Annotation of /sml/trunk/HISTORY

Parent Directory | Revision Log

---

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

---

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