Home My Page Projects Code Snippets Project Openings SML/NJ
Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

SCM Repository

[smlnj] Annotation of /sml/trunk/HISTORY
ViewVC logotype

Annotation of /sml/trunk/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log


Revision 976 - (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 : george 976 Name: Lal George
17 :     Date: 2001/11/14 09:44:04 EST
18 :     Tag:
19 :     Description:
20 :    
21 :     Fix for a backpatching bug reported by Allen.
22 :    
23 :     Because the boundary between short and long span-dependent
24 :     instructions is +/- 128, there are an astounding number of
25 :     span-dependent instructions whose size is over estimated.
26 :    
27 :     Allen came up with the idea of letting the size of span
28 :     dependent instructions be non-monotonic, for a maxIter
29 :     number of times, after which the size must be monotonically
30 :     increasing.
31 :    
32 :     This table shows the number of span-dependent instructions
33 :     whose size was over-estimated as a function of maxIter, for the
34 :     file Parse/parse/ml.grm.sml:
35 :    
36 :     maxIter # of instructions:
37 :     10 687
38 :     20 438
39 :     30 198
40 :     40 0
41 :    
42 :     In compiling the compiler, there is no significant difference in
43 :     compilation speed between maxIter=10 and maxIter=40. Actually,
44 :     my measurements showed that maxIter=40 was a tad faster than
45 :     maxIter=10! Also 96% of the files in the compiler reach a fix
46 :     point within 13 iterations, so fixing maxIter at 40, while high,
47 :     is okay.
48 :    
49 :     ----------------------------------------------------------------------
50 : blume 975 Name: Matthias Blume
51 :     Date: 2001/10/31 15:25:00 EST
52 :     Tag: blume-20011031-pgraph
53 :     Description:
54 :    
55 :     CKIT:
56 :     * Changed the "Function" constructor of type Ast.ctype to carry optional
57 :     argument identifiers.
58 :     * Changed the return type of TypeUtil.getFunction accordingly.
59 :     * Type equality ignores the argument names.
60 :     * TypeUtil.composite tries to preserve argument names but gives up quickly
61 :     if there is a mismatch.
62 :    
63 :     installation script:
64 :     * attempts to use "curl" if available (unless "wget" is available as well)
65 :    
66 :     CM:
67 :     * has an experimental implementation of "portable graphs" which I will
68 :     soon propose as an implementation-independent library format
69 :     * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
70 :    
71 :     NLFFI-LIB:
72 :     * some cleanup (all cosmetic)
73 :    
74 :     NLFFIGEN:
75 :     * temporarily disabled the mechanism that suppresses ML output for
76 :     C definitions whose identifiers start with an underscore character
77 :     * generate val bindings for enum constants
78 :     * user can request that only one style (light or heavy) is being used;
79 :     default is to use both (command-line arguments: -heavy and -light)
80 :     * fixed bug in handling of function types involving incomplete pointers
81 :     * generate ML entry points that take record arguments (i.e., using
82 :     named arguments) for C functions that have a prototype with named
83 :     arguments
84 :     (see changes to CKIT)
85 :    
86 :     ----------------------------------------------------------------------
87 : leunga 972 Name: Allen Leung
88 :     Date: 2001/10/27 20:34:00 EDT
89 :     Tag: leunga-20011027-x86-fast-fp-call
90 :     Description:
91 :    
92 :     Fixed the bug described in blume-20010920-slowfp.
93 :    
94 :     The fix involves
95 :     1. generating FCOPYs in FSTP in ia32-svid
96 :     2. marking a CALL with the appropriate annotation
97 :    
98 :     ----------------------------------------------------------------------
99 : blume 961 Name: Matthias Blume
100 :     Date: 2001/10/16 11:32:00 EDT
101 :     Tag: blume-20011016-netbsd
102 :     Description:
103 :    
104 :     Underscore patch from Chris Richards (fixing problem with compiling
105 :     runtime system under recent NetBSD).
106 :    
107 :     ----------------------------------------------------------------------
108 : george 959 Name: Allen Leung
109 :     Date: 2001/10/12 17:18:32 EDT 2001
110 :     Tag: leung-20011012-x86-printflowgraph
111 :     Description:
112 :    
113 :     X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
114 :    
115 :     ----------------------------------------------------------------------
116 : george 952 Name: Lal George
117 : george 958 Date: 2001/10/11 23:51:34 EDT
118 :     Tag: george-20011011-too-many-instrs
119 :     Description:
120 :    
121 :     The representation of a program point never expected to see more
122 :     than 65536 instructions in a basic block!
123 :    
124 :     ----------------------------------------------------------------------
125 :     Name: Lal George
126 : george 952 Date: 2001/10/09 09:41:37 EDT
127 :     Tag: george-20011008-mlrisc-labels
128 :     Description:
129 :    
130 :     Changed the machine description files to support printing of
131 :     local and global labels in assembly code, based on host assembler
132 :     conventions.
133 :    
134 :     ----------------------------------------------------------------------
135 : blume 916 Name: Matthias Blume
136 : blume 939 Date: 2001/09/25 15:25:00 EDT
137 :     Tag: blume-20010925-exninfo
138 :     Description:
139 :    
140 :     I provided a non-hook implementation of exnName (at the toplevel) and
141 :     made the "dummy" implementation of exnMessage (at the toplevel) more
142 :     useful: if nothing gets "hooked in", then at least you are going to
143 :     see the exception name and a message indicating why you don't see more.
144 :    
145 :     [For the time being, programs that need exnMessage and want to use
146 :     ml-build should either use General.exnMessage (strongly recommended) or
147 :     refer to structure General at some other point so that CM sees a
148 :     static dependency.]
149 :    
150 :     [Similar remarks go for "print" and "use": If you want to use their
151 :     functionality in stand-alone programs generated by ml-build, then use
152 :     TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
153 :    
154 :     ----------------------------------------------------------------------
155 :     Name: Matthias Blume
156 : blume 938 Date: 2001/09/20 17:28:00 EDT
157 :     Tag: blume-20010920-slowfp
158 :     Description:
159 :    
160 :     Allen says that x86-fast-fp is not safe yet, so I turned it off again...
161 :    
162 :     ----------------------------------------------------------------------
163 :     Name: Matthias Blume
164 : blume 937 Date: 2001/09/20 17:20:00 EDT
165 :     Tag: blume-20010920-canonicalpaths
166 :     Description:
167 :    
168 :     0. Updated the BOOT file (something that I forgot to do earlier).
169 :    
170 :     1. Small internal change to CM so that it avoids "/../" in filenames
171 :     as much as possible (but only where it is safe).
172 :    
173 :     2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
174 :     that arguments that contain delimiters are passed through correctly.
175 :     This change also means that all "special" arguments of the form
176 :     @SMLxxx... must come first.
177 :    
178 :     3. Changed install script to put relative anchor names for tool commands
179 :     into pathconfig.
180 :    
181 :     ----------------------------------------------------------------------
182 :     Name: Matthias Blume
183 : george 952 >>>>>>> 1.169
184 : blume 932 Date: 2001/09/18 15:35:00 EDT
185 :     Tag: blume-20010918-readme11036
186 :     Description:
187 :    
188 :     Added README files.
189 :    
190 :     ----------------------------------------------------------------------
191 :     Name: Matthias Blume
192 : blume 929 Date: 2001/09/18 11:45:00 EDT
193 :     Tag: Release_110_36 (retag)
194 :     Description:
195 :    
196 :     Fixed mistake in config/preloads. Retagged as 110.36.
197 :    
198 :     ----------------------------------------------------------------------
199 :     Name: Matthias Blume
200 : blume 928 Date: 2001/09/18 09:40:00 EDT
201 : blume 929 Tag: Release_110_36_orig (tag changed)
202 : blume 928 Description:
203 :    
204 :     New version (110.36). New bootfiles.
205 :    
206 :     ----------------------------------------------------------------------
207 :     Name: Matthias Blume
208 : blume 927 Date: 2001/09/14 16:15:00 EDT
209 :     Tag: blume-20010914-x86fastfp
210 :     Description:
211 :    
212 :     John committed some changes that Allen made, in particular a (hopefully)
213 :     correctly working version of the x86-fp module.
214 :    
215 :     I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
216 :     flag to "true". Everything seems to compile to a fixpoint ok, and
217 :     "mandelbrot" speeds up by about 15%.
218 :    
219 :     ----------------------------------------------------------------------
220 :     Name: Matthias Blume
221 : blume 923 Date: 2001/09/13 11:20:00 EDT
222 :     Tag: blume-20010913-minimal
223 :     Description:
224 :    
225 :     1. Stefan Monnier's patch to fix a miscompilation problem that
226 :     was brought to light by John Reppy's work on Moby.
227 :    
228 :     2. Implemented a minimal "structure Compiler" that contains just
229 :     "version" and "architecture". The minimal version will be
230 :     available when the full version is not. This is for backward-
231 :     compatibility with code that wants to test Compiler.version.
232 :    
233 :     ----------------------------------------------------------------------
234 :     Name: Matthias Blume
235 : blume 920 Date: 2001/08/28 14:03:00 EDT
236 :     Tag: blume-20010828-ml-lex
237 :     Description:
238 :    
239 :     Fix for bug 1581, received from Neophytos Michael.
240 :    
241 :     ----------------------------------------------------------------------
242 :     Name: Matthias Blume
243 : blume 918 Date: 2001/08/27 11:20:00 EDT
244 :     Tag: blume-20010827-readme11035
245 :     Description:
246 :    
247 :     Fleshed out the README file for 110.35.
248 :    
249 :     ----------------------------------------------------------------------
250 :     Name: Matthias Blume
251 : blume 916 Date: 2001/08/24 17:10:00 EDT
252 :     Tag: Release_110_35
253 :     Description:
254 :    
255 :     New version number (110.35). New bootfiles.
256 :    
257 :     ----------------------------------------------------------------------
258 : george 915 Name: Lal George
259 : george 909 Date: 2001/08/24 13:47:18 EDT 2001
260 :     Tag: george-20010824-MLRISC-graphs
261 :     Description:
262 :    
263 :     removed clusters from MLRISC completely and replaced with graphs.
264 :    
265 :     ----------------------------------------------------------------------
266 :     Name: Matthias Blume
267 : blume 905 Date: 2001/08/23 17:50:00 EDT
268 :     Tag: blume-20010823-toplevel
269 :     Description:
270 :    
271 :     - some reorganization of the code that implements various kinds of
272 :     environments in the compiler (static, dynamic, symbolic, combined)
273 :     - re-implemented the EnvRef module so that evalStream works properly
274 :     (if the stream contains references to "use", "CM.make", etc.)
275 :     - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
276 :    
277 :     ----------------------------------------------------------------------
278 :     Name: Matthias Blume
279 : blume 904 Date: 2001/08/20 15:50 EDT
280 :     Tag: blume20010820-slipup
281 :     Description:
282 :    
283 :     I forgot to commit a few files. Here they are...
284 :    
285 :     ----------------------------------------------------------------------
286 :     Name: Matthias Blume
287 : blume 903 Date: 2001/08/20 15:35:00 EDT
288 :     Tag: blume-20010820-debugprof
289 :     Description:
290 :    
291 :     !!!! NEW BOOTFILES !!!!
292 :    
293 :     This is another round of reorganizing the compiler sources. This
294 :     time the main goal was to factor out all the "instrumentation"
295 :     passes (for profiling and backtracing) into their own library.
296 :     The difficulty was to do it in such a way that it does not depend
297 :     on elaborate.cm but only on elabdata.cm.
298 :    
299 :     Therefore there have been further changes to both elaborate.cm and
300 :     elabdata.cm -- more "generic" things have been moved from the former
301 :     to the latter. As a result, I was forced to split the assignment
302 :     of numbers indicating "primtyc"s into two portions: SML-generic and
303 :     SML/NJ-specific. Since it would have been awkward to maintain,
304 :     I bit the bullet and actually _changed_ the mapping between these
305 :     numbers and primtycs. The bottom line of this is that you need
306 :     a new set of bin- and bootfiles.
307 :    
308 :     I have built new bootfiles for all architectures, so doing a fresh
309 :     checkout and config/install.sh should be all you need.
310 :    
311 :     The newly created library's name is
312 :    
313 :     $smlnj/viscomp/debugprof.cm
314 :    
315 :     and its sources live under
316 :    
317 :     src/compiler/DebugProf
318 :    
319 :     ----------------------------------------------------------------------
320 :     Name: Matthias Blume
321 : blume 902 Date: 2001/08/15 17:15:00 EDT
322 :     Tag: blume-20010815-compreorg
323 :     Description:
324 :    
325 :     This is a first cut at reorganizing the CM libraries that make up the
326 :     core of the compiler. The idea is to separate out pieces that could
327 :     be used independently by tools, e.g., the parser, the typechecker, etc.
328 :    
329 :     The current status is a step in this direction, but it is not quite
330 :     satisfactory yet. Expect more changes in the future.
331 :    
332 :     Here is the current (new) organization...
333 :    
334 :     What used to be $smlnj/viscomp/core.cm is now divided into
335 :     six CM libraries:
336 :    
337 :     $smlnj/viscomp/basics.cm
338 :     /parser.cm
339 :     /elabdata.cm
340 :     /elaborate.cm
341 :     /execute.cm
342 :     /core.cm
343 :    
344 :     The CM files for these libraries live under src/system/smlnj/viscomp.
345 :     All these libraries are proxy libraries that contain precisely
346 :     one CM library component. Here are the locations of the components
347 :     (all within the src/compiler tree):
348 :    
349 :     Basics/basics.cm
350 :     Parse/parser.cm
351 :     ElabData/elabdata.cm
352 :     Elaborator/elaborate.cm
353 :     Execution/execute.cm
354 :     core.cm
355 :    
356 :     [This organization is the same that has been used already
357 :     for a while for the architecture-specific parts of the visible
358 :     compiler and for the old version of core.cm.]
359 :    
360 :     As you will notice, many source files have been moved from their
361 :     respective original locations to a new home in one of the above
362 :     subtrees.
363 :    
364 :     The division of labor between the new libraries is the following:
365 :    
366 :     basics.cm:
367 :     - Simple, basic definitions that pertain to many (or all) of
368 :     the other libraries.
369 :     parser.cm:
370 :     - The SML parser, producing output of type Ast.dec.
371 :     - The type family for Ast is also defined and exported here.
372 :     elabdata.cm:
373 :     - The datatypes that describe input and output of the elaborator.
374 :     This includes types, absyn, and static environments.
375 :     elaborator.cm:
376 :     - The SML/NJ type checker and elaborator.
377 :     This maps an Ast.dec (with a given static environment) to
378 :     an Absyn.dec (with a new static environment).
379 :     - This libraries implements certain modules that used to be
380 :     structures as functors (to remove dependencies on FLINT).
381 :     execute.cm:
382 :     - Everything having to do with executing binary code objects.
383 :     - Dynamic environments.
384 :     core.cm:
385 :     - SML/NJ-specific instantiations of the elaborator and MLRISC.
386 :     - Top-level modules.
387 :     - FLINT (this should eventually become its own library)
388 :    
389 :     Notes:
390 :    
391 :     I am not 100% happy with the way I separated the elaborator (and its
392 :     data structures) from FLINT. Two instances of the same problem:
393 :    
394 :     1. Data structures contain certain fields that carry FLINT-specific
395 :     information. I hacked around this using exn and the property list
396 :     module from smlnj-lib. But the fact that there are middle-end
397 :     specific fields around at all is a bit annoying.
398 :    
399 :     2. The elaborator calculates certain FLINT-related information. I tried
400 :     to make this as abstract as I could using functorization, but, again,
401 :     the fact that the elaborator has to perform calculations on behalf
402 :     of the middle-end at all is not nice.
403 :    
404 :     3. Having to used exn and property lists is unfortunate because it
405 :     weakens type checking. The other alternative (parameterizing
406 :     nearly *everything*) is not appealing, though.
407 :    
408 :     I removed the "rebinding =" warning hack because due to the new organization
409 :     it was awkward to maintain it. As a result, the compiler now issues some of
410 :     these warnings when compiling init.cmi during bootstrap compilation. On
411 :     the plus side, you also get a warning when you do, for example:
412 :     val op = = Int32.+
413 :     which was not the case up to now.
414 :    
415 :     I placed "assign" and "deref" into the _Core structure so that the
416 :     code that deals with the "lazy" keyword can find them there. This
417 :     removes the need for having access to the primitive environment
418 :     during elaboration.
419 :    
420 :     ----------------------------------------------------------------------
421 :     Name: Matthias Blume
422 : blume 898 Date: 2001/08/13
423 :     Tag: blume-20010813-closures
424 :     Description:
425 :    
426 :     This fix was sent to us by Zhong Shao. It is supposed to improve the
427 :     performance of certain loops by avoiding needless closure allocation.
428 :    
429 :     ----------------------------------------------------------------------
430 : george 890 Name: Lal George
431 : george 896 Date: 2001/07/31 10:03:23 EDT 2001
432 :     Tag: george-20010731-x86-fmalloc
433 :     Description: Fixed bug in x86 calls
434 :    
435 :     There was a bug where call instructions would mysteriously
436 :     vanish. The call instruction had to be one that returned
437 :     a floating point value.
438 :    
439 :     ----------------------------------------------------------------------
440 :     Name: Lal George
441 : george 890 Date: 2001/07/19 16:36:29 EDT 2001
442 :     Tag: george-20010719-simple-cells
443 :     Description:
444 :    
445 :     I have dramatically simplified the interface for CELLS in MLRISC.
446 :    
447 :     In summary, the cells interface is broken up into three parts:
448 :    
449 :     1. CellsBasis : CELLS_BASIS
450 :    
451 :     CellsBasis is a top level structure and common for all
452 :     architectures. it contains the definitions of basic datatypes
453 :     and utility functions over these types.
454 :    
455 :     2. functor Cells() : CELLS
456 :    
457 :     Cells generates an interface for CELLS that incorporates the
458 :     specific resources on the target architecture, such as the
459 :     presence of special register classes, their number and size,
460 :     and various useful substructures.
461 :    
462 :     3. <ARCH>CELLS
463 :    
464 :     e.g. SparcCells: SPARCCELLS
465 :    
466 :     <ARCH>CELLS usually contains additional bindings for special
467 :     registers on the architecture, such as:
468 :    
469 :     val r0 : cell (* register zero *)
470 :     val y : cell (* Y register *)
471 :     val psr : cell (* processor status register *)
472 :     ...
473 :    
474 :     The structure returned by applying the Cells functor is opened
475 :     in this interface.
476 :    
477 :     The main implication of all this is that the datatypes for cells is
478 :     split between CellsBasis and CELLS -- a fairly simple change for user
479 :     code.
480 :    
481 :     In the old scheme the CELLS interface had a definitional binding of
482 :     the form:
483 :    
484 :     signature CELLS = sig
485 :    
486 :     structure CellsBasis = CellsBasis
487 :    
488 :     ...
489 :    
490 :     end
491 :    
492 :     With all the sharing constraints that goes on in MLRISC, this old
493 :     design quickly leads to errors such as:
494 :    
495 :     "structure definition spec inside of sharing ... "
496 :    
497 :    
498 :     and appears to require an unacceptable amount of sharing and where
499 :     constraint hackery.
500 :    
501 :     I think this error message (the interaction of definitional specs and
502 :     sharing) requires more explanation on our web page.
503 :    
504 :     ----------------------------------------------------------------------
505 : blume 826 Name: Matthias Blume
506 : blume 879 Date: 2001/07/19 15:00:00 EDT
507 :     Tag: blume-20010719-libreorg
508 :     Description:
509 :    
510 :     This update puts together a fairly extensive but straightforward change
511 :     to the way the libraries that implement the interactive system are
512 :     organized:
513 :    
514 :     The biggest change is the elimination of structure Compiler. As a
515 :     replacement for this structure, there is now a CM library
516 :     (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
517 :     that exports all the substructures of the original structure Compiler
518 :     directly. So instead of saying Compiler.Foo.bar one now simply
519 :     says Foo.bar. (The CM libraries actually export a collection of
520 :     structures that is richer than the collection of substructures of
521 :     structure Compiler.)
522 :    
523 :     To make the transition smooth, there is a separate library called
524 :     $smlnj/compiler/compiler.cm which puts together and exports the
525 :     original structure Compiler (or at least something very close to it).
526 :    
527 :     There are five members of the original structure Compiler
528 :     that are not exported directly but which instead became members
529 :     of a new structure Backend (described by signature BACKEND). These are:
530 :     structure Profile (: PROFILE), structure Compile (: COMPILE), structure
531 :     Interact (: INTERACT), structure Machine (: MACHINE), and val
532 :     architecture (: string).
533 :    
534 :     Structure Compiler.Version has become structure CompilerVersion.
535 :    
536 :     Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
537 :     by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
538 :     or x86, respectively.
539 :     Each of these exports the same frontend structures that
540 :     $smlnj/compiler.cm exports. But they do not have a structure Backend
541 :     and instead export some structure <Arch>Backend where <Arch> is Alpha32,
542 :     Hppa, PPC, Sparc, or X86, respectively.
543 :    
544 :     Library $smlnj/compiler/all.cm exports the union of the exports of
545 :     $smlnj/compiler/<arch>.cm
546 :    
547 :     There are no structures <Arch>Compiler anymore, use
548 :     $smlnj/compiler/<arch>.cm instead.
549 :    
550 :     Library host-compiler-0.cm is gone. Instead, the internal library
551 :     that instantiates CM is now called cm0.cm. Selection of the host
552 :     compiler (backend) is no longer done here but. (Responsibility for it
553 :     now lies with $smlnj/compiler/current.cm. This seems to be more
554 :     logical.)
555 :    
556 :     Many individual files have been moved or renamed. Some files have
557 :     been split into multiple files, and some "dead" files have been deleted.
558 :    
559 :     Aside from these changes to library organization, there are also changes
560 :     to the way the code itself is organized:
561 :    
562 :     Structure Binfile has been re-implemented in such a way that it no
563 :     longer needs any knowledge of the compiler. It exclusively deals
564 :     with the details of binfile layout. It no longer invokes the
565 :     compiler (for the purpose of creating new prospective binfile
566 :     content), and it no longer has any knowledge of how to interpret
567 :     pickles.
568 :    
569 :     Structure Compile (: COMPILE) has been stripped down to the bare
570 :     essentials of compilation. It no longer deals with linking/execution.
571 :     The interface has been cleaned up considerably.
572 :    
573 :     Utility routines for dealing with linking and execution have been
574 :     moved into their own substructures.
575 :    
576 :     (The ultimate goal of these changes is to provide a light-weight
577 :     binfile loader/linker (at least for, e.g., stable libraries) that
578 :     does not require CM or the compiler to be present.)
579 :    
580 :     CM documentation has been updated to reflect the changes to library
581 :     organization.
582 :    
583 :     ----------------------------------------------------------------------
584 :     Name: Matthias Blume
585 : blume 875 Date: 2001/07/10 17:30:00 EDT
586 :     Tag: Release_110_34
587 :     Description:
588 :    
589 : blume 876 Minor tweak to 110.34 (re-tagged):
590 :    
591 :     - README.html file added to CVS repository
592 :     - runtime compiles properly under FreeBSD 3.X and 4.X
593 :    
594 :     ----------------------------------------------------------------------
595 :     Name: Matthias Blume
596 :     Date: 2001/07/10 17:30:00 EDT
597 :     Tag: Release_110_34
598 :     Description:
599 :    
600 : blume 875 New version number (110.34). New bootfiles.
601 :    
602 :     ----------------------------------------------------------------------
603 :     Name: Matthias Blume
604 : blume 874 Date: 2001/07/09 16:00:00 EDT
605 :     Tag: blume-20010709-more-varargs
606 :     Description:
607 :    
608 :     I changed the handling of varargs in ml-nlffigen again:
609 :     The ellipsis ... will now simply be ignored (with an accompanying warning).
610 :    
611 :     The immediate effect is that you can actually call a varargs function
612 :     from ML -- but you can't actually supply any arguments beyond the ones
613 :     specified explicitly. (For example, you can call printf with its format
614 :     string, but you cannot pass additional arguments.)
615 :    
616 :     This behavior is only marginally more useful than the one before, but
617 :     it has the advantage that a function or, more importantly, a function
618 :     type never gets dropped on the floor, thus avoiding follow-up problems with
619 :     other types that refer to the offending one.
620 :    
621 :     ----------------------------------------------------------------------
622 :     Name: Matthias Blume
623 : blume 873 Date: 2001/07/09 11:25:00 EDT
624 :     Tag: blume-20010709-varargs
625 :     Description:
626 :    
627 :     1. ckit-lib.cm now exports structure Error
628 :     2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
629 :     with a warning accompanied by a source location. Moreover, it
630 :     merely skips the offending function or type and proceeds with the
631 :     rest of its work.u As a result, one can safely feed C code containing
632 :     "..." to ml-nlffigen.
633 :     3. There are some internal improvements to CM, providing slightly
634 :     more general string substitutions in the tools subsystem.
635 :    
636 :     ----------------------------------------------------------------------
637 :     Name: Matthias Blume
638 : blume 854 Date: 2001/06/27 15:10:00 EDT
639 :     Tag: blume-20010627-concur
640 :     Description:
641 :    
642 :     Fixed a small bug in CM's handling of parallel compilation.
643 :     (You could observe the bug by Control-C-interrupting an ordinary
644 :     CMB.make or CM.stabilize and then attaching some compile servers.
645 :     The result was that all of a sudden the previously interrupted
646 :     compilation would continue on its own. This was because of
647 :     an over-optimization: CM did not bother to clean out certain queues
648 :     when no servers were attached "anyway", resulting in the contents
649 :     of these queues to grab control when new servers did get attached.)
650 :    
651 :     There is also another minor update to the CM manual.
652 :    
653 :     ----------------------------------------------------------------------
654 :     Name: Matthias Blume
655 : blume 853 Date: 2001/06/26 16:15:00 EDT
656 :     Tag: blume-20010626-cmdoc
657 :     Description:
658 :    
659 :     Minor typo fixed in CM manual (syntax diagram for libraries).
660 :    
661 :     ----------------------------------------------------------------------
662 :     Name: Matthias Blume
663 : blume 852 Date: 2001/06/25 22:55:00 EDT
664 :     Tag: blume-20010625-x86pc
665 :     Description:
666 :    
667 :     Fixed a nasty bug in the X86 assembly code that caused signal
668 :     handlers to fail (crash) randomly.
669 :    
670 :     ----------------------------------------------------------------------
671 :     Name: Matthias Blume
672 : blume 846 Date: 2001/06/25 12:05:00 EDT
673 :     Tag: blume-20010625-nlffigen
674 :     Description:
675 :    
676 :     This update fixes a number of minor bugs in ml-nlffigen as reported by
677 :     Nick Carter <nbc@andrew.cmu.edu>.
678 :    
679 :     1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
680 :     2. Default names for generated files are now derived from the name of
681 :     the C file *without its directory*. In particular, this causes generated
682 :     files to be placed locally even if the C file is in some system directory.
683 :     3. Default names for generated signatures and structures are also derived
684 :     from the C file name without its directory. This avoids silly things
685 :     like "structure GL/GL".
686 :     (Other silly names are still possible because ml-nlffigen does not do
687 : blume 847 a thorough check of whether generated names are legal ML identifiers.
688 :     When in doubt, use command line arguments to force particular names.)
689 : blume 846
690 :     ----------------------------------------------------------------------
691 :     Name: Matthias Blume
692 : blume 845 Date: 2001/06/21 12:25:00 EDT
693 :     Tag: blume-20010621-eXene
694 :     Description:
695 :    
696 :     eXene now compiles and (sort of) works again.
697 :    
698 :     The library name (for version > 110.33) is $/eXene.cm.
699 :    
700 :     I also added an new example in src/eXene/examples/nbody. See the
701 :     README file there for details.
702 :    
703 :     ----------------------------------------------------------------------
704 :     Name: Matthias Blume
705 : blume 844 Date: 2001/06/20 16:40:00 EDT
706 :     Tag: blume-20010620-cml
707 :     Description:
708 :    
709 :     CML now compiles and works again.
710 :    
711 :     Libraries (for version > 110.33):
712 :    
713 :     $cml/cml.cm Main CML library.
714 :     $cml/basis.cm CML's version of $/basis.cm.
715 :     $cml/cml-internal.cm Internal helper library.
716 :     $cml/core-cml.cm Internal helper library.
717 :     $cml-lib/trace-cml.cm Tracing facility.
718 :     $cml-lib/smlnj-lib.cm CML's version of $/smlnj-lib.cm
719 :    
720 :     The installer (config/install.sh) has been taught how to properly
721 :     install this stuff.
722 :    
723 :     ----------------------------------------------------------------------
724 :     Name: Matthias Blume
725 : blume 843 Date: 2001/06/19 17:55:00 EDT
726 :     Tag: blume-20010619-instantiate
727 :     Description:
728 :    
729 :     This un-breaks the fix for bug 1432.
730 :     (The bug was originally fixed in 110.9 but I broke it again some
731 :     time after that.)
732 :    
733 :     ----------------------------------------------------------------------
734 :     Name: Matthias Blume
735 : blume 842 Date: 2001/06/19 17:25:00 EDT
736 :     Tag: blume-20010619-signals
737 :     Description:
738 :    
739 :     This should (hopefully) fix the long-standing signal handling bug.
740 :     (The runtime system was constructing a continuation record with an
741 :     incorrect descriptor which would cause the GC to drop data on the floor...)
742 :    
743 :     ----------------------------------------------------------------------
744 :     Name: Matthias Blume
745 : blume 840 Date: 2001/06/15 15:05:00 EDT
746 : blume 841 Tag: blume-20010615-moresparc
747 :     Description:
748 :    
749 :     Here is a short late-hour update related to Sparc c-calls:
750 :    
751 :     -- made handling of double-word arguments a bit smarter
752 :    
753 :     -- instruction selection phase tries to collapse certain clumsily
754 :     constructed ML-Trees; typical example:
755 :    
756 :     ADD(ty,ADD(_,e,LI d1),LI d2) -> ADD(ty,e,LI(d1+d2))
757 :    
758 :     This currently has no further impact on SML/NJ since mlriscGen does
759 :     not seem to generate such patterns in the first place, and c-calls
760 :     (which did generate them in the beginning) has meanwhile been fixed
761 :     so as to avoid them as well.
762 :    
763 :     ----------------------------------------------------------------------
764 :     Name: Matthias Blume
765 :     Date: 2001/06/15 15:05:00 EDT
766 : blume 840 Tag: blume-20010615-sparc
767 :     Description:
768 :    
769 :     The purpose of this update is to provide an implementation of NLFFI
770 :     on Sparc machines.
771 :    
772 :     Here are the changes in detail:
773 :    
774 :     * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
775 :     the Sparc implementation of the c-calls API.
776 :     * The Sparc backend of SML/NJ has been modified to uniformely use %fp
777 :     for accessing the ML frame. Thus, we have a real frame pointer and
778 :     can freely modify %sp without need for an omit-frame-ptr phase.
779 :     The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
780 :     this case.
781 :     * ml-nlffigen has been taught to produce code for different architectures
782 :     and calling conventions.
783 :     * In a way similar to what was done in the x86 case, the Sparc
784 :     backend uses its own specific extension to mltree. (For example,
785 :     it needs to be able to generate UNIMP instructions which are part
786 :     of the calling convention.)
787 :     * ml-nlffi-lib was reorganized to make it more modular (in particular,
788 :     to make it easier to plug in new machine- and os-dependent parts).
789 :    
790 :     There are some other fairly unrelated bug fixes and cleanups as well:
791 :    
792 :     * I further hacked the .cm files for MLRISC tools (like MDLGen) so
793 :     that they properly share their libraries with existing SML/NJ libraries.
794 :     * I fixed a minor cosmetic bug in CM, supressing certain spurious
795 :     follow-up error messages.
796 :     * Updates to CM/CMB documentation.
797 :    
798 :     TODO items:
799 :    
800 :     * MLRISC should use a different register as its asmTemp on the Sparc.
801 :     (The current %o2 is a really bad choice because it is part of the
802 :     calling conventions, so things might interfere in unexpected ways.)
803 :    
804 :     ----------------------------------------------------------------------
805 :     Name: Matthias Blume
806 : blume 839 Date: 2001/06/07
807 :     Tag: blume-20010607-calls
808 :     Description:
809 :    
810 :     A number of internal changes related to C calls and calling conventions:
811 :    
812 :     1. ML-Tree CALL statements now carry a "pops" field. It indicates the
813 :     number of bytes popped implicitly (by the callee). In most cases
814 :     this field is 0 but on x86/win32 it is some non-zero value. This
815 :     is information provided for the benefit of the "omit-frameptr" pass.
816 :     2. The CALL instruction on the x86 carries a similar "pops" field.
817 :     The instruction selection phase copies its value from the ML-Tree
818 :     CALL statement.
819 :     3. On all other architectures, the instruction selection phase checks
820 :     whether "pops=0" and complains if not.
821 :     4. The c-calls implementation for x86 now accepts two calling conventions:
822 :     "ccall" and "stdcall". When "ccall" is selected, the caller cleans
823 :     up after the call and pops is set to 0. For "stdcall", the caller
824 :     does nothing, leaving the cleanup to the callee; pops is set to
825 :     the number of bytes that were pushed onto the stack.
826 :     5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
827 :     distinguish between "ccall" and "stdcall".
828 :     6. The UNIMP instruction has been added to the supported Sparc instruction
829 :     set. (This is needed for implementing the official C calling convention
830 :     on this architecture.)
831 :     7. I fixed some of the .cm files under src/MLRISC/Tools to make them
832 :     work with the latest CM.
833 :    
834 :     ----------------------------------------------------------------------
835 :     Name: Matthias Blume
836 : blume 838 Date: 2001/06/05 15:10:00 EDT
837 :     Tag: blume-20010605-cm-index
838 :     Description:
839 :    
840 :     0. The "lambdasplit" parameter for class "sml" in CM has been documented.
841 :    
842 :     1. CM can now generate "index files". These are human-readable files
843 :     that list on a per-.cm-file basis each toplevel symbol defined or
844 :     imported. The location of the index file for
845 :     <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
846 :     To enable index-file generation, set CM.Control.generate_index to true
847 :     or export an environment-symbol: export CM_GENERATE_INDEX=true.
848 :    
849 :     The CM manual has been updated accordingly.
850 :    
851 :     2. I made some slight modifications to the c-calls API in MLRISC.
852 :    
853 :     a) There is now a callback to support saving/restoring of
854 :     dedicated but caller-save registers around the actual call
855 :     instruction.
856 :     b) One can optionally specify a comment-annotation for the
857 :     call instruction.
858 :    
859 :     3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
860 :     (For example, the comment annotation shows the C prototype of
861 :     the function being called.)
862 :    
863 :     ----------------------------------------------------------------------
864 :     Name: Matthias Blume
865 : blume 837 Date: 2001/06/01 13:30:00 EDT
866 :     Tag: blume-20010601-nlffi-cleanup
867 :     Description:
868 :    
869 :     This is mostly a cleanup of MLFFI stuff:
870 :    
871 :     - some signature files have been put into a more exposed place
872 :     - the ugly 'f type parameter is gone (simplifies types tremendously!)
873 :     - ml-nlffigen changed accordingly
874 :     - tutorial updated
875 :    
876 :     Other changes:
877 :    
878 :     - author's affiliation in CM manual(s) updated
879 :     - some more recognized keywords added to Allen's sml.sty
880 :    
881 :     ----------------------------------------------------------------------
882 :     Name: Matthias Blume
883 : blume 836 Date: 2001/05/25 15:30:00 EDT
884 :     Tag: blume-20010525-iptr
885 :     Description:
886 :    
887 :     - put the official 110.33-README (as it appears on the ftp server) under
888 :     CVS
889 :     - fixed a small bug related to incomplete pointer types in
890 :     ml-nlffigen
891 :     - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
892 :     (it does not need the 'f type parameter)
893 :    
894 :     ----------------------------------------------------------------------
895 :     Name: Matthias Blume
896 : blume 834 Date: 2001/05/23 14:30:00 EDT
897 :     Tag: Release_110_33
898 :     Description:
899 :    
900 :     New version number (110.33). New bootfiles.
901 :    
902 :     ----------------------------------------------------------------------
903 :     Name: Matthias Blume
904 : blume 833 Date: 2001/05/22 18:06:00 EDT
905 :     Tag: blume-20010522-targets
906 :     Description:
907 :    
908 :     Made install.sh use file config/targets.customized if it exists, falling
909 :     back to config/targets if it doesn't. This way one can have a customized
910 :     version of the targets file without touching the "real thing", thus
911 :     eliminating the constant fear of accidentally checking something bogus
912 :     back into the CVS repository... (File config/targets.customized must
913 :     not be added to the repository!)
914 :    
915 :     ----------------------------------------------------------------------
916 :     Name: Matthias Blume
917 : blume 832 Date: 2001/05/22 16:30:00 EDT
918 :     Tag: blume-20010522-minitut
919 :     Description:
920 :    
921 :     1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
922 :     struct returns.
923 :     2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt. This is some very
924 :     incomplete, preliminary documentation for NLFFI.
925 :    
926 :     ----------------------------------------------------------------------
927 :     Name: Matthias Blume
928 : blume 829 Date: 2001/05/14 11:30:00 EDT
929 :     Tag: blume-20010514-script
930 :     Description:
931 :    
932 :     Some bugs in install script fixed.
933 :    
934 :     In addition to that I also made a slight change to the NLFFI API:
935 :     Functors generated by ml-nlffigen now take the dynamic library as a
936 :     straight functor argument, not as a suspended one. (The original
937 :     functor code used to force the suspension right away anyway, so there
938 :     was nothing gained by this complication of the interface.)
939 :    
940 :     ----------------------------------------------------------------------
941 :     Name: Matthias Blume
942 : blume 828 Date: 2001/05/11 14:35:00 EDT
943 :     Tag: blume-20010511-ml-nlffi
944 :     Description:
945 :    
946 :     I finally took the plunge and added my new FFI code to the main
947 :     repository. For x86-linux it is now ready for prime-time.
948 :    
949 :     There are two new subdirectories of "src":
950 :    
951 :     - ml-nlffi-lib:
952 :     The utility library for programs using the FFI interface.
953 :     Here is the implementation of $/c.cm and its associated low-level
954 :     partners $/c-int.cm and $/memory.cm.
955 :     - ml-nlffigen:
956 :     A stand-alone program for generating ML glue code from C source
957 :     code.
958 :    
959 :     Building ml-nlffigen requires $/ckit-lib.cm.
960 :    
961 :     The config/install.sh script has been updates to do the Right Thing
962 :     (hopefully).
963 :    
964 :     Notice that the source tree for the C-Kit will not be put under "src"
965 :     but directly under the installation root directory. (This is the
966 :     structure that currently exists on the CVS server when you check out
967 :     module "sml".) Fortunately, config/install.sh knows about this oddity.
968 :    
969 :     Bugs: No documentation yet.
970 :    
971 :     ----------------------------------------------------------------------
972 :     Name: Matthias Blume
973 : blume 826 Date: 2001/05/09 16:35:00 EDT
974 :     Tag: blume-20010509-cpscontract
975 :     Description:
976 :    
977 :     Fixed a bug in the accounting code in cpsopt/contract.sml. (The
978 :     wrapper/unwrapper elimination did not decrement usage counts and some
979 :     dead variables got overlooked by the dead-up logic.)
980 :    
981 :     ----------------------------------------------------------------------
982 : george 824 Name: Lal George
983 :     Date: 2001/05/08 17:26:09 EDT
984 :     Tag: george-20010508-omit-frameptr
985 :     Description:
986 :    
987 :     Changes to implement the omit-frame-pointer optimization to support
988 :     raw C calls. For now, there is only support on the Intel x86, but
989 :     other architectures will follow as more experience is gained with this.
990 :    
991 :    
992 :     ----------------------------------------------------------------------
993 : blume 818 Name: Matthias Blume
994 : blume 819 Date: 2001/05/07 14:40:00 EDT
995 :     Tag: blume-20010507-proxies
996 :     Description:
997 :    
998 :     I made into "proxy libraries" all libraries that qualify for such a
999 :     change. (A qualifying library is a library that has another library or
1000 :     groups as its sole member and repeats that member's export list
1001 :     verbatim. A proxy library avoids this repetition by omitting its export
1002 :     list, effectively inheriting the list that its (only) member exports.
1003 :     See the CM manual for more explanation.)
1004 :     The main effect is that explicit export lists for these libraries
1005 :     do not have to be kepts in sync, making maintenance a bit easier.
1006 :    
1007 :     I also added copyright notices to many .cm-files.
1008 :    
1009 :     Last but not least, I made a new set of bootfiles.
1010 :    
1011 :     ----------------------------------------------------------------------
1012 :     Name: Matthias Blume
1013 : blume 818 Date: 2001/05/04 17:00:00 EDT
1014 :     Tag: blume-20010504-cm-lsplit
1015 :     Description:
1016 :    
1017 :     0. John merged pending changes to $/smlnj-lib.cm
1018 :    
1019 :     1. Allen's previous change accidentally backed out of one of Lal's
1020 :     earlier changes. I undid this mistake (re-introducing Lal's change).
1021 :    
1022 :     2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1023 :     within the compiler where applicable. There is some code simplification
1024 :     because of that.
1025 :    
1026 :     3. The "split" phase (in FLINT) is now part of the default list of phases.
1027 :     Compiler.Control.LambdaSplitting.* can be used to globally control the
1028 :     lambda-splitting (cross-module-inlining) engine. In addition to that,
1029 :     it can now also be controlled on a per-source basis: CM has been taught
1030 :     a new tool parameter applicable to ML source files.
1031 :    
1032 :     - To turn lambda-splitting off completely:
1033 :     local open Compiler.Control.LambdaSplitting in
1034 :     val _ = set Off
1035 :     end
1036 :     - To make "no lambda-splitting" the global default (but allow per-source
1037 :     overriding); this is the initial setting:
1038 :     local open Compiler.Control.LambdaSplitting in
1039 :     val _ = set (Default NONE)
1040 :     end
1041 :     - To make "lambda-splitting with aggressiveness a" the global default
1042 :     (and allow per-source overriding):
1043 :     local open Compiler.Control.LambdaSplitting in
1044 :     val _ = set (Default (SOME a))
1045 :     end
1046 :    
1047 :     - To turn lambda-splitting off for a given ML souce file (say: a.sml)
1048 :     write (in the respective .cm-file):
1049 :     a.sml (lambdasplitting:off)
1050 :     - To turn lambda-splitting for a.sml on with minimal aggressiveness:
1051 :     a.sml (lambdasplitting:on)
1052 :     - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1053 :     <a> is a decimal non-negative integer):
1054 :     a.sml (lambdasplitting:<a>)
1055 :     - To turn lambda-splitting for a.sml on with maximal aggressiveness:
1056 :     a.sml (lambdasplitting:infinity)
1057 :     - To use the global default for a.sml:
1058 :     a.sml (lambdasplitting:default)
1059 :     or simply
1060 :     a.sml
1061 :    
1062 :     ----------------------------------------------------------------------
1063 : leunga 815 Name: Allen Leung
1064 :     Date: 2001/05/04 01:57:00 EDT
1065 :     Tag: leunga-20010504-sync
1066 :     Description:
1067 : leunga 788
1068 : leunga 815 MLRISC features.
1069 :    
1070 :     1. Fix to CMPXCHG instructions.
1071 :     2. Changed RA interface to allow annotations in callbacks.
1072 :     3. Added a new method to the stream interface to allow annotations updates.
1073 :    
1074 : george 749 ----------------------------------------------------------------------
1075 : blume 812 Name: Matthias Blume
1076 : blume 814 Date: 2001/05/01 11:45:00 EDT
1077 :     Tag: blume-20010501-pcedittmp
1078 :     Description:
1079 :    
1080 :     Changed install.sh to use the current working directory instead of
1081 :     /usr/tmp for a temporary file (pcedittmp). The previous choice
1082 :     of /usr/tmp caused trouble with MacOS X because of file premission
1083 :     problems.
1084 :    
1085 :     ----------------------------------------------------------------------
1086 :     Name: Matthias Blume
1087 : blume 812 Date: 2001/04/20 11:10:00 EDT
1088 :     Tag: blume-20010420-inMLflag
1089 :     Description:
1090 :    
1091 :     - added vp_limitPtrMask to vproc-state.h
1092 :     (for use by the raw-C-calls mechanism to implement proper interrupt
1093 :     handling)
1094 :     - made the ML compiler aware of various data-structure offsets so it
1095 :     can generate code for accessing the vp_inML flag and vp_limitPtrMask
1096 :     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1097 :     raw C-calls
1098 :    
1099 :     ----------------------------------------------------------------------
1100 : george 810 Name: Lal George
1101 :     Date: 2001/04/20 09:15:28 EDT
1102 :     Tag: george-20010420-macosX
1103 :     Description:
1104 :    
1105 :     - Changes to port to Mac OS X; Darwin.
1106 :    
1107 :     - In the process I found that sqrt was broken on the PPC, because the
1108 :     fsqrt instruction is not implemented.
1109 :    
1110 :     ----------------------------------------------------------------------
1111 : blume 799 Name: Matthias Blume
1112 : blume 808 Date: 2001/04/18 12:45:00 EDT
1113 :     Tag: blume-20010418-ccalls
1114 :     Description:
1115 :    
1116 :     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1117 :     (this bug prevented structure arguments containing pointers from being
1118 :     passed correctly)
1119 :     - changed the raw-C-call code in mlriscGen.sml in such a way that
1120 :     structure arguments are represented as a pointer to the beginning
1121 :     of the structure (instead of having a series of synthesized arguments,
1122 :     one for each structure member)
1123 :    
1124 :     - made makeml script's verbosity level configurable via environment
1125 :     variable (MAKEML_VERBOSITY)
1126 :    
1127 : blume 813 - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1128 : blume 808 in rawmem-x86.sml; we are now using the real thing
1129 :    
1130 :     ----------------------------------------------------------------------
1131 :     Name: Matthias Blume
1132 : blume 806 Date: 2001/03/22 16:25:00 EST
1133 :     Tag: blume-20010322-bootfiles
1134 :     Description:
1135 :    
1136 :     Created a new set of bootfiles (for your automatic installation convenience).
1137 :    
1138 :     ----------------------------------------------------------------------
1139 :     Name: Matthias Blume
1140 : blume 805 Date: 2001/03/22 15:10:00 EST
1141 :     Tag: blume-20010322-rawmem-parcm
1142 :     Description:
1143 :    
1144 :     1. All "raw memory access" primitives for the new FFI are implemented now
1145 :     (at least on the x86).
1146 :     2. Some further cleanup of CM's parallel make mechanism.
1147 :    
1148 :     ----------------------------------------------------------------------
1149 :     Name: Matthias Blume
1150 : blume 801 Date: 2001/03/19 17:53:00 EST
1151 :     Tag: blume-20010319-parallel
1152 :     Description:
1153 :    
1154 :     Parallel make (using compile servers) now works again.
1155 :    
1156 :     To this end, CM.stabilize and CMB.make have been modified to work in
1157 :     two passes when compile servers are attached:
1158 :     1. Compile everything, do not perform stabilization; this pass
1159 :     uses compile servers
1160 :     2. Stabilize everything; this pass does not use compile servers
1161 :     If there are no compile servers, the two passes are combined into one
1162 :     (as before). Splitting the passes increases the inherent parallelism
1163 :     in the dependency graph because the entire graph including all
1164 :     libraries is available at the same time. This, in turn, improves
1165 :     server utilization. The downside is that the master process will
1166 :     have to do some extra work after compilation is done (because for
1167 :     technical reasons it must re-read all the binfiles during stabilization).
1168 :    
1169 :     ----------------------------------------------------------------------
1170 :     Name: Matthias Blume
1171 : blume 800 Date: 2001/03/16 12:22:00 EST
1172 :     Tag: blume-20010316-bootfiles
1173 :     Description:
1174 :    
1175 :     Created a new set of bootfiles (for your automatic installation convenience).
1176 :    
1177 :     ----------------------------------------------------------------------
1178 :     Name: Matthias Blume
1179 : blume 799 Date: 2001/03/16 11:00:00 EST
1180 :     Tag: blume-20010316-MLTREE-fixup
1181 :     Description:
1182 :    
1183 :     This is a minor fixup for an (untagged) earlier commit by Allen.
1184 :     (A file was missing).
1185 :    
1186 :     ----------------------------------------------------------------------
1187 : leunga 788 Name: Allen Leung
1188 : leunga 796 Date: Mon Mar 5 18:54:57 EST 2001
1189 :     Tag: leunga-20010305-cut-support
1190 :    
1191 :     1. New support for alternative control-flow in MLTREE.
1192 :     Currently we support
1193 :    
1194 :     FLOW_TO(CALL ...., [k1,...,kn])
1195 :    
1196 :     This is needed for 'cuts to' in C-- and try/handle-like constructs
1197 :     in Moby
1198 :    
1199 :     New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1200 :    
1201 :     2. Register Allocator
1202 :    
1203 :     Changes in interface [from Fermin, John]
1204 :    
1205 :     3. Alpha 8-bit SLL support [Fermin]
1206 :    
1207 :     4. All architectures
1208 :    
1209 :     A new module (ClusterExpandCopies) for expanding parallel copies.
1210 :    
1211 :     ----------------------------------------------------------------------
1212 :     Name: Allen Leung
1213 : leunga 788 Date: 2001/02/27 23:07:00 EST
1214 :     Tag: leunga-20010227-minor-stuff
1215 :    
1216 :     1. Alpha bug fix for CMOVNE
1217 :     2. Handle mltree COND(..,FCMP ...,...)
1218 :     3. Bug fix in simplifier
1219 :    
1220 :     ----------------------------------------------------------------------
1221 : blume 777 Name: Matthias Blume
1222 : blume 781 Date: 2001/01/30 17:50:00 EST
1223 :     Tag: blume-20010130-sync
1224 :     Description:
1225 :    
1226 :     This is just a minor update to sync my devel branch with the main brach.
1227 :     The only visible change is the addition of some README files.
1228 :    
1229 :     ----------------------------------------------------------------------
1230 :     Name: Matthias Blume
1231 : blume 778 Date: 2001/01/12 23:30:00 JST
1232 :     Tag: blume-20010112-bootfiles
1233 :     Description:
1234 :    
1235 :     Made a new set of bootfiles that goes with the current state of the
1236 :     repository.
1237 :    
1238 :     ----------------------------------------------------------------------
1239 :     Name: Matthias Blume
1240 : blume 777 Date: 2001/01/12 21:20:00 JST
1241 :     Tag: blume-20010112-sync
1242 :     Description:
1243 :    
1244 :     I am just flushing out some minor changes that had accumulated in
1245 :     my private branch in order to sync with the main tree. (This is
1246 :     mainly because I had CVS trouble when trying to merge _into_ my
1247 :     private branch.)
1248 :    
1249 :     Most people should be completely unaffected by this.
1250 :    
1251 :     ----------------------------------------------------------------------
1252 : leunga 775 Name: Allen Leung
1253 :     Date: Thu Jan 11 21:03:00 EST 2001
1254 :     Tag: leunga-20010111-labexp=mltree
1255 :     Description:
1256 :    
1257 :     1. Removed the type LabelExp and replace it by MLTree.
1258 :     2. Rewritten mltree-simplify with the pattern matcher tool.
1259 :     3. There were some bugs in alpha code generator which would break
1260 :     64-bit code generation.
1261 :     4. Redo the tools to generate code with the
1262 :     5. The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1263 :     are now generated by perl scripts.
1264 :    
1265 :     ----------------------------------------------------------------------
1266 : blume 771 Name: Matthias Blume
1267 : blume 774 Date: 2001/01/10 21:55:00 JST
1268 :     Tag: blume-20010110-rcc
1269 :     Description:
1270 :    
1271 :     The RCC stuff now seems to work (but only on the x86).
1272 :     This required hacking of the c-calls interface (and -implementation) in
1273 :     MLRISC.
1274 :    
1275 :     Normal compiler users should be unaffected.
1276 :    
1277 :     ----------------------------------------------------------------------
1278 :     Name: Matthias Blume
1279 : blume 773 Date: 2001/01/09 01:20:00 JST
1280 :     Tag: blume-20010109-rcc
1281 :     Description:
1282 :    
1283 :     This is a fairly big patch, flushing out a large number of pending
1284 :     changes that I made to my development copy over the last couple of days.
1285 :    
1286 :     Of practical relevance at this moment is a workaround for a pickling
1287 :     bug that Allen ran into the other day. The cause of the bug itself is
1288 :     still unknown and it might be hard to fix it properly, but the
1289 :     workaround has some merits of its own (namely somewhat reducing pickling
1290 :     overhead for certain libraries). Therefore, I think this solution should
1291 :     be satisfactory at this time.
1292 :    
1293 :     The rest of the changes (i.e., the vast majority) has to do with my
1294 :     ongoing efforts of providing direct support for C function calls from
1295 :     ML. At the moment there is a new primop "RAW_CCALL", typing magic
1296 :     in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1297 :     case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1298 :     translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1299 :     handle RCC.
1300 :    
1301 :     The last part (the changes to mlriscGen.sml) are still known to be
1302 :     wrong on the x86 and not implemented on all other architectures. But
1303 :     the infrastructure is in place. I had to change a few functor
1304 :     signatures in the backend to be able to route the CCalls interface
1305 :     from MLRISC there, and I had to specialize the mltree type (on the
1306 :     x86) to include the necessary extensions. (The extensions themselves
1307 :     were already there and redy to go in MLRISC/x86).
1308 :    
1309 :     Everything should be very happy as soon as someone helps me with
1310 :     mlriscGen.sml...
1311 :    
1312 :     In any case, nothing of this should matter to anyone as long as the
1313 :     new primop is not being used (which is going to be the case unless you
1314 :     find it where I hid it :). The rest of the compiler is completely
1315 :     unaffected.
1316 :    
1317 :     ----------------------------------------------------------------------
1318 :     Name: Matthias Blume
1319 : blume 772 Date: 2001/01/05 00:30:00 JST
1320 :     Tag: blume-20010105-primops
1321 :     Description:
1322 :    
1323 :     Added some experimental support for work that I am doing right now.
1324 :     These changes mostly concern added primops, but there is also a new
1325 :     experimental C library in the runtime system (but currently not enabled
1326 :     anywhere except on Linux/X86).
1327 :    
1328 :     In the course of adding primops (and playing with them), I discovered that
1329 :     Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1330 :     broken. (Zhong was very right he labeled this stuff as "major gross hack".)
1331 :     To recover, I made type information in INL_PRIM mandatory and changed
1332 :     prim.sml as well as built-in.sml accordingly. The InLine structure now
1333 :     has complete, correct type information (i.e., no bottom types).
1334 :    
1335 :     Since all these changes mean that we need new binfiles, I also bumped the
1336 :     version number to 110.32.1.
1337 :    
1338 :     ----------------------------------------------------------------------
1339 :     Name: Matthias Blume
1340 : blume 771 Date: 2000/12/30 22:10:00 JST
1341 :     Tag: blume-20001230-various
1342 :     Description:
1343 :    
1344 :     Added proxy libraries for MLRISC and let MLRISC libraries refer
1345 :     to each other using path anchors. (See CM manual for explanation.)
1346 :    
1347 :     Updated CM documentation.
1348 :    
1349 :     Fixed some bugs in CM.
1350 :    
1351 :     Implemented "proxy" libraries (= syntactic sugar for CM).
1352 :    
1353 :     Added "-quiet" option to makeml and changed runtime system accordingly.
1354 :    
1355 :     Added cleanup handler for exportML to reset timers and compiler stats.
1356 :    
1357 :     ----------------------------------------------------------------------
1358 : george 761 Name: Lal George
1359 :     Date: 2000/12/22 22:22:58 EST 2000
1360 :     Tag: Release_110_32
1361 :     Description:
1362 :    
1363 :     Infinite precision used throughout MLRISC.
1364 :     see MLRISC/mltree/machine-int.sig
1365 :    
1366 :     ----------------------------------------------------------------------
1367 : blume 756 Name: Matthias Blume
1368 : blume 760 Date: 2000/12/22 23:16:00 JST
1369 :     Tag: blume-20001222-warn
1370 :     Description:
1371 : blume 759
1372 : blume 760 Corrected wording and formatting of some CM warning message which I
1373 :     broke in my previous patch.
1374 : blume 759
1375 :     ----------------------------------------------------------------------
1376 :     Name: Matthias Blume
1377 : blume 763 Date: 2000/12/22 21:20:00 JST
1378 :     Tag: blume-20001222-anchorenv
1379 :     Description:
1380 :    
1381 :     Fixed CM's handling of anchor environments in connection with CMB.make.
1382 :    
1383 :     ----------------------------------------------------------------------
1384 :     Name: Matthias Blume
1385 : blume 758 Date: 2000/12/22 13:15:00 JST
1386 :     Tag: blume-20001222-cleanup
1387 :     Description:
1388 :    
1389 :     Removed src/cm/ffi which does not (and did not) belong here.
1390 :    
1391 :     ----------------------------------------------------------------------
1392 :     Name: Matthias Blume
1393 : blume 757 Date: 2000/12/21 23:55:00 JST
1394 :     Tag: blume-20001221-exn
1395 :     Description:
1396 :    
1397 :     Probably most important: CM no longer silently swallows all exceptions
1398 :     in the compiler.
1399 :     Plus: some other minor CM changes. For example, CM now reports some
1400 :     sizes for generated binfiles (code, data, envpickle, lambdapickle).
1401 :    
1402 :     ----------------------------------------------------------------------
1403 :     Name: Matthias Blume
1404 : blume 756 Date: 2000/12/15 00:01:05 JST
1405 :     Tag: blume-20001215-dirtool
1406 :     Description:
1407 :    
1408 :     - "dir" tool added.
1409 :     - improvements and cleanup to Tools structure
1410 :     - documentation updates
1411 :    
1412 :     ----------------------------------------------------------------------
1413 : leunga 752 Name: Allen Leung
1414 : leunga 755 Date: Thu Dec 14 03:45:24 EST 2000
1415 :     Description:
1416 :     Tag: leunga-20001214-int-inf
1417 :     Description:
1418 :    
1419 :     In IntInf, added these standard functions, which are missing from our
1420 :     implementation:
1421 :    
1422 :     andb : int * int -> int
1423 :     xorb : int * int -> int
1424 :     orb : int * int -> int
1425 :     notb : int -> int
1426 :     << : int * word -> int
1427 :     ~>> : int * word -> int
1428 :    
1429 :     Not tested, I hope they are correct.
1430 :    
1431 :     ----------------------------------------------------------------------
1432 :     Name: Allen Leung
1433 : leunga 752 Date: Fri Dec 8 19:23:26 EST 2000
1434 :     Description:
1435 :     Tag: leunga-20001208-nowhere
1436 :     Description:
1437 :    
1438 :     Slight improvements to the 'nowhere' tool to handle OR-patterns,
1439 :     to generate better error messages etc. Plus a brief manual.
1440 :    
1441 :     ----------------------------------------------------------------------
1442 : george 749 Name: Lal George
1443 :     Date: 2000/12/08 09:54:02 EST 2000
1444 :     Tag: Release_110_31
1445 :     Description:
1446 : george 705
1447 : george 749 - Version 110.31
1448 : leunga 591 ----------------------------------------------------------------------
1449 : leunga 744 Name: Allen Leung
1450 :     Date: Thu Dec 7 22:01:04 EST 2000
1451 :     Tag: leunga-20001207-cell-monster-hack
1452 :     Description:
1453 :    
1454 :     Major MLRISC internal changes. Affect all clients.
1455 :     Summary:
1456 :    
1457 :     1. Type CELLS.cell = int is now replaced by a datatype.
1458 :     As a result, the old regmap is now gone. Almost all interfaces
1459 :     in MLRISC change as a consequence.
1460 :    
1461 :     2. A new brand version of machine description tool (v3.0) that generates
1462 :     modules expecting the new interface. The old version is removed.
1463 :    
1464 :     3. The RA interface has been further abstracted into two new functors.
1465 :     RISC_RA and X86RA. These functors have much simpler interfaces.
1466 :     [See also directory MLRISC/demo.]
1467 :    
1468 :     4. Some other new source->source code generation tools are available:
1469 :    
1470 :     a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
1471 :     b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
1472 :     I use this tool to generate the peephole optimizers---with the new
1473 :     cell type changes, peephole rules are becoming difficult to write
1474 :     without conditional pattern matching.
1475 :    
1476 :     5. More Intmap -> IntHashTable change. Previous changes by Matthias didn't
1477 :     cover the entire MLRISC source tree so many things broke.
1478 :    
1479 :     6. CM files have been moved to the subdirectory MLRISC/cm.
1480 :     They are moved because there are a lot of them and they clutter up the
1481 :     root dir.
1482 :    
1483 :     7. More detailed documentation to come...
1484 :    
1485 :     NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
1486 :     a makeml -rebuild first. This is because of other other
1487 :     changes that Matthias has made (see below).
1488 :    
1489 :    
1490 :     ----------------------------------------------------------------------
1491 : blume 733 Name: Matthias Blume
1492 : blume 742 Date: 2000/11/30 23:12:00 JST
1493 :     Tag: blume-20001130-filereorg
1494 :     Description:
1495 :    
1496 :     Some manual updates and some file reorganizations in CM.
1497 :    
1498 :     ----------------------------------------------------------------------
1499 :     Name: Matthias Blume
1500 : blume 737 Date: 2000/11/24 17:45:00 JST
1501 :     Tag: blume-20001124-link
1502 :     Description:
1503 :    
1504 :     Drastically improved link traversal code for the case that the dynamic
1505 :     value was already loaded at bootstrap time. As a result, CM and CMB
1506 :     now both load blazingly fast -- even on a very slow machine. Also,
1507 :     memory consumption has been further reduced by this.
1508 :    
1509 :     Warning: The format of the PIDMAP file has changed. THerefore, to
1510 :     bootstrap you have to do this:
1511 :    
1512 :     1. Run CMB.make
1513 :     2. Make a symbolic link for the boot directory:
1514 :     ln -s sml.boot.ARCH-OS xxx
1515 :     3. "Rebuild" the boot directory:
1516 :     ./makeml -boot xxx -rebuild sml ; rm xxx
1517 :     4. Boot normally:
1518 :     ./makeml
1519 :    
1520 :     ----------------------------------------------------------------------
1521 :     Name: Matthias Blume
1522 : blume 735 Date: 2000/11/21 21:20:00 JST
1523 :     Tag: blume-20001121-tools
1524 :     Description:
1525 :    
1526 :     Continued hacking on autoloading problem -- with success this time.
1527 :     Also changed tool-plugin mechanism. See new CM manual.
1528 :    
1529 :     ----------------------------------------------------------------------
1530 :     Name: Matthias Blume
1531 : blume 734 Date: 2000/11/19 14:30:00 JST
1532 :     Tag: blume-20001119-autoload
1533 :     Description:
1534 :    
1535 :     Some hacking to make autoloading faster. Success for CMB, no success
1536 :     so far for CM. There is a reduced structure CM' that autoloads faster.
1537 :     (This is a temporary, non-documented hack to be eliminated again when
1538 :     the general problem is solved.)
1539 :    
1540 :     ----------------------------------------------------------------------
1541 :     Name: Matthias Blume
1542 : blume 733 Date: 2000/11/17 14:10:00 JST
1543 :     Tag: blume-20001117-pickle-lib
1544 :     Description:
1545 :    
1546 :     1. Eliminated comp-lib.cm
1547 :     2. Made pickle-lib.cm
1548 :     3. Eliminated all uses of intset.sml (from comp-lib.cm)
1549 :     4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
1550 :     equivalent constructs from smlnj-lib.cm (INtHashTable).
1551 :     5. Point 4. also goes for those uses of intmap.* in MLRISC.
1552 :     Duplicated intmap modules thrown out.
1553 :     6. Hunted down all duplicated SCC code and replaced it with
1554 :     equivalent stuff (GraphSCCFn from smlnj-lib.cm).
1555 :     7. Rewrote Feedback module.
1556 :     8. Moved sortedlist.sml into viscomp-lib.cm. Eventually it
1557 :     should be thrown out and equivalent modules from smlnj-lib.cm
1558 :     should be used (IntRedBlackSet, IntListSet, ...).
1559 :    
1560 :     Confirmed that compiler compiles to fixpoint.
1561 :    
1562 :     ----------------------------------------------------------------------
1563 : leunga 731 Name: Allen Leung
1564 :     Date: 2000/11/10 18:00:00
1565 :     Tag: leunga-20001110-new-x86-fp
1566 :    
1567 :     A new x86 floating point code generator has been added.
1568 :     By default this is turned off. To turn this on, do:
1569 :    
1570 :     CM.autoload "$smlnj/compiler.cm";
1571 :     Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
1572 :    
1573 :     Changes:
1574 :    
1575 :     1. Changed FTAN to FPTAN so that the assembly output is correct.
1576 :     2. Changed the extension callback for FTANGENT to generate:
1577 :    
1578 :     fptan
1579 :     fstp %st(0)
1580 :     instead of
1581 :     fptan
1582 :     fstpl ftempmem
1583 :    
1584 :     3. Numerous assembly fixes for x86.
1585 :    
1586 :     5. Cleaned up the machine code output module x86/x86MC.sml and added
1587 :     support for a whole bunch of instructions and addressing modes:
1588 :    
1589 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st, %st(n)
1590 :     faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp %st, %st(n)
1591 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st(n), %st
1592 :     fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
1593 :     fxch %st(n)
1594 :     fld %st(n)
1595 :     fst %st(n)
1596 :     fst mem
1597 :     fstp %st(n)
1598 :     fucom %st(n)
1599 :     fucomp %st(n)
1600 :    
1601 :     All these are now generated when the fast fp mode is turned on.
1602 :    
1603 :     6. Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
1604 :    
1605 :     ----------------------------------------------------------------------
1606 : blume 729 Name: Matthias Blume
1607 :     Date: 2000/11/09 11:20:00 JST
1608 :     Tag: blume-20001109-scc
1609 :     Description:
1610 :    
1611 :     Eliminated some code duplication:
1612 :    
1613 :     1. Added "where" clause to GraphSCCFn in SML/NJ Library.
1614 :     (Otherwise the functor is useless.)
1615 :     2. Used GraphSCCFn where SCCUtilFun was used previously.
1616 :     3. Got rid of SCCUtilFun (in comp-lib.cm).
1617 :    
1618 :     ----------------------------------------------------------------------
1619 : george 721 Name: Lal George
1620 :     Date: 2000/11/06 09:02:21 EST 2000
1621 :     Tag: Release_110_30
1622 :     Description:
1623 :    
1624 :     - Version 110.30
1625 :     ----------------------------------------------------------------------
1626 : blume 715 Name: Matthias Blume
1627 : blume 716 Date: 2000/11/04 14:45:00
1628 :     Tag: blume-20001104-mlbuild
1629 :     Description:
1630 :    
1631 :     - Made ml-build faster on startup.
1632 :     - Documentation fixes.
1633 :    
1634 :     ----------------------------------------------------------------------
1635 :     Name: Matthias Blume
1636 : blume 715 Date: 2000/11/02 17:00:00 JST
1637 :     Tag: blume-20001102-condcomp
1638 :     Description:
1639 :    
1640 :     - Small tweaks to pickler -- new BOOTFILES!
1641 :     - Version bumped to 110.29.2.
1642 :     - Added conditional compilation facility to init.cmi (see comment there).
1643 :     ----------------------------------------------------------------------
1644 : leunga 713 Name: Allen Leung
1645 :     Date: 2000/10/23 19:31:00
1646 :     Tag: leunga-20001023-demo-ra
1647 :    
1648 :     1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
1649 :     2. Test programs for the graph library updated
1650 :     3. Some new MLRISC demo programs added
1651 :    
1652 :     ----------------------------------------------------------------------
1653 : blume 710 Name: Matthias Blume
1654 : blume 711 Date: 2000/08/31 22:15:00 JST
1655 :     Tag: blume-20001017-errmsg
1656 :     Description:
1657 :    
1658 :     More error message grief: Where there used to be no messages, there
1659 :     now were some that had bogus error regions. Fixed.
1660 :    
1661 :     ----------------------------------------------------------------------
1662 :     Name: Matthias Blume
1663 : blume 710 Date: 2000/08/31 17:30:00 JST
1664 :     Tag: blume-20001017-v110p29p1
1665 :     Description:
1666 :    
1667 :     I made a version 110.29.1 with new bootfiles.
1668 :    
1669 :     Changes: Modified pickler/unpickler for faster and leaner unpickling.
1670 :     CM documentation changes and a small bugfix in CM's error reporting.
1671 :    
1672 :     ----------------------------------------------------------------------
1673 : george 705 Name: Lal George
1674 :     Date: 2000/09/27 14:42:35 EDT
1675 :     Tag: george-20000927-nodestatus
1676 :     Description:
1677 :    
1678 :     Changed the type of the nodestatus, so that:
1679 :    
1680 :     SPILLED(~1) is now SPILLED
1681 :     SPILLED(m) where m>=0 is now MEMREG(m)
1682 :     SPILLED(s) where s<~1 is now SPILL_LOC(~s)
1683 :    
1684 :     ----------------------------------------------------------------------
1685 : blume 697 Name: Matthias Blume
1686 : blume 703 Date: 2000/09/07 14:45:00 JST
1687 :     Tag: blume-20000907-cmerrmsg
1688 :     Description:
1689 :    
1690 :     Small tweak to CM to avoid getting ML syntax error messages twice.
1691 :    
1692 :     ----------------------------------------------------------------------
1693 :     Name: Matthias Blume
1694 : blume 701 Date: 2000/08/31 18:00:00 JST
1695 :     Tag: blume-20000831-cvsbootfiles
1696 :     Description:
1697 :    
1698 :     New URL for boot files (because the 110.29 files on the BL server do
1699 :     now work correctly with my updated install scripts for yacc and lex).
1700 :    
1701 :     ----------------------------------------------------------------------
1702 :     Name: Matthias Blume
1703 : blume 697 Date: 2000/08/08 12:33:00 JST
1704 :     Tag: blume-20000808-manual
1705 :     Description:
1706 :    
1707 :     Tiny update to CM manual.
1708 :    
1709 :     ----------------------------------------------------------------------
1710 : leunga 695 Name: Allen Leung
1711 :     Date: 2000/08/7 19:31:00
1712 :     Tag: leunga-20000807-a-whole-bunch-of-stuff
1713 :    
1714 :     Moby, C--, SSA, x86, machine descriptions etc. Should only affect C--
1715 :     and Mobdy.
1716 :    
1717 :     1. x86
1718 :    
1719 :     a. Fixes to peephole module by John and Dan.
1720 :     b. Assembly fix to SETcc by Allen.
1721 :     c. Fix to c-call by John.
1722 :     d. Fix to spilling by John. (This one deals with the missing FSTPT case)
1723 :     e. Instruction selection optimization to SETcc as suggested by John.
1724 :    
1725 :     For example,
1726 :    
1727 :     MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
1728 :    
1729 :     should generate:
1730 :    
1731 :     MOVL a, x
1732 :     SUBL b, x
1733 :     SHRL 31, x
1734 :    
1735 :     2. IR stuff
1736 :    
1737 :     A bunch of new DJ-graph related algorithms added. These
1738 :     speed up SSA construction.
1739 :    
1740 :     3. SSA + Scheduling
1741 :    
1742 :     Added code for SSA and scheduling to the repository
1743 :    
1744 :     ----------------------------------------------------------------------
1745 : george 694 Name: Lal George
1746 :     Date: 2000/07/27 11:53:14 EDT
1747 :    
1748 :     Tag: lal-20000727-linux-ppc
1749 :     Description:
1750 :    
1751 :     Made changes to support Linux PPC.
1752 :     p.s. I have confirmation that the 110.29 boot files work fine.
1753 :    
1754 :     ----------------------------------------------------------------------
1755 : blume 684 Name: Matthias Blume
1756 : blume 692 Date: 2000/07/27 17:40:00 JST
1757 :     Tag: blume-20000727-scripts
1758 :     Description:
1759 :    
1760 :     !!!! WARNING !!!!
1761 :     You must recompile the runtime system!
1762 :     !!!! WARNING !!!!
1763 :    
1764 :     This is basically another round of script-enhancements:
1765 :    
1766 :     1. sml, ml-build, and ml-makedepend accept options -D and -U to define
1767 :     and undefine CM preprocessor symbols.
1768 :    
1769 :     2. ml-build avoids generating a new heap image if it finds that the
1770 :     existing one is still ok. (The condition is that no ML file had to
1771 :     be recompiled and all ML files are found to be older that the heap
1772 :     file.)
1773 :    
1774 :     To make this work smoothly, I also hacked the runtime system as
1775 :     well as SMLofNJ.SysInfo to get access to the heap image suffix
1776 :     (.sparc-solaris, ...) that is currently being used.
1777 :    
1778 :     Moreover, the signature of CM.mk_standalone has changed. See the
1779 :     CM manual.
1780 :    
1781 :     3. ml-makedepend accepts additional options -n, -a, and -o. (See the
1782 :     CM manual for details.)
1783 :    
1784 :     4. More CM manual updates:
1785 :     - all of the above has been documented.
1786 :     - there is now a section describing the (CM-related) command line
1787 :     arguments that are accepted by the "sml" command
1788 :    
1789 :     ----------------------------------------------------------------------
1790 :     Name: Matthias Blume
1791 : blume 691 Date: 2000/07/25 16:20:00 JST
1792 :     Tag: blume-20000725-makedepend
1793 :     Description:
1794 :    
1795 :     Added a script called ml-makedepend. This can be used in makefiles
1796 :     for Unix' make in a way very similar to the "makedepend" command for
1797 :     C.
1798 :    
1799 :     The script internally uses function CM.sources.
1800 :    
1801 :     Synopsis:
1802 :    
1803 :     ml-makedepend [-f makefile] cmfile targetname
1804 :    
1805 :     The default for the makefile is "makefile" (or "Makefile" should
1806 :     "makefile" not exist).
1807 :    
1808 :     ml-makedepend adds a cmfile/targetname-specific section to this
1809 :     makefile (after removing the previous version of this section). The
1810 :     section contains a single dependency specification with targetname on
1811 :     the LHS (targetname is an arbitrary name), and a list of files derived
1812 :     from the cmfile on the RHS. Some of the files on the RHS are
1813 :     ARCH/OPSYS-specific. Therefore, ml-makedepend inserts references to
1814 :     "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
1815 :     path names. The makefile writer is responsible for making sure that
1816 :     these variables have correct at the time "make" is invoked.
1817 :    
1818 :     ----------------------------------------------------------------------
1819 :     Name: Matthias Blume
1820 : blume 690 Date: 2000/07/22 23:30:00 JST
1821 :     Tag: blume-20000722-urlupdate
1822 :     Description:
1823 :    
1824 :     Changed BOOT and config/srcarchiveurl to point to BL server:
1825 :    
1826 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
1827 :    
1828 :     ----------------------------------------------------------------------
1829 :     Name: Matthias Blume
1830 : blume 686 Date: 2000/07/18 18:00:00 JST
1831 :     Tag: blume-20000718-Version_110_29
1832 :     Description:
1833 :    
1834 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
1835 :    
1836 :     2. Updated config/version to 110.29
1837 :    
1838 :     3. Updated config/srcarchiveurl
1839 :    
1840 :     3. New boot files!
1841 :     ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
1842 :    
1843 :     ----------------------------------------------------------------------
1844 :     Name: Matthias Blume
1845 : blume 684 Date: 2000/07/11 13:58:00 JST
1846 :     Tag: blume-20000711-doctypo
1847 :     Description:
1848 :    
1849 :     Fixed a few typos in CM manual.
1850 :    
1851 :     ----------------------------------------------------------------------
1852 : leunga 683 Name: Allen Leung
1853 :     Date: 2000/06/15 00:38:00
1854 :     Tag: leunga-20000704-sparc-x86
1855 :    
1856 :     1. x86 peephole improvement sp += k; sp -= k => nop [from John]
1857 :     2. fix to x86 RET bug [found by Dan Grossman]
1858 :     3. sparc assembly bug fix for ticc instructions [found by Fermin]
1859 :    
1860 :     Affects c-- and moby only
1861 :    
1862 :     ----------------------------------------------------------------------
1863 : blume 680 Name: Matthias Blume
1864 : blume 682 Date: 2000/07/04 15:26:00
1865 :     Tag: blume-20000704-trigger
1866 :     Description:
1867 :    
1868 :     1. Improvements to CM manual.
1869 :     2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
1870 :     of getting a back-trace. The function, when called, raises an
1871 :     internal exception which explicitly carries the full back-trace history,
1872 :     so it is unaffected by any intervening handle-raise pairs ("trivial"
1873 :     or not). The interactive loop will print that history once it arrives
1874 :     at top level.
1875 :     Short of having all exceptions implicitly carry the full history, the
1876 :     recommended way of using this facility is:
1877 :     - compile your program with instrumentation "on"
1878 :     - run it, when it raises an exception, look at the history
1879 :     - if the history is "cut off" because of some handler, go and modify
1880 :     your program so that it explicitly calls BTrace.trigger
1881 :     - recompile (still instrumented), and rerun; look at the full history
1882 :    
1883 :     ----------------------------------------------------------------------
1884 :     Name: Matthias Blume
1885 :     Date: 2000/07/03 15:36:00 JST
1886 : blume 680 Tag: blume-20000702-manual
1887 :     Description:
1888 :    
1889 :     Small corrections and updates to CM manual.
1890 :    
1891 :     ----------------------------------------------------------------------
1892 : blume 682 Name: Matthias Blume
1893 : blume 679 Date: 2000/06/29 16:04:00 JST
1894 :     Tag: blume-20000629-yacctool
1895 :     Description:
1896 :    
1897 :     Changes:
1898 :    
1899 :     1. Class "mlyacc" now takes separate arguments to pass options to
1900 :     generated .sml- and .sig-files independently.
1901 :     2. Corresponding CM manual updates.
1902 :     3. BTrace module now also reports call sites. (However, for loop clusters
1903 :     it only shows from where the cluster was entered.) There are associated
1904 :     modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1905 :    
1906 :     ----------------------------------------------------------------------
1907 : blume 675 Name: Matthias Blume
1908 : blume 678 Date: 2000/06/27 16:51:00 JST
1909 :     Tag: blume-20000627-noweb
1910 :     Description:
1911 :    
1912 :     Changes:
1913 :    
1914 :     1. Implemented "subdir" and "witness" options for noweb tool.
1915 :     This caused some slight internal changes in CM's tool implementation.
1916 :     2. Fixed bug in "tool plugin" mechanism. This is essentially cleaning
1917 :     some remaining issues from earlier path anchor changes.
1918 :     3. Updated CM manual accordingly.
1919 :    
1920 :     4. Changed implementation of back-tracing so that I now consider it
1921 :     ready for prime-time.
1922 :    
1923 :     In particular, you don't have to explicitly trigger the back-trace
1924 :     anymore. Instead, if you are running BTrace-instrumented code and
1925 :     there is an uncaught exception (regardless of whether or not it was
1926 :     raised in instrumented code), the top-level evalloop will print
1927 :     the back-trace.
1928 :    
1929 :     Features:
1930 :    
1931 :     - Instrumented and uninstrumented code work together seemlessly.
1932 :     (Of course, uninstrumented code is never mentioned in actual
1933 :     back-traces.)
1934 :    
1935 :     - Asymptotic time- and space-complexity of instrumented code is
1936 :     equal to that of uninstrumented code. (This means that
1937 :     tail-recursion is preserved by the instrumentation phase.)
1938 :    
1939 :     - Modules whose code has been instrumented in different sessions
1940 :     work together without problem.
1941 :    
1942 :     - There is no penalty whatsoever on uninstrumented code.
1943 :    
1944 :     - There is no penalty on "raise" expressions, even in
1945 :     instrumented code.
1946 :    
1947 :     A potential bug (or perhaps it is a feature, too):
1948 :    
1949 :     A back-trace reaches no further than the outermost instrumented
1950 :     non-trivial "raise". Here, a "trivial" raise is one that is the
1951 :     sole RHS of a "handle" rule. Thus, back-traces reach trough
1952 :    
1953 :     <exp> handle e => raise e
1954 :    
1955 :     and even
1956 :    
1957 :     <exp> handle Foo => raise Bar
1958 :    
1959 :     and, of course, through
1960 :    
1961 :     <exp> handle Foo => ...
1962 :    
1963 :     if the exception was not Foo.
1964 :    
1965 :     Back-traces always reach right through any un-instrumented code
1966 :     including any of its "handle" expressions, trivial or not.
1967 :    
1968 :     To try this out, do the following:
1969 :    
1970 :     - Erase all existing binfiles for your program.
1971 :     (You may keep binfiles for those modules where you think you
1972 :     definitely don't need back-tracing.)
1973 :     - Turn on back-trace instrumentation:
1974 :     SMLofNJ.Internals.BTrace.mode (SOME true);
1975 :     - Recompile your program. (I.e., run "CM.make" or "use".)
1976 :     - You may now turn instrumentation off again (if you want):
1977 :     SMLofNJ.Internals.BTrace.mode (SOME false);
1978 :     - Run your program as usual. If it raises an exception that
1979 :     reaches the interactive toplevel, then a back-trace will
1980 :     automatically be printed. After that, the toplevel loop
1981 :     will print the exception history as usual.
1982 :    
1983 :     ----------------------------------------------------------------------
1984 :     Name: Matthias Blume
1985 : blume 677 Date: 2000/06/26 09:56:46 JST
1986 :     Tag: blume-20000626-setup
1987 :     Description:
1988 :    
1989 :     CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1990 :     ML code before and after compiling a file (e.g., to set compiler
1991 :     flags)
1992 :    
1993 :     Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1994 :     - associated changes to btrace.sml (BTrace instrumentation pass)
1995 :     - cleaner implementation of btimp.sml (BTrace tracing and report
1996 :     module)
1997 :    
1998 :     CM manual: * new path encoding documented
1999 :     * description of setup-parameter to "sml" added
2000 :    
2001 :     The biggest user-visible change to back-tracing is that it is no
2002 :     longer necessary to compile all traced modules within the same
2003 :     session. (This was a real limitation.)
2004 :    
2005 :     ----------------------------------------------------------------------
2006 :     Name: Matthias Blume
2007 : blume 676 Date: 2000/06/24 12:40:00 JST
2008 :     Tag: blume-20000624-startup
2009 :     Description:
2010 :    
2011 :     Fixes startup slowdown problem. (I was calling SrcPath.sync a _tad_
2012 :     bit too often -- to put it mildly. :)
2013 :    
2014 :     ----------------------------------------------------------------------
2015 :     Name: Matthias Blume
2016 : blume 675 Date: 2000/06/23 18:20:00 JST
2017 :     Tag: blume-20000623-btrace
2018 :     Description:
2019 :    
2020 :     This updates adds a backtrace facility to aid programmers in debugging
2021 :     their programs. This involves the following changes:
2022 :    
2023 :     1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
2024 :     keeping track of the current call stack. When programs are compiled
2025 :     in a special mode, the compiler will insert calls to these hooks
2026 :     into the user program.
2027 :     "Hook" means that it is possible for different implementations of
2028 :     back-tracing to register themselves (at different times).
2029 :    
2030 :     2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
2031 :     as an Absyn.dec->Absyn.dec rewrite. Normally this phase is turned off.
2032 :     It can be turned on using this call:
2033 :     SMLofNJ.Internals.BTrace.mode (SOME true);
2034 :     Turning it off again:
2035 :     SMLofNJ.Internals.BTrace.mode (SOME false);
2036 :     Querying the current status:
2037 :     SMLofNJ.Internals.BTrace.mode NONE;
2038 :     Annotated programs are about twice as big as normal ones, and they
2039 :     run a factor of 2 to 4 slower with a dummy back-trace plugin (one
2040 :     where all hooks do nothing). The slowdown with a plugin that is
2041 :     actually useful (such as the one supplied by default) is even greater,
2042 :     but in the case of the default plugin it is still only an constant
2043 :     factor (amortized).
2044 :    
2045 :     3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
2046 :     with a sub-structure BTrace for controlling back-tracing. In particular,
2047 :     the above-mentioned function "mode" controls whether the annotation
2048 :     phase is invoked by the compiler. Another important function is
2049 :     "trigger": when called it aborts the current execution and causes
2050 :     the top-level loop to print a full back-trace.
2051 :    
2052 :     4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
2053 :     for back-tracing. It keeps track of the dynamic call stack and in
2054 :     addition to that it keeps a partial history at each "level" of that
2055 :     stack. For example, if a tail-calls b, b tail-calls c, and c tail-calls
2056 :     d and b (at separate times, dynamically), then the report will show:
2057 :    
2058 :     GOTO d
2059 :     /c
2060 :     GOTO \b
2061 :     CALL a
2062 :    
2063 :     This shows that there was an initial non-tail call of a, then a
2064 :     tail-call to b or c, looping behavior in a cluster of functions that
2065 :     consist of b and c, and then a goto from that cluster (i.e., either from
2066 :     b or from c) to d.
2067 :    
2068 :     Note that (depending on the user program) the amount of information
2069 :     that the back-trace module has to keep track of at each level is bounded
2070 :     by a constant. Thus, the whole implementation has the same asymptotical
2071 :     complexity as the original program (both in space and in time).
2072 :    
2073 :     5. compiler/TopLevel/interact/evalloop.sml has been modified to
2074 :     handle the special exception SMLofNJ.Internals.BTrace.BTrace
2075 :     which is raised by the "trigger" function mentioned above.
2076 :    
2077 :     Notes on usage:
2078 :    
2079 :     - Annotated code works well together with unannotated code:
2080 :     Unannotated calls simply do not show up at all in the backtrace.
2081 :    
2082 :     - It is not a good idea to let modules that were annotated during
2083 :     different sessions run at the same time. This is because the compiler
2084 :     chooses small integers to identify individual functions, and there
2085 :     will be clashes if different modules were compiled in separate sessions.
2086 :     (Nothing will crash, and you will even be told about the clashes, but
2087 :     back-trace information will in general not be useful.)
2088 :    
2089 :     - Back-tracing can be confused by callcc and capture.
2090 :    
2091 :     - The only way of getting a back-trace right now is to explicitly
2092 :     invoke the "trigger" function from your user program. Eventually, we
2093 :     should make every exception carry back-trace information (if
2094 :     available). But since this creates more overhead at "raise"-time
2095 :     (similar to the current exnHistory overhead), I have not yet
2096 :     implemented this. (The implementation will be rather easy.) With
2097 :     exceptions carrying back-trace information, this facility will be even
2098 :     more useful because users don't need to modify their programs...
2099 :    
2100 :     - While it is possible to compile the compiler with back-trace
2101 :     annotations turned on (I did it to get some confidence in
2102 :     correctness), you must make absolutely sure that core.sml and
2103 :     btimp.sml are compiled WITHOUT annotation! (core.sml cannot actually
2104 :     be compiled with annotation because there is no core access yet, but
2105 :     if you compile btimp.sml with annotation, then the system will go into
2106 :     an infinite recursion and crash.)
2107 :     Since CM currently does not know about BTrace, the only way to turn
2108 :     annotations on and off for different modules of the compiler is to
2109 :     interrupt CMB.make, change the settings, and re-invoke it. Of course,
2110 :     this is awkward and clumsy.
2111 :    
2112 :     Sample sessions:
2113 :    
2114 :     Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2115 :     - SMLofNJ.Internals.BTrace.mode (SOME true);
2116 :     [autoloading]
2117 :     [autoloading done]
2118 :     val it = false : bool
2119 :     - structure X = struct
2120 :     - fun main n = let
2121 :     - fun a (x, 0) = d x
2122 :     - | a (x, n) = b (x, n - 1)
2123 :     - and b (x, n) = c (x, n)
2124 :     - and c (x, n) = a (x, n)
2125 :     - and d x = e (x, 3)
2126 :     - and e (x, 0) = f x
2127 :     - | e (x, n) = e (x, n - 1)
2128 :     - and f 0 = SMLofNJ.Internals.BTrace.trigger ()
2129 :     - | f n = n * g (n - 1)
2130 :     - and g n = a (n, 3)
2131 :     - in
2132 :     - f n
2133 :     - end
2134 :     - end;
2135 :     structure X : sig val main : int -> int end
2136 :     - X.main 3;
2137 :     *** BACK-TRACE ***
2138 :     GOTO stdIn:4.2-13.20: X.main[2].f
2139 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
2140 :     GOTO stdIn:4.2-13.20: X.main[2].d
2141 :     / stdIn:4.2-13.20: X.main[2].a
2142 :     | stdIn:4.2-13.20: X.main[2].b
2143 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
2144 :     CALL stdIn:4.2-13.20: X.main[2].g
2145 :     GOTO stdIn:4.2-13.20: X.main[2].f
2146 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
2147 :     GOTO stdIn:4.2-13.20: X.main[2].d
2148 :     / stdIn:4.2-13.20: X.main[2].a
2149 :     | stdIn:4.2-13.20: X.main[2].b
2150 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
2151 :     CALL stdIn:4.2-13.20: X.main[2].g
2152 :     GOTO stdIn:4.2-13.20: X.main[2].f
2153 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
2154 :     GOTO stdIn:4.2-13.20: X.main[2].d
2155 :     / stdIn:4.2-13.20: X.main[2].a
2156 :     | stdIn:4.2-13.20: X.main[2].b
2157 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
2158 :     CALL stdIn:4.2-13.20: X.main[2].g
2159 :     GOTO stdIn:4.2-13.20: X.main[2].f
2160 :     CALL stdIn:2.15-17.4: X.main[2]
2161 :     -
2162 :    
2163 :     (Note that because of a FLINt bug the above code currently does not
2164 :     compile without BTrace turned on.)
2165 :    
2166 :     Here is another example, using my modified Tiger compiler:
2167 :    
2168 :     Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2169 :     - SMLofNJ.Internals.BTrace.mode (SOME true);
2170 :     [autoloading]
2171 :     [autoloading done]
2172 :     val it = false : bool
2173 :     - CM.make "sources.cm";
2174 :     [autoloading]
2175 :     ...
2176 :     [autoloading done]
2177 :     [scanning sources.cm]
2178 :     [parsing (sources.cm):parse.sml]
2179 :     [creating directory CM/SKEL ...]
2180 :     [parsing (sources.cm):tiger.lex.sml]
2181 :     ...
2182 :     [wrote CM/sparc-unix/semant.sml]
2183 :     [compiling (sources.cm):main.sml]
2184 :     [wrote CM/sparc-unix/main.sml]
2185 :     [New bindings added.]
2186 :     val it = true : bool
2187 :     - Main.compile ("../testcases/merge.tig", "foo.out");
2188 :     *** BACK-TRACE ***
2189 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2190 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2191 :     CALL lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2192 :     GOTO lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2193 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2194 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2195 :     CALL lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2196 :     / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2197 :     CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2198 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2199 :     CALL lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2200 :     CALL main.sml:1.18-118.4: Main.compile[2]
2201 :     -
2202 :    
2203 :     ----------------------------------------------------------------------
2204 : blume 673 Name: Matthias Blumen
2205 :     Date: 2000/06/21 18:00:00 JST
2206 :     Tag: blume-20000621-manual
2207 :     Description:
2208 :    
2209 :     CM manual update: Path environments documented.
2210 :    
2211 :     ----------------------------------------------------------------------
2212 : blume 666 Name: Matthias Blume
2213 : blume 672 Date: 2000/06/19 13:40:00
2214 :     Tag: blume-20000619-manual
2215 :     Description:
2216 :    
2217 :     CM manual and system/README update. This only covers the fact that
2218 :     there are no more implicit anchors. (Path environments and the "bind"
2219 :     option to "cm" have yet to be documented.)
2220 :    
2221 :     ----------------------------------------------------------------------
2222 :     Name: Matthias Blume
2223 : blume 671 Date: 2000/06/19 11:05:00 JST
2224 :     Tag: blume-20000619-chdir-bugfix
2225 :     Description:
2226 :    
2227 :     Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2228 :    
2229 :     ----------------------------------------------------------------------
2230 :     Name: Matthias Blume
2231 : blume 670 Date: 2000/06/18 22:00:10 JST
2232 :     Tag: blume-20000618-implicit-anchors-really-gone
2233 :     Description:
2234 :    
2235 :     I updates the previous HISTORY entry where I forgot to mention that
2236 :     implicit anchors are no longer with us.
2237 :    
2238 :     The current update also gets rid of the (now useless) controller
2239 :     CM.Control.implicit_anchors.
2240 :    
2241 :     ----------------------------------------------------------------------
2242 :     Name: Matthias Blume
2243 : blume 666 Date: 2000/06/16 17:30:00 JST
2244 :     Tag: blume-20000616-anchorenv
2245 :     Description:
2246 :    
2247 :     This patch implements the long anticipated (just kidding :) "anchor
2248 :     environment" mechanism. In the course of doing this, I also
2249 :     re-implemented CM's internal "SrcPath" module from scratch. The new
2250 :     one should be more robust in certain boundary cases. In any case, it
2251 :     is a lot cleaner than its predecessor (IMHO).
2252 :    
2253 :     This time, although there is yet another boot file format change, I
2254 :     kept the unpickler backward-compatible. As a result, no new bootfiles
2255 :     are necessary and bootstrapping is straightforward. (You cannot read
2256 :     new bootfiles into an old system, but the other way around is no
2257 :     problem.)
2258 :    
2259 :     Visible changes:
2260 :    
2261 : blume 670 ** 0. Implicit path anchors (without the leading $-symbol) are no
2262 :     longer recognized at all. This means that such path names are not
2263 :     illegal either. For example, the name basis.cm simply refers to a
2264 :     local file called "basis.cm" (i.e, the name is an ordinary path
2265 :     relative to .cm-files directory). Or, to put it differently, only
2266 :     names that start with $ are anchored paths.
2267 : blume 666
2268 : blume 670 ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
2269 :     vanished.
2270 :    
2271 :     John (Reppy) had critizised this as soon as I originally proposed and
2272 : blume 666 implemented it, but at that time I did not really deeply believe
2273 :     him. :) Now I came full-circle because I need the $<singlearc> syntax
2274 :     in another place where it cannot be seen as an abbreviation for
2275 :     $/<singlearc>. To avoid the confusion, $<singlearc> now means what it
2276 :     seems to mean (i.e., it "expands" into the corresponding anchor
2277 :     value).
2278 :    
2279 :     However, when paths are used as members in CM description files, it
2280 :     continues to be true that there must be at least another arc after the
2281 :     anchor. This is now enforced separately during semantic analysis
2282 : blume 670 (i.e., from a lexical/syntactical point of view, the notation is ok.)
2283 : blume 666
2284 :     ** 2. The "cm" class now accepts an option "bind". The option's value
2285 :     is a sub-option list of precisely two items -- one labeled "anchor"
2286 :     and the other one labeled "value". As you might expect, "anchor" is
2287 :     used to specify an anchor name to be bound, and "value" specifies what
2288 :     the anchor is being bound to.
2289 :    
2290 :     The value must be a directory name and can be given in either standard
2291 :     syntax (including the possibility that it is itself an anchored path)
2292 :     or native syntax.
2293 :    
2294 :     Examples:
2295 :    
2296 :     foo.cm (bind:(anchor:bar value:$mystuff/bar))
2297 :     lib.cm (bind:(anchor:a value:"H:\\x\\y\\z")) (* only works under windows *)
2298 :    
2299 :     and so on.
2300 :    
2301 :     The meaning of this is that the .cm-file will be processed with an
2302 :     augmented anchor environment where the given anchor(s) is/are bound to
2303 :     the given values(s).
2304 :    
2305 :     The rationale for having this feature is this: Suppose you are trying
2306 :     to use two different (already stable) libraries a.cm and b.cm (that
2307 :     you perhaps didn't write yourself). Further, suppose each of these
2308 :     two libraries internally uses its own auxiliary library $aux/lib.cm.
2309 :     Normally you would now have a problem because the anchor "lib" can not
2310 :     be bound to more than one value globally. Therefore, the project that
2311 :     uses both a.cm and b.cm must locally redirect the anchor to some other
2312 :     place:
2313 :    
2314 :     a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
2315 :     b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
2316 :    
2317 :     This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
2318 :     /usr/lib/smlnj/b-stuff/aux.cm, respectively.
2319 :    
2320 :     Hard-wiring path names is a bit inflexible (and CM will verbosely warn
2321 :     you when you do so at the time of CM.stabilize). Therefore, you can
2322 :     also use an anchored path as the value:
2323 :    
2324 :     a.cm (bind:(anchor:lib value:$a-lib))
2325 :     b.cm (bind:(anchor:lib value:$b-lib))
2326 :    
2327 :     Now you can globally configure (using the usual CM.Anchor.anchor or
2328 :     pathconfig machinery) bindings for "a-lib" and "b-lib". Since "lib"
2329 :     itself is always locally bound, setting it globally is no longer
2330 :     meaningful or necessary (but it does not hurt either). In fact, "lib"
2331 :     can still be used as a global anchor for separate purposes. As a
2332 :     matter of fact, one can locally define "lib" in terms of a global
2333 :     "lib":
2334 :    
2335 :     a.cm (bind:(anchor:lib value:$lib/a))
2336 :     b.cm (bind:(anchor:lib value:$lib/b))
2337 :    
2338 :     ** 3: The encoding of path names has changed. This affects the way
2339 :     path names are shown in CM's progress report and also the internal
2340 :     protocol encoding used for parallel make.
2341 :    
2342 :     The encoding now uses one or more ':'-separated segments. Each
2343 :     segments corresponds to a file that has been specified relative to the
2344 :     file given by its preceding segment. The first segment is either
2345 :     relative to the CWD, absolute, or anchored. Each segment itself is
2346 :     basically a Unix pathname; all segments but the first are relative.
2347 :    
2348 :     Example:
2349 :    
2350 :     $foo/bar/baz.cm:a/b/c.sml
2351 :    
2352 :     This path denotes the file bar/a/b/c.sml relative to the directory
2353 :     denoted by anchor "foo". Notice that the encoding also includes
2354 :     baz.cm which is the .cm-file that listed a/b/c.sml. As usual, such
2355 :     paths are resolved relative to the .cm-files directory, so baz.cm must
2356 :     be ignored to get the "real" pathname.
2357 :    
2358 :     To make this fact more obvious, CM puts the names of such "virtual
2359 :     arcs" into parentheses when they appear in progress reports. (No
2360 :     parentheses will appear in the internal protocol encoding.) Thus,
2361 :     what you really see is:
2362 :    
2363 :     $foo/bar/(baz.cm):a/b/c.sml
2364 :    
2365 :     I find this notation to be much more informative than before.
2366 :    
2367 :     Another new feature of the encoding is that special characters
2368 :     including parentheses, colons, (back)slashes, and white space are
2369 :     written as \ddd (where ddd is the decimal encoding of the character).
2370 :    
2371 : blume 670 *** The CM manual still needs to be updated.
2372 :    
2373 : blume 666 ----------------------------------------------------------------------
2374 : leunga 660 Name: Allen Leung
2375 :     Date: 2000/06/15 00:38:00
2376 :     Tag: leunga-20000615-x86-peephole
2377 :    
2378 :     x86 Peephole fix by Fermin. Affects c-- and moby only.
2379 :    
2380 :     ----------------------------------------------------------------------
2381 : blume 658 Name: Matthias Blume
2382 :     Date: 2000/06/12 11:40:00
2383 :     Tag: blume-20000612-parmakefix
2384 :     Description:
2385 :    
2386 :     More cleanup after changing the file naming scheme: This time I
2387 :     repaired the parallel make mechanism for CMB.make which I broke earlier.
2388 :    
2389 :     ----------------------------------------------------------------------
2390 : leunga 657 Name: Allen Leung
2391 :     Date: 2000/06/09 01:25:00
2392 :     Tag: leunga-20000609-various
2393 :    
2394 :     None of these things should affect normal SML/NJ operations
2395 :    
2396 :     1. Peephole improvements provided by Fermin (c--)
2397 :     2. New annotation DEFUSE for adding extra dependence (moby)
2398 :     3. New X86 LOCK instructions (moby)
2399 :     4. New machine description language for reservation tables (scheduling)
2400 :     5. Fixes to various optimization/analysis modules (branch chaining, dominator
2401 :     trees etc.)
2402 :     6. I've changed the CM files so that they can work with versions
2403 :     110.0.6, 110.25 and 110.28
2404 :    
2405 :     ----------------------------------------------------------------------
2406 : blume 652 Name: Matthias Blume
2407 : blume 656 Date: 2000/06/09 12:40:00
2408 :     Tag: blume-20000609-log
2409 :     Description:
2410 :    
2411 :     - Removed all(?) remaining RCS Log entries from sources.
2412 :    
2413 :     - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
2414 :     anchored paths).
2415 :    
2416 :     ----------------------------------------------------------------------
2417 :     Name: Matthias Blume
2418 : blume 653 Date: 2000/06/07 17:00:00 JST
2419 :     Tag: blume-20000607-no-implicit-anchors
2420 :     Description:
2421 :    
2422 :     1. This update changes the default setting for
2423 :     CM.Control.implicit_anchors from true to false. This means that
2424 :     implicit anchors are no longer permitted by default. I also tried to
2425 :     make sure that nothing else still relies on implicit anchors.
2426 :     (This is the next step on the schedule towards a CM that does not even
2427 :     have the notion of implicit anchors anymore.)
2428 :    
2429 :     2. More CM manual updates.
2430 :    
2431 :     3. I managed to track down and fix the pickling bug I mentioned last
2432 :     time. Because of the previously existing workaround, this entails no
2433 :     immediate practical changes.
2434 :    
2435 :     ----------------------------------------------------------------------
2436 :     Name: Matthias Blume
2437 : blume 652 Date: 2000/06/06 11:15:00 JST
2438 :     Tag: blume-20000606-lazierpickle
2439 :     Description:
2440 :    
2441 :     !!!! NEW BOOT FILES !!!!
2442 :    
2443 :     * The main purpose of this update is to make library pickles lazier in
2444 :     order to reduce the initial space penalty for autoloading a library.
2445 :     As a result, it is now possible to have $smlnj/compiler.cm
2446 :     pre-registered. This should take care of the many complaints or
2447 :     inquiries about missing structure Compiler. This required changes to
2448 :     CM's internal data structures and small tweaks to some algorithms.
2449 :    
2450 :     As a neat additional effect, it is no longer necessary (for the sake
2451 :     of lean heap image files) to distinguish between a "minimal" CM and a
2452 :     "full" CM. Now, there is only one CM (i.e., the "full" version:
2453 :     $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
2454 :     interactive top level. ($smlnj/cm/minimal.cm is gone.)
2455 :    
2456 :     To make the life of compiler-hackers easier, "makeml" now also
2457 :     pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm). In other
2458 :     words, after you bootstrap a new sml for the first time, you will not
2459 :     have to autoload $smlnj/cmb.cm again afterwards. (The first time
2460 :     around you will still have to do it, though.)
2461 :    
2462 :     * A second change consists of major updates to the CM manual. There
2463 :     are now several appendices with summary information and also a full
2464 :     specification of the CM description file syntax.
2465 :    
2466 :     * In directory src/system I added the script "allcross". This script
2467 :     invokes sml and cross-compiles the compiler for all supported
2468 :     architectures. (Useful when providing a new set of boot files.)
2469 :    
2470 :     * There seems to be a latent bug in my "lazy pickles" mechanism. I
2471 :     added a small tweak to pickle-util.sml to work around this problem,
2472 :     but it is not a proper fix yet. I will investigate further. (The
2473 :     effect of the bug was an inflation of library pickle size.)
2474 :    
2475 :     * Version number increased to 110.28.1 (to avoid compatibility problems).
2476 :    
2477 :     ----------------------------------------------------------------------
2478 : leunga 646 Name: Allen Leung
2479 : leunga 648 Date: 2000/05/25 17:28 EDT
2480 :     Tag: leunga-20000525-ra
2481 :     Description:
2482 :    
2483 :     Fixed a bug in freezing phase of the register allocator.
2484 :    
2485 :     ----------------------------------------------------------------------
2486 :     Name: Allen Leung
2487 : leunga 647 Date: 2000/05/15 22:53 EDT
2488 : leunga 646 Tag: leunga-20000515-alpha-x86-ra
2489 :     Description:
2490 :    
2491 :     1. Alpha
2492 :    
2493 :     Slight cleanup. Removed the instruction SGNXL
2494 :    
2495 :     2. X86
2496 :    
2497 :     Added the following instructions to the instruction set:
2498 :    
2499 :     ROLx, RORx,
2500 :     BTx, BTSx, BTLx, BTRx,
2501 :     XCHGx, and variants with the LOCK prefix
2502 :    
2503 :     3. Register Allocation
2504 :    
2505 :     The module ra-rewrite-with-renaming has been improved.
2506 :    
2507 : leunga 647 These have no effect on SML/NJ.
2508 :    
2509 : leunga 646 ----------------------------------------------------------------------
2510 : blume 642 Name: Matthias Blume
2511 : blume 645 Date: 2000/05/15 16:20:00 JST
2512 :     Tag: blume-20000515-lightrebuild
2513 :     Description:
2514 :    
2515 :     1. I added an alternative to "-rebuild" to "makeml". The difference is
2516 :     that prior to calling CMB.make' the CM-variable "LIGHT" will be
2517 :     defined. In effect, the command will not build any cross-compiler
2518 :     backends and therefore finish more quickly.
2519 :    
2520 :     The "fixpt" script also takes a "-light" switch to be able to use
2521 :     this new facility while compiling for a fixpoint.
2522 :    
2523 :     2. I replaced all mentions of anchored paths in group owner specifications
2524 :     with simple relative paths (usually starting with "..").
2525 :     The rationale is that a library's internal workings should not be
2526 :     compromised by the lack of some anchor. (An anchor is necessary
2527 :     for someone who wants to refer to the library by an anchored path,
2528 :     but it should not be necessary to build the same library in the first
2529 :     place.)
2530 :    
2531 :     3. I changed the way CM's tool mechanism determines the shell command
2532 :     string used for things like ml-yacc etc. so that it does not break
2533 :     when CM.Control.implicit_anchors is turned off.
2534 :    
2535 :     ----------------------------------------------------------------------
2536 :     Name: Matthias Blume
2537 : blume 644 Date: 2000/05/12 18:20:00 JST
2538 :     Tag: blume-20000512-ml-build
2539 :     Description:
2540 :    
2541 :     Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
2542 :     getting installed properly (by config/install.sh).
2543 :    
2544 :     ----------------------------------------------------------------------
2545 :     Name: Matthias Blume
2546 : blume 643 Date: 2000/05/12 17:30:00 JST
2547 :     Tag: blume-20000512-anchors
2548 :     Description:
2549 :    
2550 :     !!! NEW BOOT FILES !!!
2551 :    
2552 :     This change is in preparation of fading out support for "implicitly
2553 :     anchored path names". I went through all sources and used the
2554 :     explicit (and relatively new) $-notation. See system/README and the
2555 :     CM manual for more info on this.
2556 :    
2557 :     I also modified the anchoring scheme for some things such as "smlnj",
2558 :     "MLRISC", "cm", etc. to take advantage of the fact that explicit
2559 :     anchors are more expressive: anchor name and first arc do not have to
2560 :     coincide. This entails the following user-visible change:
2561 :    
2562 :     You have to write $smlnj/foo/bar instead of smlnj/foo/bar. In
2563 :     particular, when you fire up sml with a command-line argument, say,
2564 :     e.g.:
2565 :    
2566 :     sml '$smlnj/cmb.cm'
2567 :    
2568 :     At the ML toplevel prompt:
2569 :    
2570 :     CM.autoload "$smlnj/cmb.cm";
2571 :    
2572 :     There is also a new controller in CM.Control that can be used to turn
2573 :     off all remaining support for implicit anchors by saying:
2574 :    
2575 :     CM.autoload "$smlnj/
2576 :     #set CM.Control.implicit_anchors false;
2577 :    
2578 :     This causes CM to reject implicitly anchored paths. This is (for the
2579 :     time being) less permissive than the "final" version where there will
2580 :     be no more such implicit anchors and relative paths will be just that:
2581 :     relative.
2582 :    
2583 :     The next step (version after next version?) will be to make the
2584 :     default for CM.Control.implicit_anchors false. After the dust has
2585 :     settled, I can then produce the "final" version of this...
2586 :    
2587 :     Note: Since bootstrapping is a bit tricky, I provided new boot files.
2588 :    
2589 :     ----------------------------------------------------------------------
2590 :     Name: Matthias Blume
2591 : blume 642 Date: 2000/05/11 16:30:00 JST
2592 :     Tag: blume-20000511-sources
2593 :     Description:
2594 :    
2595 :     The main change is that I added function CM.sources as a generalized
2596 :     version of the earlier CM.makedepend. This entails the following
2597 :     additional changes:
2598 :    
2599 :     - CM.makedepend has been dropped.
2600 :    
2601 :     - CM manual has been updated.
2602 :    
2603 :     - TOOLS signature and API have been changed.
2604 :    
2605 :     ----------------------------------------------------------------------
2606 : leunga 640 Name: Allen Leung
2607 : leunga 641 Date: 2000/05/10 21:17 EDT
2608 :     Tag: leunga-20000510-moby-c--ssa
2609 :     Description:
2610 :    
2611 :     Various bug fixes and new features for C--, Moby and MLRISC optimizations.
2612 :     None of these affect SML/NJ.
2613 :    
2614 :     1. Register Allocation
2615 :    
2616 :     a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
2617 :     This module tries to remove local (i.e. basic block level) redundancies
2618 :     during spilling.
2619 :    
2620 :     b. A new framework for performing region based register allocation.
2621 :     Not yet entirely functional.
2622 :    
2623 :     2. X86
2624 :    
2625 :     a. DefUse for POP was missing the stack pointer [found by Lal]
2626 :     b. Reload for CALL was incorrect in X86Spill [found by John]
2627 :     c. Various fixes in X86Spill so that it can be used correctly for
2628 :     the new spilling module.
2629 :    
2630 :     3. SSA/IR
2631 :    
2632 :     a. New module ir/dj-dataflow.sml implements elimination based
2633 :     data flow analysis.
2634 :    
2635 :     4. MLRiscGen
2636 :    
2637 :     a. Fix for gc type annotation
2638 :    
2639 :     5. MDGen
2640 :    
2641 :     Various fixes for machine description -> ml code translation. For ssa
2642 :     only.
2643 :    
2644 :     ----------------------------------------------------------------------
2645 :     Name: Allen Leung
2646 : leunga 640 Date: 2000/05/08 22:17 EDT
2647 :     Tag: leunga-20000508-labexp
2648 :     Description:
2649 :    
2650 :     Fermin has found a few assembly problems with constant expressions
2651 :     generated in LabelExp. Mostly, the problems involve extra parentheses,
2652 :     which choke on dumb assemblers. This is his fix.
2653 :    
2654 :     ----------------------------------------------------------------------
2655 : dbm 635 Name: Dave MacQueen
2656 :     Date: 2000/04/09 14:00 EDT
2657 :     Tag: dbm-20000502-Version_110_28
2658 :     Description:
2659 :    
2660 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
2661 :    
2662 :     2. Updated config/version to 110.28
2663 :    
2664 :     3. Updated config/srcarchiveurl
2665 :    
2666 :     3. New boot files!
2667 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
2668 :    
2669 :     ----------------------------------------------------------------------
2670 : blume 634 Name: Matthias Blume
2671 :     Date: 2000/05/01 19:05:00 JST
2672 :     Tag: blume-20000501-noweb
2673 :     Description:
2674 :    
2675 :     A new noweb tool has been added. The existing system is entirely
2676 :     unaffected by this, but some CM users have asked for renewed noweb
2677 :     support. Everything is documented in the CM manual.
2678 :    
2679 :     New (plugin) libraries:
2680 :    
2681 :     noweb-tool.cm
2682 :     nw-ext.cm
2683 :    
2684 :     ----------------------------------------------------------------------
2685 : dbm 633 Name: Dave MacQueen
2686 :     Date: 2000/04/30 12:40PM EDT
2687 :     Tag: dbm-20000430-bug_fixes
2688 :     Description:
2689 :    
2690 :     1. Fix for bug 1498
2691 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sig
2692 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sml
2693 :     added toRealArray function
2694 :     smlnj/src/compiler/MiscUtil/print/ppobj.sml
2695 :     added check for tag Obj.RealArray to array printing case in ppObj
2696 :    
2697 :     2. Fix for bug 1510
2698 :     smlnj/src/compiler/Semant/types/typesutil.sml
2699 :     fixed definition of dummyargs (used by equalTycon) so that
2700 :     dummy args are distinct types
2701 :    
2702 :     ----------------------------------------------------------------------
2703 : blume 629 Name: Matthias Blume
2704 : blume 632 Date: 2000/04/30 01:00:00 JST
2705 :     Tag: blume-20000430-versions
2706 :     Description:
2707 :    
2708 :     1. CM version numbering added. This is an implementation of Lal's
2709 :     proposal for adding version numbers and version checking to .cm
2710 :     files. Lal said that his proposal was just that -- a proposal.
2711 :     For the time being I went ahead and implemented it so that people
2712 :     can comment on it. Everything is completely backward-compatible
2713 :     (except for the stable library format, i.e., new bootfiles!).
2714 :    
2715 :     As usual, see the CM manual for details.
2716 :    
2717 :     2. An alternative syntax for anchored paths has been implemented.
2718 :     Dave has recently voiced the same concerns that I had when I did
2719 :     this, so there should be some support. My take is that eventually
2720 :     I will let support for the current syntax (where anchors are
2721 :     "implicit") fade out in favor of the new, explicit syntax.
2722 :     In order to be backward-compatible, both old and new syntax are
2723 :     currently supported.
2724 :    
2725 :     Again, see the CM manual for details.
2726 :    
2727 :     3. Parallel make is trying to be slightly smarter: When the master
2728 :     process finds a "bottleneck", i.e., when there is only one
2729 :     compilation unit that can be compiled and everybody else is
2730 :     waiting on it, then it will simply compile it directly instead
2731 :     of clumsily telling one of the slaves to do it.
2732 :    
2733 :     4. Support for "unsharing" added. This is necessary in order to be
2734 :     able to have two different versions of the same library running
2735 :     at the same time (e.g., for trying out a new MLRISC while still
2736 :     having the old MLRISC linked into the current compiler, etc.)
2737 :     See the CM manual.
2738 :    
2739 :     5. Simple "makedepend" functionality added for generating Makefile
2740 :     dependency information. (This is rather crude at the moment.
2741 :     Expect some changes here in the future.)
2742 :    
2743 :     6. ".fun" added as a recognized suffix for ML files. Also documented
2744 :     explicitly in the manual that the fallback behavior (unknown suffix
2745 :     -> ML file) is not an official feature!
2746 :    
2747 :     7. Small changes to the pickler for stable libraries.
2748 :    
2749 :     8. Several internal changes to CM (for cleanup/improvement).
2750 :    
2751 :    
2752 :     !!!! NEW BINFILES !!!!
2753 :    
2754 :     ----------------------------------------------------------------------
2755 :     Name: Matthias Blume
2756 : blume 631 Date: 2000/04/28 17:30:00 JST
2757 :     Tag: blume-20000428-pathconfig
2758 :     Description:
2759 :    
2760 :     1. I changed config/install.sh to remove duplicate entries from the
2761 :     lib/pathconfig file at the end. Moreover, the final version of
2762 :     lib/pathconfig is sorted alphabetically. The same (sorting) is done
2763 :     in src/system/installml.
2764 :    
2765 :     2. The config/install.sh script now consistently uses relative
2766 :     pathnames in lib/pathconfig whenever the anchor is in the lib
2767 :     directory. (So far this was true for the libraries that come
2768 :     pre-compiled and bundled as part of the bootfiles but not for
2769 :     libraries that are compiled by the script itself.)
2770 :    
2771 :     ----------------------------------------------------------------------
2772 :     Name: Matthias Blume
2773 : blume 629 Date: 2000/04/26 13:10:00 JST
2774 :     Tag: blume-20000426-fun_suffix
2775 :     Description:
2776 :    
2777 :     Added ".fun" as a recognized file name suffix (for ML code).
2778 :    
2779 :     ----------------------------------------------------------------------
2780 : leunga 628 Name: Allen Leung
2781 :     Date: 2000/04/25 17:00:00 EST
2782 :     Tag: leunga-20000425-alpha-ra
2783 :     Description:
2784 :    
2785 :     1. Alpha
2786 :    
2787 :     PSEUDOARITH was missing in AlphaRewrite. This causes an endless loop
2788 :     in C--.
2789 :    
2790 :     2. RA
2791 :    
2792 :     Added a flag "ra-dump-size" to print out the size of the flowgraph
2793 :     and the interference graph.
2794 :    
2795 :     ----------------------------------------------------------------------
2796 : dbm 627 Name: Dave MacQueen
2797 :     Date: 2000/04/25/
2798 :     Tag: dbm-20000425-mlyacc_doc_examples
2799 :     Description:
2800 :     Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
2801 :     Updated all three examples in src/ml-yacc/examples to run
2802 :     under 110.* using CM.make.
2803 :    
2804 :     ----------------------------------------------------------------------
2805 : leunga 624 Name: Allen Leung
2806 :     Date: 2000/04/20 23:04:00 EST
2807 :     Tag: leunga-20000420-ssa-c---stuff
2808 :     Description:
2809 :    
2810 :     This update synchronizes my repository with Yale's. Most of these
2811 :     changes, however, do not affect SML/NJ at all (the RA is an exception).
2812 :    
2813 :     1. Register Allocator
2814 :    
2815 :     a. An improvement in the interference graph construction:
2816 :     Given a copy
2817 :    
2818 :     s <- t
2819 :    
2820 : leunga 628 no interference edge between s and t is added for this definition of s.
2821 : leunga 624
2822 :     b. I've added two new spill heuristic modules that Fermin and I developed
2823 :     (in the new library RA.cm). These are unused in SML/NJ but maybe
2824 :     useful for others (Moby?)
2825 :    
2826 :     2. X86
2827 :    
2828 :     a. Various fixes in the backend provided by Fermin [C--] and Lal.
2829 :    
2830 :     3. Alpha
2831 :    
2832 :     a. Added the BSR instruction and code generation that goes with it [C--]
2833 :     b. Other fixes too numerous to recount provided by Fermin [C--]
2834 :    
2835 :     4. Regmaps
2836 :    
2837 :     a. The regmaps are not initialized with the identity physical bindings
2838 :     at creation time. This is unneeded.
2839 :    
2840 :     5. MLRISC Optimizations
2841 :    
2842 :     a. The DJ-Graph module can now compute the iterated dominance frontiers
2843 :     intersects with liveness incrementally in linear time! Woohoo!
2844 :     This is now used in my new SSA construction algorithm.
2845 :    
2846 :     b. THe branch reorganization module is now smarter about linear chains of
2847 :     basic blocks.
2848 :    
2849 :    
2850 :     ----------------------------------------------------------------------
2851 : blume 615 Name: Matthias Blume
2852 :     Date: 2000/04/12 13:52:00 JST
2853 :     Tag: blume_main_v110p27_1
2854 :     Description:
2855 :    
2856 :     Changed install.sh script to handle archive files without version number
2857 :     and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
2858 :     name of the boot file archive.
2859 :    
2860 :     ----------------------------------------------------------------------
2861 : dbm 607 Name: Dave MacQueen
2862 :     Date: 2000/04/09 14:00 EDT
2863 :     Tag: dbm-20000410-Version_110_27
2864 :     Description:
2865 :    
2866 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
2867 :    
2868 :     2. Updated src/config/version to 110.27
2869 :    
2870 :     3. New boot files!
2871 :    
2872 :     ----------------------------------------------------------------------
2873 : leunga 606 Name: Allen Leung
2874 :     Date: 2000/04/09 19:09:00 EST
2875 :     Tag: leunga-20000409-misc
2876 :     Description:
2877 :    
2878 :     1. Yet another fix for x86 assembly for idivl, imull, mull and friends.
2879 :    
2880 :     2. Miscellaneous improvements to MLRISC (unused in sml/nj)
2881 :    
2882 :     ----------------------------------------------------------------------
2883 : monnier 605 Name: Stefan
2884 :     Date: 2000/04/07 10:00:00 EDT
2885 :     Tag: monnier-20000406-branch-handling
2886 :     Description:
2887 :    
2888 :     Improved handling of branches (mostly those generated from
2889 :     polymorphic equality), removed switchoff and changed the
2890 :     default optimization settings (more cpsopt and less flintopt).
2891 :    
2892 :     ----------------------------------------------------------------------
2893 : leunga 601 Name: Allen Leung
2894 : leunga 602 Date: 2000/04/06 01:30:00 EST
2895 :     Tag: leunga-20000406-peephole-x86-SSA-2
2896 :     Description:
2897 :    
2898 :     Forgot a few files.
2899 :    
2900 :     ----------------------------------------------------------------------
2901 :     Name: Allen Leung
2902 : leunga 601 Date: 2000/04/06 00:36:00 EST
2903 :     Tag: leunga-20000406-peephole-x86-SSA
2904 :     Description:
2905 :    
2906 :     1. New Peephole code
2907 :    
2908 :     2. Minor improvement to X86 instruction selection
2909 :    
2910 :     3. Various fixes to SSA and machine description -> code translator
2911 :    
2912 :     ----------------------------------------------------------------------
2913 : blume 595 Name: Matthias Blume
2914 :     Date: 2000/04/05 12:30:00 JST
2915 :     Tag: blume_main_v110p26p2_3
2916 :     Description:
2917 :    
2918 :     This update just merges three minor cosmetic updates to CM's sources
2919 :     to get ready for the 110.27 code freeze on Friday. No functionality
2920 :     has changed.
2921 :    
2922 :     ----------------------------------------------------------------------
2923 : leunga 593 Name: Allen Leung
2924 : leunga 594 Date: 2000/04/04 19:39:00 EST
2925 :     Tag: leunga-20000404-x86-asm
2926 :     Description:
2927 :    
2928 :     1. Fixed a problem in X86 assembly.
2929 :    
2930 :     Things like
2931 :    
2932 :     jmp %eax
2933 :     jmp (%eax)
2934 :    
2935 :     should be output as
2936 :    
2937 :     jmp *%eax
2938 :     jmp *(%eax)
2939 :    
2940 :     2. Assembly output
2941 :    
2942 :     Added a new flag
2943 :    
2944 :     "asm-indent-copies" (default to false)
2945 :    
2946 :     When this flag is on, parallel copies will be indented an extra level.
2947 :    
2948 :     ----------------------------------------------------------------------
2949 :     Name: Allen Leung
2950 : leunga 593 Date: 2000/04/04 03:18:00 EST
2951 :     Tag: leunga-20000404-C--Moby
2952 :     Description:
2953 :    
2954 :     All of these fixes are related to C--, Moby, and my own optimization
2955 :     stuff; so they shouldn't affect SML/NJ.
2956 :    
2957 :     1. X86
2958 :    
2959 :     Various fixes related floating point, and extensions.
2960 :    
2961 :     2. Alpha
2962 :    
2963 :     Some extra patterns related to loads with signed/zero extension
2964 :     provided by Fermin.
2965 :    
2966 :     3. Assembly
2967 :    
2968 : monnier 659 When generating assembly, resolve the value of client defined constants,
2969 : leunga 593 instead of generating symbolic values. This is controlled by the
2970 :     new flag "asm-resolve-constants", which is default to true.
2971 :    
2972 :     4. Machine Descriptions
2973 :    
2974 :     a. The precedence parser was slightly broken when parsing infixr symbols.
2975 :     b. The type generalizing code had the bound variables reversed, resulting
2976 :     in a problem during arity raising.
2977 :     c. Various fixes in machine descriptions.
2978 :    
2979 :     ----------------------------------------------------------------------
2980 : blume 592 Name: Matthias Blume
2981 :     Date: 2000/04/03 16:05:00 JST
2982 :     Tag: blume_main_v110p26p2_2
2983 :     Description:
2984 :    
2985 :     I eliminated coreEnv from compInfo. Access to the "Core" structure is
2986 :     now done via the ordinary static environment that is context to each
2987 :     compilation unit.
2988 :    
2989 :     To this end, I arranged that instead of "structure Core" as "structure
2990 :     _Core" is bound in the pervasive environment. Core access is done via
2991 : monnier 659 _Core (which can never be accidentally rebound because _Core is not a
2992 : blume 592 legal surface-syntax symbol).
2993 :    
2994 :     The current solution is much cleaner because the core environment is
2995 :     now simply part of the pervasive environment which is part of every
2996 :     compilation unit's context anyway. In particular, this eliminates all
2997 :     special-case handling that was necessary until now in order to deal
2998 :     with dynamic and symbolic parts of the core environment.
2999 :    
3000 :     Remaining hackery (to bind the "magic" symbol _Core) is localized in the
3001 : monnier 659 compilation manager's bootstrap compiler (actually: in the "init group"
3002 : blume 592 handling). See the comments in src/system/smlnj/init/init.cmi for
3003 :     more details.
3004 :    
3005 :     I also tried to track down all mentions of "Core" (as string argument
3006 :     to Symbol.strSymbol) in the compiler and replaced them with a
3007 :     reference to the new CoreSym.coreSym. Seems cleaner since the actual
3008 :     name appears in one place only.
3009 :    
3010 :     Binfile and bootfile format have not changed, but the switchover from
3011 :     the old "init.cmi" to the new one is a bit tricky, so I supplied new
3012 :     bootfiles anyway.
3013 :    
3014 :     ----------------------------------------------------------------------
3015 : leunga 591 Name: Allen Leung
3016 :     Date: 2000/04/02 21:17:00 EST
3017 :     Tag: leunga-20000402-mltree
3018 :     Description:
3019 : leunga 585
3020 : leunga 591 1. Renamed the constructor CALL in MLTREE by popular demand.
3021 :     2. Added a bunch of files from my repository. These are currently
3022 :     used by other non-SMLNJ backends.
3023 :    
3024 : leunga 576 ----------------------------------------------------------------------
3025 : leunga 590 Name: Allen Leung
3026 :     Date: 2000/03/31 21:15:00 EST
3027 :     Tag: leunga-20000331-aliasing
3028 :     Description:
3029 :    
3030 :     This update contains a rewritten (and hopefully more correct) module
3031 :     for extracting aliasing information from CPS.
3032 :    
3033 :     To turn on this feature:
3034 :    
3035 :     Compiler.Control.CG.memDisambiguate := true
3036 :    
3037 :     To pretty print the region information with assembly
3038 :    
3039 :     Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
3040 :    
3041 :     To control how many levels of aliasing information are printed, use:
3042 :    
3043 :     Compiler.Control.MLRISC.getInt "points-to-show-level" := n
3044 :    
3045 :     The default of n is 3.
3046 :    
3047 :     ----------------------------------------------------------------------
3048 : dbm 589 Name: David MacQueen
3049 :     Date: 2000/03/31 11:15:00 EST
3050 :     Tag: dbm-20000331-runtime_fix
3051 :     Description:
3052 :    
3053 :     This update contains:
3054 :    
3055 :     1. runtime/c-lib/c-libraries.c
3056 :     includes added in revision 1.2 caused compilation errors on hppa-hpux
3057 :    
3058 :     2. fix for bug 1556
3059 :     system/Basis/Implementation/NJ/internal-signals.sml
3060 :    
3061 :     ----------------------------------------------------------------------
3062 : blume 587 Name: Matthias Blume
3063 : blume 588 Date: 2000/03/31 18:00:00 JST
3064 :     Tag: blume_main_v110p26p2_1
3065 :     Description:
3066 :    
3067 :     This update contains:
3068 :    
3069 :     1. A small change to CM's handling of stable libraries:
3070 :     CM now maintains one "global" modmap that is used for all stable
3071 :     libraries. The use of such a global modmap maximizes sharing and
3072 :     minimizes the need for re-traversing parts of environments during
3073 :     modmap construction. (However, this has minor impact since modmap
3074 :     construction seems to account for just one percent or less of total
3075 :     compile time.)
3076 :    
3077 :     2. I added a "genmap" phase to the statistics. This is where I got the
3078 :     "one percent" number (see above).
3079 :    
3080 :     3. CM's new tool parameter mechanism just became _even_ better. :)
3081 :     - The parser understands named parameters and recursive options.
3082 :     - The "make" and "shell" tools use these new features.
3083 :     (This makes it a lot easier to cascade these tools.)
3084 :     - There is a small syntax change: named parameters use a
3085 :    
3086 :     <name> : ( <option> ... ) or
3087 :     <name> : <string>
3088 :    
3089 :     syntax. Previously, named parameters were implemented in an
3090 :     ad-hoc fashion by each tool individually (by parsing strings)
3091 :     and had the form
3092 :    
3093 :     <name>=<string>
3094 :    
3095 :     See the CM manual for a full description of these issues.
3096 :    
3097 :     ----------------------------------------------------------------------
3098 :     Name: Matthias Blume
3099 : blume 587 Date: 2000/03/30 18:00:00 JST
3100 :     Tag: blume_main_v110p26p2_0
3101 :     Description:
3102 :    
3103 :     !!!!! WARNING !!!!!!
3104 :     !! New binfiles !!
3105 :     !!!!!!!!!!!!!!!!!!!!
3106 :    
3107 :     This update contains:
3108 :    
3109 :     1. Moderate changes to CM:
3110 :    
3111 :     - Changes to CM's tools mechanism. In particular, it is now possible
3112 :     to have tools that accept additional "command line" parameters
3113 :     (specified in the .cm file at each instance where the tool's class is
3114 :     used).
3115 :    
3116 : monnier 659 This was done to accommodate the new "make" and "shell" tools which
3117 :     facilitate fairly seamless hookup to portions of code managed using
3118 : blume 587 Makefiles or Shell scripts.
3119 :    
3120 :     There are no classes "shared" or "private" anymore. Instead, the
3121 :     sharing annotation is now a parameter to the "sml" class.
3122 :    
3123 :     There is a bit of generic machinery for implementing one's own
3124 :     tools that accept command-line parameters. However, I am not yet fully
3125 :     satisfied with that part, so expect changes here in the future.
3126 :    
3127 :     All existing tools are described in the CM manual.
3128 :    
3129 : monnier 659 - Slightly better error handling. (CM now suppresses many followup
3130 : blume 587 error messages that tended to be more annoying than helpful.)
3131 :    
3132 :     2. Major changes to the compiler's static environment data structures.
3133 :    
3134 :     - no CMStaticEnv anymore.
3135 :     - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
3136 :     but it is called Environment), no conversions between different
3137 :     kinds of static environments
3138 :    
3139 :     - There is still a notion of a "modmap", but such modmaps are generated
3140 :     on demand at the time when they are needed. This sounds slow, but I
3141 :     sped up the code that generates modmaps enough for this not to lead to
3142 :     a slowdown of the compiler (at least I didn't detect any).
3143 :    
3144 :     - To facilitate rapid modmap generation, static environments now
3145 :     contain an (optional) "modtree" structure. Modtree annotations are
3146 :     constructed by the unpickler during unpickling. (This means that
3147 :     the elaborator does not have to worry about modtrees at all.)
3148 :     Modtrees have the advantage that they are compositional in the same
3149 :     way as the environment data structure itself is compositional.
3150 :     As a result, modtrees never hang on to parts of an environment that
3151 :     has already been rendered "stale" by filtering or rebinding.
3152 :    
3153 :     - I went through many, many trials and errors before arriving at the
3154 :     current solution. (The initial idea of "linkpaths" did not work.)
3155 :     But the result of all this is that I have touched a lot of files that
3156 :     depend on the "modules" and "types" data structures (most of the
3157 :     elaborator). There were a lot of changes during my "linkpath" trials
3158 :     that could have been reverted to their original state but weren't.
3159 :     Please, don't be too harsh on me for messing with this code a bit more
3160 : monnier 921 than what was strictly necessary... (I _did_ resist the temptation
3161 : blume 587 of doing any "global reformatting" to avoid an untimely death at
3162 :     Dave's hands. :)
3163 :    
3164 :     - One positive aspect of the previous point: At least I made sure that
3165 :     all files that I touched now compile without warnings (other than
3166 :     "polyEqual").
3167 :    
3168 :     - compiler now tends to run "leaner" (i.e., ties up less memory in
3169 :     redundant modmaps)
3170 :    
3171 :     ----------------------------------------------------------------------
3172 : leunga 580 Name: Allen Leung
3173 : leunga 585 Date: 2000/03/29 18:00:00
3174 :     Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
3175 : leunga 586 Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
3176 : leunga 585 Description:
3177 :    
3178 :     This update contains *MAJOR* changes to the way code is generated from CPS
3179 :     in the module mlriscGen, and in various backend modules.
3180 :    
3181 :     CHANGES
3182 :     =======
3183 :    
3184 :     1. MLRiscGen: forward propagation fix.
3185 :    
3186 :     There was a bug in forward propagation introduced at about the same time
3187 :     as the MLRISC x86 backend, which prohibits coalescing to be
3188 :     performed effectively in loops.
3189 :    
3190 :     Effect: speed up of loops in RISC architectures.
3191 :     By itself, this actually slowed down certain benchmarks on the x86.
3192 :    
3193 :     2. MLRiscGen: forward propagating addresses from consing.
3194 :    
3195 :     I've changed the way consing code is generated. Basically I separated
3196 :     out the initialization part:
3197 :    
3198 :     store tag, offset(allocptr)
3199 :     store elem1, offset+4(allocptr)
3200 :     store elem2, offset+8(allocptr)
3201 :     ...
3202 :     store elemn, offset+4n(allocptr)
3203 :    
3204 :     and the address computation part:
3205 :    
3206 :     celladdr <- offset+4+alloctpr
3207 :    
3208 :     and move the address computation part
3209 :    
3210 :     Effect: register pressure is generally lower as a result. This
3211 :     makes compilation of certain expressions much faster, such as
3212 :     long lists with non-trivial elements.
3213 :    
3214 :     [(0,0), (0,0), .... (0,0)]
3215 :    
3216 :     3. MLRiscGen: base pointer elimination.
3217 :    
3218 :     As part of the linkage mechanism, we generate the sequence:
3219 :    
3220 :     L: ... <- start of the code fragment
3221 :    
3222 :     L1:
3223 :     base pointer <- linkreg - L1 + L
3224 :    
3225 :     The base pointer was then used for computing relocatable addresses
3226 :     in the code fragment. Frequently (such as in lots of continuations)
3227 :     this is not needed. We now eliminate this sequence whenever possible.
3228 :    
3229 :     For compile time efficiency, I'm using a very stupid local heuristic.
3230 :     But in general, this should be done as a control flow analysis.
3231 :    
3232 :     Effect: Smaller code size. Speed up of most programs.
3233 :    
3234 :     4. Hppa back end
3235 :    
3236 :     Long jumps in span dependence resolution used to depend on the existence
3237 :     of the base pointer.
3238 :    
3239 :     A jump to a long label L was expanded into the following sequence:
3240 :    
3241 :     LDIL %hi(L-8192), %r29
3242 :     LDO %lo(L-8192)(%r29), %r29
3243 :     ADD %r29, baseptr, %r29
3244 :     BV,n %r0(%r29)
3245 :    
3246 :     In the presence of change (3) above, this will not work. I've changed
3247 :     it so that the following sequence of instructions are generated, which
3248 :     doesn't mention the base pointer at all:
3249 :    
3250 :     BL,n L', %r29 /* branch and link, L' + 4 -> %r29 */
3251 :     L': ADDIL L-(L'+4), %r29 /* Compute address of L */
3252 :     BV,n %r0(%r29) /* Jump */
3253 :    
3254 :     5. Alpha back end
3255 :    
3256 :     New alpha instructions LDB/LDW have been added, as per Fermin's
3257 :     suggestions. This is unrelated to all other changes.
3258 :    
3259 :     6. X86 back end
3260 :    
3261 :     I've changed andl to testl in the floating point test sequence
3262 :     whenever appropriate. The Intel optimization guide states that
3263 : monnier 659 testl is preferable to andl.
3264 : leunga 585
3265 :     7. RA (x86 only)
3266 :    
3267 :     I've improved the spill propagation algorithm, using an approximation
3268 :     of maximal weighted independent sets. This seems to be necessary to
3269 :     alleviate the negative effect in light of the slow down in (1).
3270 :    
3271 :     I'll write down the algorithm one of these days.
3272 :    
3273 :     8. MLRiscGen: frequencies
3274 :    
3275 :     I've added an annotation that states that all call gc blocks have zero
3276 :     execution frequencies. This improves register allocation on the x86.
3277 :    
3278 :     BENCHMARKS
3279 :     ==========
3280 :    
3281 :     I've only perform the comparison on 110.25.
3282 :    
3283 :     The platforms are:
3284 :    
3285 :     HPPA A four processor HP machine (E9000) with 5G of memory.
3286 :     X86 A 300Hhz Pentium II with 128M of memory, and
3287 :     SPARC An Ultra sparc 2 with 512M of memory.
3288 :    
3289 :     I used the following parameters for the SML benchmarks:
3290 :    
3291 :     @SMLalloc
3292 :     HPPA 256k
3293 :     SPARC 512k
3294 :     X86 256k
3295 :    
3296 :     COMPILATION TIME
3297 :     ----------------
3298 :     Here are the numbers comparing the compilation times of the compilers.
3299 :     I've only compared 110.25 compiling the new sources versus
3300 :     a fixpoint version of the new compiler compiling the same.
3301 :    
3302 :     110.25 New
3303 :     Total Time in RA Spill+Reload Total Time In RA Spill+Reload
3304 :     HPPA 627s 116s 2684+3584 599s 95s 1003+1879
3305 :     SPARC 892s 173s 2891+3870 708s 116s 1004+1880
3306 :     X86 999s 315s 94006+130691 987s 296s 108877+141957
3307 :    
3308 :     110.25 New
3309 :     Code Size Code Size
3310 :     HPPA 8596736 8561421
3311 :     SPARC 8974299 8785143
3312 :     X86 9029180 8716783
3313 :    
3314 :     So in summary, things are at least as good as before. Dramatic
3315 :     reduction in compilation is obtained on the Sparc; I can't explain it,
3316 :     but it is reproducible. Perhaps someone should try to reproduce this
3317 :     on their own machines.
3318 :    
3319 :     SML BENCHMARKS
3320 :     --------------
3321 :    
3322 :     On the average, all benchmarks perform at least as well as before.
3323 :    
3324 :     HPPA Compilation Time Spill+Reload Run Time
3325 :     110.25 New 110.25 New 110.25 New
3326 :    
3327 :     barnesHut 3.158 3.015 4.75% 1+1 0+0 2.980 2.922 2.00%
3328 :     boyer 6.152 5.708 7.77% 0+0 0+0 0.218 0.213 2.34%
3329 :     count-graphs 1.168 1.120 4.32% 0+0 0+0 22.705 23.073 -1.60%
3330 :     fft 0.877 0.792 10.74% 1+3 1+3 0.602 0.587 2.56%
3331 :     knuthBendix 3.180 2.857 11.32% 0+0 0+0 0.675 0.662 2.02%
3332 :     lexgen 6.190 5.290 17.01% 0+0 0+0 0.913 0.788 15.86%
3333 :     life 0.803 0.703 14.22% 25+25 0+0 0.153 0.140 9.52%
3334 :     logic 2.048 2.007 2.08% 6+6 1+1 4.133 4.008 3.12%
3335 :     mandelbrot 0.077 0.080 -4.17% 0+0 0+0 0.765 0.712 7.49%
3336 :     mlyacc 22.932 20.937 9.53% 154+181 32+57 0.468 0.430 8.91%
3337 :     nucleic 5.183 5.060 2.44% 2+2 0+0 0.125 0.120 4.17%
3338 :     ratio-regions 3.357 3.142 6.84% 0+0 0+0 116.225 113.173 2.70%
3339 :     ray 1.283 1.290 -0.52% 0+0 0+0 2.887 2.855 1.11%
3340 :     simple 6.307 6.032 4.56% 28+30 5+7 3.705 3.658 1.28%
3341 :     tsp 0.888 0.862 3.09% 0+0 0+0 7.040 6.893 2.13%
3342 :     vliw 24.378 23.455 3.94% 106+127 25+45 2.758 2.707 1.91%
3343 :     --------------------------------------------------------------------------
3344 :     Average 6.12% 4.09%
3345 :    
3346 :     SPARC Compilation Time Spill+Reload Run Time
3347 :     110.25 New 110.25 New 110.25 New
3348 :    
3349 :     barnesHut 3.778 3.592 5.20% 2+2 0+0 3.648 3.453 5.65%
3350 :     boyer 6.632 6.110 8.54% 0+0 0+0 0.258 0.242 6.90%
3351 :     count-graphs 1.435 1.325 8.30% 0+0 0+0 33.672 34.737 -3.07%
3352 :     fft 0.980 0.940 4.26% 3+9 2+6 0.838 0.827 1.41%
3353 :     knuthBendix 3.590 3.138 14.39% 0+0 0+0 0.962 0.967 -0.52%
3354 :     lexgen 6.593 6.072 8.59% 1+1 0+0 1.077 1.078 -0.15%
3355 :     life 0.972 0.868 11.90% 26+26 0+0 0.143 0.140 2.38%
3356 :     logic 2.525 2.387 5.80% 7+7 1+1 5.625 5.158 9.05%
3357 :     mandelbrot 0.090 0.093 -3.57% 0+0 0+0 0.855 0.728 17.39%
3358 :     mlyacc 26.732 23.827 12.19% 162+189 32+57 0.550 0.560 -1.79%
3359 :     nucleic 6.233 6.197 0.59% 3+3 0+0 0.163 0.173 -5.77%
3360 :     ratio-regions 3.780 3.507 7.79% 0+0 0+0 133.993 131.035 2.26%
3361 :     ray 1.595 1.550 2.90% 1+1 0+0 3.440 3.418 0.63%
3362 :     simple 6.972 6.487 7.48% 29+32 5+7 3.523 3.525 -0.05%
3363 :     tsp 1.115 1.063 4.86% 0+0 0+0 7.393 7.265 1.77%
3364 :     vliw 27.765 24.818 11.87% 110+135 25+45 2.265 2.135 6.09%
3365 :     ----------------------------------------------------------------------------
3366 :     Average 6.94% 2.64%
3367 :    
3368 :     X86 Compilation Time Spill+Reload Run Time
3369 :     110.25 New 110.25 New 110.25 New
3370 :    
3371 :     barnesHut 5.530 5.420 2.03% 593+893 597+915 3.532 3.440 2.66%
3372 :     boyer 8.768 7.747 13.19% 493+199 301+289 0.327 0.297 10.11%
3373 :     count-graphs 2.040 2.010 1.49% 298+394 315+457 26.578 28.660 -7.26%
3374 :     fft 1.327 1.302 1.92% 112+209 115+210 1.055 0.962 9.71%
3375 :     knuthBendix 5.218 5.475 -4.69% 451+598 510+650 0.928 0.932 -0.36%
3376 :     lexgen 9.970 9.623 3.60% 1014+841 1157+885 0.947 0.928 1.97%
3377 :     life 1.183 1.183 0.00% 162+182 145+148 0.127 0.103 22.58%
3378 :     logic 3.285 3.512 -6.45% 514+684 591+836 5.682 5.577 1.88%
3379 :     mandelbrot 0.147 0.143 2.33% 38+41 33+54 0.703 0.690 1.93%
3380 :     mlyacc 35.457 32.763 8.22% 3496+4564 3611+4860 0.552 0.550 0.30%
3381 :     nucleic 7.100 6.888 3.07% 239+168 201+158 0.175 0.173 0.96%
3382 :     ratio-regions 6.388 6.843 -6.65% 1182+257 981+300 120.142 120.345 -0.17%
3383 :     ray 2.332 2.338 -0.29% 346+398 402+494 3.593 3.540 1.51%
3384 :     simple 9.912 9.903 0.08% 1475+941 1579+1168 3.057 3.178 -3.83%
3385 :     tsp 1.623 1.532 5.98% 266+200 250+211 8.045 7.878 2.12%
3386 :     vliw 33.947 35.470 -4.29% 2629+2774 2877+3171 2.072 1.890 9.61%
3387 :     ----------------------------------------------------------------------------
3388 :     Average 1.22% 3.36%
3389 :    
3390 :     ----------------------------------------------------------------------
3391 :     Name: Allen Leung
3392 : leunga 583 Date: 2000/03/23 16:25:00
3393 :     Tag: leunga-20000323-fix_x86_alpha
3394 :     Description:
3395 :    
3396 :     1. X86 fixes/changes
3397 :    
3398 :     a. The old code generated for SETcc was completely wrong.
3399 :     The Intel optimization guide is VERY misleading.
3400 :    
3401 :     2. ALPHA fixes/changes
3402 :    
3403 :     a. Added the instructions LDBU, LDWU, STB, STW as per Fermin's suggestion.
3404 :     b. Added a new mode byteWordLoadStores to the functor parameter to Alpha()
3405 :     c. Added reassociation code for address computation.
3406 :    
3407 :     ----------------------------------------------------------------------
3408 :     Name: Allen Leung
3409 : leunga 580 Date: 2000/03/22 01:23:00
3410 :     Tag: leunga-20000322-fix_x86_hppa_ra
3411 :     Description:
3412 :    
3413 :     1. X86 fixes/changes
3414 :    
3415 :     a. x86Rewrite bug with MUL3 (found by Lal)
3416 :     b. Added the instructions FSTS, FSTL
3417 :    
3418 :     2. PA-RISC fixes/changes
3419 :    
3420 :     a. B label should not be a delay slot candidate! Why did this work?
3421 :     b. ADDT(32, REG(32, r), LI n) now generates one instruction instead of two,
3422 :     as it should be.
3423 :     c. The assembly syntax for fstds and fstdd was wrong.
3424 :     d. Added the composite instruction COMICLR/LDO, which is the immediate
3425 :     operand variant of COMCLR/LDO.
3426 :    
3427 :     3. Generic MLRISC
3428 :    
3429 :     a. shuffle.sml rewritten to be slightly more efficient
3430 :     b. DIV bug in mltree-simplify fixed (found by Fermin)
3431 :    
3432 :     4. Register Allocator
3433 :    
3434 :     a. I now release the interference graph earlier during spilling.
3435 :     May improve memory usage.
3436 :    
3437 :     ----------------------------------------------------------------------
3438 : blume 577 Name: Matthias Blume
3439 : blume 578 Date: 2000/03/14 14:15:32
3440 :     Tag: blume_main_v110p26p1_2
3441 :     Description:
3442 :    
3443 :     1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
3444 :     additional argument called "template" which is an optional string that
3445 : monnier 659 specifies the layout of the tool command line. See the CM manual for
3446 : blume 578 explanation.
3447 :    
3448 : monnier 659 2. A special-purpose tool can be "registered" by simply dropping the
3449 : blume 578 corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
3450 :     directory where the .cm file lives that uses this tool. (The
3451 :     behavior/misfeature until now was to look for the tool description
3452 :     files in the current working directory.) As before, tool description
3453 :     files could also be anchored -- in which case they can live anywhere
3454 :     they like. Following the recent e-mail discussion, this change should
3455 :     make it easier to have special-purpose tools that are shipped together
3456 :     with the sources of the program that uses them.
3457 :    
3458 :     ----------------------------------------------------------------------
3459 :     Name: Matthias Blume
3460 : blume 577 Date: 2000/03/10 07:48:34
3461 :     Tag: blume_main_v110p26p1_1
3462 :     Description:
3463 :    
3464 :     I added a re-written version of Dave's fixpt script to src/system.
3465 :     Changes relative to the original version:
3466 :     - sh-ified (not everybody has ksh)
3467 :     - automatically figures out which architecture it runs on
3468 :     - uses ./makeml a bit more cleverly
3469 :     - never invokes ./installml (and, thus, does not clobber your
3470 :     good and working installation of sml in case something goes wrong)
3471 :     - accepts max iteration count using option "-iter <n>"
3472 :     - accepts a "base" name using option "-base <base>"
3473 :    
3474 :     It does not build any extraneous heap images but directly rebuilds
3475 :     bin- and boot-hierarchies using makeml's "-rebuild" switch. Finally,
3476 :     it can incorporate existing bin- and boot- hierarchies. For example,
3477 :     suppose the base is set to "sml" (which is the default). Then it
3478 :     successively builds
3479 :    
3480 :     sml.bin.<arch>-unix and sml.boot.<arch>-unix
3481 :     then sml1.bin.<arch>-unix and sml1.boot.<arch>-unix
3482 :     then sml2.bin.<arch>-unix and sml2.boot.<arch>-unix
3483 :     ...
3484 :     then sml<n>.bin.<arch>-unix and sml<n>.boot.<arch>-unix
3485 :    
3486 :     and so on. If any of these already exist, it will just use what's
3487 :     there. In particular, many people will have the initial set of bin
3488 :     and boot files around, so this saves time for at least one full
3489 :     rebuild. Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
3490 :     <k>=1,2,... is normally not a good idea when invoking fixpt. However,
3491 :     they might be the result of an earlier partial run of fixpt (which
3492 : monnier 659 perhaps got accidentally killed). In this case, fixpt will quickly
3493 : blume 577 move through what exists before continuing where it left off earlier,
3494 :     and, thus, saves a lot of time.
3495 :    
3496 :     ----------------------------------------------------------------------
3497 : leunga 576 Name: Allen Leung
3498 :     Date: 00/03/10 02:20:00
3499 :     Tag: leunga-20000310-fix_x86_asm_ra
3500 :     Description:
3501 : dbm 570
3502 : leunga 576 More assembly output problems involving the indexed addressing mode
3503 :     on the x86 have been found and corrected. Thanks to Fermin Reig for the
3504 :     fix.
3505 :    
3506 :     The interface and implementation of the register allocator have been changed
3507 :     slightly to accommodate the possibility to skip the register allocation
3508 :     phases completely and go directly to memory allocation. This is needed
3509 :     for C-- use.
3510 :    
3511 : dbm 570 ----------------------------------------------------------------------
3512 : blume 572 Name: Matthias Blume
3513 : blume 575 Date: 00/03/09 10:23:53
3514 :     Tag: blume_main_v110p26p1_0
3515 :     Description:
3516 :    
3517 :     * Complete re-organization of library names. Many libraries have been
3518 :     consolidated so that they share the same path anchor. For example,
3519 :     all MLRISC-related libraries are anchored at MLRISC, most libraries that
3520 :     are SML/NJ-specific are under "smlnj". Notice that names like
3521 :     host-cmb.cm or host-compiler.cm no longer exist. See system/README
3522 :     for a complete description of the new naming scheme. Quick reference:
3523 :    
3524 :     host-cmb.cm -> smlnj/cmb.cm
3525 :     host-compiler.cm -> smlnj/compiler.cm
3526 :     full-cm.cm -> smlnj/cm.cm
3527 :     <arch>-<os>.cm -> smlnj/cmb/<arch>-<os>.cm
3528 :     <arch>-compiler.cm -> smlnj/compiler/<arch>.cm
3529 :    
3530 :     * Bug fixes in CM.
3531 :     - exceptions in user code are being passed through (i.e., reach top level)
3532 :     - more bugs in paranoia mode fixed
3533 :     - bug related to checking group owners fixed
3534 :    
3535 :     * New install.sh script that automagically fetches archive files:
3536 :     The new file config/srcarchiveurl must contain the URL of the
3537 :     (remote) directory that contains bin files (or other source archives).
3538 :     If install.sh does not find the archive locally, it tries to get
3539 :     it from that remote directory.
3540 :     This should simplify installation further: For machines that have
3541 :     access to the internet, just fetch <version>-config.tgz, unpack it,
3542 : monnier 659 edit config/targets, and go (run config/install.sh). The script will
3543 : blume 575 fetch everything else that it might need all by itself.
3544 :    
3545 :     For CVS users, this mechanism is not relevant for source archives, but
3546 :     it is convenient for getting new sets of binfiles.
3547 :    
3548 :     Archives should be tar files compressed with either gzip, compress, or
3549 :     bzip2. The script recognizes .tgz, .tar, tar.gz, tz, .tar.Z, and .tar.bz2.
3550 :    
3551 :     ----------------------------------------------------------------------
3552 :     Name: Matthias Blume
3553 : blume 572 Date: 2000/03/07 04:01:04
3554 :     Tag: blume_main_v110_26_2
3555 : dbm 570 Description:
3556 : blume 572 - size info in BOOTLIST
3557 :     * no fixed upper limits for number of bootfiles or length of
3558 :     bootfile names in runtime
3559 :     * falling back to old behavior if no BOOTLIST size info found
3560 :     - allocation size heuristics in .run-sml
3561 :     * tries to read cache size from /proc/cpuinfo (this is important for
3562 :     small-cache Celeron systems!)
3563 :     - install.sh robustified
3564 :     - CM manual updates
3565 :     - paranoid mode
3566 :     * no more CMB.deliver() (i.e., all done by CMB.make())
3567 :     * can re-use existing sml.boot.* files
3568 :     * init.cmi now treated as library
3569 :     * library stamps for consistency checks
3570 :     - sml.boot.<arch>-<os>/PIDMAP file
3571 :     * This file is read by the CM startup code. This is used to minimize
3572 :     the amount of dynamic state that needs to be stowed away for the
3573 :     purpose of sharing between interactive system and user code.
3574 :     - CM.Anchor.anchor instead of CM.Anchor.{set,cancel}
3575 :     * Upon request by Elsa. Anchors now controlled by get-set-pair
3576 :     like most other CM state variables.
3577 :     - Compiler.CMSA eliminated
3578 :     * No longer supported by CM anyway.
3579 :     - fixed bugs in pickler that kept biting Stefan
3580 :     * past refs to past refs (was caused by the possibility that
3581 :     ad-hoc sharing is more discriminating than hash-cons sharing)
3582 :     * integer overflow on LargeInt.minInt
3583 :     - ml-{lex,yacc} build scripts now use new mechanism
3584 :     for building standalone programs
3585 :     - fixed several gcc -Wall warnings that were caused by missing header
3586 :     files, missing initializations, etc., in runtime (not all warnings
3587 :     eliminated, though)

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