commit CVS tag>
13 :			Description:
14 :	george	824
15 :	leunga	815	----------------------------------------------------------------------
16 :	blume	916	Name: Matthias Blume
17 :	blume	929	Date: 2001/09/18 11:45:00 EDT
18 :			Tag: Release_110_36 (retag)
19 :			Description:
20 :
21 :			Fixed mistake in config/preloads. Retagged as 110.36.
22 :
23 :			----------------------------------------------------------------------
24 :			Name: Matthias Blume
25 :	blume	928	Date: 2001/09/18 09:40:00 EDT
26 :	blume	929	Tag: Release_110_36_orig (tag changed)
27 :	blume	928	Description:
28 :
29 :			New version (110.36). New bootfiles.
30 :
31 :			----------------------------------------------------------------------
32 :			Name: Matthias Blume
33 :	blume	927	Date: 2001/09/14 16:15:00 EDT
34 :			Tag: blume-20010914-x86fastfp
35 :			Description:
36 :
37 :			John committed some changes that Allen made, in particular a (hopefully)
38 :			correctly working version of the x86-fp module.
39 :
40 :			I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
41 :			flag to "true". Everything seems to compile to a fixpoint ok, and
42 :			"mandelbrot" speeds up by about 15%.
43 :
44 :			----------------------------------------------------------------------
45 :			Name: Matthias Blume
46 :	blume	923	Date: 2001/09/13 11:20:00 EDT
47 :			Tag: blume-20010913-minimal
48 :			Description:
49 :
50 :			1. Stefan Monnier's patch to fix a miscompilation problem that
51 :			was brought to light by John Reppy's work on Moby.
52 :
53 :			2. Implemented a minimal "structure Compiler" that contains just
54 :			"version" and "architecture". The minimal version will be
55 :			available when the full version is not. This is for backward-
56 :			compatibility with code that wants to test Compiler.version.
57 :
58 :			----------------------------------------------------------------------
59 :			Name: Matthias Blume
60 :	blume	920	Date: 2001/08/28 14:03:00 EDT
61 :			Tag: blume-20010828-ml-lex
62 :			Description:
63 :
64 :			Fix for bug 1581, received from Neophytos Michael.
65 :
66 :			----------------------------------------------------------------------
67 :			Name: Matthias Blume
68 :	blume	918	Date: 2001/08/27 11:20:00 EDT
69 :			Tag: blume-20010827-readme11035
70 :			Description:
71 :
72 :			Fleshed out the README file for 110.35.
73 :
74 :			----------------------------------------------------------------------
75 :			Name: Matthias Blume
76 :	blume	916	Date: 2001/08/24 17:10:00 EDT
77 :			Tag: Release_110_35
78 :			Description:
79 :
80 :			New version number (110.35). New bootfiles.
81 :
82 :			----------------------------------------------------------------------
83 :	george	915	Name: Lal George
84 :	george	909	Date: 2001/08/24 13:47:18 EDT 2001
85 :			Tag: george-20010824-MLRISC-graphs
86 :			Description:
87 :
88 :			removed clusters from MLRISC completely and replaced with graphs.
89 :
90 :			----------------------------------------------------------------------
91 :			Name: Matthias Blume
92 :	blume	905	Date: 2001/08/23 17:50:00 EDT
93 :			Tag: blume-20010823-toplevel
94 :			Description:
95 :
96 :			- some reorganization of the code that implements various kinds of
97 :			environments in the compiler (static, dynamic, symbolic, combined)
98 :			- re-implemented the EnvRef module so that evalStream works properly
99 :			(if the stream contains references to "use", "CM.make", etc.)
100 :			- cleaned up evalloop.sml and interact.sml (but they need more cleaning)
101 :
102 :			----------------------------------------------------------------------
103 :			Name: Matthias Blume
104 :	blume	904	Date: 2001/08/20 15:50 EDT
105 :			Tag: blume20010820-slipup
106 :			Description:
107 :
108 :			I forgot to commit a few files. Here they are...
109 :
110 :			----------------------------------------------------------------------
111 :			Name: Matthias Blume
112 :	blume	903	Date: 2001/08/20 15:35:00 EDT
113 :			Tag: blume-20010820-debugprof
114 :			Description:
115 :
116 :			!!!! NEW BOOTFILES !!!!
117 :
118 :			This is another round of reorganizing the compiler sources. This
119 :			time the main goal was to factor out all the "instrumentation"
120 :			passes (for profiling and backtracing) into their own library.
121 :			The difficulty was to do it in such a way that it does not depend
122 :			on elaborate.cm but only on elabdata.cm.
123 :
124 :			Therefore there have been further changes to both elaborate.cm and
125 :			elabdata.cm -- more "generic" things have been moved from the former
126 :			to the latter. As a result, I was forced to split the assignment
127 :			of numbers indicating "primtyc"s into two portions: SML-generic and
128 :			SML/NJ-specific. Since it would have been awkward to maintain,
129 :			I bit the bullet and actually _changed_ the mapping between these
130 :			numbers and primtycs. The bottom line of this is that you need
131 :			a new set of bin- and bootfiles.
132 :
133 :			I have built new bootfiles for all architectures, so doing a fresh
134 :			checkout and config/install.sh should be all you need.
135 :
136 :			The newly created library's name is
137 :
138 :			$smlnj/viscomp/debugprof.cm
139 :
140 :			and its sources live under
141 :
142 :			src/compiler/DebugProf
143 :
144 :			----------------------------------------------------------------------
145 :			Name: Matthias Blume
146 :	blume	902	Date: 2001/08/15 17:15:00 EDT
147 :			Tag: blume-20010815-compreorg
148 :			Description:
149 :
150 :			This is a first cut at reorganizing the CM libraries that make up the
151 :			core of the compiler. The idea is to separate out pieces that could
152 :			be used independently by tools, e.g., the parser, the typechecker, etc.
153 :
154 :			The current status is a step in this direction, but it is not quite
155 :			satisfactory yet. Expect more changes in the future.
156 :
157 :			Here is the current (new) organization...
158 :
159 :			What used to be $smlnj/viscomp/core.cm is now divided into
160 :			six CM libraries:
161 :
162 :			$smlnj/viscomp/basics.cm
163 :			/parser.cm
164 :			/elabdata.cm
165 :			/elaborate.cm
166 :			/execute.cm
167 :			/core.cm
168 :
169 :			The CM files for these libraries live under src/system/smlnj/viscomp.
170 :			All these libraries are proxy libraries that contain precisely
171 :			one CM library component. Here are the locations of the components
172 :			(all within the src/compiler tree):
173 :
174 :			Basics/basics.cm
175 :			Parse/parser.cm
176 :			ElabData/elabdata.cm
177 :			Elaborator/elaborate.cm
178 :			Execution/execute.cm
179 :			core.cm
180 :
181 :			[This organization is the same that has been used already
182 :			for a while for the architecture-specific parts of the visible
183 :			compiler and for the old version of core.cm.]
184 :
185 :			As you will notice, many source files have been moved from their
186 :			respective original locations to a new home in one of the above
187 :			subtrees.
188 :
189 :			The division of labor between the new libraries is the following:
190 :
191 :			basics.cm:
192 :			- Simple, basic definitions that pertain to many (or all) of
193 :			the other libraries.
194 :			parser.cm:
195 :			- The SML parser, producing output of type Ast.dec.
196 :			- The type family for Ast is also defined and exported here.
197 :			elabdata.cm:
198 :			- The datatypes that describe input and output of the elaborator.
199 :			This includes types, absyn, and static environments.
200 :			elaborator.cm:
201 :			- The SML/NJ type checker and elaborator.
202 :			This maps an Ast.dec (with a given static environment) to
203 :			an Absyn.dec (with a new static environment).
204 :			- This libraries implements certain modules that used to be
205 :			structures as functors (to remove dependencies on FLINT).
206 :			execute.cm:
207 :			- Everything having to do with executing binary code objects.
208 :			- Dynamic environments.
209 :			core.cm:
210 :			- SML/NJ-specific instantiations of the elaborator and MLRISC.
211 :			- Top-level modules.
212 :			- FLINT (this should eventually become its own library)
213 :
214 :			Notes:
215 :
216 :			I am not 100% happy with the way I separated the elaborator (and its
217 :			data structures) from FLINT. Two instances of the same problem:
218 :
219 :			1. Data structures contain certain fields that carry FLINT-specific
220 :			information. I hacked around this using exn and the property list
221 :			module from smlnj-lib. But the fact that there are middle-end
222 :			specific fields around at all is a bit annoying.
223 :
224 :			2. The elaborator calculates certain FLINT-related information. I tried
225 :			to make this as abstract as I could using functorization, but, again,
226 :			the fact that the elaborator has to perform calculations on behalf
227 :			of the middle-end at all is not nice.
228 :
229 :			3. Having to used exn and property lists is unfortunate because it
230 :			weakens type checking. The other alternative (parameterizing
231 :			nearly *everything*) is not appealing, though.
232 :
233 :			I removed the "rebinding =" warning hack because due to the new organization
234 :			it was awkward to maintain it. As a result, the compiler now issues some of
235 :			these warnings when compiling init.cmi during bootstrap compilation. On
236 :			the plus side, you also get a warning when you do, for example:
237 :			val op = = Int32.+
238 :			which was not the case up to now.
239 :
240 :			I placed "assign" and "deref" into the _Core structure so that the
241 :			code that deals with the "lazy" keyword can find them there. This
242 :			removes the need for having access to the primitive environment
243 :			during elaboration.
244 :
245 :			----------------------------------------------------------------------
246 :			Name: Matthias Blume
247 :	blume	898	Date: 2001/08/13
248 :			Tag: blume-20010813-closures
249 :			Description:
250 :
251 :			This fix was sent to us by Zhong Shao. It is supposed to improve the
252 :			performance of certain loops by avoiding needless closure allocation.
253 :
254 :			----------------------------------------------------------------------
255 :	george	890	Name: Lal George
256 :	george	896	Date: 2001/07/31 10:03:23 EDT 2001
257 :			Tag: george-20010731-x86-fmalloc
258 :			Description: Fixed bug in x86 calls
259 :
260 :			There was a bug where call instructions would mysteriously
261 :			vanish. The call instruction had to be one that returned
262 :			a floating point value.
263 :
264 :			----------------------------------------------------------------------
265 :			Name: Lal George
266 :	george	890	Date: 2001/07/19 16:36:29 EDT 2001
267 :			Tag: george-20010719-simple-cells
268 :			Description:
269 :
270 :			I have dramatically simplified the interface for CELLS in MLRISC.
271 :
272 :			In summary, the cells interface is broken up into three parts:
273 :
274 :			1. CellsBasis : CELLS_BASIS
275 :
276 :			CellsBasis is a top level structure and common for all
277 :			architectures. it contains the definitions of basic datatypes
278 :			and utility functions over these types.
279 :
280 :			2. functor Cells() : CELLS
281 :
282 :			Cells generates an interface for CELLS that incorporates the
283 :			specific resources on the target architecture, such as the
284 :			presence of special register classes, their number and size,
285 :			and various useful substructures.
286 :
287 :			3. <ARCH>CELLS
288 :
289 :			e.g. SparcCells: SPARCCELLS
290 :
291 :			<ARCH>CELLS usually contains additional bindings for special
292 :			registers on the architecture, such as:
293 :
294 :			val r0 : cell (* register zero *)
295 :			val y : cell (* Y register *)
296 :			val psr : cell (* processor status register *)
297 :			...
298 :
299 :			The structure returned by applying the Cells functor is opened
300 :			in this interface.
301 :
302 :			The main implication of all this is that the datatypes for cells is
303 :			split between CellsBasis and CELLS -- a fairly simple change for user
304 :			code.
305 :
306 :			In the old scheme the CELLS interface had a definitional binding of
307 :			the form:
308 :
309 :			signature CELLS = sig
310 :
311 :			structure CellsBasis = CellsBasis
312 :
313 :			...
314 :
315 :			end
316 :
317 :			With all the sharing constraints that goes on in MLRISC, this old
318 :			design quickly leads to errors such as:
319 :
320 :			"structure definition spec inside of sharing ... "
321 :
322 :
323 :			and appears to require an unacceptable amount of sharing and where
324 :			constraint hackery.
325 :
326 :			I think this error message (the interaction of definitional specs and
327 :			sharing) requires more explanation on our web page.
328 :
329 :			----------------------------------------------------------------------
330 :	blume	826	Name: Matthias Blume
331 :	blume	879	Date: 2001/07/19 15:00:00 EDT
332 :			Tag: blume-20010719-libreorg
333 :			Description:
334 :
335 :			This update puts together a fairly extensive but straightforward change
336 :			to the way the libraries that implement the interactive system are
337 :			organized:
338 :
339 :			The biggest change is the elimination of structure Compiler. As a
340 :			replacement for this structure, there is now a CM library
341 :			(known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
342 :			that exports all the substructures of the original structure Compiler
343 :			directly. So instead of saying Compiler.Foo.bar one now simply
344 :			says Foo.bar. (The CM libraries actually export a collection of
345 :			structures that is richer than the collection of substructures of
346 :			structure Compiler.)
347 :
348 :			To make the transition smooth, there is a separate library called
349 :			$smlnj/compiler/compiler.cm which puts together and exports the
350 :			original structure Compiler (or at least something very close to it).
351 :
352 :			There are five members of the original structure Compiler
353 :			that are not exported directly but which instead became members
354 :			of a new structure Backend (described by signature BACKEND). These are:
355 :			structure Profile (: PROFILE), structure Compile (: COMPILE), structure
356 :			Interact (: INTERACT), structure Machine (: MACHINE), and val
357 :			architecture (: string).
358 :
359 :			Structure Compiler.Version has become structure CompilerVersion.
360 :
361 :			Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
362 :			by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
363 :			or x86, respectively.
364 :			Each of these exports the same frontend structures that
365 :			$smlnj/compiler.cm exports. But they do not have a structure Backend
366 :			and instead export some structure <Arch>Backend where <Arch> is Alpha32,
367 :			Hppa, PPC, Sparc, or X86, respectively.
368 :
369 :			Library $smlnj/compiler/all.cm exports the union of the exports of
370 :			$smlnj/compiler/<arch>.cm
371 :
372 :			There are no structures <Arch>Compiler anymore, use
373 :			$smlnj/compiler/<arch>.cm instead.
374 :
375 :			Library host-compiler-0.cm is gone. Instead, the internal library
376 :			that instantiates CM is now called cm0.cm. Selection of the host
377 :			compiler (backend) is no longer done here but. (Responsibility for it
378 :			now lies with $smlnj/compiler/current.cm. This seems to be more
379 :			logical.)
380 :
381 :			Many individual files have been moved or renamed. Some files have
382 :			been split into multiple files, and some "dead" files have been deleted.
383 :
384 :			Aside from these changes to library organization, there are also changes
385 :			to the way the code itself is organized:
386 :
387 :			Structure Binfile has been re-implemented in such a way that it no
388 :			longer needs any knowledge of the compiler. It exclusively deals
389 :			with the details of binfile layout. It no longer invokes the
390 :			compiler (for the purpose of creating new prospective binfile
391 :			content), and it no longer has any knowledge of how to interpret
392 :			pickles.
393 :
394 :			Structure Compile (: COMPILE) has been stripped down to the bare
395 :			essentials of compilation. It no longer deals with linking/execution.
396 :			The interface has been cleaned up considerably.
397 :
398 :			Utility routines for dealing with linking and execution have been
399 :			moved into their own substructures.
400 :
401 :			(The ultimate goal of these changes is to provide a light-weight
402 :			binfile loader/linker (at least for, e.g., stable libraries) that
403 :			does not require CM or the compiler to be present.)
404 :
405 :			CM documentation has been updated to reflect the changes to library
406 :			organization.
407 :
408 :			----------------------------------------------------------------------
409 :			Name: Matthias Blume
410 :	blume	875	Date: 2001/07/10 17:30:00 EDT
411 :			Tag: Release_110_34
412 :			Description:
413 :
414 :	blume	876	Minor tweak to 110.34 (re-tagged):
415 :
416 :			- README.html file added to CVS repository
417 :			- runtime compiles properly under FreeBSD 3.X and 4.X
418 :
419 :			----------------------------------------------------------------------
420 :			Name: Matthias Blume
421 :			Date: 2001/07/10 17:30:00 EDT
422 :			Tag: Release_110_34
423 :			Description:
424 :
425 :	blume	875	New version number (110.34). New bootfiles.
426 :
427 :			----------------------------------------------------------------------
428 :			Name: Matthias Blume
429 :	blume	874	Date: 2001/07/09 16:00:00 EDT
430 :			Tag: blume-20010709-more-varargs
431 :			Description:
432 :
433 :			I changed the handling of varargs in ml-nlffigen again:
434 :			The ellipsis ... will now simply be ignored (with an accompanying warning).
435 :
436 :			The immediate effect is that you can actually call a varargs function
437 :			from ML -- but you can't actually supply any arguments beyond the ones
438 :			specified explicitly. (For example, you can call printf with its format
439 :			string, but you cannot pass additional arguments.)
440 :
441 :			This behavior is only marginally more useful than the one before, but
442 :			it has the advantage that a function or, more importantly, a function
443 :			type never gets dropped on the floor, thus avoiding follow-up problems with
444 :			other types that refer to the offending one.
445 :
446 :			----------------------------------------------------------------------
447 :			Name: Matthias Blume
448 :	blume	873	Date: 2001/07/09 11:25:00 EDT
449 :			Tag: blume-20010709-varargs
450 :			Description:
451 :
452 :			1. ckit-lib.cm now exports structure Error
453 :			2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
454 :			with a warning accompanied by a source location. Moreover, it
455 :			merely skips the offending function or type and proceeds with the
456 :			rest of its work.u As a result, one can safely feed C code containing
457 :			"..." to ml-nlffigen.
458 :			3. There are some internal improvements to CM, providing slightly
459 :			more general string substitutions in the tools subsystem.
460 :
461 :			----------------------------------------------------------------------
462 :			Name: Matthias Blume
463 :	blume	854	Date: 2001/06/27 15:10:00 EDT
464 :			Tag: blume-20010627-concur
465 :			Description:
466 :
467 :			Fixed a small bug in CM's handling of parallel compilation.
468 :			(You could observe the bug by Control-C-interrupting an ordinary
469 :			CMB.make or CM.stabilize and then attaching some compile servers.
470 :			The result was that all of a sudden the previously interrupted
471 :			compilation would continue on its own. This was because of
472 :			an over-optimization: CM did not bother to clean out certain queues
473 :			when no servers were attached "anyway", resulting in the contents
474 :			of these queues to grab control when new servers did get attached.)
475 :
476 :			There is also another minor update to the CM manual.
477 :
478 :			----------------------------------------------------------------------
479 :			Name: Matthias Blume
480 :	blume	853	Date: 2001/06/26 16:15:00 EDT
481 :			Tag: blume-20010626-cmdoc
482 :			Description:
483 :
484 :			Minor typo fixed in CM manual (syntax diagram for libraries).
485 :
486 :			----------------------------------------------------------------------
487 :			Name: Matthias Blume
488 :	blume	852	Date: 2001/06/25 22:55:00 EDT
489 :			Tag: blume-20010625-x86pc
490 :			Description:
491 :
492 :			Fixed a nasty bug in the X86 assembly code that caused signal
493 :			handlers to fail (crash) randomly.
494 :
495 :			----------------------------------------------------------------------
496 :			Name: Matthias Blume
497 :	blume	846	Date: 2001/06/25 12:05:00 EDT
498 :			Tag: blume-20010625-nlffigen
499 :			Description:
500 :
501 :			This update fixes a number of minor bugs in ml-nlffigen as reported by
502 :			Nick Carter <nbc@andrew.cmu.edu>.
503 :
504 :			1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
505 :			2. Default names for generated files are now derived from the name of
506 :			the C file *without its directory*. In particular, this causes generated
507 :			files to be placed locally even if the C file is in some system directory.
508 :			3. Default names for generated signatures and structures are also derived
509 :			from the C file name without its directory. This avoids silly things
510 :			like "structure GL/GL".
511 :			(Other silly names are still possible because ml-nlffigen does not do
512 :	blume	847	a thorough check of whether generated names are legal ML identifiers.
513 :			When in doubt, use command line arguments to force particular names.)
514 :	blume	846
515 :			----------------------------------------------------------------------
516 :			Name: Matthias Blume
517 :	blume	845	Date: 2001/06/21 12:25:00 EDT
518 :			Tag: blume-20010621-eXene
519 :			Description:
520 :
521 :			eXene now compiles and (sort of) works again.
522 :
523 :			The library name (for version > 110.33) is $/eXene.cm.
524 :
525 :			I also added an new example in src/eXene/examples/nbody. See the
526 :			README file there for details.
527 :
528 :			----------------------------------------------------------------------
529 :			Name: Matthias Blume
530 :	blume	844	Date: 2001/06/20 16:40:00 EDT
531 :			Tag: blume-20010620-cml
532 :			Description:
533 :
534 :			CML now compiles and works again.
535 :
536 :			Libraries (for version > 110.33):
537 :
538 :			$cml/cml.cm Main CML library.
539 :			$cml/basis.cm CML's version of $/basis.cm.
540 :			$cml/cml-internal.cm Internal helper library.
541 :			$cml/core-cml.cm Internal helper library.
542 :			$cml-lib/trace-cml.cm Tracing facility.
543 :			$cml-lib/smlnj-lib.cm CML's version of $/smlnj-lib.cm
544 :
545 :			The installer (config/install.sh) has been taught how to properly
546 :			install this stuff.
547 :
548 :			----------------------------------------------------------------------
549 :			Name: Matthias Blume
550 :	blume	843	Date: 2001/06/19 17:55:00 EDT
551 :			Tag: blume-20010619-instantiate
552 :			Description:
553 :
554 :			This un-breaks the fix for bug 1432.
555 :			(The bug was originally fixed in 110.9 but I broke it again some
556 :			time after that.)
557 :
558 :			----------------------------------------------------------------------
559 :			Name: Matthias Blume
560 :	blume	842	Date: 2001/06/19 17:25:00 EDT
561 :			Tag: blume-20010619-signals
562 :			Description:
563 :
564 :			This should (hopefully) fix the long-standing signal handling bug.
565 :			(The runtime system was constructing a continuation record with an
566 :			incorrect descriptor which would cause the GC to drop data on the floor...)
567 :
568 :			----------------------------------------------------------------------
569 :			Name: Matthias Blume
570 :	blume	840	Date: 2001/06/15 15:05:00 EDT
571 :	blume	841	Tag: blume-20010615-moresparc
572 :			Description:
573 :
574 :			Here is a short late-hour update related to Sparc c-calls:
575 :
576 :			-- made handling of double-word arguments a bit smarter
577 :
578 :			-- instruction selection phase tries to collapse certain clumsily
579 :			constructed ML-Trees; typical example:
580 :
581 :			ADD(ty,ADD(_,e,LI d1),LI d2) -> ADD(ty,e,LI(d1+d2))
582 :
583 :			This currently has no further impact on SML/NJ since mlriscGen does
584 :			not seem to generate such patterns in the first place, and c-calls
585 :			(which did generate them in the beginning) has meanwhile been fixed
586 :			so as to avoid them as well.
587 :
588 :			----------------------------------------------------------------------
589 :			Name: Matthias Blume
590 :			Date: 2001/06/15 15:05:00 EDT
591 :	blume	840	Tag: blume-20010615-sparc
592 :			Description:
593 :
594 :			The purpose of this update is to provide an implementation of NLFFI
595 :			on Sparc machines.
596 :
597 :			Here are the changes in detail:
598 :
599 :			* src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
600 :			the Sparc implementation of the c-calls API.
601 :			* The Sparc backend of SML/NJ has been modified to uniformely use %fp
602 :			for accessing the ML frame. Thus, we have a real frame pointer and
603 :			can freely modify %sp without need for an omit-frame-ptr phase.
604 :			The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
605 :			this case.
606 :			* ml-nlffigen has been taught to produce code for different architectures
607 :			and calling conventions.
608 :			* In a way similar to what was done in the x86 case, the Sparc
609 :			backend uses its own specific extension to mltree. (For example,
610 :			it needs to be able to generate UNIMP instructions which are part
611 :			of the calling convention.)
612 :			* ml-nlffi-lib was reorganized to make it more modular (in particular,
613 :			to make it easier to plug in new machine- and os-dependent parts).
614 :
615 :			There are some other fairly unrelated bug fixes and cleanups as well:
616 :
617 :			* I further hacked the .cm files for MLRISC tools (like MDLGen) so
618 :			that they properly share their libraries with existing SML/NJ libraries.
619 :			* I fixed a minor cosmetic bug in CM, supressing certain spurious
620 :			follow-up error messages.
621 :			* Updates to CM/CMB documentation.
622 :
623 :			TODO items:
624 :
625 :			* MLRISC should use a different register as its asmTemp on the Sparc.
626 :			(The current %o2 is a really bad choice because it is part of the
627 :			calling conventions, so things might interfere in unexpected ways.)
628 :
629 :			----------------------------------------------------------------------
630 :			Name: Matthias Blume
631 :	blume	839	Date: 2001/06/07
632 :			Tag: blume-20010607-calls
633 :			Description:
634 :
635 :			A number of internal changes related to C calls and calling conventions:
636 :
637 :			1. ML-Tree CALL statements now carry a "pops" field. It indicates the
638 :			number of bytes popped implicitly (by the callee). In most cases
639 :			this field is 0 but on x86/win32 it is some non-zero value. This
640 :			is information provided for the benefit of the "omit-frameptr" pass.
641 :			2. The CALL instruction on the x86 carries a similar "pops" field.
642 :			The instruction selection phase copies its value from the ML-Tree
643 :			CALL statement.
644 :			3. On all other architectures, the instruction selection phase checks
645 :			whether "pops=0" and complains if not.
646 :			4. The c-calls implementation for x86 now accepts two calling conventions:
647 :			"ccall" and "stdcall". When "ccall" is selected, the caller cleans
648 :			up after the call and pops is set to 0. For "stdcall", the caller
649 :			does nothing, leaving the cleanup to the callee; pops is set to
650 :			the number of bytes that were pushed onto the stack.
651 :			5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
652 :			distinguish between "ccall" and "stdcall".
653 :			6. The UNIMP instruction has been added to the supported Sparc instruction
654 :			set. (This is needed for implementing the official C calling convention
655 :			on this architecture.)
656 :			7. I fixed some of the .cm files under src/MLRISC/Tools to make them
657 :			work with the latest CM.
658 :
659 :			----------------------------------------------------------------------
660 :			Name: Matthias Blume
661 :	blume	838	Date: 2001/06/05 15:10:00 EDT
662 :			Tag: blume-20010605-cm-index
663 :			Description:
664 :
665 :			0. The "lambdasplit" parameter for class "sml" in CM has been documented.
666 :
667 :			1. CM can now generate "index files". These are human-readable files
668 :			that list on a per-.cm-file basis each toplevel symbol defined or
669 :			imported. The location of the index file for
670 :			<p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
671 :			To enable index-file generation, set CM.Control.generate_index to true
672 :			or export an environment-symbol: export CM_GENERATE_INDEX=true.
673 :
674 :			The CM manual has been updated accordingly.
675 :
676 :			2. I made some slight modifications to the c-calls API in MLRISC.
677 :
678 :			a) There is now a callback to support saving/restoring of
679 :			dedicated but caller-save registers around the actual call
680 :			instruction.
681 :			b) One can optionally specify a comment-annotation for the
682 :			call instruction.
683 :
684 :			3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
685 :			(For example, the comment annotation shows the C prototype of
686 :			the function being called.)
687 :
688 :			----------------------------------------------------------------------
689 :			Name: Matthias Blume
690 :	blume	837	Date: 2001/06/01 13:30:00 EDT
691 :			Tag: blume-20010601-nlffi-cleanup
692 :			Description:
693 :
694 :			This is mostly a cleanup of MLFFI stuff:
695 :
696 :			- some signature files have been put into a more exposed place
697 :			- the ugly 'f type parameter is gone (simplifies types tremendously!)
698 :			- ml-nlffigen changed accordingly
699 :			- tutorial updated
700 :
701 :			Other changes:
702 :
703 :			- author's affiliation in CM manual(s) updated
704 :			- some more recognized keywords added to Allen's sml.sty
705 :
706 :			----------------------------------------------------------------------
707 :			Name: Matthias Blume
708 :	blume	836	Date: 2001/05/25 15:30:00 EDT
709 :			Tag: blume-20010525-iptr
710 :			Description:
711 :
712 :			- put the official 110.33-README (as it appears on the ftp server) under
713 :			CVS
714 :			- fixed a small bug related to incomplete pointer types in
715 :			ml-nlffigen
716 :			- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
717 :			(it does not need the 'f type parameter)
718 :
719 :			----------------------------------------------------------------------
720 :			Name: Matthias Blume
721 :	blume	834	Date: 2001/05/23 14:30:00 EDT
722 :			Tag: Release_110_33
723 :			Description:
724 :
725 :			New version number (110.33). New bootfiles.
726 :
727 :			----------------------------------------------------------------------
728 :			Name: Matthias Blume
729 :	blume	833	Date: 2001/05/22 18:06:00 EDT
730 :			Tag: blume-20010522-targets
731 :			Description:
732 :
733 :			Made install.sh use file config/targets.customized if it exists, falling
734 :			back to config/targets if it doesn't. This way one can have a customized
735 :			version of the targets file without touching the "real thing", thus
736 :			eliminating the constant fear of accidentally checking something bogus
737 :			back into the CVS repository... (File config/targets.customized must
738 :			not be added to the repository!)
739 :
740 :			----------------------------------------------------------------------
741 :			Name: Matthias Blume
742 :	blume	832	Date: 2001/05/22 16:30:00 EDT
743 :			Tag: blume-20010522-minitut
744 :			Description:
745 :
746 :			1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
747 :			struct returns.
748 :			2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt. This is some very
749 :			incomplete, preliminary documentation for NLFFI.
750 :
751 :			----------------------------------------------------------------------
752 :			Name: Matthias Blume
753 :	blume	829	Date: 2001/05/14 11:30:00 EDT
754 :			Tag: blume-20010514-script
755 :			Description:
756 :
757 :			Some bugs in install script fixed.
758 :
759 :			In addition to that I also made a slight change to the NLFFI API:
760 :			Functors generated by ml-nlffigen now take the dynamic library as a
761 :			straight functor argument, not as a suspended one. (The original
762 :			functor code used to force the suspension right away anyway, so there
763 :			was nothing gained by this complication of the interface.)
764 :
765 :			----------------------------------------------------------------------
766 :			Name: Matthias Blume
767 :	blume	828	Date: 2001/05/11 14:35:00 EDT
768 :			Tag: blume-20010511-ml-nlffi
769 :			Description:
770 :
771 :			I finally took the plunge and added my new FFI code to the main
772 :			repository. For x86-linux it is now ready for prime-time.
773 :
774 :			There are two new subdirectories of "src":
775 :
776 :			- ml-nlffi-lib:
777 :			The utility library for programs using the FFI interface.
778 :			Here is the implementation of $/c.cm and its associated low-level
779 :			partners $/c-int.cm and $/memory.cm.
780 :			- ml-nlffigen:
781 :			A stand-alone program for generating ML glue code from C source
782 :			code.
783 :
784 :			Building ml-nlffigen requires $/ckit-lib.cm.
785 :
786 :			The config/install.sh script has been updates to do the Right Thing
787 :			(hopefully).
788 :
789 :			Notice that the source tree for the C-Kit will not be put under "src"
790 :			but directly under the installation root directory. (This is the
791 :			structure that currently exists on the CVS server when you check out
792 :			module "sml".) Fortunately, config/install.sh knows about this oddity.
793 :
794 :			Bugs: No documentation yet.
795 :
796 :			----------------------------------------------------------------------
797 :			Name: Matthias Blume
798 :	blume	826	Date: 2001/05/09 16:35:00 EDT
799 :			Tag: blume-20010509-cpscontract
800 :			Description:
801 :
802 :			Fixed a bug in the accounting code in cpsopt/contract.sml. (The
803 :			wrapper/unwrapper elimination did not decrement usage counts and some
804 :			dead variables got overlooked by the dead-up logic.)
805 :
806 :			----------------------------------------------------------------------
807 :	george	824	Name: Lal George
808 :			Date: 2001/05/08 17:26:09 EDT
809 :			Tag: george-20010508-omit-frameptr
810 :			Description:
811 :
812 :			Changes to implement the omit-frame-pointer optimization to support
813 :			raw C calls. For now, there is only support on the Intel x86, but
814 :			other architectures will follow as more experience is gained with this.
815 :
816 :
817 :			----------------------------------------------------------------------
818 :	blume	818	Name: Matthias Blume
819 :	blume	819	Date: 2001/05/07 14:40:00 EDT
820 :			Tag: blume-20010507-proxies
821 :			Description:
822 :
823 :			I made into "proxy libraries" all libraries that qualify for such a
824 :			change. (A qualifying library is a library that has another library or
825 :			groups as its sole member and repeats that member's export list
826 :			verbatim. A proxy library avoids this repetition by omitting its export
827 :			list, effectively inheriting the list that its (only) member exports.
828 :			See the CM manual for more explanation.)
829 :			The main effect is that explicit export lists for these libraries
830 :			do not have to be kepts in sync, making maintenance a bit easier.
831 :
832 :			I also added copyright notices to many .cm-files.
833 :
834 :			Last but not least, I made a new set of bootfiles.
835 :
836 :			----------------------------------------------------------------------
837 :			Name: Matthias Blume
838 :	blume	818	Date: 2001/05/04 17:00:00 EDT
839 :			Tag: blume-20010504-cm-lsplit
840 :			Description:
841 :
842 :			0. John merged pending changes to $/smlnj-lib.cm
843 :
844 :			1. Allen's previous change accidentally backed out of one of Lal's
845 :			earlier changes. I undid this mistake (re-introducing Lal's change).
846 :
847 :			2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
848 :			within the compiler where applicable. There is some code simplification
849 :			because of that.
850 :
851 :			3. The "split" phase (in FLINT) is now part of the default list of phases.
852 :			Compiler.Control.LambdaSplitting.* can be used to globally control the
853 :			lambda-splitting (cross-module-inlining) engine. In addition to that,
854 :			it can now also be controlled on a per-source basis: CM has been taught
855 :			a new tool parameter applicable to ML source files.
856 :
857 :			- To turn lambda-splitting off completely:
858 :			local open Compiler.Control.LambdaSplitting in
859 :			val _ = set Off
860 :			end
861 :			- To make "no lambda-splitting" the global default (but allow per-source
862 :			overriding); this is the initial setting:
863 :			local open Compiler.Control.LambdaSplitting in
864 :			val _ = set (Default NONE)
865 :			end
866 :			- To make "lambda-splitting with aggressiveness a" the global default
867 :			(and allow per-source overriding):
868 :			local open Compiler.Control.LambdaSplitting in
869 :			val _ = set (Default (SOME a))
870 :			end
871 :
872 :			- To turn lambda-splitting off for a given ML souce file (say: a.sml)
873 :			write (in the respective .cm-file):
874 :			a.sml (lambdasplitting:off)
875 :			- To turn lambda-splitting for a.sml on with minimal aggressiveness:
876 :			a.sml (lambdasplitting:on)
877 :			- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
878 :			<a> is a decimal non-negative integer):
879 :			a.sml (lambdasplitting:<a>)
880 :			- To turn lambda-splitting for a.sml on with maximal aggressiveness:
881 :			a.sml (lambdasplitting:infinity)
882 :			- To use the global default for a.sml:
883 :			a.sml (lambdasplitting:default)
884 :			or simply
885 :			a.sml
886 :
887 :			----------------------------------------------------------------------
888 :	leunga	815	Name: Allen Leung
889 :			Date: 2001/05/04 01:57:00 EDT
890 :			Tag: leunga-20010504-sync
891 :			Description:
892 :	leunga	788
893 :	leunga	815	MLRISC features.
894 :
895 :			1. Fix to CMPXCHG instructions.
896 :			2. Changed RA interface to allow annotations in callbacks.
897 :			3. Added a new method to the stream interface to allow annotations updates.
898 :
899 :	george	749	----------------------------------------------------------------------
900 :	blume	812	Name: Matthias Blume
901 :	blume	814	Date: 2001/05/01 11:45:00 EDT
902 :			Tag: blume-20010501-pcedittmp
903 :			Description:
904 :
905 :			Changed install.sh to use the current working directory instead of
906 :			/usr/tmp for a temporary file (pcedittmp). The previous choice
907 :			of /usr/tmp caused trouble with MacOS X because of file premission
908 :			problems.
909 :
910 :			----------------------------------------------------------------------
911 :			Name: Matthias Blume
912 :	blume	812	Date: 2001/04/20 11:10:00 EDT
913 :			Tag: blume-20010420-inMLflag
914 :			Description:
915 :
916 :			- added vp_limitPtrMask to vproc-state.h
917 :			(for use by the raw-C-calls mechanism to implement proper interrupt
918 :			handling)
919 :			- made the ML compiler aware of various data-structure offsets so it
920 :			can generate code for accessing the vp_inML flag and vp_limitPtrMask
921 :			- tweaked mlriscGen.sml to have it emit interrupt-handling code for
922 :			raw C-calls
923 :
924 :			----------------------------------------------------------------------
925 :	george	810	Name: Lal George
926 :			Date: 2001/04/20 09:15:28 EDT
927 :			Tag: george-20010420-macosX
928 :			Description:
929 :
930 :			- Changes to port to Mac OS X; Darwin.
931 :
932 :			- In the process I found that sqrt was broken on the PPC, because the
933 :			fsqrt instruction is not implemented.
934 :
935 :			----------------------------------------------------------------------
936 :	blume	799	Name: Matthias Blume
937 :	blume	808	Date: 2001/04/18 12:45:00 EDT
938 :			Tag: blume-20010418-ccalls
939 :			Description:
940 :
941 :			- fixed two off-by-4 errors in the x86-specific c-calls implementation
942 :			(this bug prevented structure arguments containing pointers from being
943 :			passed correctly)
944 :			- changed the raw-C-call code in mlriscGen.sml in such a way that
945 :			structure arguments are represented as a pointer to the beginning
946 :			of the structure (instead of having a series of synthesized arguments,
947 :			one for each structure member)
948 :
949 :			- made makeml script's verbosity level configurable via environment
950 :			variable (MAKEML_VERBOSITY)
951 :
952 :	blume	813	- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
953 :	blume	808	in rawmem-x86.sml; we are now using the real thing
954 :
955 :			----------------------------------------------------------------------
956 :			Name: Matthias Blume
957 :	blume	806	Date: 2001/03/22 16:25:00 EST
958 :			Tag: blume-20010322-bootfiles
959 :			Description:
960 :
961 :			Created a new set of bootfiles (for your automatic installation convenience).
962 :
963 :			----------------------------------------------------------------------
964 :			Name: Matthias Blume
965 :	blume	805	Date: 2001/03/22 15:10:00 EST
966 :			Tag: blume-20010322-rawmem-parcm
967 :			Description:
968 :
969 :			1. All "raw memory access" primitives for the new FFI are implemented now
970 :			(at least on the x86).
971 :			2. Some further cleanup of CM's parallel make mechanism.
972 :
973 :			----------------------------------------------------------------------
974 :			Name: Matthias Blume
975 :	blume	801	Date: 2001/03/19 17:53:00 EST
976 :			Tag: blume-20010319-parallel
977 :			Description:
978 :
979 :			Parallel make (using compile servers) now works again.
980 :
981 :			To this end, CM.stabilize and CMB.make have been modified to work in
982 :			two passes when compile servers are attached:
983 :			1. Compile everything, do not perform stabilization; this pass
984 :			uses compile servers
985 :			2. Stabilize everything; this pass does not use compile servers
986 :			If there are no compile servers, the two passes are combined into one
987 :			(as before). Splitting the passes increases the inherent parallelism
988 :			in the dependency graph because the entire graph including all
989 :			libraries is available at the same time. This, in turn, improves
990 :			server utilization. The downside is that the master process will
991 :			have to do some extra work after compilation is done (because for
992 :			technical reasons it must re-read all the binfiles during stabilization).
993 :
994 :			----------------------------------------------------------------------
995 :			Name: Matthias Blume
996 :	blume	800	Date: 2001/03/16 12:22:00 EST
997 :			Tag: blume-20010316-bootfiles
998 :			Description:
999 :
1000 :			Created a new set of bootfiles (for your automatic installation convenience).
1001 :
1002 :			----------------------------------------------------------------------
1003 :			Name: Matthias Blume
1004 :	blume	799	Date: 2001/03/16 11:00:00 EST
1005 :			Tag: blume-20010316-MLTREE-fixup
1006 :			Description:
1007 :
1008 :			This is a minor fixup for an (untagged) earlier commit by Allen.
1009 :			(A file was missing).
1010 :
1011 :			----------------------------------------------------------------------
1012 :	leunga	788	Name: Allen Leung
1013 :	leunga	796	Date: Mon Mar 5 18:54:57 EST 2001
1014 :			Tag: leunga-20010305-cut-support
1015 :
1016 :			1. New support for alternative control-flow in MLTREE.
1017 :			Currently we support
1018 :
1019 :			FLOW_TO(CALL ...., [k1,...,kn])
1020 :
1021 :			This is needed for 'cuts to' in C-- and try/handle-like constructs
1022 :			in Moby
1023 :
1024 :			New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1025 :
1026 :			2. Register Allocator
1027 :
1028 :			Changes in interface [from Fermin, John]
1029 :
1030 :			3. Alpha 8-bit SLL support [Fermin]
1031 :
1032 :			4. All architectures
1033 :
1034 :			A new module (ClusterExpandCopies) for expanding parallel copies.
1035 :
1036 :			----------------------------------------------------------------------
1037 :			Name: Allen Leung
1038 :	leunga	788	Date: 2001/02/27 23:07:00 EST
1039 :			Tag: leunga-20010227-minor-stuff
1040 :
1041 :			1. Alpha bug fix for CMOVNE
1042 :			2. Handle mltree COND(..,FCMP ...,...)
1043 :			3. Bug fix in simplifier
1044 :
1045 :			----------------------------------------------------------------------
1046 :	blume	777	Name: Matthias Blume
1047 :	blume	781	Date: 2001/01/30 17:50:00 EST
1048 :			Tag: blume-20010130-sync
1049 :			Description:
1050 :
1051 :			This is just a minor update to sync my devel branch with the main brach.
1052 :			The only visible change is the addition of some README files.
1053 :
1054 :			----------------------------------------------------------------------
1055 :			Name: Matthias Blume
1056 :	blume	778	Date: 2001/01/12 23:30:00 JST
1057 :			Tag: blume-20010112-bootfiles
1058 :			Description:
1059 :
1060 :			Made a new set of bootfiles that goes with the current state of the
1061 :			repository.
1062 :
1063 :			----------------------------------------------------------------------
1064 :			Name: Matthias Blume
1065 :	blume	777	Date: 2001/01/12 21:20:00 JST
1066 :			Tag: blume-20010112-sync
1067 :			Description:
1068 :
1069 :			I am just flushing out some minor changes that had accumulated in
1070 :			my private branch in order to sync with the main tree. (This is
1071 :			mainly because I had CVS trouble when trying to merge _into_ my
1072 :			private branch.)
1073 :
1074 :			Most people should be completely unaffected by this.
1075 :
1076 :			----------------------------------------------------------------------
1077 :	leunga	775	Name: Allen Leung
1078 :			Date: Thu Jan 11 21:03:00 EST 2001
1079 :			Tag: leunga-20010111-labexp=mltree
1080 :			Description:
1081 :
1082 :			1. Removed the type LabelExp and replace it by MLTree.
1083 :			2. Rewritten mltree-simplify with the pattern matcher tool.
1084 :			3. There were some bugs in alpha code generator which would break
1085 :			64-bit code generation.
1086 :			4. Redo the tools to generate code with the
1087 :			5. The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1088 :			are now generated by perl scripts.
1089 :
1090 :			----------------------------------------------------------------------
1091 :	blume	771	Name: Matthias Blume
1092 :	blume	774	Date: 2001/01/10 21:55:00 JST
1093 :			Tag: blume-20010110-rcc
1094 :			Description:
1095 :
1096 :			The RCC stuff now seems to work (but only on the x86).
1097 :			This required hacking of the c-calls interface (and -implementation) in
1098 :			MLRISC.
1099 :
1100 :			Normal compiler users should be unaffected.
1101 :
1102 :			----------------------------------------------------------------------
1103 :			Name: Matthias Blume
1104 :	blume	773	Date: 2001/01/09 01:20:00 JST
1105 :			Tag: blume-20010109-rcc
1106 :			Description:
1107 :
1108 :			This is a fairly big patch, flushing out a large number of pending
1109 :			changes that I made to my development copy over the last couple of days.
1110 :
1111 :			Of practical relevance at this moment is a workaround for a pickling
1112 :			bug that Allen ran into the other day. The cause of the bug itself is
1113 :			still unknown and it might be hard to fix it properly, but the
1114 :			workaround has some merits of its own (namely somewhat reducing pickling
1115 :			overhead for certain libraries). Therefore, I think this solution should
1116 :			be satisfactory at this time.
1117 :
1118 :			The rest of the changes (i.e., the vast majority) has to do with my
1119 :			ongoing efforts of providing direct support for C function calls from
1120 :			ML. At the moment there is a new primop "RAW_CCALL", typing magic
1121 :			in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1122 :			case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1123 :			translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1124 :			handle RCC.
1125 :
1126 :			The last part (the changes to mlriscGen.sml) are still known to be
1127 :			wrong on the x86 and not implemented on all other architectures. But
1128 :			the infrastructure is in place. I had to change a few functor
1129 :			signatures in the backend to be able to route the CCalls interface
1130 :			from MLRISC there, and I had to specialize the mltree type (on the
1131 :			x86) to include the necessary extensions. (The extensions themselves
1132 :			were already there and redy to go in MLRISC/x86).
1133 :
1134 :			Everything should be very happy as soon as someone helps me with
1135 :			mlriscGen.sml...
1136 :
1137 :			In any case, nothing of this should matter to anyone as long as the
1138 :			new primop is not being used (which is going to be the case unless you
1139 :			find it where I hid it :). The rest of the compiler is completely
1140 :			unaffected.
1141 :
1142 :			----------------------------------------------------------------------
1143 :			Name: Matthias Blume
1144 :	blume	772	Date: 2001/01/05 00:30:00 JST
1145 :			Tag: blume-20010105-primops
1146 :			Description:
1147 :
1148 :			Added some experimental support for work that I am doing right now.
1149 :			These changes mostly concern added primops, but there is also a new
1150 :			experimental C library in the runtime system (but currently not enabled
1151 :			anywhere except on Linux/X86).
1152 :
1153 :			In the course of adding primops (and playing with them), I discovered that
1154 :			Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1155 :			broken. (Zhong was very right he labeled this stuff as "major gross hack".)
1156 :			To recover, I made type information in INL_PRIM mandatory and changed
1157 :			prim.sml as well as built-in.sml accordingly. The InLine structure now
1158 :			has complete, correct type information (i.e., no bottom types).
1159 :
1160 :			Since all these changes mean that we need new binfiles, I also bumped the
1161 :			version number to 110.32.1.
1162 :
1163 :			----------------------------------------------------------------------
1164 :			Name: Matthias Blume
1165 :	blume	771	Date: 2000/12/30 22:10:00 JST
1166 :			Tag: blume-20001230-various
1167 :			Description:
1168 :
1169 :			Added proxy libraries for MLRISC and let MLRISC libraries refer
1170 :			to each other using path anchors. (See CM manual for explanation.)
1171 :
1172 :			Updated CM documentation.
1173 :
1174 :			Fixed some bugs in CM.
1175 :
1176 :			Implemented "proxy" libraries (= syntactic sugar for CM).
1177 :
1178 :			Added "-quiet" option to makeml and changed runtime system accordingly.
1179 :
1180 :			Added cleanup handler for exportML to reset timers and compiler stats.
1181 :
1182 :			----------------------------------------------------------------------
1183 :	george	761	Name: Lal George
1184 :			Date: 2000/12/22 22:22:58 EST 2000
1185 :			Tag: Release_110_32
1186 :			Description:
1187 :
1188 :			Infinite precision used throughout MLRISC.
1189 :			see MLRISC/mltree/machine-int.sig
1190 :
1191 :			----------------------------------------------------------------------
1192 :	blume	756	Name: Matthias Blume
1193 :	blume	760	Date: 2000/12/22 23:16:00 JST
1194 :			Tag: blume-20001222-warn
1195 :			Description:
1196 :	blume	759
1197 :	blume	760	Corrected wording and formatting of some CM warning message which I
1198 :			broke in my previous patch.
1199 :	blume	759
1200 :			----------------------------------------------------------------------
1201 :			Name: Matthias Blume
1202 :	blume	763	Date: 2000/12/22 21:20:00 JST
1203 :			Tag: blume-20001222-anchorenv
1204 :			Description:
1205 :
1206 :			Fixed CM's handling of anchor environments in connection with CMB.make.
1207 :
1208 :			----------------------------------------------------------------------
1209 :			Name: Matthias Blume
1210 :	blume	758	Date: 2000/12/22 13:15:00 JST
1211 :			Tag: blume-20001222-cleanup
1212 :			Description:
1213 :
1214 :			Removed src/cm/ffi which does not (and did not) belong here.
1215 :
1216 :			----------------------------------------------------------------------
1217 :			Name: Matthias Blume
1218 :	blume	757	Date: 2000/12/21 23:55:00 JST
1219 :			Tag: blume-20001221-exn
1220 :			Description:
1221 :
1222 :			Probably most important: CM no longer silently swallows all exceptions
1223 :			in the compiler.
1224 :			Plus: some other minor CM changes. For example, CM now reports some
1225 :			sizes for generated binfiles (code, data, envpickle, lambdapickle).
1226 :
1227 :			----------------------------------------------------------------------
1228 :			Name: Matthias Blume
1229 :	blume	756	Date: 2000/12/15 00:01:05 JST
1230 :			Tag: blume-20001215-dirtool
1231 :			Description:
1232 :
1233 :			- "dir" tool added.
1234 :			- improvements and cleanup to Tools structure
1235 :			- documentation updates
1236 :
1237 :			----------------------------------------------------------------------
1238 :	leunga	752	Name: Allen Leung
1239 :	leunga	755	Date: Thu Dec 14 03:45:24 EST 2000
1240 :			Description:
1241 :			Tag: leunga-20001214-int-inf
1242 :			Description:
1243 :
1244 :			In IntInf, added these standard functions, which are missing from our
1245 :			implementation:
1246 :
1247 :			andb : int * int -> int
1248 :			xorb : int * int -> int
1249 :			orb : int * int -> int
1250 :			notb : int -> int
1251 :			<< : int * word -> int
1252 :			~>> : int * word -> int
1253 :
1254 :			Not tested, I hope they are correct.
1255 :
1256 :			----------------------------------------------------------------------
1257 :			Name: Allen Leung
1258 :	leunga	752	Date: Fri Dec 8 19:23:26 EST 2000
1259 :			Description:
1260 :			Tag: leunga-20001208-nowhere
1261 :			Description:
1262 :
1263 :			Slight improvements to the 'nowhere' tool to handle OR-patterns,
1264 :			to generate better error messages etc. Plus a brief manual.
1265 :
1266 :			----------------------------------------------------------------------
1267 :	george	749	Name: Lal George
1268 :			Date: 2000/12/08 09:54:02 EST 2000
1269 :			Tag: Release_110_31
1270 :			Description:
1271 :	george	705
1272 :	george	749	- Version 110.31
1273 :	leunga	591	----------------------------------------------------------------------
1274 :	leunga	744	Name: Allen Leung
1275 :			Date: Thu Dec 7 22:01:04 EST 2000
1276 :			Tag: leunga-20001207-cell-monster-hack
1277 :			Description:
1278 :
1279 :			Major MLRISC internal changes. Affect all clients.
1280 :			Summary:
1281 :
1282 :			1. Type CELLS.cell = int is now replaced by a datatype.
1283 :			As a result, the old regmap is now gone. Almost all interfaces
1284 :			in MLRISC change as a consequence.
1285 :
1286 :			2. A new brand version of machine description tool (v3.0) that generates
1287 :			modules expecting the new interface. The old version is removed.
1288 :
1289 :			3. The RA interface has been further abstracted into two new functors.
1290 :			RISC_RA and X86RA. These functors have much simpler interfaces.
1291 :			[See also directory MLRISC/demo.]
1292 :
1293 :			4. Some other new source->source code generation tools are available:
1294 :
1295 :			a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
1296 :			b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
1297 :			I use this tool to generate the peephole optimizers---with the new
1298 :			cell type changes, peephole rules are becoming difficult to write
1299 :			without conditional pattern matching.
1300 :
1301 :			5. More Intmap -> IntHashTable change. Previous changes by Matthias didn't
1302 :			cover the entire MLRISC source tree so many things broke.
1303 :
1304 :			6. CM files have been moved to the subdirectory MLRISC/cm.
1305 :			They are moved because there are a lot of them and they clutter up the
1306 :			root dir.
1307 :
1308 :			7. More detailed documentation to come...
1309 :
1310 :			NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
1311 :			a makeml -rebuild first. This is because of other other
1312 :			changes that Matthias has made (see below).
1313 :
1314 :
1315 :			----------------------------------------------------------------------
1316 :	blume	733	Name: Matthias Blume
1317 :	blume	742	Date: 2000/11/30 23:12:00 JST
1318 :			Tag: blume-20001130-filereorg
1319 :			Description:
1320 :
1321 :			Some manual updates and some file reorganizations in CM.
1322 :
1323 :			----------------------------------------------------------------------
1324 :			Name: Matthias Blume
1325 :	blume	737	Date: 2000/11/24 17:45:00 JST
1326 :			Tag: blume-20001124-link
1327 :			Description:
1328 :
1329 :			Drastically improved link traversal code for the case that the dynamic
1330 :			value was already loaded at bootstrap time. As a result, CM and CMB
1331 :			now both load blazingly fast -- even on a very slow machine. Also,
1332 :			memory consumption has been further reduced by this.
1333 :
1334 :			Warning: The format of the PIDMAP file has changed. THerefore, to
1335 :			bootstrap you have to do this:
1336 :
1337 :			1. Run CMB.make
1338 :			2. Make a symbolic link for the boot directory:
1339 :			ln -s sml.boot.ARCH-OS xxx
1340 :			3. "Rebuild" the boot directory:
1341 :			./makeml -boot xxx -rebuild sml ; rm xxx
1342 :			4. Boot normally:
1343 :			./makeml
1344 :
1345 :			----------------------------------------------------------------------
1346 :			Name: Matthias Blume
1347 :	blume	735	Date: 2000/11/21 21:20:00 JST
1348 :			Tag: blume-20001121-tools
1349 :			Description:
1350 :
1351 :			Continued hacking on autoloading problem -- with success this time.
1352 :			Also changed tool-plugin mechanism. See new CM manual.
1353 :
1354 :			----------------------------------------------------------------------
1355 :			Name: Matthias Blume
1356 :	blume	734	Date: 2000/11/19 14:30:00 JST
1357 :			Tag: blume-20001119-autoload
1358 :			Description:
1359 :
1360 :			Some hacking to make autoloading faster. Success for CMB, no success
1361 :			so far for CM. There is a reduced structure CM' that autoloads faster.
1362 :			(This is a temporary, non-documented hack to be eliminated again when
1363 :			the general problem is solved.)
1364 :
1365 :			----------------------------------------------------------------------
1366 :			Name: Matthias Blume
1367 :	blume	733	Date: 2000/11/17 14:10:00 JST
1368 :			Tag: blume-20001117-pickle-lib
1369 :			Description:
1370 :
1371 :			1. Eliminated comp-lib.cm
1372 :			2. Made pickle-lib.cm
1373 :			3. Eliminated all uses of intset.sml (from comp-lib.cm)
1374 :			4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
1375 :			equivalent constructs from smlnj-lib.cm (INtHashTable).
1376 :			5. Point 4. also goes for those uses of intmap.* in MLRISC.
1377 :			Duplicated intmap modules thrown out.
1378 :			6. Hunted down all duplicated SCC code and replaced it with
1379 :			equivalent stuff (GraphSCCFn from smlnj-lib.cm).
1380 :			7. Rewrote Feedback module.
1381 :			8. Moved sortedlist.sml into viscomp-lib.cm. Eventually it
1382 :			should be thrown out and equivalent modules from smlnj-lib.cm
1383 :			should be used (IntRedBlackSet, IntListSet, ...).
1384 :
1385 :			Confirmed that compiler compiles to fixpoint.
1386 :
1387 :			----------------------------------------------------------------------
1388 :	leunga	731	Name: Allen Leung
1389 :			Date: 2000/11/10 18:00:00
1390 :			Tag: leunga-20001110-new-x86-fp
1391 :
1392 :			A new x86 floating point code generator has been added.
1393 :			By default this is turned off. To turn this on, do:
1394 :
1395 :			CM.autoload "$smlnj/compiler.cm";
1396 :			Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
1397 :
1398 :			Changes:
1399 :
1400 :			1. Changed FTAN to FPTAN so that the assembly output is correct.
1401 :			2. Changed the extension callback for FTANGENT to generate:
1402 :
1403 :			fptan
1404 :			fstp %st(0)
1405 :			instead of
1406 :			fptan
1407 :			fstpl ftempmem
1408 :
1409 :			3. Numerous assembly fixes for x86.
1410 :
1411 :			5. Cleaned up the machine code output module x86/x86MC.sml and added
1412 :			support for a whole bunch of instructions and addressing modes:
1413 :
1414 :			fadd/fsub/fsubr/fmul/fdiv/fdivr %st, %st(n)
1415 :			faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp %st, %st(n)
1416 :			fadd/fsub/fsubr/fmul/fdiv/fdivr %st(n), %st
1417 :			fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
1418 :			fxch %st(n)
1419 :			fld %st(n)
1420 :			fst %st(n)
1421 :			fst mem
1422 :			fstp %st(n)
1423 :			fucom %st(n)
1424 :			fucomp %st(n)
1425 :
1426 :			All these are now generated when the fast fp mode is turned on.
1427 :
1428 :			6. Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
1429 :
1430 :			----------------------------------------------------------------------
1431 :	blume	729	Name: Matthias Blume
1432 :			Date: 2000/11/09 11:20:00 JST
1433 :			Tag: blume-20001109-scc
1434 :			Description:
1435 :
1436 :			Eliminated some code duplication:
1437 :
1438 :			1. Added "where" clause to GraphSCCFn in SML/NJ Library.
1439 :			(Otherwise the functor is useless.)
1440 :			2. Used GraphSCCFn where SCCUtilFun was used previously.
1441 :			3. Got rid of SCCUtilFun (in comp-lib.cm).
1442 :
1443 :			----------------------------------------------------------------------
1444 :	george	721	Name: Lal George
1445 :			Date: 2000/11/06 09:02:21 EST 2000
1446 :			Tag: Release_110_30
1447 :			Description:
1448 :
1449 :			- Version 110.30
1450 :			----------------------------------------------------------------------
1451 :	blume	715	Name: Matthias Blume
1452 :	blume	716	Date: 2000/11/04 14:45:00
1453 :			Tag: blume-20001104-mlbuild
1454 :			Description:
1455 :
1456 :			- Made ml-build faster on startup.
1457 :			- Documentation fixes.
1458 :
1459 :			----------------------------------------------------------------------
1460 :			Name: Matthias Blume
1461 :	blume	715	Date: 2000/11/02 17:00:00 JST
1462 :			Tag: blume-20001102-condcomp
1463 :			Description:
1464 :
1465 :			- Small tweaks to pickler -- new BOOTFILES!
1466 :			- Version bumped to 110.29.2.
1467 :			- Added conditional compilation facility to init.cmi (see comment there).
1468 :			----------------------------------------------------------------------
1469 :	leunga	713	Name: Allen Leung
1470 :			Date: 2000/10/23 19:31:00
1471 :			Tag: leunga-20001023-demo-ra
1472 :
1473 :			1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
1474 :			2. Test programs for the graph library updated
1475 :			3. Some new MLRISC demo programs added
1476 :
1477 :			----------------------------------------------------------------------
1478 :	blume	710	Name: Matthias Blume
1479 :	blume	711	Date: 2000/08/31 22:15:00 JST
1480 :			Tag: blume-20001017-errmsg
1481 :			Description:
1482 :
1483 :			More error message grief: Where there used to be no messages, there
1484 :			now were some that had bogus error regions. Fixed.
1485 :
1486 :			----------------------------------------------------------------------
1487 :			Name: Matthias Blume
1488 :	blume	710	Date: 2000/08/31 17:30:00 JST
1489 :			Tag: blume-20001017-v110p29p1
1490 :			Description:
1491 :
1492 :			I made a version 110.29.1 with new bootfiles.
1493 :
1494 :			Changes: Modified pickler/unpickler for faster and leaner unpickling.
1495 :			CM documentation changes and a small bugfix in CM's error reporting.
1496 :
1497 :			----------------------------------------------------------------------
1498 :	george	705	Name: Lal George
1499 :			Date: 2000/09/27 14:42:35 EDT
1500 :			Tag: george-20000927-nodestatus
1501 :			Description:
1502 :
1503 :			Changed the type of the nodestatus, so that:
1504 :
1505 :			SPILLED(~1) is now SPILLED
1506 :			SPILLED(m) where m>=0 is now MEMREG(m)
1507 :			SPILLED(s) where s<~1 is now SPILL_LOC(~s)
1508 :
1509 :			----------------------------------------------------------------------
1510 :	blume	697	Name: Matthias Blume
1511 :	blume	703	Date: 2000/09/07 14:45:00 JST
1512 :			Tag: blume-20000907-cmerrmsg
1513 :			Description:
1514 :
1515 :			Small tweak to CM to avoid getting ML syntax error messages twice.
1516 :
1517 :			----------------------------------------------------------------------
1518 :			Name: Matthias Blume
1519 :	blume	701	Date: 2000/08/31 18:00:00 JST
1520 :			Tag: blume-20000831-cvsbootfiles
1521 :			Description:
1522 :
1523 :			New URL for boot files (because the 110.29 files on the BL server do
1524 :			now work correctly with my updated install scripts for yacc and lex).
1525 :
1526 :			----------------------------------------------------------------------
1527 :			Name: Matthias Blume
1528 :	blume	697	Date: 2000/08/08 12:33:00 JST
1529 :			Tag: blume-20000808-manual
1530 :			Description:
1531 :
1532 :			Tiny update to CM manual.
1533 :
1534 :			----------------------------------------------------------------------
1535 :	leunga	695	Name: Allen Leung
1536 :			Date: 2000/08/7 19:31:00
1537 :			Tag: leunga-20000807-a-whole-bunch-of-stuff
1538 :
1539 :			Moby, C--, SSA, x86, machine descriptions etc. Should only affect C--
1540 :			and Mobdy.
1541 :
1542 :			1. x86
1543 :
1544 :			a. Fixes to peephole module by John and Dan.
1545 :			b. Assembly fix to SETcc by Allen.
1546 :			c. Fix to c-call by John.
1547 :			d. Fix to spilling by John. (This one deals with the missing FSTPT case)
1548 :			e. Instruction selection optimization to SETcc as suggested by John.
1549 :
1550 :			For example,
1551 :
1552 :			MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1553 :
1554 :			should generate:
1555 :
1556 :			MOVL a, x
1557 :			SUBL b, x
1558 :			SHRL 31, x
1559 :
1560 :			2. IR stuff
1561 :
1562 :			A bunch of new DJ-graph related algorithms added. These
1563 :			speed up SSA construction.
1564 :
1565 :			3. SSA + Scheduling
1566 :
1567 :			Added code for SSA and scheduling to the repository
1568 :
1569 :			----------------------------------------------------------------------
1570 :	george	694	Name: Lal George
1571 :			Date: 2000/07/27 11:53:14 EDT
1572 :
1573 :			Tag: lal-20000727-linux-ppc
1574 :			Description:
1575 :
1576 :			Made changes to support Linux PPC.
1577 :			p.s. I have confirmation that the 110.29 boot files work fine.
1578 :
1579 :			----------------------------------------------------------------------
1580 :	blume	684	Name: Matthias Blume
1581 :	blume	692	Date: 2000/07/27 17:40:00 JST
1582 :			Tag: blume-20000727-scripts
1583 :			Description:
1584 :
1585 :			!!!! WARNING !!!!
1586 :			You must recompile the runtime system!
1587 :			!!!! WARNING !!!!
1588 :
1589 :			This is basically another round of script-enhancements:
1590 :
1591 :			1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1592 :			and undefine CM preprocessor symbols.
1593 :
1594 :			2. ml-build avoids generating a new heap image if it finds that the
1595 :			existing one is still ok. (The condition is that no ML file had to
1596 :			be recompiled and all ML files are found to be older that the heap
1597 :			file.)
1598 :
1599 :			To make this work smoothly, I also hacked the runtime system as
1600 :			well as SMLofNJ.SysInfo to get access to the heap image suffix
1601 :			(.sparc-solaris, ...) that is currently being used.
1602 :
1603 :			Moreover, the signature of CM.mk_standalone has changed. See the
1604 :			CM manual.
1605 :
1606 :			3. ml-makedepend accepts additional options -n, -a, and -o. (See the
1607 :			CM manual for details.)
1608 :
1609 :			4. More CM manual updates:
1610 :			- all of the above has been documented.
1611 :			- there is now a section describing the (CM-related) command line
1612 :			arguments that are accepted by the "sml" command
1613 :
1614 :			----------------------------------------------------------------------
1615 :			Name: Matthias Blume
1616 :	blume	691	Date: 2000/07/25 16:20:00 JST
1617 :			Tag: blume-20000725-makedepend
1618 :			Description:
1619 :
1620 :			Added a script called ml-makedepend. This can be used in makefiles
1621 :			for Unix' make in a way very similar to the "makedepend" command for
1622 :			C.
1623 :
1624 :			The script internally uses function CM.sources.
1625 :
1626 :			Synopsis:
1627 :
1628 :			ml-makedepend [-f makefile] cmfile targetname
1629 :
1630 :			The default for the makefile is "makefile" (or "Makefile" should
1631 :			"makefile" not exist).
1632 :
1633 :			ml-makedepend adds a cmfile/targetname-specific section to this
1634 :			makefile (after removing the previous version of this section). The
1635 :			section contains a single dependency specification with targetname on
1636 :			the LHS (targetname is an arbitrary name), and a list of files derived
1637 :			from the cmfile on the RHS. Some of the files on the RHS are
1638 :			ARCH/OPSYS-specific. Therefore, ml-makedepend inserts references to
1639 :			"make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1640 :			path names. The makefile writer is responsible for making sure that
1641 :			these variables have correct at the time "make" is invoked.
1642 :
1643 :			----------------------------------------------------------------------
1644 :			Name: Matthias Blume
1645 :	blume	690	Date: 2000/07/22 23:30:00 JST
1646 :			Tag: blume-20000722-urlupdate
1647 :			Description:
1648 :
1649 :			Changed BOOT and config/srcarchiveurl to point to BL server:
1650 :
1651 :			ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1652 :
1653 :			----------------------------------------------------------------------
1654 :			Name: Matthias Blume
1655 :	blume	686	Date: 2000/07/18 18:00:00 JST
1656 :			Tag: blume-20000718-Version_110_29
1657 :			Description:
1658 :
1659 :			1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1660 :
1661 :			2. Updated config/version to 110.29
1662 :
1663 :			3. Updated config/srcarchiveurl
1664 :
1665 :			3. New boot files!
1666 :			ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1667 :
1668 :			----------------------------------------------------------------------
1669 :			Name: Matthias Blume
1670 :	blume	684	Date: 2000/07/11 13:58:00 JST
1671 :			Tag: blume-20000711-doctypo
1672 :			Description:
1673 :
1674 :			Fixed a few typos in CM manual.
1675 :
1676 :			----------------------------------------------------------------------
1677 :	leunga	683	Name: Allen Leung
1678 :			Date: 2000/06/15 00:38:00
1679 :			Tag: leunga-20000704-sparc-x86
1680 :
1681 :			1. x86 peephole improvement sp += k; sp -= k => nop [from John]
1682 :			2. fix to x86 RET bug [found by Dan Grossman]
1683 :			3. sparc assembly bug fix for ticc instructions [found by Fermin]
1684 :
1685 :			Affects c-- and moby only
1686 :
1687 :			----------------------------------------------------------------------
1688 :	blume	680	Name: Matthias Blume
1689 :	blume	682	Date: 2000/07/04 15:26:00
1690 :			Tag: blume-20000704-trigger
1691 :			Description:
1692 :
1693 :			1. Improvements to CM manual.
1694 :			2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1695 :			of getting a back-trace. The function, when called, raises an
1696 :			internal exception which explicitly carries the full back-trace history,
1697 :			so it is unaffected by any intervening handle-raise pairs ("trivial"
1698 :			or not). The interactive loop will print that history once it arrives
1699 :			at top level.
1700 :			Short of having all exceptions implicitly carry the full history, the
1701 :			recommended way of using this facility is:
1702 :			- compile your program with instrumentation "on"
1703 :			- run it, when it raises an exception, look at the history
1704 :			- if the history is "cut off" because of some handler, go and modify
1705 :			your program so that it explicitly calls BTrace.trigger
1706 :			- recompile (still instrumented), and rerun; look at the full history
1707 :
1708 :			----------------------------------------------------------------------
1709 :			Name: Matthias Blume
1710 :			Date: 2000/07/03 15:36:00 JST
1711 :	blume	680	Tag: blume-20000702-manual
1712 :			Description:
1713 :
1714 :			Small corrections and updates to CM manual.
1715 :
1716 :			----------------------------------------------------------------------
1717 :	blume	682	Name: Matthias Blume
1718 :	blume	679	Date: 2000/06/29 16:04:00 JST
1719 :			Tag: blume-20000629-yacctool
1720 :			Description:
1721 :
1722 :			Changes:
1723 :
1724 :			1. Class "mlyacc" now takes separate arguments to pass options to
1725 :			generated .sml- and .sig-files independently.
1726 :			2. Corresponding CM manual updates.
1727 :			3. BTrace module now also reports call sites. (However, for loop clusters
1728 :			it only shows from where the cluster was entered.) There are associated
1729 :			modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1730 :
1731 :			----------------------------------------------------------------------
1732 :	blume	675	Name: Matthias Blume
1733 :	blume	678	Date: 2000/06/27 16:51:00 JST
1734 :			Tag: blume-20000627-noweb
1735 :			Description:
1736 :
1737 :			Changes:
1738 :
1739 :			1. Implemented "subdir" and "witness" options for noweb tool.
1740 :			This caused some slight internal changes in CM's tool implementation.
1741 :			2. Fixed bug in "tool plugin" mechanism. This is essentially cleaning
1742 :			some remaining issues from earlier path anchor changes.
1743 :			3. Updated CM manual accordingly.
1744 :
1745 :			4. Changed implementation of back-tracing so that I now consider it
1746 :			ready for prime-time.
1747 :
1748 :			In particular, you don't have to explicitly trigger the back-trace
1749 :			anymore. Instead, if you are running BTrace-instrumented code and
1750 :			there is an uncaught exception (regardless of whether or not it was
1751 :			raised in instrumented code), the top-level evalloop will print
1752 :			the back-trace.
1753 :
1754 :			Features:
1755 :
1756 :			- Instrumented and uninstrumented code work together seemlessly.
1757 :			(Of course, uninstrumented code is never mentioned in actual
1758 :			back-traces.)
1759 :
1760 :			- Asymptotic time- and space-complexity of instrumented code is
1761 :			equal to that of uninstrumented code. (This means that
1762 :			tail-recursion is preserved by the instrumentation phase.)
1763 :
1764 :			- Modules whose code has been instrumented in different sessions
1765 :			work together without problem.
1766 :
1767 :			- There is no penalty whatsoever on uninstrumented code.
1768 :
1769 :			- There is no penalty on "raise" expressions, even in
1770 :			instrumented code.
1771 :
1772 :			A potential bug (or perhaps it is a feature, too):
1773 :
1774 :			A back-trace reaches no further than the outermost instrumented
1775 :			non-trivial "raise". Here, a "trivial" raise is one that is the
1776 :			sole RHS of a "handle" rule. Thus, back-traces reach trough
1777 :
1778 :			<exp> handle e => raise e
1779 :
1780 :			and even
1781 :
1782 :			<exp> handle Foo => raise Bar
1783 :
1784 :			and, of course, through
1785 :
1786 :			<exp> handle Foo => ...
1787 :
1788 :			if the exception was not Foo.
1789 :
1790 :			Back-traces always reach right through any un-instrumented code
1791 :			including any of its "handle" expressions, trivial or not.
1792 :
1793 :			To try this out, do the following:
1794 :
1795 :			- Erase all existing binfiles for your program.
1796 :			(You may keep binfiles for those modules where you think you
1797 :			definitely don't need back-tracing.)
1798 :			- Turn on back-trace instrumentation:
1799 :			SMLofNJ.Internals.BTrace.mode (SOME true);
1800 :			- Recompile your program. (I.e., run "CM.make" or "use".)
1801 :			- You may now turn instrumentation off again (if you want):
1802 :			SMLofNJ.Internals.BTrace.mode (SOME false);
1803 :			- Run your program as usual. If it raises an exception that
1804 :			reaches the interactive toplevel, then a back-trace will
1805 :			automatically be printed. After that, the toplevel loop
1806 :			will print the exception history as usual.
1807 :
1808 :			----------------------------------------------------------------------
1809 :			Name: Matthias Blume
1810 :	blume	677	Date: 2000/06/26 09:56:46 JST
1811 :			Tag: blume-20000626-setup
1812 :			Description:
1813 :
1814 :			CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1815 :			ML code before and after compiling a file (e.g., to set compiler
1816 :			flags)
1817 :
1818 :			Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1819 :			- associated changes to btrace.sml (BTrace instrumentation pass)
1820 :			- cleaner implementation of btimp.sml (BTrace tracing and report
1821 :			module)
1822 :
1823 :			CM manual: * new path encoding documented
1824 :			* description of setup-parameter to "sml" added
1825 :
1826 :			The biggest user-visible change to back-tracing is that it is no
1827 :			longer necessary to compile all traced modules within the same
1828 :			session. (This was a real limitation.)
1829 :
1830 :			----------------------------------------------------------------------
1831 :			Name: Matthias Blume
1832 :	blume	676	Date: 2000/06/24 12:40:00 JST
1833 :			Tag: blume-20000624-startup
1834 :			Description:
1835 :
1836 :			Fixes startup slowdown problem. (I was calling SrcPath.sync a _tad_
1837 :			bit too often -- to put it mildly. :)
1838 :
1839 :			----------------------------------------------------------------------
1840 :			Name: Matthias Blume
1841 :	blume	675	Date: 2000/06/23 18:20:00 JST
1842 :			Tag: blume-20000623-btrace
1843 :			Description:
1844 :
1845 :			This updates adds a backtrace facility to aid programmers in debugging
1846 :			their programs. This involves the following changes:
1847 :
1848 :			1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1849 :			keeping track of the current call stack. When programs are compiled
1850 :			in a special mode, the compiler will insert calls to these hooks
1851 :			into the user program.
1852 :			"Hook" means that it is possible for different implementations of
1853 :			back-tracing to register themselves (at different times).
1854 :
1855 :			2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1856 :			as an Absyn.dec->Absyn.dec rewrite. Normally this phase is turned off.
1857 :			It can be turned on using this call:
1858 :			SMLofNJ.Internals.BTrace.mode (SOME true);
1859 :			Turning it off again:
1860 :			SMLofNJ.Internals.BTrace.mode (SOME false);
1861 :			Querying the current status:
1862 :			SMLofNJ.Internals.BTrace.mode NONE;
1863 :			Annotated programs are about twice as big as normal ones, and they
1864 :			run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1865 :			where all hooks do nothing). The slowdown with a plugin that is
1866 :			actually useful (such as the one supplied by default) is even greater,
1867 :			but in the case of the default plugin it is still only an constant
1868 :			factor (amortized).
1869 :
1870 :			3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1871 :			with a sub-structure BTrace for controlling back-tracing. In particular,
1872 :			the above-mentioned function "mode" controls whether the annotation
1873 :			phase is invoked by the compiler. Another important function is
1874 :			"trigger": when called it aborts the current execution and causes
1875 :			the top-level loop to print a full back-trace.
1876 :
1877 :			4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1878 :			for back-tracing. It keeps track of the dynamic call stack and in
1879 :			addition to that it keeps a partial history at each "level" of that
1880 :			stack. For example, if a tail-calls b, b tail-calls c, and c tail-calls
1881 :			d and b (at separate times, dynamically), then the report will show:
1882 :
1883 :			GOTO d
1884 :			/c
1885 :			GOTO \b
1886 :			CALL a
1887 :
1888 :			This shows that there was an initial non-tail call of a, then a
1889 :			tail-call to b or c, looping behavior in a cluster of functions that
1890 :			consist of b and c, and then a goto from that cluster (i.e., either from
1891 :			b or from c) to d.
1892 :
1893 :			Note that (depending on the user program) the amount of information
1894 :			that the back-trace module has to keep track of at each level is bounded
1895 :			by a constant. Thus, the whole implementation has the same asymptotical
1896 :			complexity as the original program (both in space and in time).
1897 :
1898 :			5. compiler/TopLevel/interact/evalloop.sml has been modified to
1899 :			handle the special exception SMLofNJ.Internals.BTrace.BTrace
1900 :			which is raised by the "trigger" function mentioned above.
1901 :
1902 :			Notes on usage:
1903 :
1904 :			- Annotated code works well together with unannotated code:
1905 :			Unannotated calls simply do not show up at all in the backtrace.
1906 :
1907 :			- It is not a good idea to let modules that were annotated during
1908 :			different sessions run at the same time. This is because the compiler
1909 :			chooses small integers to identify individual functions, and there
1910 :			will be clashes if different modules were compiled in separate sessions.
1911 :			(Nothing will crash, and you will even be told about the clashes, but
1912 :			back-trace information will in general not be useful.)
1913 :
1914 :			- Back-tracing can be confused by callcc and capture.
1915 :
1916 :			- The only way of getting a back-trace right now is to explicitly
1917 :			invoke the "trigger" function from your user program. Eventually, we
1918 :			should make every exception carry back-trace information (if
1919 :			available). But since this creates more overhead at "raise"-time
1920 :			(similar to the current exnHistory overhead), I have not yet
1921 :			implemented this. (The implementation will be rather easy.) With
1922 :			exceptions carrying back-trace information, this facility will be even
1923 :			more useful because users don't need to modify their programs...
1924 :
1925 :			- While it is possible to compile the compiler with back-trace
1926 :			annotations turned on (I did it to get some confidence in
1927 :			correctness), you must make absolutely sure that core.sml and
1928 :			btimp.sml are compiled WITHOUT annotation! (core.sml cannot actually
1929 :			be compiled with annotation because there is no core access yet, but
1930 :			if you compile btimp.sml with annotation, then the system will go into
1931 :			an infinite recursion and crash.)
1932 :			Since CM currently does not know about BTrace, the only way to turn
1933 :			annotations on and off for different modules of the compiler is to
1934 :			interrupt CMB.make, change the settings, and re-invoke it. Of course,
1935 :			this is awkward and clumsy.
1936 :
1937 :			Sample sessions:
1938 :
1939 :			Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1940 :			- SMLofNJ.Internals.BTrace.mode (SOME true);
1941 :			[autoloading]
1942 :			[autoloading done]
1943 :			val it = false : bool
1944 :			- structure X = struct
1945 :			- fun main n = let
1946 :			- fun a (x, 0) = d x
1947 :			- | a (x, n) = b (x, n - 1)
1948 :			- and b (x, n) = c (x, n)
1949 :			- and c (x, n) = a (x, n)
1950 :			- and d x = e (x, 3)
1951 :			- and e (x, 0) = f x
1952 :			- | e (x, n) = e (x, n - 1)
1953 :			- and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1954 :			- | f n = n * g (n - 1)
1955 :			- and g n = a (n, 3)
1956 :			- in
1957 :			- f n
1958 :			- end
1959 :			- end;
1960 :			structure X : sig val main : int -> int end
1961 :			- X.main 3;
1962 :			*** BACK-TRACE ***
1963 :			GOTO stdIn:4.2-13.20: X.main[2].f
1964 :			GOTO-( stdIn:4.2-13.20: X.main[2].e
1965 :			GOTO stdIn:4.2-13.20: X.main[2].d
1966 :			/ stdIn:4.2-13.20: X.main[2].a
1967 :			| stdIn:4.2-13.20: X.main[2].b
1968 :			GOTO-\ stdIn:4.2-13.20: X.main[2].c
1969 :			CALL stdIn:4.2-13.20: X.main[2].g
1970 :			GOTO stdIn:4.2-13.20: X.main[2].f
1971 :			GOTO-( stdIn:4.2-13.20: X.main[2].e
1972 :			GOTO stdIn:4.2-13.20: X.main[2].d
1973 :			/ stdIn:4.2-13.20: X.main[2].a
1974 :			| stdIn:4.2-13.20: X.main[2].b
1975 :			GOTO-\ stdIn:4.2-13.20: X.main[2].c
1976 :			CALL stdIn:4.2-13.20: X.main[2].g
1977 :			GOTO stdIn:4.2-13.20: X.main[2].f
1978 :			GOTO-( stdIn:4.2-13.20: X.main[2].e
1979 :			GOTO stdIn:4.2-13.20: X.main[2].d
1980 :			/ stdIn:4.2-13.20: X.main[2].a
1981 :			| stdIn:4.2-13.20: X.main[2].b
1982 :			GOTO-\ stdIn:4.2-13.20: X.main[2].c
1983 :			CALL stdIn:4.2-13.20: X.main[2].g
1984 :			GOTO stdIn:4.2-13.20: X.main[2].f
1985 :			CALL stdIn:2.15-17.4: X.main[2]
1986 :			-
1987 :
1988 :			(Note that because of a FLINt bug the above code currently does not
1989 :			compile without BTrace turned on.)
1990 :
1991 :			Here is another example, using my modified Tiger compiler:
1992 :
1993 :			Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1994 :			- SMLofNJ.Internals.BTrace.mode (SOME true);
1995 :			[autoloading]
1996 :			[autoloading done]
1997 :			val it = false : bool
1998 :			- CM.make "sources.cm";
1999 :			[autoloading]
2000 :			...
2001 :			[autoloading done]
2002 :			[scanning sources.cm]
2003 :			[parsing (sources.cm):parse.sml]
2004 :			[creating directory CM/SKEL ...]
2005 :			[parsing (sources.cm):tiger.lex.sml]
2006 :			...
2007 :			[wrote CM/sparc-unix/semant.sml]
2008 :			[compiling (sources.cm):main.sml]
2009 :			[wrote CM/sparc-unix/main.sml]
2010 :			[New bindings added.]
2011 :			val it = true : bool
2012 :			- Main.compile ("../testcases/merge.tig", "foo.out");
2013 :			*** BACK-TRACE ***
2014 :			CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2015 :			CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2016 :			CALL lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2017 :			GOTO lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2018 :			CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2019 :			CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2020 :			CALL lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2021 :			/ lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2022 :			CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2023 :			CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2024 :			CALL lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2025 :			CALL main.sml:1.18-118.4: Main.compile[2]
2026 :			-
2027 :
2028 :			----------------------------------------------------------------------
2029 :	blume	673	Name: Matthias Blumen
2030 :			Date: 2000/06/21 18:00:00 JST
2031 :			Tag: blume-20000621-manual
2032 :			Description:
2033 :
2034 :			CM manual update: Path environments documented.
2035 :
2036 :			----------------------------------------------------------------------
2037 :	blume	666	Name: Matthias Blume
2038 :	blume	672	Date: 2000/06/19 13:40:00
2039 :			Tag: blume-20000619-manual
2040 :			Description:
2041 :
2042 :			CM manual and system/README update. This only covers the fact that
2043 :			there are no more implicit anchors. (Path environments and the "bind"
2044 :			option to "cm" have yet to be documented.)
2045 :
2046 :			----------------------------------------------------------------------
2047 :			Name: Matthias Blume
2048 :	blume	671	Date: 2000/06/19 11:05:00 JST
2049 :			Tag: blume-20000619-chdir-bugfix
2050 :			Description:
2051 :
2052 :			Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2053 :
2054 :			----------------------------------------------------------------------
2055 :			Name: Matthias Blume
2056 :	blume	670	Date: 2000/06/18 22:00:10 JST
2057 :			Tag: blume-20000618-implicit-anchors-really-gone
2058 :			Description:
2059 :
2060 :			I updates the previous HISTORY entry where I forgot to mention that
2061 :			implicit anchors are no longer with us.
2062 :
2063 :			The current update also gets rid of the (now useless) controller
2064 :			CM.Control.implicit_anchors.
2065 :
2066 :			----------------------------------------------------------------------
2067 :			Name: Matthias Blume
2068 :	blume	666	Date: 2000/06/16 17:30:00 JST
2069 :			Tag: blume-20000616-anchorenv
2070 :			Description:
2071 :
2072 :			This patch implements the long anticipated (just kidding :) "anchor
2073 :			environment" mechanism. In the course of doing this, I also
2074 :			re-implemented CM's internal "SrcPath" module from scratch. The new
2075 :			one should be more robust in certain boundary cases. In any case, it
2076 :			is a lot cleaner than its predecessor (IMHO).
2077 :
2078 :			This time, although there is yet another boot file format change, I
2079 :			kept the unpickler backward-compatible. As a result, no new bootfiles
2080 :			are necessary and bootstrapping is straightforward. (You cannot read
2081 :			new bootfiles into an old system, but the other way around is no
2082 :			problem.)
2083 :
2084 :			Visible changes:
2085 :
2086 :	blume	670	** 0. Implicit path anchors (without the leading $-symbol) are no
2087 :			longer recognized at all. This means that such path names are not
2088 :			illegal either. For example, the name basis.cm simply refers to a
2089 :			local file called "basis.cm" (i.e, the name is an ordinary path
2090 :			relative to .cm-files directory). Or, to put it differently, only
2091 :			names that start with $ are anchored paths.
2092 :	blume	666
2093 :	blume	670	** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
2094 :			vanished.
2095 :
2096 :			John (Reppy) had critizised this as soon as I originally proposed and
2097 :	blume	666	implemented it, but at that time I did not really deeply believe
2098 :			him. :) Now I came full-circle because I need the $<singlearc> syntax
2099 :			in another place where it cannot be seen as an abbreviation for
2100 :			$/<singlearc>. To avoid the confusion, $<singlearc> now means what it
2101 :			seems to mean (i.e., it "expands" into the corresponding anchor
2102 :			value).
2103 :
2104 :			However, when paths are used as members in CM description files, it
2105 :			continues to be true that there must be at least another arc after the
2106 :			anchor. This is now enforced separately during semantic analysis
2107 :	blume	670	(i.e., from a lexical/syntactical point of view, the notation is ok.)
2108 :	blume	666
2109 :			** 2. The "cm" class now accepts an option "bind". The option's value
2110 :			is a sub-option list of precisely two items -- one labeled "anchor"
2111 :			and the other one labeled "value". As you might expect, "anchor" is
2112 :			used to specify an anchor name to be bound, and "value" specifies what
2113 :			the anchor is being bound to.
2114 :
2115 :			The value must be a directory name and can be given in either standard
2116 :			syntax (including the possibility that it is itself an anchored path)
2117 :			or native syntax.
2118 :
2119 :			Examples:
2120 :
2121 :			foo.cm (bind:(anchor:bar value:$mystuff/bar))
2122 :			lib.cm (bind:(anchor:a value:"H:\\x\\y\\z")) (* only works under windows *)
2123 :
2124 :			and so on.
2125 :
2126 :			The meaning of this is that the .cm-file will be processed with an
2127 :			augmented anchor environment where the given anchor(s) is/are bound to
2128 :			the given values(s).
2129 :
2130 :			The rationale for having this feature is this: Suppose you are trying
2131 :			to use two different (already stable) libraries a.cm and b.cm (that
2132 :			you perhaps didn't write yourself). Further, suppose each of these
2133 :			two libraries internally uses its own auxiliary library $aux/lib.cm.
2134 :			Normally you would now have a problem because the anchor "lib" can not
2135 :			be bound to more than one value globally. Therefore, the project that
2136 :			uses both a.cm and b.cm must locally redirect the anchor to some other
2137 :			place:
2138 :
2139 :			a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
2140 :			b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
2141 :
2142 :			This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
2143 :			/usr/lib/smlnj/b-stuff/aux.cm, respectively.
2144 :
2145 :			Hard-wiring path names is a bit inflexible (and CM will verbosely warn
2146 :			you when you do so at the time of CM.stabilize). Therefore, you can
2147 :			also use an anchored path as the value:
2148 :
2149 :			a.cm (bind:(anchor:lib value:$a-lib))
2150 :			b.cm (bind:(anchor:lib value:$b-lib))
2151 :
2152 :			Now you can globally configure (using the usual CM.Anchor.anchor or
2153 :			pathconfig machinery) bindings for "a-lib" and "b-lib". Since "lib"
2154 :			itself is always locally bound, setting it globally is no longer
2155 :			meaningful or necessary (but it does not hurt either). In fact, "lib"
2156 :			can still be used as a global anchor for separate purposes. As a
2157 :			matter of fact, one can locally define "lib" in terms of a global
2158 :			"lib":
2159 :
2160 :			a.cm (bind:(anchor:lib value:$lib/a))
2161 :			b.cm (bind:(anchor:lib value:$lib/b))
2162 :
2163 :			** 3: The encoding of path names has changed. This affects the way
2164 :			path names are shown in CM's progress report and also the internal
2165 :			protocol encoding used for parallel make.
2166 :
2167 :			The encoding now uses one or more ':'-separated segments. Each
2168 :			segments corresponds to a file that has been specified relative to the
2169 :			file given by its preceding segment. The first segment is either
2170 :			relative to the CWD, absolute, or anchored. Each segment itself is
2171 :			basically a Unix pathname; all segments but the first are relative.
2172 :
2173 :			Example:
2174 :
2175 :			$foo/bar/baz.cm:a/b/c.sml
2176 :
2177 :			This path denotes the file bar/a/b/c.sml relative to the directory
2178 :			denoted by anchor "foo". Notice that the encoding also includes
2179 :			baz.cm which is the .cm-file that listed a/b/c.sml. As usual, such
2180 :			paths are resolved relative to the .cm-files directory, so baz.cm must
2181 :			be ignored to get the "real" pathname.
2182 :
2183 :			To make this fact more obvious, CM puts the names of such "virtual
2184 :			arcs" into parentheses when they appear in progress reports. (No
2185 :			parentheses will appear in the internal protocol encoding.) Thus,
2186 :			what you really see is:
2187 :
2188 :			$foo/bar/(baz.cm):a/b/c.sml
2189 :
2190 :			I find this notation to be much more informative than before.
2191 :
2192 :			Another new feature of the encoding is that special characters
2193 :			including parentheses, colons, (back)slashes, and white space are
2194 :			written as \ddd (where ddd is the decimal encoding of the character).
2195 :
2196 :	blume	670	*** The CM manual still needs to be updated.
2197 :
2198 :	blume	666	----------------------------------------------------------------------
2199 :	leunga	660	Name: Allen Leung
2200 :			Date: 2000/06/15 00:38:00
2201 :			Tag: leunga-20000615-x86-peephole
2202 :
2203 :			x86 Peephole fix by Fermin. Affects c-- and moby only.
2204 :
2205 :			----------------------------------------------------------------------
2206 :	blume	658	Name: Matthias Blume
2207 :			Date: 2000/06/12 11:40:00
2208 :			Tag: blume-20000612-parmakefix
2209 :			Description:
2210 :
2211 :			More cleanup after changing the file naming scheme: This time I
2212 :			repaired the parallel make mechanism for CMB.make which I broke earlier.
2213 :
2214 :			----------------------------------------------------------------------
2215 :	leunga	657	Name: Allen Leung
2216 :			Date: 2000/06/09 01:25:00
2217 :			Tag: leunga-20000609-various
2218 :
2219 :			None of these things should affect normal SML/NJ operations
2220 :
2221 :			1. Peephole improvements provided by Fermin (c--)
2222 :			2. New annotation DEFUSE for adding extra dependence (moby)
2223 :			3. New X86 LOCK instructions (moby)
2224 :			4. New machine description language for reservation tables (scheduling)
2225 :			5. Fixes to various optimization/analysis modules (branch chaining, dominator
2226 :			trees etc.)
2227 :			6. I've changed the CM files so that they can work with versions
2228 :			110.0.6, 110.25 and 110.28
2229 :
2230 :			----------------------------------------------------------------------
2231 :	blume	652	Name: Matthias Blume
2232 :	blume	656	Date: 2000/06/09 12:40:00
2233 :			Tag: blume-20000609-log
2234 :			Description:
2235 :
2236 :			- Removed all(?) remaining RCS Log entries from sources.
2237 :
2238 :			- Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
2239 :			anchored paths).
2240 :
2241 :			----------------------------------------------------------------------
2242 :			Name: Matthias Blume
2243 :	blume	653	Date: 2000/06/07 17:00:00 JST
2244 :			Tag: blume-20000607-no-implicit-anchors
2245 :			Description:
2246 :
2247 :			1. This update changes the default setting for
2248 :			CM.Control.implicit_anchors from true to false. This means that
2249 :			implicit anchors are no longer permitted by default. I also tried to
2250 :			make sure that nothing else still relies on implicit anchors.
2251 :			(This is the next step on the schedule towards a CM that does not even
2252 :			have the notion of implicit anchors anymore.)
2253 :
2254 :			2. More CM manual updates.
2255 :
2256 :			3. I managed to track down and fix the pickling bug I mentioned last
2257 :			time. Because of the previously existing workaround, this entails no
2258 :			immediate practical changes.
2259 :
2260 :			----------------------------------------------------------------------
2261 :			Name: Matthias Blume
2262 :	blume	652	Date: 2000/06/06 11:15:00 JST
2263 :			Tag: blume-20000606-lazierpickle
2264 :			Description:
2265 :
2266 :			!!!! NEW BOOT FILES !!!!
2267 :
2268 :			* The main purpose of this update is to make library pickles lazier in
2269 :			order to reduce the initial space penalty for autoloading a library.
2270 :			As a result, it is now possible to have $smlnj/compiler.cm
2271 :			pre-registered. This should take care of the many complaints or
2272 :			inquiries about missing structure Compiler. This required changes to
2273 :			CM's internal data structures and small tweaks to some algorithms.
2274 :
2275 :			As a neat additional effect, it is no longer necessary (for the sake
2276 :			of lean heap image files) to distinguish between a "minimal" CM and a
2277 :			"full" CM. Now, there is only one CM (i.e., the "full" version:
2278 :			$smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
2279 :			interactive top level. ($smlnj/cm/minimal.cm is gone.)
2280 :
2281 :			To make the life of compiler-hackers easier, "makeml" now also
2282 :			pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm). In other
2283 :			words, after you bootstrap a new sml for the first time, you will not
2284 :			have to autoload $smlnj/cmb.cm again afterwards. (The first time
2285 :			around you will still have to do it, though.)
2286 :
2287 :			* A second change consists of major updates to the CM manual. There
2288 :			are now several appendices with summary information and also a full
2289 :			specification of the CM description file syntax.
2290 :
2291 :			* In directory src/system I added the script "allcross". This script
2292 :			invokes sml and cross-compiles the compiler for all supported
2293 :			architectures. (Useful when providing a new set of boot files.)
2294 :
2295 :			* There seems to be a latent bug in my "lazy pickles" mechanism. I
2296 :			added a small tweak to pickle-util.sml to work around this problem,
2297 :			but it is not a proper fix yet. I will investigate further. (The
2298 :			effect of the bug was an inflation of library pickle size.)
2299 :
2300 :			* Version number increased to 110.28.1 (to avoid compatibility problems).
2301 :
2302 :			----------------------------------------------------------------------
2303 :	leunga	646	Name: Allen Leung
2304 :	leunga	648	Date: 2000/05/25 17:28 EDT
2305 :			Tag: leunga-20000525-ra
2306 :			Description:
2307 :
2308 :			Fixed a bug in freezing phase of the register allocator.
2309 :
2310 :			----------------------------------------------------------------------
2311 :			Name: Allen Leung
2312 :	leunga	647	Date: 2000/05/15 22:53 EDT
2313 :	leunga	646	Tag: leunga-20000515-alpha-x86-ra
2314 :			Description:
2315 :
2316 :			1. Alpha
2317 :
2318 :			Slight cleanup. Removed the instruction SGNXL
2319 :
2320 :			2. X86
2321 :
2322 :			Added the following instructions to the instruction set:
2323 :
2324 :			ROLx, RORx,
2325 :			BTx, BTSx, BTLx, BTRx,
2326 :			XCHGx, and variants with the LOCK prefix
2327 :
2328 :			3. Register Allocation
2329 :
2330 :			The module ra-rewrite-with-renaming has been improved.
2331 :
2332 :	leunga	647	These have no effect on SML/NJ.
2333 :
2334 :	leunga	646	----------------------------------------------------------------------
2335 :	blume	642	Name: Matthias Blume
2336 :	blume	645	Date: 2000/05/15 16:20:00 JST
2337 :			Tag: blume-20000515-lightrebuild
2338 :			Description:
2339 :
2340 :			1. I added an alternative to "-rebuild" to "makeml". The difference is
2341 :			that prior to calling CMB.make' the CM-variable "LIGHT" will be
2342 :			defined. In effect, the command will not build any cross-compiler
2343 :			backends and therefore finish more quickly.
2344 :
2345 :			The "fixpt" script also takes a "-light" switch to be able to use
2346 :			this new facility while compiling for a fixpoint.
2347 :
2348 :			2. I replaced all mentions of anchored paths in group owner specifications
2349 :			with simple relative paths (usually starting with "..").
2350 :			The rationale is that a library's internal workings should not be
2351 :			compromised by the lack of some anchor. (An anchor is necessary
2352 :			for someone who wants to refer to the library by an anchored path,
2353 :			but it should not be necessary to build the same library in the first
2354 :			place.)
2355 :
2356 :			3. I changed the way CM's tool mechanism determines the shell command
2357 :			string used for things like ml-yacc etc. so that it does not break
2358 :			when CM.Control.implicit_anchors is turned off.
2359 :
2360 :			----------------------------------------------------------------------
2361 :			Name: Matthias Blume
2362 :	blume	644	Date: 2000/05/12 18:20:00 JST
2363 :			Tag: blume-20000512-ml-build
2364 :			Description:
2365 :
2366 :			Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
2367 :			getting installed properly (by config/install.sh).
2368 :
2369 :			----------------------------------------------------------------------
2370 :			Name: Matthias Blume
2371 :	blume	643	Date: 2000/05/12 17:30:00 JST
2372 :			Tag: blume-20000512-anchors
2373 :			Description:
2374 :
2375 :			!!! NEW BOOT FILES !!!
2376 :
2377 :			This change is in preparation of fading out support for "implicitly
2378 :			anchored path names". I went through all sources and used the
2379 :			explicit (and relatively new) $-notation. See system/README and the
2380 :			CM manual for more info on this.
2381 :
2382 :			I also modified the anchoring scheme for some things such as "smlnj",
2383 :			"MLRISC", "cm", etc. to take advantage of the fact that explicit
2384 :			anchors are more expressive: anchor name and first arc do not have to
2385 :			coincide. This entails the following user-visible change:
2386 :
2387 :			You have to write $smlnj/foo/bar instead of smlnj/foo/bar. In
2388 :			particular, when you fire up sml with a command-line argument, say,
2389 :			e.g.:
2390 :
2391 :			sml '$smlnj/cmb.cm'
2392 :
2393 :			At the ML toplevel prompt:
2394 :
2395 :			CM.autoload "$smlnj/cmb.cm";
2396 :
2397 :			There is also a new controller in CM.Control that can be used to turn
2398 :			off all remaining support for implicit anchors by saying:
2399 :
2400 :			CM.autoload "$smlnj/
2401 :			#set CM.Control.implicit_anchors false;
2402 :
2403 :			This causes CM to reject implicitly anchored paths. This is (for the
2404 :			time being) less permissive than the "final" version where there will
2405 :			be no more such implicit anchors and relative paths will be just that:
2406 :			relative.
2407 :
2408 :			The next step (version after next version?) will be to make the
2409 :			default for CM.Control.implicit_anchors false. After the dust has
2410 :			settled, I can then produce the "final" version of this...
2411 :
2412 :			Note: Since bootstrapping is a bit tricky, I provided new boot files.
2413 :
2414 :			----------------------------------------------------------------------
2415 :			Name: Matthias Blume
2416 :	blume	642	Date: 2000/05/11 16:30:00 JST
2417 :			Tag: blume-20000511-sources
2418 :			Description:
2419 :
2420 :			The main change is that I added function CM.sources as a generalized
2421 :			version of the earlier CM.makedepend. This entails the following
2422 :			additional changes:
2423 :
2424 :			- CM.makedepend has been dropped.
2425 :
2426 :			- CM manual has been updated.
2427 :
2428 :			- TOOLS signature and API have been changed.
2429 :
2430 :			----------------------------------------------------------------------
2431 :	leunga	640	Name: Allen Leung
2432 :	leunga	641	Date: 2000/05/10 21:17 EDT
2433 :			Tag: leunga-20000510-moby-c--ssa
2434 :			Description:
2435 :
2436 :			Various bug fixes and new features for C--, Moby and MLRISC optimizations.
2437 :			None of these affect SML/NJ.
2438 :
2439 :			1. Register Allocation
2440 :
2441 :			a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
2442 :			This module tries to remove local (i.e. basic block level) redundancies
2443 :			during spilling.
2444 :
2445 :			b. A new framework for performing region based register allocation.
2446 :			Not yet entirely functional.
2447 :
2448 :			2. X86
2449 :
2450 :			a. DefUse for POP was missing the stack pointer [found by Lal]
2451 :			b. Reload for CALL was incorrect in X86Spill [found by John]
2452 :			c. Various fixes in X86Spill so that it can be used correctly for
2453 :			the new spilling module.
2454 :
2455 :			3. SSA/IR
2456 :
2457 :			a. New module ir/dj-dataflow.sml implements elimination based
2458 :			data flow analysis.
2459 :
2460 :			4. MLRiscGen
2461 :
2462 :			a. Fix for gc type annotation
2463 :
2464 :			5. MDGen
2465 :
2466 :			Various fixes for machine description -> ml code translation. For ssa
2467 :			only.
2468 :
2469 :			----------------------------------------------------------------------
2470 :			Name: Allen Leung
2471 :	leunga	640	Date: 2000/05/08 22:17 EDT
2472 :			Tag: leunga-20000508-labexp
2473 :			Description:
2474 :
2475 :			Fermin has found a few assembly problems with constant expressions
2476 :			generated in LabelExp. Mostly, the problems involve extra parentheses,
2477 :			which choke on dumb assemblers. This is his fix.
2478 :
2479 :			----------------------------------------------------------------------
2480 :	dbm	635	Name: Dave MacQueen
2481 :			Date: 2000/04/09 14:00 EDT
2482 :			Tag: dbm-20000502-Version_110_28
2483 :			Description:
2484 :
2485 :			1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
2486 :
2487 :			2. Updated config/version to 110.28
2488 :
2489 :			3. Updated config/srcarchiveurl
2490 :
2491 :			3. New boot files!
2492 :			ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
2493 :
2494 :			----------------------------------------------------------------------
2495 :	blume	634	Name: Matthias Blume
2496 :			Date: 2000/05/01 19:05:00 JST
2497 :			Tag: blume-20000501-noweb
2498 :			Description:
2499 :
2500 :			A new noweb tool has been added. The existing system is entirely
2501 :			unaffected by this, but some CM users have asked for renewed noweb
2502 :			support. Everything is documented in the CM manual.
2503 :
2504 :			New (plugin) libraries:
2505 :
2506 :			noweb-tool.cm
2507 :			nw-ext.cm
2508 :
2509 :			----------------------------------------------------------------------
2510 :	dbm	633	Name: Dave MacQueen
2511 :			Date: 2000/04/30 12:40PM EDT
2512 :			Tag: dbm-20000430-bug_fixes
2513 :			Description:
2514 :
2515 :			1. Fix for bug 1498
2516 :			smlnj/src/system/Basis/Implementation/Unsafe/object.sig
2517 :			smlnj/src/system/Basis/Implementation/Unsafe/object.sml
2518 :			added toRealArray function
2519 :			smlnj/src/compiler/MiscUtil/print/ppobj.sml
2520 :			added check for tag Obj.RealArray to array printing case in ppObj
2521 :
2522 :			2. Fix for bug 1510
2523 :			smlnj/src/compiler/Semant/types/typesutil.sml
2524 :			fixed definition of dummyargs (used by equalTycon) so that
2525 :			dummy args are distinct types
2526 :
2527 :			----------------------------------------------------------------------
2528 :	blume	629	Name: Matthias Blume
2529 :	blume	632	Date: 2000/04/30 01:00:00 JST
2530 :			Tag: blume-20000430-versions
2531 :			Description:
2532 :
2533 :			1. CM version numbering added. This is an implementation of Lal's
2534 :			proposal for adding version numbers and version checking to .cm
2535 :			files. Lal said that his proposal was just that -- a proposal.
2536 :			For the time being I went ahead and implemented it so that people
2537 :			can comment on it. Everything is completely backward-compatible
2538 :			(except for the stable library format, i.e., new bootfiles!).
2539 :
2540 :			As usual, see the CM manual for details.
2541 :
2542 :			2. An alternative syntax for anchored paths has been implemented.
2543 :			Dave has recently voiced the same concerns that I had when I did
2544 :			this, so there should be some support. My take is that eventually
2545 :			I will let support for the current syntax (where anchors are
2546 :			"implicit") fade out in favor of the new, explicit syntax.
2547 :			In order to be backward-compatible, both old and new syntax are
2548 :			currently supported.
2549 :
2550 :			Again, see the CM manual for details.
2551 :
2552 :			3. Parallel make is trying to be slightly smarter: When the master
2553 :			process finds a "bottleneck", i.e., when there is only one
2554 :			compilation unit that can be compiled and everybody else is
2555 :			waiting on it, then it will simply compile it directly instead
2556 :			of clumsily telling one of the slaves to do it.
2557 :
2558 :			4. Support for "unsharing" added. This is necessary in order to be
2559 :			able to have two different versions of the same library running
2560 :			at the same time (e.g., for trying out a new MLRISC while still
2561 :			having the old MLRISC linked into the current compiler, etc.)
2562 :			See the CM manual.
2563 :
2564 :			5. Simple "makedepend" functionality added for generating Makefile
2565 :			dependency information. (This is rather crude at the moment.
2566 :			Expect some changes here in the future.)
2567 :
2568 :			6. ".fun" added as a recognized suffix for ML files. Also documented
2569 :			explicitly in the manual that the fallback behavior (unknown suffix
2570 :			-> ML file) is not an official feature!
2571 :
2572 :			7. Small changes to the pickler for stable libraries.
2573 :
2574 :			8. Several internal changes to CM (for cleanup/improvement).
2575 :
2576 :
2577 :			!!!! NEW BINFILES !!!!
2578 :
2579 :			----------------------------------------------------------------------
2580 :			Name: Matthias Blume
2581 :	blume	631	Date: 2000/04/28 17:30:00 JST
2582 :			Tag: blume-20000428-pathconfig
2583 :			Description:
2584 :
2585 :			1. I changed config/install.sh to remove duplicate entries from the
2586 :			lib/pathconfig file at the end. Moreover, the final version of
2587 :			lib/pathconfig is sorted alphabetically. The same (sorting) is done
2588 :			in src/system/installml.
2589 :
2590 :			2. The config/install.sh script now consistently uses relative
2591 :			pathnames in lib/pathconfig whenever the anchor is in the lib
2592 :			directory. (So far this was true for the libraries that come
2593 :			pre-compiled and bundled as part of the bootfiles but not for
2594 :			libraries that are compiled by the script itself.)
2595 :
2596 :			----------------------------------------------------------------------
2597 :			Name: Matthias Blume
2598 :	blume	629	Date: 2000/04/26 13:10:00 JST
2599 :			Tag: blume-20000426-fun_suffix
2600 :			Description:
2601 :
2602 :			Added ".fun" as a recognized file name suffix (for ML code).
2603 :
2604 :			----------------------------------------------------------------------
2605 :	leunga	628	Name: Allen Leung
2606 :			Date: 2000/04/25 17:00:00 EST
2607 :			Tag: leunga-20000425-alpha-ra
2608 :			Description:
2609 :
2610 :			1. Alpha
2611 :
2612 :			PSEUDOARITH was missing in AlphaRewrite. This causes an endless loop
2613 :			in C--.
2614 :
2615 :			2. RA
2616 :
2617 :			Added a flag "ra-dump-size" to print out the size of the flowgraph
2618 :			and the interference graph.
2619 :
2620 :			----------------------------------------------------------------------
2621 :	dbm	627	Name: Dave MacQueen
2622 :			Date: 2000/04/25/
2623 :			Tag: dbm-20000425-mlyacc_doc_examples
2624 :			Description:
2625 :			Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
2626 :			Updated all three examples in src/ml-yacc/examples to run
2627 :			under 110.* using CM.make.
2628 :
2629 :			----------------------------------------------------------------------
2630 :	leunga	624	Name: Allen Leung
2631 :			Date: 2000/04/20 23:04:00 EST
2632 :			Tag: leunga-20000420-ssa-c---stuff
2633 :			Description:
2634 :
2635 :			This update synchronizes my repository with Yale's. Most of these
2636 :			changes, however, do not affect SML/NJ at all (the RA is an exception).
2637 :
2638 :			1. Register Allocator
2639 :
2640 :			a. An improvement in the interference graph construction:
2641 :			Given a copy
2642 :
2643 :			s <- t
2644 :
2645 :	leunga	628	no interference edge between s and t is added for this definition of s.
2646 :	leunga	624
2647 :			b. I've added two new spill heuristic modules that Fermin and I developed
2648 :			(in the new library RA.cm). These are unused in SML/NJ but maybe
2649 :			useful for others (Moby?)
2650 :
2651 :			2. X86
2652 :
2653 :			a. Various fixes in the backend provided by Fermin [C--] and Lal.
2654 :
2655 :			3. Alpha
2656 :
2657 :			a. Added the BSR instruction and code generation that goes with it [C--]
2658 :			b. Other fixes too numerous to recount provided by Fermin [C--]
2659 :
2660 :			4. Regmaps
2661 :
2662 :			a. The regmaps are not initialized with the identity physical bindings
2663 :			at creation time. This is unneeded.
2664 :
2665 :			5. MLRISC Optimizations
2666 :
2667 :			a. The DJ-Graph module can now compute the iterated dominance frontiers
2668 :			intersects with liveness incrementally in linear time! Woohoo!
2669 :			This is now used in my new SSA construction algorithm.
2670 :
2671 :			b. THe branch reorganization module is now smarter about linear chains of
2672 :			basic blocks.
2673 :
2674 :
2675 :			----------------------------------------------------------------------
2676 :	blume	615	Name: Matthias Blume
2677 :			Date: 2000/04/12 13:52:00 JST
2678 :			Tag: blume_main_v110p27_1
2679 :			Description:
2680 :
2681 :			Changed install.sh script to handle archive files without version number
2682 :			and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
2683 :			name of the boot file archive.
2684 :
2685 :			----------------------------------------------------------------------
2686 :	dbm	607	Name: Dave MacQueen
2687 :			Date: 2000/04/09 14:00 EDT
2688 :			Tag: dbm-20000410-Version_110_27
2689 :			Description:
2690 :
2691 :			1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
2692 :
2693 :			2. Updated src/config/version to 110.27
2694 :
2695 :			3. New boot files!
2696 :
2697 :			----------------------------------------------------------------------
2698 :	leunga	606	Name: Allen Leung
2699 :			Date: 2000/04/09 19:09:00 EST
2700 :			Tag: leunga-20000409-misc
2701 :			Description:
2702 :
2703 :			1. Yet another fix for x86 assembly for idivl, imull, mull and friends.
2704 :
2705 :			2. Miscellaneous improvements to MLRISC (unused in sml/nj)
2706 :
2707 :			----------------------------------------------------------------------
2708 :	monnier	605	Name: Stefan
2709 :			Date: 2000/04/07 10:00:00 EDT
2710 :			Tag: monnier-20000406-branch-handling
2711 :			Description:
2712 :
2713 :			Improved handling of branches (mostly those generated from
2714 :			polymorphic equality), removed switchoff and changed the
2715 :			default optimization settings (more cpsopt and less flintopt).
2716 :
2717 :			----------------------------------------------------------------------
2718 :	leunga	601	Name: Allen Leung
2719 :	leunga	602	Date: 2000/04/06 01:30:00 EST
2720 :			Tag: leunga-20000406-peephole-x86-SSA-2
2721 :			Description:
2722 :
2723 :			Forgot a few files.
2724 :
2725 :			----------------------------------------------------------------------
2726 :			Name: Allen Leung
2727 :	leunga	601	Date: 2000/04/06 00:36:00 EST
2728 :			Tag: leunga-20000406-peephole-x86-SSA
2729 :			Description:
2730 :
2731 :			1. New Peephole code
2732 :
2733 :			2. Minor improvement to X86 instruction selection
2734 :
2735 :			3. Various fixes to SSA and machine description -> code translator
2736 :
2737 :			----------------------------------------------------------------------
2738 :	blume	595	Name: Matthias Blume
2739 :			Date: 2000/04/05 12:30:00 JST
2740 :			Tag: blume_main_v110p26p2_3
2741 :			Description:
2742 :
2743 :			This update just merges three minor cosmetic updates to CM's sources
2744 :			to get ready for the 110.27 code freeze on Friday. No functionality
2745 :			has changed.
2746 :
2747 :			----------------------------------------------------------------------
2748 :	leunga	593	Name: Allen Leung
2749 :	leunga	594	Date: 2000/04/04 19:39:00 EST
2750 :			Tag: leunga-20000404-x86-asm
2751 :			Description:
2752 :
2753 :			1. Fixed a problem in X86 assembly.
2754 :
2755 :			Things like
2756 :
2757 :			jmp %eax
2758 :			jmp (%eax)
2759 :
2760 :			should be output as
2761 :
2762 :			jmp *%eax
2763 :			jmp *(%eax)
2764 :
2765 :			2. Assembly output
2766 :
2767 :			Added a new flag
2768 :
2769 :			"asm-indent-copies" (default to false)
2770 :
2771 :			When this flag is on, parallel copies will be indented an extra level.
2772 :
2773 :			----------------------------------------------------------------------
2774 :			Name: Allen Leung
2775 :	leunga	593	Date: 2000/04/04 03:18:00 EST
2776 :			Tag: leunga-20000404-C--Moby
2777 :			Description:
2778 :
2779 :			All of these fixes are related to C--, Moby, and my own optimization
2780 :			stuff; so they shouldn't affect SML/NJ.
2781 :
2782 :			1. X86
2783 :
2784 :			Various fixes related floating point, and extensions.
2785 :
2786 :			2. Alpha
2787 :
2788 :			Some extra patterns related to loads with signed/zero extension
2789 :			provided by Fermin.
2790 :
2791 :			3. Assembly
2792 :
2793 :	monnier	659	When generating assembly, resolve the value of client defined constants,
2794 :	leunga	593	instead of generating symbolic values. This is controlled by the
2795 :			new flag "asm-resolve-constants", which is default to true.
2796 :
2797 :			4. Machine Descriptions
2798 :
2799 :			a. The precedence parser was slightly broken when parsing infixr symbols.
2800 :			b. The type generalizing code had the bound variables reversed, resulting
2801 :			in a problem during arity raising.
2802 :			c. Various fixes in machine descriptions.
2803 :
2804 :			----------------------------------------------------------------------
2805 :	blume	592	Name: Matthias Blume
2806 :			Date: 2000/04/03 16:05:00 JST
2807 :			Tag: blume_main_v110p26p2_2
2808 :			Description:
2809 :
2810 :			I eliminated coreEnv from compInfo. Access to the "Core" structure is
2811 :			now done via the ordinary static environment that is context to each
2812 :			compilation unit.
2813 :
2814 :			To this end, I arranged that instead of "structure Core" as "structure
2815 :			_Core" is bound in the pervasive environment. Core access is done via
2816 :	monnier	659	_Core (which can never be accidentally rebound because _Core is not a
2817 :	blume	592	legal surface-syntax symbol).
2818 :
2819 :			The current solution is much cleaner because the core environment is
2820 :			now simply part of the pervasive environment which is part of every
2821 :			compilation unit's context anyway. In particular, this eliminates all
2822 :			special-case handling that was necessary until now in order to deal
2823 :			with dynamic and symbolic parts of the core environment.
2824 :
2825 :			Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2826 :	monnier	659	compilation manager's bootstrap compiler (actually: in the "init group"
2827 :	blume	592	handling). See the comments in src/system/smlnj/init/init.cmi for
2828 :			more details.
2829 :
2830 :			I also tried to track down all mentions of "Core" (as string argument
2831 :			to Symbol.strSymbol) in the compiler and replaced them with a
2832 :			reference to the new CoreSym.coreSym. Seems cleaner since the actual
2833 :			name appears in one place only.
2834 :
2835 :			Binfile and bootfile format have not changed, but the switchover from
2836 :			the old "init.cmi" to the new one is a bit tricky, so I supplied new
2837 :			bootfiles anyway.
2838 :
2839 :			----------------------------------------------------------------------
2840 :	leunga	591	Name: Allen Leung
2841 :			Date: 2000/04/02 21:17:00 EST
2842 :			Tag: leunga-20000402-mltree
2843 :			Description:
2844 :	leunga	585
2845 :	leunga	591	1. Renamed the constructor CALL in MLTREE by popular demand.
2846 :			2. Added a bunch of files from my repository. These are currently
2847 :			used by other non-SMLNJ backends.
2848 :
2849 :	leunga	576	----------------------------------------------------------------------
2850 :	leunga	590	Name: Allen Leung
2851 :			Date: 2000/03/31 21:15:00 EST
2852 :			Tag: leunga-20000331-aliasing
2853 :			Description:
2854 :
2855 :			This update contains a rewritten (and hopefully more correct) module
2856 :			for extracting aliasing information from CPS.
2857 :
2858 :			To turn on this feature:
2859 :
2860 :			Compiler.Control.CG.memDisambiguate := true
2861 :
2862 :			To pretty print the region information with assembly
2863 :
2864 :			Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
2865 :
2866 :			To control how many levels of aliasing information are printed, use:
2867 :
2868 :			Compiler.Control.MLRISC.getInt "points-to-show-level" := n
2869 :
2870 :			The default of n is 3.
2871 :
2872 :			----------------------------------------------------------------------
2873 :	dbm	589	Name: David MacQueen
2874 :			Date: 2000/03/31 11:15:00 EST
2875 :			Tag: dbm-20000331-runtime_fix
2876 :			Description:
2877 :
2878 :			This update contains:
2879 :
2880 :			1. runtime/c-lib/c-libraries.c
2881 :			includes added in revision 1.2 caused compilation errors on hppa-hpux
2882 :
2883 :			2. fix for bug 1556
2884 :			system/Basis/Implementation/NJ/internal-signals.sml
2885 :
2886 :			----------------------------------------------------------------------
2887 :	blume	587	Name: Matthias Blume
2888 :	blume	588	Date: 2000/03/31 18:00:00 JST
2889 :			Tag: blume_main_v110p26p2_1
2890 :			Description:
2891 :
2892 :			This update contains:
2893 :
2894 :			1. A small change to CM's handling of stable libraries:
2895 :			CM now maintains one "global" modmap that is used for all stable
2896 :			libraries. The use of such a global modmap maximizes sharing and
2897 :			minimizes the need for re-traversing parts of environments during
2898 :			modmap construction. (However, this has minor impact since modmap
2899 :			construction seems to account for just one percent or less of total
2900 :			compile time.)
2901 :
2902 :			2. I added a "genmap" phase to the statistics. This is where I got the
2903 :			"one percent" number (see above).
2904 :
2905 :			3. CM's new tool parameter mechanism just became _even_ better. :)
2906 :			- The parser understands named parameters and recursive options.
2907 :			- The "make" and "shell" tools use these new features.
2908 :			(This makes it a lot easier to cascade these tools.)
2909 :			- There is a small syntax change: named parameters use a
2910 :
2911 :			<name> : ( <option> ... ) or
2912 :			<name> : <string>
2913 :
2914 :			syntax. Previously, named parameters were implemented in an
2915 :			ad-hoc fashion by each tool individually (by parsing strings)
2916 :			and had the form
2917 :
2918 :			<name>=<string>
2919 :
2920 :			See the CM manual for a full description of these issues.
2921 :
2922 :			----------------------------------------------------------------------
2923 :			Name: Matthias Blume
2924 :	blume	587	Date: 2000/03/30 18:00:00 JST
2925 :			Tag: blume_main_v110p26p2_0
2926 :			Description:
2927 :
2928 :			!!!!! WARNING !!!!!!
2929 :			!! New binfiles !!
2930 :			!!!!!!!!!!!!!!!!!!!!
2931 :
2932 :			This update contains:
2933 :
2934 :			1. Moderate changes to CM:
2935 :
2936 :			- Changes to CM's tools mechanism. In particular, it is now possible
2937 :			to have tools that accept additional "command line" parameters
2938 :			(specified in the .cm file at each instance where the tool's class is
2939 :			used).
2940 :
2941 :	monnier	659	This was done to accommodate the new "make" and "shell" tools which
2942 :			facilitate fairly seamless hookup to portions of code managed using
2943 :	blume	587	Makefiles or Shell scripts.
2944 :
2945 :			There are no classes "shared" or "private" anymore. Instead, the
2946 :			sharing annotation is now a parameter to the "sml" class.
2947 :
2948 :			There is a bit of generic machinery for implementing one's own
2949 :			tools that accept command-line parameters. However, I am not yet fully
2950 :			satisfied with that part, so expect changes here in the future.
2951 :
2952 :			All existing tools are described in the CM manual.
2953 :
2954 :	monnier	659	- Slightly better error handling. (CM now suppresses many followup
2955 :	blume	587	error messages that tended to be more annoying than helpful.)
2956 :
2957 :			2. Major changes to the compiler's static environment data structures.
2958 :
2959 :			- no CMStaticEnv anymore.
2960 :			- no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
2961 :			but it is called Environment), no conversions between different
2962 :			kinds of static environments
2963 :
2964 :			- There is still a notion of a "modmap", but such modmaps are generated
2965 :			on demand at the time when they are needed. This sounds slow, but I
2966 :			sped up the code that generates modmaps enough for this not to lead to
2967 :			a slowdown of the compiler (at least I didn't detect any).
2968 :
2969 :			- To facilitate rapid modmap generation, static environments now
2970 :			contain an (optional) "modtree" structure. Modtree annotations are
2971 :			constructed by the unpickler during unpickling. (This means that
2972 :			the elaborator does not have to worry about modtrees at all.)
2973 :			Modtrees have the advantage that they are compositional in the same
2974 :			way as the environment data structure itself is compositional.
2975 :			As a result, modtrees never hang on to parts of an environment that
2976 :			has already been rendered "stale" by filtering or rebinding.
2977 :
2978 :			- I went through many, many trials and errors before arriving at the
2979 :			current solution. (The initial idea of "linkpaths" did not work.)
2980 :			But the result of all this is that I have touched a lot of files that
2981 :			depend on the "modules" and "types" data structures (most of the
2982 :			elaborator). There were a lot of changes during my "linkpath" trials
2983 :			that could have been reverted to their original state but weren't.
2984 :			Please, don't be too harsh on me for messing with this code a bit more
2985 :	monnier	921	than what was strictly necessary... (I _did_ resist the temptation
2986 :	blume	587	of doing any "global reformatting" to avoid an untimely death at
2987 :			Dave's hands. :)
2988 :
2989 :			- One positive aspect of the previous point: At least I made sure that
2990 :			all files that I touched now compile without warnings (other than
2991 :			"polyEqual").
2992 :
2993 :			- compiler now tends to run "leaner" (i.e., ties up less memory in
2994 :			redundant modmaps)
2995 :
2996 :			----------------------------------------------------------------------
2997 :	leunga	580	Name: Allen Leung
2998 :	leunga	585	Date: 2000/03/29 18:00:00
2999 :			Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
3000 :	leunga	586	Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
3001 :	leunga	585	Description:
3002 :
3003 :			This update contains *MAJOR* changes to the way code is generated from CPS
3004 :			in the module mlriscGen, and in various backend modules.
3005 :
3006 :			CHANGES
3007 :			=======
3008 :
3009 :			1. MLRiscGen: forward propagation fix.
3010 :
3011 :			There was a bug in forward propagation introduced at about the same time
3012 :			as the MLRISC x86 backend, which prohibits coalescing to be
3013 :			performed effectively in loops.
3014 :
3015 :			Effect: speed up of loops in RISC architectures.
3016 :			By itself, this actually slowed down certain benchmarks on the x86.
3017 :
3018 :			2. MLRiscGen: forward propagating addresses from consing.
3019 :
3020 :			I've changed the way consing code is generated. Basically I separated
3021 :			out the initialization part:
3022 :
3023 :			store tag, offset(allocptr)
3024 :			store elem1, offset+4(allocptr)
3025 :			store elem2, offset+8(allocptr)
3026 :			...
3027 :			store elemn, offset+4n(allocptr)
3028 :
3029 :			and the address computation part:
3030 :
3031 :			celladdr <- offset+4+alloctpr
3032 :
3033 :			and move the address computation part
3034 :
3035 :			Effect: register pressure is generally lower as a result. This
3036 :			makes compilation of certain expressions much faster, such as
3037 :			long lists with non-trivial elements.
3038 :
3039 :			[(0,0), (0,0), .... (0,0)]
3040 :
3041 :			3. MLRiscGen: base pointer elimination.
3042 :
3043 :			As part of the linkage mechanism, we generate the sequence:
3044 :
3045 :			L: ... <- start of the code fragment
3046 :
3047 :			L1:
3048 :			base pointer <- linkreg - L1 + L
3049 :
3050 :			The base pointer was then used for computing relocatable addresses
3051 :			in the code fragment. Frequently (such as in lots of continuations)
3052 :			this is not needed. We now eliminate this sequence whenever possible.
3053 :
3054 :			For compile time efficiency, I'm using a very stupid local heuristic.
3055 :			But in general, this should be done as a control flow analysis.
3056 :
3057 :			Effect: Smaller code size. Speed up of most programs.
3058 :
3059 :			4. Hppa back end
3060 :
3061 :			Long jumps in span dependence resolution used to depend on the existence
3062 :			of the base pointer.
3063 :
3064 :			A jump to a long label L was expanded into the following sequence:
3065 :
3066 :			LDIL %hi(L-8192), %r29
3067 :			LDO %lo(L-8192)(%r29), %r29
3068 :			ADD %r29, baseptr, %r29
3069 :			BV,n %r0(%r29)
3070 :
3071 :			In the presence of change (3) above, this will not work. I've changed
3072 :			it so that the following sequence of instructions are generated, which
3073 :			doesn't mention the base pointer at all:
3074 :
3075 :			BL,n L', %r29 /* branch and link, L' + 4 -> %r29 */
3076 :			L': ADDIL L-(L'+4), %r29 /* Compute address of L */
3077 :			BV,n %r0(%r29) /* Jump */
3078 :
3079 :			5. Alpha back end
3080 :
3081 :			New alpha instructions LDB/LDW have been added, as per Fermin's
3082 :			suggestions. This is unrelated to all other changes.
3083 :
3084 :			6. X86 back end
3085 :
3086 :			I've changed andl to testl in the floating point test sequence
3087 :			whenever appropriate. The Intel optimization guide states that
3088 :	monnier	659	testl is preferable to andl.
3089 :	leunga	585
3090 :			7. RA (x86 only)
3091 :
3092 :			I've improved the spill propagation algorithm, using an approximation
3093 :			of maximal weighted independent sets. This seems to be necessary to
3094 :			alleviate the negative effect in light of the slow down in (1).
3095 :
3096 :			I'll write down the algorithm one of these days.
3097 :
3098 :			8. MLRiscGen: frequencies
3099 :
3100 :			I've added an annotation that states that all call gc blocks have zero
3101 :			execution frequencies. This improves register allocation on the x86.
3102 :
3103 :			BENCHMARKS
3104 :			==========
3105 :
3106 :			I've only perform the comparison on 110.25.
3107 :
3108 :			The platforms are:
3109 :
3110 :			HPPA A four processor HP machine (E9000) with 5G of memory.
3111 :			X86 A 300Hhz Pentium II with 128M of memory, and
3112 :			SPARC An Ultra sparc 2 with 512M of memory.
3113 :
3114 :			I used the following parameters for the SML benchmarks:
3115 :
3116 :			@SMLalloc
3117 :			HPPA 256k
3118 :			SPARC 512k
3119 :			X86 256k
3120 :
3121 :			COMPILATION TIME
3122 :			----------------
3123 :			Here are the numbers comparing the compilation times of the compilers.
3124 :			I've only compared 110.25 compiling the new sources versus
3125 :			a fixpoint version of the new compiler compiling the same.
3126 :
3127 :			110.25 New
3128 :			Total Time in RA Spill+Reload Total Time In RA Spill+Reload
3129 :			HPPA 627s 116s 2684+3584 599s 95s 1003+1879
3130 :			SPARC 892s 173s 2891+3870 708s 116s 1004+1880
3131 :			X86 999s 315s 94006+130691 987s 296s 108877+141957
3132 :
3133 :			110.25 New
3134 :			Code Size Code Size
3135 :			HPPA 8596736 8561421
3136 :			SPARC 8974299 8785143
3137 :			X86 9029180 8716783
3138 :
3139 :			So in summary, things are at least as good as before. Dramatic
3140 :			reduction in compilation is obtained on the Sparc; I can't explain it,
3141 :			but it is reproducible. Perhaps someone should try to reproduce this
3142 :			on their own machines.
3143 :
3144 :			SML BENCHMARKS
3145 :			--------------
3146 :
3147 :			On the average, all benchmarks perform at least as well as before.
3148 :
3149 :			HPPA Compilation Time Spill+Reload Run Time
3150 :			110.25 New 110.25 New 110.25 New
3151 :
3152 :			barnesHut 3.158 3.015 4.75% 1+1 0+0 2.980 2.922 2.00%
3153 :			boyer 6.152 5.708 7.77% 0+0 0+0 0.218 0.213 2.34%
3154 :			count-graphs 1.168 1.120 4.32% 0+0 0+0 22.705 23.073 -1.60%
3155 :			fft 0.877 0.792 10.74% 1+3 1+3 0.602 0.587 2.56%
3156 :			knuthBendix 3.180 2.857 11.32% 0+0 0+0 0.675 0.662 2.02%
3157 :			lexgen 6.190 5.290 17.01% 0+0 0+0 0.913 0.788 15.86%
3158 :			life 0.803 0.703 14.22% 25+25 0+0 0.153 0.140 9.52%
3159 :			logic 2.048 2.007 2.08% 6+6 1+1 4.133 4.008 3.12%
3160 :			mandelbrot 0.077 0.080 -4.17% 0+0 0+0 0.765 0.712 7.49%
3161 :			mlyacc 22.932 20.937 9.53% 154+181 32+57 0.468 0.430 8.91%
3162 :			nucleic 5.183 5.060 2.44% 2+2 0+0 0.125 0.120 4.17%
3163 :			ratio-regions 3.357 3.142 6.84% 0+0 0+0 116.225 113.173 2.70%
3164 :			ray 1.283 1.290 -0.52% 0+0 0+0 2.887 2.855 1.11%
3165 :			simple 6.307 6.032 4.56% 28+30 5+7 3.705 3.658 1.28%
3166 :			tsp 0.888 0.862 3.09% 0+0 0+0 7.040 6.893 2.13%
3167 :			vliw 24.378 23.455 3.94% 106+127 25+45 2.758 2.707 1.91%
3168 :			--------------------------------------------------------------------------
3169 :			Average 6.12% 4.09%
3170 :
3171 :			SPARC Compilation Time Spill+Reload Run Time
3172 :			110.25 New 110.25 New 110.25 New
3173 :
3174 :			barnesHut 3.778 3.592 5.20% 2+2 0+0 3.648 3.453 5.65%
3175 :			boyer 6.632 6.110 8.54% 0+0 0+0 0.258 0.242 6.90%
3176 :			count-graphs 1.435 1.325 8.30% 0+0 0+0 33.672 34.737 -3.07%
3177 :			fft 0.980 0.940 4.26% 3+9 2+6 0.838 0.827 1.41%
3178 :			knuthBendix 3.590 3.138 14.39% 0+0 0+0 0.962 0.967 -0.52%
3179 :			lexgen 6.593 6.072 8.59% 1+1 0+0 1.077 1.078 -0.15%
3180 :			life 0.972 0.868 11.90% 26+26 0+0 0.143 0.140 2.38%
3181 :			logic 2.525 2.387 5.80% 7+7 1+1 5.625 5.158 9.05%
3182 :			mandelbrot 0.090 0.093 -3.57% 0+0 0+0 0.855 0.728 17.39%
3183 :			mlyacc 26.732 23.827 12.19% 162+189 32+57 0.550 0.560 -1.79%
3184 :			nucleic 6.233 6.197 0.59% 3+3 0+0 0.163 0.173 -5.77%
3185 :			ratio-regions 3.780 3.507 7.79% 0+0 0+0 133.993 131.035 2.26%
3186 :			ray 1.595 1.550 2.90% 1+1 0+0 3.440 3.418 0.63%
3187 :			simple 6.972 6.487 7.48% 29+32 5+7 3.523 3.525 -0.05%
3188 :			tsp 1.115 1.063 4.86% 0+0 0+0 7.393 7.265 1.77%
3189 :			vliw 27.765 24.818 11.87% 110+135 25+45 2.265 2.135 6.09%
3190 :			----------------------------------------------------------------------------
3191 :			Average 6.94% 2.64%
3192 :
3193 :			X86 Compilation Time Spill+Reload Run Time
3194 :			110.25 New 110.25 New 110.25 New
3195 :
3196 :			barnesHut 5.530 5.420 2.03% 593+893 597+915 3.532 3.440 2.66%
3197 :			boyer 8.768 7.747 13.19% 493+199 301+289 0.327 0.297 10.11%
3198 :			count-graphs 2.040 2.010 1.49% 298+394 315+457 26.578 28.660 -7.26%
3199 :			fft 1.327 1.302 1.92% 112+209 115+210 1.055 0.962 9.71%
3200 :			knuthBendix 5.218 5.475 -4.69% 451+598 510+650 0.928 0.932 -0.36%
3201 :			lexgen 9.970 9.623 3.60% 1014+841 1157+885 0.947 0.928 1.97%
3202 :			life 1.183 1.183 0.00% 162+182 145+148 0.127 0.103 22.58%
3203 :			logic 3.285 3.512 -6.45% 514+684 591+836 5.682 5.577 1.88%
3204 :			mandelbrot 0.147 0.143 2.33% 38+41 33+54 0.703 0.690 1.93%
3205 :			mlyacc 35.457 32.763 8.22% 3496+4564 3611+4860 0.552 0.550 0.30%
3206 :			nucleic 7.100 6.888 3.07% 239+168 201+158 0.175 0.173 0.96%
3207 :			ratio-regions 6.388 6.843 -6.65% 1182+257 981+300 120.142 120.345 -0.17%
3208 :			ray 2.332 2.338 -0.29% 346+398 402+494 3.593 3.540 1.51%
3209 :			simple 9.912 9.903 0.08% 1475+941 1579+1168 3.057 3.178 -3.83%
3210 :			tsp 1.623 1.532 5.98% 266+200 250+211 8.045 7.878 2.12%
3211 :			vliw 33.947 35.470 -4.29% 2629+2774 2877+3171 2.072 1.890 9.61%
3212 :			----------------------------------------------------------------------------
3213 :			Average 1.22% 3.36%
3214 :
3215 :			----------------------------------------------------------------------
3216 :			Name: Allen Leung
3217 :	leunga	583	Date: 2000/03/23 16:25:00
3218 :			Tag: leunga-20000323-fix_x86_alpha
3219 :			Description:
3220 :
3221 :			1. X86 fixes/changes
3222 :
3223 :			a. The old code generated for SETcc was completely wrong.
3224 :			The Intel optimization guide is VERY misleading.
3225 :
3226 :			2. ALPHA fixes/changes
3227 :
3228 :			a. Added the instructions LDBU, LDWU, STB, STW as per Fermin's suggestion.
3229 :			b. Added a new mode byteWordLoadStores to the functor parameter to Alpha()
3230 :			c. Added reassociation code for address computation.
3231 :
3232 :			----------------------------------------------------------------------
3233 :			Name: Allen Leung
3234 :	leunga	580	Date: 2000/03/22 01:23:00
3235 :			Tag: leunga-20000322-fix_x86_hppa_ra
3236 :			Description:
3237 :
3238 :			1. X86 fixes/changes
3239 :
3240 :			a. x86Rewrite bug with MUL3 (found by Lal)
3241 :			b. Added the instructions FSTS, FSTL
3242 :
3243 :			2. PA-RISC fixes/changes
3244 :
3245 :			a. B label should not be a delay slot candidate! Why did this work?
3246 :			b. ADDT(32, REG(32, r), LI n) now generates one instruction instead of two,
3247 :			as it should be.
3248 :			c. The assembly syntax for fstds and fstdd was wrong.
3249 :			d. Added the composite instruction COMICLR/LDO, which is the immediate
3250 :			operand variant of COMCLR/LDO.
3251 :
3252 :			3. Generic MLRISC
3253 :
3254 :			a. shuffle.sml rewritten to be slightly more efficient
3255 :			b. DIV bug in mltree-simplify fixed (found by Fermin)
3256 :
3257 :			4. Register Allocator
3258 :
3259 :			a. I now release the interference graph earlier during spilling.
3260 :			May improve memory usage.
3261 :
3262 :			----------------------------------------------------------------------
3263 :	blume	577	Name: Matthias Blume
3264 :	blume	578	Date: 2000/03/14 14:15:32
3265 :			Tag: blume_main_v110p26p1_2
3266 :			Description:
3267 :
3268 :			1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
3269 :			additional argument called "template" which is an optional string that
3270 :	monnier	659	specifies the layout of the tool command line. See the CM manual for
3271 :	blume	578	explanation.
3272 :
3273 :	monnier	659	2. A special-purpose tool can be "registered" by simply dropping the
3274 :	blume	578	corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
3275 :			directory where the .cm file lives that uses this tool. (The
3276 :			behavior/misfeature until now was to look for the tool description
3277 :			files in the current working directory.) As before, tool description
3278 :			files could also be anchored -- in which case they can live anywhere
3279 :			they like. Following the recent e-mail discussion, this change should
3280 :			make it easier to have special-purpose tools that are shipped together
3281 :			with the sources of the program that uses them.
3282 :
3283 :			----------------------------------------------------------------------
3284 :			Name: Matthias Blume
3285 :	blume	577	Date: 2000/03/10 07:48:34
3286 :			Tag: blume_main_v110p26p1_1
3287 :			Description:
3288 :
3289 :			I added a re-written version of Dave's fixpt script to src/system.
3290 :			Changes relative to the original version:
3291 :			- sh-ified (not everybody has ksh)
3292 :			- automatically figures out which architecture it runs on
3293 :			- uses ./makeml a bit more cleverly
3294 :			- never invokes ./installml (and, thus, does not clobber your
3295 :			good and working installation of sml in case something goes wrong)
3296 :			- accepts max iteration count using option "-iter <n>"
3297 :			- accepts a "base" name using option "-base <base>"
3298 :
3299 :			It does not build any extraneous heap images but directly rebuilds
3300 :			bin- and boot-hierarchies using makeml's "-rebuild" switch. Finally,
3301 :			it can incorporate existing bin- and boot- hierarchies. For example,
3302 :			suppose the base is set to "sml" (which is the default). Then it
3303 :			successively builds
3304 :
3305 :			sml.bin.<arch>-unix and sml.boot.<arch>-unix
3306 :			then sml1.bin.<arch>-unix and sml1.boot.<arch>-unix
3307 :			then sml2.bin.<arch>-unix and sml2.boot.<arch>-unix
3308 :			...
3309 :			then sml<n>.bin.<arch>-unix and sml<n>.boot.<arch>-unix
3310 :
3311 :			and so on. If any of these already exist, it will just use what's
3312 :			there. In particular, many people will have the initial set of bin
3313 :			and boot files around, so this saves time for at least one full
3314 :			rebuild. Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
3315 :			<k>=1,2,... is normally not a good idea when invoking fixpt. However,
3316 :			they might be the result of an earlier partial run of fixpt (which
3317 :	monnier	659	perhaps got accidentally killed). In this case, fixpt will quickly
3318 :	blume	577	move through what exists before continuing where it left off earlier,
3319 :			and, thus, saves a lot of time.
3320 :
3321 :			----------------------------------------------------------------------
3322 :	leunga	576	Name: Allen Leung
3323 :			Date: 00/03/10 02:20:00
3324 :			Tag: leunga-20000310-fix_x86_asm_ra
3325 :			Description:
3326 :	dbm	570
3327 :	leunga	576	More assembly output problems involving the indexed addressing mode
3328 :			on the x86 have been found and corrected. Thanks to Fermin Reig for the
3329 :			fix.
3330 :
3331 :			The interface and implementation of the register allocator have been changed
3332 :			slightly to accommodate the possibility to skip the register allocation
3333 :			phases completely and go directly to memory allocation. This is needed
3334 :			for C-- use.
3335 :
3336 :	dbm	570	----------------------------------------------------------------------
3337 :	blume	572	Name: Matthias Blume
3338 :	blume	575	Date: 00/03/09 10:23:53
3339 :			Tag: blume_main_v110p26p1_0
3340 :			Description:
3341 :
3342 :			* Complete re-organization of library names. Many libraries have been
3343 :			consolidated so that they share the same path anchor. For example,
3344 :			all MLRISC-related libraries are anchored at MLRISC, most libraries that
3345 :			are SML/NJ-specific are under "smlnj". Notice that names like
3346 :			host-cmb.cm or host-compiler.cm no longer exist. See system/README
3347 :			for a complete description of the new naming scheme. Quick reference:
3348 :
3349 :			host-cmb.cm -> smlnj/cmb.cm
3350 :			host-compiler.cm -> smlnj/compiler.cm
3351 :			full-cm.cm -> smlnj/cm.cm
3352 :			<arch>-<os>.cm -> smlnj/cmb/<arch>-<os>.cm
3353 :			<arch>-compiler.cm -> smlnj/compiler/<arch>.cm
3354 :
3355 :			* Bug fixes in CM.
3356 :			- exceptions in user code are being passed through (i.e., reach top level)
3357 :			- more bugs in paranoia mode fixed
3358 :			- bug related to checking group owners fixed
3359 :
3360 :			* New install.sh script that automagically fetches archive files:
3361 :			The new file config/srcarchiveurl must contain the URL of the
3362 :			(remote) directory that contains bin files (or other source archives).
3363 :			If install.sh does not find the archive locally, it tries to get
3364 :			it from that remote directory.
3365 :			This should simplify installation further: For machines that have
3366 :			access to the internet, just fetch <version>-config.tgz, unpack it,
3367 :	monnier	659	edit config/targets, and go (run config/install.sh). The script will
3368 :	blume	575	fetch everything else that it might need all by itself.
3369 :
3370 :			For CVS users, this mechanism is not relevant for source archives, but
3371 :			it is convenient for getting new sets of binfiles.
3372 :
3373 :			Archives should be tar files compressed with either gzip, compress, or
3374 :			bzip2. The script recognizes .tgz, .tar, tar.gz, tz, .tar.Z, and .tar.bz2.
3375 :
3376 :			----------------------------------------------------------------------
3377 :			Name: Matthias Blume
3378 :	blume	572	Date: 2000/03/07 04:01:04
3379 :			Tag: blume_main_v110_26_2
3380 :	dbm	570	Description:
3381 :	blume	572	- size info in BOOTLIST
3382 :			* no fixed upper limits for number of bootfiles or length of
3383 :			bootfile names in runtime
3384 :			* falling back to old behavior if no BOOTLIST size info found
3385 :			- allocation size heuristics in .run-sml
3386 :			* tries to read cache size from /proc/cpuinfo (this is important for
3387 :			small-cache Celeron systems!)
3388 :			- install.sh robustified
3389 :			- CM manual updates
3390 :			- paranoid mode
3391 :			* no more CMB.deliver() (i.e., all done by CMB.make())
3392 :			* can re-use existing sml.boot.* files
3393 :			* init.cmi now treated as library
3394 :			* library stamps for consistency checks
3395 :			- sml.boot.<arch>-<os>/PIDMAP file
3396 :			* This file is read by the CM startup code. This is used to minimize
3397 :			the amount of dynamic state that needs to be stowed away for the
3398 :			purpose of sharing between interactive system and user code.
3399 :			- CM.Anchor.anchor instead of CM.Anchor.{set,cancel}
3400 :			* Upon request by Elsa. Anchors now controlled by get-set-pair
3401 :			like most other CM state variables.
3402 :			- Compiler.CMSA eliminated
3403 :			* No longer supported by CM anyway.
3404 :			- fixed bugs in pickler that kept biting Stefan
3405 :			* past refs to past refs (was caused by the possibility that
3406 :			ad-hoc sharing is more discriminating than hash-cons sharing)
3407 :			* integer overflow on LargeInt.minInt
3408 :			- ml-{lex,yacc} build scripts now use new mechanism
3409 :			for building standalone programs
3410 :			- fixed several gcc -Wall warnings that were caused by missing header
3411 :			files, missing initializations, etc., in runtime (not all warnings
3412 :			eliminated, though)

---

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