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/branches/primop-branch-3/NOTES/HISTORY
ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1069 - (view) (download)
Original Path: sml/trunk/HISTORY

1 : dbm 570 This is the HISTORY file for the Yale SML/NJ CVS repository.
2 :    
3 :     An entry should be made for _every_ commit to the repository.
4 :     The entries in this file will be used when creating the README
5 :     for new versions, so keep that in mind when writing the
6 :     description.
7 :    
8 :     The form of an entry should be:
9 :    
10 :     Name:
11 : george 1003 Date: yyyy/mm/dd
12 : dbm 570 Tag: <post-commit CVS tag>
13 :     Description:
14 : leunga 1025
15 : george 1003 ----------------------------------------------------------------------
16 : blume 1066 Name: Matthias Blume
17 : blume 1069 Date: 2002/02/15 16:00:00 EST
18 :     Tag: Release_110_39
19 :     Description:
20 :    
21 :     Working release 110.39. New bootfiles.
22 :    
23 :     ----------------------------------------------------------------------
24 :     Name: Matthias Blume
25 : blume 1068 Date: 2002/02/15 14:17:00 EST
26 :     Tag: blume-20020215-showbindings
27 :     Description:
28 :    
29 :     Added EnvRef.listBoundSymbols and CM.State.showBindings. Especially
30 :     the latter can be useful for exploring what bindings are available at
31 :     the interactive prompt. (The first function returns only the list
32 :     of symbols that are really bound, the second prints those but also the
33 :     ones that CM's autoloading mechanism knows about.)
34 :    
35 :     ----------------------------------------------------------------------
36 :     Name: Matthias Blume
37 : blume 1067 Date: 2002/02/15 12:08:00 EST
38 :     Tag: blume-20020215-iptrs
39 :     Description:
40 :    
41 :     Two improvements to ml-nlffigen:
42 :    
43 :     1. Write files only if they do not exist or if their current contents
44 :     do not coincide with what's being written. (That is, avoid messing
45 :     with the time stamps unless absolutely necessary.)
46 :    
47 :     2. Implement a "repository" mechanism for generated files related
48 :     to "incomplete pointer types". See the README file for details.
49 :    
50 :     ----------------------------------------------------------------------
51 :     Name: Matthias Blume
52 : blume 1066 Date: 2002/02/14 11:50:00 EST
53 :     Tag: blume-20020214-quote
54 :     Description:
55 :    
56 :     Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm). This is required
57 :     because of the new and improved tag generation scheme. (Thanks to Allen
58 :     Leung for pointing it out.)
59 :    
60 :     ----------------------------------------------------------------------
61 : george 1065 Name: Lal George
62 :     Date: 2002/02/14 09:55:27 EST 2002
63 :     Tag: george-20020214-isabelle-bug
64 :     Description:
65 :    
66 :     Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
67 :     of Isabelle on the x86.
68 :    
69 :     From Allen:
70 :     -----------
71 :     I've found the problem:
72 :    
73 :     in ra-core.sml, I use the counter "blocked" to keep track of the
74 :     true number of elements in the freeze queue. When the counter goes
75 :     to zero, I skip examining the queue. But I've messed up the
76 :     bookkeeping in combine():
77 :    
78 :     else ();
79 :     case !ucol of
80 :     PSEUDO => (if !cntv > 0 then
81 :     (if !cntu > 0 then blocked := !blocked - 1 else ();
82 :     ^^^^^^^^^^^^^^^^^^^^^^^
83 :     moveu := mergeMoveList(!movev, !moveu)
84 :     )
85 :     else ();
86 :    
87 :     combine() is called to coalesce two nodes u and v.
88 :     I think I was thinking that if the move counts of u and v are both
89 :     greater than zero then after they are coalesced then one node is
90 :     removed from the freeze queue. Apparently I was thinking that
91 :     both u and v are of low degree, but that's clearly not necessarily true.
92 :    
93 :    
94 :     02/12/2002:
95 :     Here's the patch. HOL now compiles.
96 :    
97 :     I don't know how this impact on performance (compile
98 :     time or runtime). This bug caused the RA (especially on the x86)
99 :     to go thru the potential spill phase when there are still nodes on the
100 :     freeze queue.
101 :    
102 :    
103 :    
104 :    
105 :     ----------------------------------------------------------------------
106 : blume 1058 Name: Matthias Blume
107 : blume 1064 Date: 2002/02/13 22:40:00 EST
108 :     Tag: blume-20020213-fptr-rtti
109 :     Description:
110 :    
111 :     Fixed a bug in ml-nlffigen that was introduced with one of the previous
112 :     updates.
113 :    
114 :     ----------------------------------------------------------------------
115 :     Name: Matthias Blume
116 : blume 1063 Date: 2002/02/13 16:41:00 EST
117 :     Tag: blume-20020213-cmlpq
118 :     Description:
119 :    
120 :     Added new priority queue export symbols (which have just been added to
121 :     smlnj-lib.cm) to CML's version of smlnj-lib.cm. (Otherwise CML would
122 :     not compile and the installer would choke.)
123 :    
124 :     ----------------------------------------------------------------------
125 :     Name: Matthias Blume
126 : blume 1062 Date: 2002/02/13 16:15:00 EST
127 :     Tag: blume-20020213-various
128 :     Description:
129 :    
130 :     1. More tweaks to ml-nlffigen:
131 :    
132 :     - better internal datastructures (resulting in slight speedup)
133 :     - "-match" option requires exact match
134 :     - "localized" gensym counters (untagged structs/unions nested within
135 :     other structs/unions or within typedefs get a fresh counter; their
136 :     tag will be prefixed by a concatenation of their parents' tags)
137 :     - bug fixes (related to calculation of transitive closure of types
138 :     to be included in the output)
139 :    
140 :     2. Minor Basis updates:
141 :    
142 :     - added implementations for List.collate and Option.app
143 :    
144 :     ----------------------------------------------------------------------
145 :     Name: Matthias Blume
146 : blume 1060 Date: 2002/02/11 15:55:00 EST
147 :     Tag: blume-20020211-gensym
148 :     Description:
149 :    
150 :     Added a "-gensym" option to command line of ml-nlffigen. This can be
151 :     used to specify a "stem" -- a string that is inserted in all "gensym'd"
152 :     names (ML structure names that correspond to unnamed C structs, unions,
153 :     and enums), so that separate runs of ml-nlffigen do not clash.
154 :    
155 :     ----------------------------------------------------------------------
156 :     Name: Matthias Blume
157 : blume 1059 Date: 2002/02/11 12:05:00 EST
158 :     Tag: blume-20020211-gensml
159 :     Description:
160 :    
161 :     A quick fix for a problem with GenSML (in the pgraph-util library):
162 :     Make generation of toplevel "local" optional. (Strictly speaking,
163 :     signature definitions within "local" are not legal SML.)
164 :    
165 :     Other than that: updates to INSTALL and cm/TODO.
166 :    
167 :     ----------------------------------------------------------------------
168 :     Name: Matthias Blume
169 : blume 1058 Date: 2002/02/08 15:00:00 EST
170 :     Tag: blume-20020208-uniquepid
171 :     Description:
172 :    
173 :     0. Version number has been bumped to 110.38.1. NEW BOOTFILES!!!
174 :    
175 :     1. The installer (config/install.sh) has gotten smarter:
176 :    
177 :     - Configuration options are a bit easier to specify now
178 :     (in config/targets).
179 :     - Bug in recognizing .tar.bz2 files fixed.
180 :     - Installer automatically resolves dependencies between
181 :     configuration options (e.g., if you ask for eXene, you will
182 :     also get cml -- regardless whether you asked for it or not).
183 :     - Installer can run in "quieter mode" by setting the environment
184 :     variable INSTALL_QUIETLY to "true". "Quieter" does not mean
185 :     "completely silent", though.
186 :     - Build HashCons library as part of smlnj-lib.
187 :    
188 :     2. A new scheme for assigning persistent identifiers to compilation
189 :     units (and, by extension, to types etc.) has been put into place.
190 :     This fixes a long-standing bug where types and even dynamic values
191 :     can get internally confused, thereby compromising type safety
192 :     (abstraction) and dynamic correctness. See
193 :    
194 :     http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
195 :    
196 :     for an example of how things could go wrong until now.
197 :    
198 :     The downside of the new scheme is that pids are not quite as
199 :     persistent as they used to be: CM will generate a fresh pid
200 :     for every compilation unit that it thinks it sees for the first
201 :     time. That means that if you compile starting from a clean, fresh
202 :     source tree at two different times, you end up with different
203 :     binaries.
204 :    
205 :     Cutoff recompilation, however, has not been compromised because
206 :     CM keeps pid information in special caches between runs.
207 :    
208 :     ----------------------------------------------------------------------
209 : george 1053 Name: Lal George
210 : george 1055 Date: 2002/02/07 15:34:13 EST 2002
211 :     Tag: <none>
212 :     Description:
213 :    
214 :     Compilers that generate assembly code may produce global labels
215 :     whose value is resolved at link time. The various peephole optimization
216 :     modules did not take this in account.
217 :    
218 :     TODO. The Labels.addrOf function should really return an option
219 :     type so that clients are forced to deal with this issue, rather
220 :     than an exception being raised.
221 :    
222 :     ----------------------------------------------------------------------
223 :     Name: Lal George
224 : george 1053 Date: 2002/02/06 13:55:02 EST
225 :     Tag: george-20020206-ra-breakup
226 :     Description:
227 :    
228 :     1. A bug fix from Allen.
229 :    
230 :     A typo causes extra fstp %st(0)'s to be generated at compensation
231 :     edges, which might cause stack underflow traps at runtime. This
232 :     occurs in fft where there are extraneous fstps right before the 'into'
233 :     trap instruction (in this case they are harmless since none of the
234 :     integers overflow.)
235 :    
236 : george 1054 2. Pulled out various utility modules that were embedded in the modules
237 : george 1053 of the register allocator. I need these modules for other purposes, but
238 : george 1054 they are not complete enough to put into a library (just yet).
239 : george 1053 ----------------------------------------------------------------------
240 : blume 1049 Name: Matthias Blume
241 :     Date: 2002/01/31 16:05:00 EST
242 :     Tag: blume-20020131-sparc-ccalls
243 :     Description:
244 :    
245 :     1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
246 :     of extra stack space by mistake. Fixed.
247 :    
248 :     2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
249 :    
250 :     ----------------------------------------------------------------------
251 : leunga 1048 Name: Allen Leung
252 :     Date: 2002/01/30
253 :     Tag: leunga-20020130-nowhere-bug-fix
254 :     Description:
255 :    
256 :     MLRISC bug fixes:
257 :     1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
258 :     2. MachineInt.fromString was negating its value.
259 :    
260 :     ----------------------------------------------------------------------
261 : blume 1039 Name: Matthias Blume
262 : blume 1047 Date: 2002/01/29
263 :     Tag: blume-20020129-INSTALL
264 :     Description:
265 :    
266 :     - Added somewhat detailed installation instructions (file INSTALL).
267 :     - Fixed curl-detection bug in config/install.sh.
268 :     - It is now possible to select the URL getter using the URLGETTER
269 :     environment variable:
270 :    
271 :     not set / "unknown" --> automatic detection (script tries wget,
272 :     curl, and lynx)
273 :     "wget" / "curl" / "lynx" --> use the specified program (script "knows"
274 :     how to properly invoke them)
275 :     other --> use $URLGETTER directly, it must take
276 :     precisely two command-line arguments
277 :     (source URL and destination file name)
278 :    
279 :     ----------------------------------------------------------------------
280 :     Name: Matthias Blume
281 : blume 1044 Date: 2002/01/28
282 : blume 1045 Tag: blume-20020128-sparc-ccalls
283 :     Description:
284 :    
285 :     - Fixed problem with calculation of "used" registers in sparc-c-calls.
286 :     - Make use of the allocParam argument in sparc-c-calls.
287 :    
288 :     ----------------------------------------------------------------------
289 :     Name: Matthias Blume
290 :     Date: 2002/01/28
291 : blume 1044 Tag: blume-20020128-allocParam
292 :     Description:
293 :    
294 :     John Reppy: Changes c-calls API to accept client-callback for
295 :     allocating extra stack space.
296 :     me: Corresponding changes to mlriscGen (using a dummy argument that
297 :     does not change the current behavior).
298 :    
299 :     ----------------------------------------------------------------------
300 :     Name: Matthias Blume
301 : blume 1040 Date: 2002/01/28 12:00:00
302 :     Tag: Release_110_38
303 :     Description:
304 :    
305 :     This time for real!!!
306 :    
307 :     ----------------------------------------------------------------------
308 :     Name: Matthias Blume
309 : blume 1039 Date: 2002/01/28 10:56:00 EST
310 :     Tag: blume-20020128-retraction
311 :     Description:
312 :    
313 :     0. Retracted earlier 110.38. (The Release_110_38 tag has been replaced
314 :     with blume-Release_110_38-retracted.)
315 :    
316 :     1. Fixed a problem with incorrect rounding modes in real64.sml.
317 :     (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
318 :    
319 :     2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
320 :     and enums fixed. The naming of corresponding ML identifiers should
321 :     now be consistent again.
322 :    
323 :     ----------------------------------------------------------------------
324 : leunga 1038 Name: Allen Leung
325 :     Date: 2002/01/27
326 :     Tag: leunga-20020127-nowhere
327 :     Description:
328 :    
329 :     Added a target called nowhere in the configuration scripts.
330 :     Enabling this will build the MLRISC 'nowhere' tool (for translating
331 :     programs with where-clauses into legal SML code) during installation.
332 :    
333 :     ----------------------------------------------------------------------
334 : blume 1036 Name: Matthias Blume
335 : leunga 1038 Date: 2002/01/25 21:27:00 EST
336 : blume 1039 Tag: blume-Release_110_38-retracted
337 : blume 1037 Description:
338 :    
339 :     Call it a (working) release! Version is 110.38. Bootfiles are ready.
340 :    
341 :     README will be added later.
342 :    
343 : blume 1039 !!! NOTE: Re-tagged as blume-Release_110_38-retracted. Original tag
344 :     (Release_110_38) removed. Reason: Last-minute bug fixes.
345 :    
346 : blume 1037 ----------------------------------------------------------------------
347 :     Name: Matthias Blume
348 : blume 1036 Date: 2002/01/25
349 :     Tag: blume-20020125-ffi
350 :     Description:
351 :    
352 :     A large number of tweaks and improvements to ml-nlffi-lib and
353 :     ml-nlffigen:
354 :    
355 :     - ML represenation types have been streamlined
356 :     - getter and setter functions work with concrete values, not abstract
357 :     ones where possible
358 :     - ml-nlffigen command line more flexible (see README file there)
359 :     - some bugs have been fixed (hopefully)
360 :    
361 :     ----------------------------------------------------------------------
362 : george 1034 Name: Lal George
363 :     Date: 2002/01/24
364 :     Tag: george-20020124-risc-ra-interface
365 :     Description:
366 :    
367 :     There is a dramatic simplification in the interface to the
368 :     register allocator for RISC architectures as a result of making
369 :     parallel copy instructions explicit.
370 :    
371 :     ----------------------------------------------------------------------
372 : blume 1028 Name: Matthias Blume
373 : blume 1032 Date: 2002/01/22
374 :     Tag: blume-20020122-x86-ccalls
375 :     Description:
376 :    
377 :     Bug fix for c-calls on x86 (having to do with how char- and
378 :     short-arguments are being handled).
379 :    
380 :     ----------------------------------------------------------------------
381 :     Name: Matthias Blume
382 : blume 1031 Date: 2002/01/21
383 :     Tag: blume-20020121-ff
384 :     Description:
385 :    
386 :     Another day of fiddling with the FFI...
387 :    
388 :     1. Bug fix/workaround: CKIT does not complain about negative array
389 :     dimensions, so ml-nlffigen has to guard itself against this possibility.
390 :     (Otherwise a negative dimension would send it into an infinite loop.)
391 :    
392 :     2. Some of the abstract types (light objects, light pointers, most "base"
393 :     types) in structure C are now eqtypes.
394 :    
395 :     3. Added constructors and test functions for NULL function pointers.
396 :    
397 :     ----------------------------------------------------------------------
398 :     Name: Matthias Blume
399 : blume 1028 Date: 2002/01/18
400 : blume 1030 Tag: blume-20020118-ready-for-new-release
401 :     Description:
402 :    
403 :     Made config/srcarchiveurl point to a new place. (Will provide boot
404 :     files shortly.)
405 :    
406 :     Maybe we christen this to be 110.38?
407 :    
408 :     ----------------------------------------------------------------------
409 :     Name: Matthias Blume
410 :     Date: 2002/01/18
411 : blume 1029 Tag: blume-20020118-more-ffifiddle
412 :     Description:
413 :    
414 :     Today's FFI fiddling:
415 :    
416 :     - Provided a structure CGetSet with "convenient" versions of C.Get.* and
417 :     C.Set.* that use concrete (MLRep.*) arguments and results instead
418 :     of abstract ones.
419 :    
420 :     - Provided word-style bit operations etc. for "int" representation
421 :     types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
422 :     and Long.
423 :    
424 :     ----------------------------------------------------------------------
425 :     Name: Matthias Blume
426 :     Date: 2002/01/18
427 : blume 1028 Tag: blume-20020118-use-x86-fp
428 :     Description:
429 :    
430 :     Now that x86-fast-fp seems to be working, I turned it back on again
431 :     by default. (Seems to work fine now, even with the FFI.)
432 :    
433 :     Other than that, I added some documentation about the FFI to
434 :     src/ml-nlffigen/README and updated the FFI test examples in
435 :     src/ml-nlffi-lib/Tests/*.
436 :    
437 :     ----------------------------------------------------------------------
438 : leunga 1025 Name: Allen Leung
439 :     Date: 2002/01/17
440 :     Tag: leunga-20020117-x86-fast-fp-call
441 :     Description:
442 :    
443 :     1. Fixed a problem with handling return fp values when x86's fast fp
444 :     mode is turned on.
445 :    
446 :     2. Minor pretty printing fix for cellset. Print %st(0) as %st(0) instead
447 :     of %f32.
448 :    
449 :     3. Added a constructor INT32lit to the ast of MLRISC tools.
450 :    
451 :     ----------------------------------------------------------------------
452 : blume 1021 Name: Matthias Blume
453 :     Date: 2002/01/16
454 :     Tag: blume-20020116-ffifiddle
455 :     Description:
456 :    
457 :     More fiddling with the FFI interface:
458 :    
459 :     - Make constness 'c instead of rw wherever possible. This eliminates
460 :     the need for certain explicit coercions. (However, due to ML's
461 :     value polymorphism, there will still be many cases where explicit
462 :     coercions are necessary. Phantom types are not the whole answer
463 :     to modeling a subtyping relationship in ML.)
464 :    
465 :     - ro/rw coersions for pointers added. (Avoids the detour through */&.)
466 :    
467 :     - "printf" test example added to src/ml-nlffi-lib/Tests. (Demonstrates
468 :     clumsy workaround for varargs problem.)
469 :    
470 :     ----------------------------------------------------------------------
471 : george 1016 Name: Lal George
472 :     Date: 2002/01/15
473 :     Tag: <none>
474 :     Description:
475 :    
476 :     1. Since COPY instructions are no longer native to the architecture,
477 :     a generic functor can be used to implement the expandCopies function.
478 :    
479 :     2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
480 :     TEXT segment.
481 :    
482 :     ----------------------------------------------------------------------
483 : blume 1011 Name: Matthias Blume
484 : blume 1015 Date: 2002/01/15
485 :     Tag: blume-20020115-ffiupdates
486 :     Description:
487 :    
488 :     1. Fix for bug resulting in single-precision float values being returned
489 :     incorrectly from FFI calls.
490 :    
491 :     2. Small modifications to C FFI API:
492 :    
493 :     - memory-allocation routines return straight objects (no options)
494 :     and raise an exception in out-of-memory situations
495 :     - unsafe extensions to cast between function pointers and pointers
496 :     from/to ints
497 :     - added structure C_Debug as an alternative to structure C where
498 :     pointer-dereferencing (|*| and |*!) always check for null-pointers
499 :     - added open_lib' to DynLinkage; open_lib' works like open_lib
500 :     but also takes a (possibly empty) list of existing library handles
501 :     that the current library depends on
502 :    
503 :     ----------------------------------------------------------------------
504 :     Name: Matthias Blume
505 : blume 1011 Date: 2002/01/10
506 :     Tag: blume-20020110-newffigen
507 :     Description:
508 :    
509 :     1. Updates to portable graph code.
510 :    
511 :     2. Major update to ml-nlffigen and ml-nlffi-lib. Things are much
512 :     more scalable now so that even huge interfaces such as the one
513 :     for GTK compile in finite time and space. :-)
514 :     See src/ml-nlffigen/README for details on what's new.
515 :    
516 :     ----------------------------------------------------------------------
517 : george 1003 Name: Lal George
518 : george 1009 Date: 2001/01/09 14:31:35 EST 2002
519 :     Tag: george-20011206-rm-native-copy
520 :     Description:
521 :    
522 :     Removed the native COPY and FCOPY instructions
523 :     from all the architectures and replaced it with the
524 :     explicit COPY instruction from the previous commit.
525 :    
526 :     It is now possible to simplify many of the optimizations
527 :     modules that manipulate copies. This has not been
528 :     done in this change.
529 :    
530 :     ----------------------------------------------------------------------
531 :     Name: Lal George
532 : george 1003 Date: 2001/12/06 16:50:13 EST 2001
533 :     Tag: george-20011206-mlrisc-instruction
534 :     Description:
535 : george 824
536 : george 1003 Changed the representation of instructions from being fully abstract
537 :     to being partially concrete. That is to say:
538 :    
539 :     from
540 :     type instruction
541 :    
542 :     to
543 :     type instr (* machine instruction *)
544 :    
545 :     datatype instruction =
546 :     LIVE of {regs: C.cellset, spilled: C.cellset}
547 :     | KILL of {regs: C.cellset, spilled: C.cellset}
548 :     | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
549 :     | ANNOTATION of {i: instruction, a: Annotations.annotation}
550 :     | INSTR of instr
551 :    
552 :     This makes the handling of certain special instructions that appear on
553 :     all architectures easier and uniform.
554 :    
555 :     LIVE and KILL say that a list of registers are live or killed at the
556 :     program point where they appear. No spill code is generated when an
557 :     element of the 'regs' field is spilled, but the register is moved to
558 :     the 'spilled' (which is present, more for debugging than anything else).
559 :    
560 :     LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
561 :     We used to generate:
562 :    
563 :     DEFFREG f1
564 :     f1 := f2 + f3
565 :     trapb
566 :    
567 :     but now generate:
568 :    
569 :     f1 := f2 + f3
570 :     trapb
571 :     LIVE {regs=[f1,f2,f3], spilled=[]}
572 :    
573 :     Furthermore, the DEFFREG (hack) required that all floating point instruction
574 :     use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
575 :     defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
576 :     in a cleaner alpha implementation. (Hopefully, intel will not get rid of
577 :     this architecture).
578 :    
579 :     COPYXXX is intended to replace the parallel COPY and FCOPY available on
580 :     all the architectures. This will result in further simplification of the
581 :     register allocator that must be aware of them for coalescing purposes, and
582 :     will also simplify certain aspects of the machine description that provides
583 :     callbacks related to parallel copies.
584 :    
585 :     ANNOTATION should be obvious, and now INSTR represents the honest to God
586 :     machine instruction set!
587 :    
588 :     The <arch>/instructions/<arch>Instr.sml files define certain utility
589 :     functions for making porting easier -- essentially converting upper case
590 :     to lower case. All machine instructions (of type instr) are in upper case,
591 :     and the lower case form generates an MLRISC instruction. For example on
592 :     the alpha we have:
593 :    
594 :     datatype instr =
595 :     LDA of {r:cell, b:cell, d:operand}
596 :     | ...
597 :    
598 :     val lda : {r:cell, b:cell, d:operand} -> instruction
599 :     ...
600 :    
601 :     where lda is just (INSTR o LDA), etc.
602 :    
603 : leunga 815 ----------------------------------------------------------------------
604 : blume 986 Name: Matthias Blume
605 : blume 993 Date: 2001/11/22 21:40:00 EST
606 :     Tag: Release_110_37
607 :     Description:
608 :    
609 :     Release 110.37. This time for real.
610 :    
611 :     ----------------------------------------------------------------------
612 :     Name: Matthias Blume
613 : blume 988 Date: 2001/11/21 16:35:00 EST
614 :     Tag: blume-20011121-foot-in-mouth
615 :     Description:
616 :    
617 :     Removed the "Release_110_37" tag because of a serious bug.
618 :     This will be re-tagged once the bug is fixed.
619 :    
620 :     ----------------------------------------------------------------------
621 :     Name: Matthias Blume
622 : blume 987 Date: 2001/11/21 16:14:00 EST
623 :     Tag: blume-20011121-forgottenfile
624 :     Description:
625 :    
626 :     Forgot to add a file. (Just a .tex-file -- part of
627 :     the CM manual source.)
628 :    
629 :     ----------------------------------------------------------------------
630 :     Name: Matthias Blume
631 : blume 986 Date: 2001/11/21 16:10:00 EST
632 : blume 988 Tag: blume-20011121-invalid_110_37
633 : blume 986 Description:
634 :    
635 : blume 988 Note: I removed the original tag "Release_110_37" from this commit
636 :     because we found a serious bug in all non-x86 backends.
637 :     - Matthias
638 :    
639 : blume 986 1. Modifications to the SML/NJ code generator and to the runtime system
640 :     so that code object name strings are directly inserted into code
641 :     objects at code generation time. The only business the runtime system
642 :     has with this is now to read the name strings on occasions.
643 :     (The encoding of the name string has also changed somewhat.)
644 :    
645 :     2. CM now implements a simple "set calculus" for specifying export lists.
646 :     In particular, it is now possible to refer to the export lists of
647 :     other libraries/groups/sources and form unions as well as differences.
648 :     See the latest CM manual for details.
649 :    
650 :     3. An separate notion of "proxy" libraries has again be eliminated from
651 :     CM's model. (Proxy libraries are now simply a special case of using
652 :     the export list calculus.)
653 :    
654 :     4. Some of the existing libraries now take advantage of the new set
655 :     calculus.
656 :     (Notice that not all libraries have been converted because some
657 :     of the existing .cm-files are supposed to be backward compatible
658 :     with 110.0.x.)
659 :    
660 :     5. Some cleanup in stand-alone programs. (Don't use "exnMessage" -- use
661 :     "General.exnMessage"! The former relies on a certain hook to be
662 :     initialized, and that often does not happen in the stand-alone case.)
663 :    
664 :     ----------------------------------------------------------------------
665 : george 984 Name: Lal George
666 :     Date: 2001/11/21 13:56:18 EST
667 :     Tag: george-2001121-pseudo-ops
668 :     Description:
669 :    
670 :     Implemented a complete redesign of MLRISC pseudo-ops. Now there
671 :     ought to never be any question of incompatabilities with
672 :     pseudo-op syntax expected by host assemblers.
673 :    
674 :     For now, only modules supporting GAS syntax are implemented
675 :     but more should follow, such as MASM, and vendor assembler
676 :     syntax, e.g. IBM as, Sun as, etc.
677 :    
678 :     ----------------------------------------------------------------------
679 : blume 977 Name: Matthias Blume
680 :     Date: 2001/11/14 11:52:00 EST
681 :     Tag: blume-20011114-srcname
682 :     Description:
683 :    
684 :     1. Routed the name of the current source file to mlriscgen where it
685 :     should be directly emitted into the code object. (This last part
686 :     is yet to be done.)
687 :    
688 :     2. Some cleanup of the pgraph code to make it match the proposal that
689 :     I put out the other day. (The proposal notwithstanding, things are
690 :     still in flux here.)
691 :    
692 :     ----------------------------------------------------------------------
693 : george 976 Name: Lal George
694 :     Date: 2001/11/14 09:44:04 EST
695 :     Tag:
696 :     Description:
697 :    
698 :     Fix for a backpatching bug reported by Allen.
699 :    
700 :     Because the boundary between short and long span-dependent
701 :     instructions is +/- 128, there are an astounding number of
702 :     span-dependent instructions whose size is over estimated.
703 :    
704 :     Allen came up with the idea of letting the size of span
705 :     dependent instructions be non-monotonic, for a maxIter
706 :     number of times, after which the size must be monotonically
707 :     increasing.
708 :    
709 :     This table shows the number of span-dependent instructions
710 :     whose size was over-estimated as a function of maxIter, for the
711 :     file Parse/parse/ml.grm.sml:
712 :    
713 :     maxIter # of instructions:
714 :     10 687
715 :     20 438
716 :     30 198
717 :     40 0
718 :    
719 :     In compiling the compiler, there is no significant difference in
720 :     compilation speed between maxIter=10 and maxIter=40. Actually,
721 :     my measurements showed that maxIter=40 was a tad faster than
722 :     maxIter=10! Also 96% of the files in the compiler reach a fix
723 :     point within 13 iterations, so fixing maxIter at 40, while high,
724 :     is okay.
725 :    
726 :     ----------------------------------------------------------------------
727 : blume 975 Name: Matthias Blume
728 :     Date: 2001/10/31 15:25:00 EST
729 :     Tag: blume-20011031-pgraph
730 :     Description:
731 :    
732 :     CKIT:
733 :     * Changed the "Function" constructor of type Ast.ctype to carry optional
734 :     argument identifiers.
735 :     * Changed the return type of TypeUtil.getFunction accordingly.
736 :     * Type equality ignores the argument names.
737 :     * TypeUtil.composite tries to preserve argument names but gives up quickly
738 :     if there is a mismatch.
739 :    
740 :     installation script:
741 :     * attempts to use "curl" if available (unless "wget" is available as well)
742 :    
743 :     CM:
744 :     * has an experimental implementation of "portable graphs" which I will
745 :     soon propose as an implementation-independent library format
746 :     * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
747 :    
748 :     NLFFI-LIB:
749 :     * some cleanup (all cosmetic)
750 :    
751 :     NLFFIGEN:
752 :     * temporarily disabled the mechanism that suppresses ML output for
753 :     C definitions whose identifiers start with an underscore character
754 :     * generate val bindings for enum constants
755 :     * user can request that only one style (light or heavy) is being used;
756 :     default is to use both (command-line arguments: -heavy and -light)
757 :     * fixed bug in handling of function types involving incomplete pointers
758 :     * generate ML entry points that take record arguments (i.e., using
759 :     named arguments) for C functions that have a prototype with named
760 :     arguments
761 :     (see changes to CKIT)
762 :    
763 :     ----------------------------------------------------------------------
764 : leunga 972 Name: Allen Leung
765 :     Date: 2001/10/27 20:34:00 EDT
766 :     Tag: leunga-20011027-x86-fast-fp-call
767 :     Description:
768 :    
769 :     Fixed the bug described in blume-20010920-slowfp.
770 :    
771 :     The fix involves
772 :     1. generating FCOPYs in FSTP in ia32-svid
773 :     2. marking a CALL with the appropriate annotation
774 :    
775 :     ----------------------------------------------------------------------
776 : blume 961 Name: Matthias Blume
777 :     Date: 2001/10/16 11:32:00 EDT
778 :     Tag: blume-20011016-netbsd
779 :     Description:
780 :    
781 :     Underscore patch from Chris Richards (fixing problem with compiling
782 :     runtime system under recent NetBSD).
783 :    
784 :     ----------------------------------------------------------------------
785 : george 959 Name: Allen Leung
786 :     Date: 2001/10/12 17:18:32 EDT 2001
787 :     Tag: leung-20011012-x86-printflowgraph
788 :     Description:
789 :    
790 :     X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
791 :    
792 :     ----------------------------------------------------------------------
793 : george 952 Name: Lal George
794 : george 958 Date: 2001/10/11 23:51:34 EDT
795 :     Tag: george-20011011-too-many-instrs
796 :     Description:
797 :    
798 :     The representation of a program point never expected to see more
799 :     than 65536 instructions in a basic block!
800 :    
801 :     ----------------------------------------------------------------------
802 :     Name: Lal George
803 : george 952 Date: 2001/10/09 09:41:37 EDT
804 :     Tag: george-20011008-mlrisc-labels
805 :     Description:
806 :    
807 :     Changed the machine description files to support printing of
808 :     local and global labels in assembly code, based on host assembler
809 :     conventions.
810 :    
811 :     ----------------------------------------------------------------------
812 : blume 916 Name: Matthias Blume
813 : blume 939 Date: 2001/09/25 15:25:00 EDT
814 :     Tag: blume-20010925-exninfo
815 :     Description:
816 :    
817 :     I provided a non-hook implementation of exnName (at the toplevel) and
818 :     made the "dummy" implementation of exnMessage (at the toplevel) more
819 :     useful: if nothing gets "hooked in", then at least you are going to
820 :     see the exception name and a message indicating why you don't see more.
821 :    
822 :     [For the time being, programs that need exnMessage and want to use
823 :     ml-build should either use General.exnMessage (strongly recommended) or
824 :     refer to structure General at some other point so that CM sees a
825 :     static dependency.]
826 :    
827 :     [Similar remarks go for "print" and "use": If you want to use their
828 :     functionality in stand-alone programs generated by ml-build, then use
829 :     TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
830 :    
831 :     ----------------------------------------------------------------------
832 :     Name: Matthias Blume
833 : blume 938 Date: 2001/09/20 17:28:00 EDT
834 :     Tag: blume-20010920-slowfp
835 :     Description:
836 :    
837 :     Allen says that x86-fast-fp is not safe yet, so I turned it off again...
838 :    
839 :     ----------------------------------------------------------------------
840 :     Name: Matthias Blume
841 : blume 937 Date: 2001/09/20 17:20:00 EDT
842 :     Tag: blume-20010920-canonicalpaths
843 :     Description:
844 :    
845 :     0. Updated the BOOT file (something that I forgot to do earlier).
846 :    
847 :     1. Small internal change to CM so that it avoids "/../" in filenames
848 :     as much as possible (but only where it is safe).
849 :    
850 :     2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
851 :     that arguments that contain delimiters are passed through correctly.
852 :     This change also means that all "special" arguments of the form
853 :     @SMLxxx... must come first.
854 :    
855 :     3. Changed install script to put relative anchor names for tool commands
856 :     into pathconfig.
857 :    
858 :     ----------------------------------------------------------------------
859 :     Name: Matthias Blume
860 : blume 932 Date: 2001/09/18 15:35:00 EDT
861 :     Tag: blume-20010918-readme11036
862 :     Description:
863 :    
864 :     Added README files.
865 :    
866 :     ----------------------------------------------------------------------
867 :     Name: Matthias Blume
868 : blume 929 Date: 2001/09/18 11:45:00 EDT
869 :     Tag: Release_110_36 (retag)
870 :     Description:
871 :    
872 :     Fixed mistake in config/preloads. Retagged as 110.36.
873 :    
874 :     ----------------------------------------------------------------------
875 :     Name: Matthias Blume
876 : blume 928 Date: 2001/09/18 09:40:00 EDT
877 : blume 929 Tag: Release_110_36_orig (tag changed)
878 : blume 928 Description:
879 :    
880 :     New version (110.36). New bootfiles.
881 :    
882 :     ----------------------------------------------------------------------
883 :     Name: Matthias Blume
884 : blume 927 Date: 2001/09/14 16:15:00 EDT
885 :     Tag: blume-20010914-x86fastfp
886 :     Description:
887 :    
888 :     John committed some changes that Allen made, in particular a (hopefully)
889 :     correctly working version of the x86-fp module.
890 :    
891 :     I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
892 :     flag to "true". Everything seems to compile to a fixpoint ok, and
893 :     "mandelbrot" speeds up by about 15%.
894 :    
895 :     ----------------------------------------------------------------------
896 :     Name: Matthias Blume
897 : blume 923 Date: 2001/09/13 11:20:00 EDT
898 :     Tag: blume-20010913-minimal
899 :     Description:
900 :    
901 :     1. Stefan Monnier's patch to fix a miscompilation problem that
902 :     was brought to light by John Reppy's work on Moby.
903 :    
904 :     2. Implemented a minimal "structure Compiler" that contains just
905 :     "version" and "architecture". The minimal version will be
906 :     available when the full version is not. This is for backward-
907 :     compatibility with code that wants to test Compiler.version.
908 :    
909 :     ----------------------------------------------------------------------
910 :     Name: Matthias Blume
911 : blume 920 Date: 2001/08/28 14:03:00 EDT
912 :     Tag: blume-20010828-ml-lex
913 :     Description:
914 :    
915 :     Fix for bug 1581, received from Neophytos Michael.
916 :    
917 :     ----------------------------------------------------------------------
918 :     Name: Matthias Blume
919 : blume 918 Date: 2001/08/27 11:20:00 EDT
920 :     Tag: blume-20010827-readme11035
921 :     Description:
922 :    
923 :     Fleshed out the README file for 110.35.
924 :    
925 :     ----------------------------------------------------------------------
926 :     Name: Matthias Blume
927 : blume 916 Date: 2001/08/24 17:10:00 EDT
928 :     Tag: Release_110_35
929 :     Description:
930 :    
931 :     New version number (110.35). New bootfiles.
932 :    
933 :     ----------------------------------------------------------------------
934 : george 915 Name: Lal George
935 : george 909 Date: 2001/08/24 13:47:18 EDT 2001
936 :     Tag: george-20010824-MLRISC-graphs
937 :     Description:
938 :    
939 :     removed clusters from MLRISC completely and replaced with graphs.
940 :    
941 :     ----------------------------------------------------------------------
942 :     Name: Matthias Blume
943 : blume 905 Date: 2001/08/23 17:50:00 EDT
944 :     Tag: blume-20010823-toplevel
945 :     Description:
946 :    
947 :     - some reorganization of the code that implements various kinds of
948 :     environments in the compiler (static, dynamic, symbolic, combined)
949 :     - re-implemented the EnvRef module so that evalStream works properly
950 :     (if the stream contains references to "use", "CM.make", etc.)
951 :     - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
952 :    
953 :     ----------------------------------------------------------------------
954 :     Name: Matthias Blume
955 : blume 904 Date: 2001/08/20 15:50 EDT
956 :     Tag: blume20010820-slipup
957 :     Description:
958 :    
959 :     I forgot to commit a few files. Here they are...
960 :    
961 :     ----------------------------------------------------------------------
962 :     Name: Matthias Blume
963 : blume 903 Date: 2001/08/20 15:35:00 EDT
964 :     Tag: blume-20010820-debugprof
965 :     Description:
966 :    
967 :     !!!! NEW BOOTFILES !!!!
968 :    
969 :     This is another round of reorganizing the compiler sources. This
970 :     time the main goal was to factor out all the "instrumentation"
971 :     passes (for profiling and backtracing) into their own library.
972 :     The difficulty was to do it in such a way that it does not depend
973 :     on elaborate.cm but only on elabdata.cm.
974 :    
975 :     Therefore there have been further changes to both elaborate.cm and
976 :     elabdata.cm -- more "generic" things have been moved from the former
977 :     to the latter. As a result, I was forced to split the assignment
978 :     of numbers indicating "primtyc"s into two portions: SML-generic and
979 :     SML/NJ-specific. Since it would have been awkward to maintain,
980 :     I bit the bullet and actually _changed_ the mapping between these
981 :     numbers and primtycs. The bottom line of this is that you need
982 :     a new set of bin- and bootfiles.
983 :    
984 :     I have built new bootfiles for all architectures, so doing a fresh
985 :     checkout and config/install.sh should be all you need.
986 :    
987 :     The newly created library's name is
988 :    
989 :     $smlnj/viscomp/debugprof.cm
990 :    
991 :     and its sources live under
992 :    
993 :     src/compiler/DebugProf
994 :    
995 :     ----------------------------------------------------------------------
996 :     Name: Matthias Blume
997 : blume 902 Date: 2001/08/15 17:15:00 EDT
998 :     Tag: blume-20010815-compreorg
999 :     Description:
1000 :    
1001 :     This is a first cut at reorganizing the CM libraries that make up the
1002 :     core of the compiler. The idea is to separate out pieces that could
1003 :     be used independently by tools, e.g., the parser, the typechecker, etc.
1004 :    
1005 :     The current status is a step in this direction, but it is not quite
1006 :     satisfactory yet. Expect more changes in the future.
1007 :    
1008 :     Here is the current (new) organization...
1009 :    
1010 :     What used to be $smlnj/viscomp/core.cm is now divided into
1011 :     six CM libraries:
1012 :    
1013 :     $smlnj/viscomp/basics.cm
1014 :     /parser.cm
1015 :     /elabdata.cm
1016 :     /elaborate.cm
1017 :     /execute.cm
1018 :     /core.cm
1019 :    
1020 :     The CM files for these libraries live under src/system/smlnj/viscomp.
1021 :     All these libraries are proxy libraries that contain precisely
1022 :     one CM library component. Here are the locations of the components
1023 :     (all within the src/compiler tree):
1024 :    
1025 :     Basics/basics.cm
1026 :     Parse/parser.cm
1027 :     ElabData/elabdata.cm
1028 :     Elaborator/elaborate.cm
1029 :     Execution/execute.cm
1030 :     core.cm
1031 :    
1032 :     [This organization is the same that has been used already
1033 :     for a while for the architecture-specific parts of the visible
1034 :     compiler and for the old version of core.cm.]
1035 :    
1036 :     As you will notice, many source files have been moved from their
1037 :     respective original locations to a new home in one of the above
1038 :     subtrees.
1039 :    
1040 :     The division of labor between the new libraries is the following:
1041 :    
1042 :     basics.cm:
1043 :     - Simple, basic definitions that pertain to many (or all) of
1044 :     the other libraries.
1045 :     parser.cm:
1046 :     - The SML parser, producing output of type Ast.dec.
1047 :     - The type family for Ast is also defined and exported here.
1048 :     elabdata.cm:
1049 :     - The datatypes that describe input and output of the elaborator.
1050 :     This includes types, absyn, and static environments.
1051 :     elaborator.cm:
1052 :     - The SML/NJ type checker and elaborator.
1053 :     This maps an Ast.dec (with a given static environment) to
1054 :     an Absyn.dec (with a new static environment).
1055 :     - This libraries implements certain modules that used to be
1056 :     structures as functors (to remove dependencies on FLINT).
1057 :     execute.cm:
1058 :     - Everything having to do with executing binary code objects.
1059 :     - Dynamic environments.
1060 :     core.cm:
1061 :     - SML/NJ-specific instantiations of the elaborator and MLRISC.
1062 :     - Top-level modules.
1063 :     - FLINT (this should eventually become its own library)
1064 :    
1065 :     Notes:
1066 :    
1067 :     I am not 100% happy with the way I separated the elaborator (and its
1068 :     data structures) from FLINT. Two instances of the same problem:
1069 :    
1070 :     1. Data structures contain certain fields that carry FLINT-specific
1071 :     information. I hacked around this using exn and the property list
1072 :     module from smlnj-lib. But the fact that there are middle-end
1073 :     specific fields around at all is a bit annoying.
1074 :    
1075 :     2. The elaborator calculates certain FLINT-related information. I tried
1076 :     to make this as abstract as I could using functorization, but, again,
1077 :     the fact that the elaborator has to perform calculations on behalf
1078 :     of the middle-end at all is not nice.
1079 :    
1080 :     3. Having to used exn and property lists is unfortunate because it
1081 :     weakens type checking. The other alternative (parameterizing
1082 :     nearly *everything*) is not appealing, though.
1083 :    
1084 :     I removed the "rebinding =" warning hack because due to the new organization
1085 :     it was awkward to maintain it. As a result, the compiler now issues some of
1086 :     these warnings when compiling init.cmi during bootstrap compilation. On
1087 :     the plus side, you also get a warning when you do, for example:
1088 :     val op = = Int32.+
1089 :     which was not the case up to now.
1090 :    
1091 :     I placed "assign" and "deref" into the _Core structure so that the
1092 :     code that deals with the "lazy" keyword can find them there. This
1093 :     removes the need for having access to the primitive environment
1094 :     during elaboration.
1095 :    
1096 :     ----------------------------------------------------------------------
1097 :     Name: Matthias Blume
1098 : blume 898 Date: 2001/08/13
1099 :     Tag: blume-20010813-closures
1100 :     Description:
1101 :    
1102 :     This fix was sent to us by Zhong Shao. It is supposed to improve the
1103 :     performance of certain loops by avoiding needless closure allocation.
1104 :    
1105 :     ----------------------------------------------------------------------
1106 : george 890 Name: Lal George
1107 : george 896 Date: 2001/07/31 10:03:23 EDT 2001
1108 :     Tag: george-20010731-x86-fmalloc
1109 :     Description: Fixed bug in x86 calls
1110 :    
1111 :     There was a bug where call instructions would mysteriously
1112 :     vanish. The call instruction had to be one that returned
1113 :     a floating point value.
1114 :    
1115 :     ----------------------------------------------------------------------
1116 :     Name: Lal George
1117 : george 890 Date: 2001/07/19 16:36:29 EDT 2001
1118 :     Tag: george-20010719-simple-cells
1119 :     Description:
1120 :    
1121 :     I have dramatically simplified the interface for CELLS in MLRISC.
1122 :    
1123 :     In summary, the cells interface is broken up into three parts:
1124 :    
1125 :     1. CellsBasis : CELLS_BASIS
1126 :    
1127 :     CellsBasis is a top level structure and common for all
1128 :     architectures. it contains the definitions of basic datatypes
1129 :     and utility functions over these types.
1130 :    
1131 :     2. functor Cells() : CELLS
1132 :    
1133 :     Cells generates an interface for CELLS that incorporates the
1134 :     specific resources on the target architecture, such as the
1135 :     presence of special register classes, their number and size,
1136 :     and various useful substructures.
1137 :    
1138 :     3. <ARCH>CELLS
1139 :    
1140 :     e.g. SparcCells: SPARCCELLS
1141 :    
1142 :     <ARCH>CELLS usually contains additional bindings for special
1143 :     registers on the architecture, such as:
1144 :    
1145 :     val r0 : cell (* register zero *)
1146 :     val y : cell (* Y register *)
1147 :     val psr : cell (* processor status register *)
1148 :     ...
1149 :    
1150 :     The structure returned by applying the Cells functor is opened
1151 :     in this interface.
1152 :    
1153 :     The main implication of all this is that the datatypes for cells is
1154 :     split between CellsBasis and CELLS -- a fairly simple change for user
1155 :     code.
1156 :    
1157 :     In the old scheme the CELLS interface had a definitional binding of
1158 :     the form:
1159 :    
1160 :     signature CELLS = sig
1161 :    
1162 :     structure CellsBasis = CellsBasis
1163 :    
1164 :     ...
1165 :    
1166 :     end
1167 :    
1168 :     With all the sharing constraints that goes on in MLRISC, this old
1169 :     design quickly leads to errors such as:
1170 :    
1171 :     "structure definition spec inside of sharing ... "
1172 :    
1173 :    
1174 :     and appears to require an unacceptable amount of sharing and where
1175 :     constraint hackery.
1176 :    
1177 :     I think this error message (the interaction of definitional specs and
1178 :     sharing) requires more explanation on our web page.
1179 :    
1180 :     ----------------------------------------------------------------------
1181 : blume 826 Name: Matthias Blume
1182 : blume 879 Date: 2001/07/19 15:00:00 EDT
1183 :     Tag: blume-20010719-libreorg
1184 :     Description:
1185 :    
1186 :     This update puts together a fairly extensive but straightforward change
1187 :     to the way the libraries that implement the interactive system are
1188 :     organized:
1189 :    
1190 :     The biggest change is the elimination of structure Compiler. As a
1191 :     replacement for this structure, there is now a CM library
1192 :     (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1193 :     that exports all the substructures of the original structure Compiler
1194 :     directly. So instead of saying Compiler.Foo.bar one now simply
1195 :     says Foo.bar. (The CM libraries actually export a collection of
1196 :     structures that is richer than the collection of substructures of
1197 :     structure Compiler.)
1198 :    
1199 :     To make the transition smooth, there is a separate library called
1200 :     $smlnj/compiler/compiler.cm which puts together and exports the
1201 :     original structure Compiler (or at least something very close to it).
1202 :    
1203 :     There are five members of the original structure Compiler
1204 :     that are not exported directly but which instead became members
1205 :     of a new structure Backend (described by signature BACKEND). These are:
1206 :     structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1207 :     Interact (: INTERACT), structure Machine (: MACHINE), and val
1208 :     architecture (: string).
1209 :    
1210 :     Structure Compiler.Version has become structure CompilerVersion.
1211 :    
1212 :     Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1213 :     by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1214 :     or x86, respectively.
1215 :     Each of these exports the same frontend structures that
1216 :     $smlnj/compiler.cm exports. But they do not have a structure Backend
1217 :     and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1218 :     Hppa, PPC, Sparc, or X86, respectively.
1219 :    
1220 :     Library $smlnj/compiler/all.cm exports the union of the exports of
1221 :     $smlnj/compiler/<arch>.cm
1222 :    
1223 :     There are no structures <Arch>Compiler anymore, use
1224 :     $smlnj/compiler/<arch>.cm instead.
1225 :    
1226 :     Library host-compiler-0.cm is gone. Instead, the internal library
1227 :     that instantiates CM is now called cm0.cm. Selection of the host
1228 :     compiler (backend) is no longer done here but. (Responsibility for it
1229 :     now lies with $smlnj/compiler/current.cm. This seems to be more
1230 :     logical.)
1231 :    
1232 :     Many individual files have been moved or renamed. Some files have
1233 :     been split into multiple files, and some "dead" files have been deleted.
1234 :    
1235 :     Aside from these changes to library organization, there are also changes
1236 :     to the way the code itself is organized:
1237 :    
1238 :     Structure Binfile has been re-implemented in such a way that it no
1239 :     longer needs any knowledge of the compiler. It exclusively deals
1240 :     with the details of binfile layout. It no longer invokes the
1241 :     compiler (for the purpose of creating new prospective binfile
1242 :     content), and it no longer has any knowledge of how to interpret
1243 :     pickles.
1244 :    
1245 :     Structure Compile (: COMPILE) has been stripped down to the bare
1246 :     essentials of compilation. It no longer deals with linking/execution.
1247 :     The interface has been cleaned up considerably.
1248 :    
1249 :     Utility routines for dealing with linking and execution have been
1250 :     moved into their own substructures.
1251 :    
1252 :     (The ultimate goal of these changes is to provide a light-weight
1253 :     binfile loader/linker (at least for, e.g., stable libraries) that
1254 :     does not require CM or the compiler to be present.)
1255 :    
1256 :     CM documentation has been updated to reflect the changes to library
1257 :     organization.
1258 :    
1259 :     ----------------------------------------------------------------------
1260 :     Name: Matthias Blume
1261 : blume 875 Date: 2001/07/10 17:30:00 EDT
1262 :     Tag: Release_110_34
1263 :     Description:
1264 :    
1265 : blume 876 Minor tweak to 110.34 (re-tagged):
1266 :    
1267 :     - README.html file added to CVS repository
1268 :     - runtime compiles properly under FreeBSD 3.X and 4.X
1269 :    
1270 :     ----------------------------------------------------------------------
1271 :     Name: Matthias Blume
1272 :     Date: 2001/07/10 17:30:00 EDT
1273 :     Tag: Release_110_34
1274 :     Description:
1275 :    
1276 : blume 875 New version number (110.34). New bootfiles.
1277 :    
1278 :     ----------------------------------------------------------------------
1279 :     Name: Matthias Blume
1280 : blume 874 Date: 2001/07/09 16:00:00 EDT
1281 :     Tag: blume-20010709-more-varargs
1282 :     Description:
1283 :    
1284 :     I changed the handling of varargs in ml-nlffigen again:
1285 :     The ellipsis ... will now simply be ignored (with an accompanying warning).
1286 :    
1287 :     The immediate effect is that you can actually call a varargs function
1288 :     from ML -- but you can't actually supply any arguments beyond the ones
1289 :     specified explicitly. (For example, you can call printf with its format
1290 :     string, but you cannot pass additional arguments.)
1291 :    
1292 :     This behavior is only marginally more useful than the one before, but
1293 :     it has the advantage that a function or, more importantly, a function
1294 :     type never gets dropped on the floor, thus avoiding follow-up problems with
1295 :     other types that refer to the offending one.
1296 :    
1297 :     ----------------------------------------------------------------------
1298 :     Name: Matthias Blume
1299 : blume 873 Date: 2001/07/09 11:25:00 EDT
1300 :     Tag: blume-20010709-varargs
1301 :     Description:
1302 :    
1303 :     1. ckit-lib.cm now exports structure Error
1304 :     2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1305 :     with a warning accompanied by a source location. Moreover, it
1306 :     merely skips the offending function or type and proceeds with the
1307 :     rest of its work.u As a result, one can safely feed C code containing
1308 :     "..." to ml-nlffigen.
1309 :     3. There are some internal improvements to CM, providing slightly
1310 :     more general string substitutions in the tools subsystem.
1311 :    
1312 :     ----------------------------------------------------------------------
1313 :     Name: Matthias Blume
1314 : blume 854 Date: 2001/06/27 15:10:00 EDT
1315 :     Tag: blume-20010627-concur
1316 :     Description:
1317 :    
1318 :     Fixed a small bug in CM's handling of parallel compilation.
1319 :     (You could observe the bug by Control-C-interrupting an ordinary
1320 :     CMB.make or CM.stabilize and then attaching some compile servers.
1321 :     The result was that all of a sudden the previously interrupted
1322 :     compilation would continue on its own. This was because of
1323 :     an over-optimization: CM did not bother to clean out certain queues
1324 :     when no servers were attached "anyway", resulting in the contents
1325 :     of these queues to grab control when new servers did get attached.)
1326 :    
1327 :     There is also another minor update to the CM manual.
1328 :    
1329 :     ----------------------------------------------------------------------
1330 :     Name: Matthias Blume
1331 : blume 853 Date: 2001/06/26 16:15:00 EDT
1332 :     Tag: blume-20010626-cmdoc
1333 :     Description:
1334 :    
1335 :     Minor typo fixed in CM manual (syntax diagram for libraries).
1336 :    
1337 :     ----------------------------------------------------------------------
1338 :     Name: Matthias Blume
1339 : blume 852 Date: 2001/06/25 22:55:00 EDT
1340 :     Tag: blume-20010625-x86pc
1341 :     Description:
1342 :    
1343 :     Fixed a nasty bug in the X86 assembly code that caused signal
1344 :     handlers to fail (crash) randomly.
1345 :    
1346 :     ----------------------------------------------------------------------
1347 :     Name: Matthias Blume
1348 : blume 846 Date: 2001/06/25 12:05:00 EDT
1349 :     Tag: blume-20010625-nlffigen
1350 :     Description:
1351 :    
1352 :     This update fixes a number of minor bugs in ml-nlffigen as reported by
1353 :     Nick Carter <nbc@andrew.cmu.edu>.
1354 :    
1355 :     1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1356 :     2. Default names for generated files are now derived from the name of
1357 :     the C file *without its directory*. In particular, this causes generated
1358 :     files to be placed locally even if the C file is in some system directory.
1359 :     3. Default names for generated signatures and structures are also derived
1360 :     from the C file name without its directory. This avoids silly things
1361 :     like "structure GL/GL".
1362 :     (Other silly names are still possible because ml-nlffigen does not do
1363 : blume 847 a thorough check of whether generated names are legal ML identifiers.
1364 :     When in doubt, use command line arguments to force particular names.)
1365 : blume 846
1366 :     ----------------------------------------------------------------------
1367 :     Name: Matthias Blume
1368 : blume 845 Date: 2001/06/21 12:25:00 EDT
1369 :     Tag: blume-20010621-eXene
1370 :     Description:
1371 :    
1372 :     eXene now compiles and (sort of) works again.
1373 :    
1374 :     The library name (for version > 110.33) is $/eXene.cm.
1375 :    
1376 :     I also added an new example in src/eXene/examples/nbody. See the
1377 :     README file there for details.
1378 :    
1379 :     ----------------------------------------------------------------------
1380 :     Name: Matthias Blume
1381 : blume 844 Date: 2001/06/20 16:40:00 EDT
1382 :     Tag: blume-20010620-cml
1383 :     Description:
1384 :    
1385 :     CML now compiles and works again.
1386 :    
1387 :     Libraries (for version > 110.33):
1388 :    
1389 :     $cml/cml.cm Main CML library.
1390 :     $cml/basis.cm CML's version of $/basis.cm.
1391 :     $cml/cml-internal.cm Internal helper library.
1392 :     $cml/core-cml.cm Internal helper library.
1393 :     $cml-lib/trace-cml.cm Tracing facility.
1394 :     $cml-lib/smlnj-lib.cm CML's version of $/smlnj-lib.cm
1395 :    
1396 :     The installer (config/install.sh) has been taught how to properly
1397 :     install this stuff.
1398 :    
1399 :     ----------------------------------------------------------------------
1400 :     Name: Matthias Blume
1401 : blume 843 Date: 2001/06/19 17:55:00 EDT
1402 :     Tag: blume-20010619-instantiate
1403 :     Description:
1404 :    
1405 :     This un-breaks the fix for bug 1432.
1406 :     (The bug was originally fixed in 110.9 but I broke it again some
1407 :     time after that.)
1408 :    
1409 :     ----------------------------------------------------------------------
1410 :     Name: Matthias Blume
1411 : blume 842 Date: 2001/06/19 17:25:00 EDT
1412 :     Tag: blume-20010619-signals
1413 :     Description:
1414 :    
1415 :     This should (hopefully) fix the long-standing signal handling bug.
1416 :     (The runtime system was constructing a continuation record with an
1417 :     incorrect descriptor which would cause the GC to drop data on the floor...)
1418 :    
1419 :     ----------------------------------------------------------------------
1420 :     Name: Matthias Blume
1421 : blume 840 Date: 2001/06/15 15:05:00 EDT
1422 : blume 841 Tag: blume-20010615-moresparc
1423 :     Description:
1424 :    
1425 :     Here is a short late-hour update related to Sparc c-calls:
1426 :    
1427 :     -- made handling of double-word arguments a bit smarter
1428 :    
1429 :     -- instruction selection phase tries to collapse certain clumsily
1430 :     constructed ML-Trees; typical example:
1431 :    
1432 :     ADD(ty,ADD(_,e,LI d1),LI d2) -> ADD(ty,e,LI(d1+d2))
1433 :    
1434 :     This currently has no further impact on SML/NJ since mlriscGen does
1435 :     not seem to generate such patterns in the first place, and c-calls
1436 :     (which did generate them in the beginning) has meanwhile been fixed
1437 :     so as to avoid them as well.
1438 :    
1439 :     ----------------------------------------------------------------------
1440 :     Name: Matthias Blume
1441 :     Date: 2001/06/15 15:05:00 EDT
1442 : blume 840 Tag: blume-20010615-sparc
1443 :     Description:
1444 :    
1445 :     The purpose of this update is to provide an implementation of NLFFI
1446 :     on Sparc machines.
1447 :    
1448 :     Here are the changes in detail:
1449 :    
1450 :     * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1451 :     the Sparc implementation of the c-calls API.
1452 :     * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1453 :     for accessing the ML frame. Thus, we have a real frame pointer and
1454 :     can freely modify %sp without need for an omit-frame-ptr phase.
1455 :     The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1456 :     this case.
1457 :     * ml-nlffigen has been taught to produce code for different architectures
1458 :     and calling conventions.
1459 :     * In a way similar to what was done in the x86 case, the Sparc
1460 :     backend uses its own specific extension to mltree. (For example,
1461 :     it needs to be able to generate UNIMP instructions which are part
1462 :     of the calling convention.)
1463 :     * ml-nlffi-lib was reorganized to make it more modular (in particular,
1464 :     to make it easier to plug in new machine- and os-dependent parts).
1465 :    
1466 :     There are some other fairly unrelated bug fixes and cleanups as well:
1467 :    
1468 :     * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1469 :     that they properly share their libraries with existing SML/NJ libraries.
1470 :     * I fixed a minor cosmetic bug in CM, supressing certain spurious
1471 :     follow-up error messages.
1472 :     * Updates to CM/CMB documentation.
1473 :    
1474 :     TODO items:
1475 :    
1476 :     * MLRISC should use a different register as its asmTemp on the Sparc.
1477 :     (The current %o2 is a really bad choice because it is part of the
1478 :     calling conventions, so things might interfere in unexpected ways.)
1479 :    
1480 :     ----------------------------------------------------------------------
1481 :     Name: Matthias Blume
1482 : blume 839 Date: 2001/06/07
1483 :     Tag: blume-20010607-calls
1484 :     Description:
1485 :    
1486 :     A number of internal changes related to C calls and calling conventions:
1487 :    
1488 :     1. ML-Tree CALL statements now carry a "pops" field. It indicates the
1489 :     number of bytes popped implicitly (by the callee). In most cases
1490 :     this field is 0 but on x86/win32 it is some non-zero value. This
1491 :     is information provided for the benefit of the "omit-frameptr" pass.
1492 :     2. The CALL instruction on the x86 carries a similar "pops" field.
1493 :     The instruction selection phase copies its value from the ML-Tree
1494 :     CALL statement.
1495 :     3. On all other architectures, the instruction selection phase checks
1496 :     whether "pops=0" and complains if not.
1497 :     4. The c-calls implementation for x86 now accepts two calling conventions:
1498 :     "ccall" and "stdcall". When "ccall" is selected, the caller cleans
1499 :     up after the call and pops is set to 0. For "stdcall", the caller
1500 :     does nothing, leaving the cleanup to the callee; pops is set to
1501 :     the number of bytes that were pushed onto the stack.
1502 :     5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1503 :     distinguish between "ccall" and "stdcall".
1504 :     6. The UNIMP instruction has been added to the supported Sparc instruction
1505 :     set. (This is needed for implementing the official C calling convention
1506 :     on this architecture.)
1507 :     7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1508 :     work with the latest CM.
1509 :    
1510 :     ----------------------------------------------------------------------
1511 :     Name: Matthias Blume
1512 : blume 838 Date: 2001/06/05 15:10:00 EDT
1513 :     Tag: blume-20010605-cm-index
1514 :     Description:
1515 :    
1516 :     0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1517 :    
1518 :     1. CM can now generate "index files". These are human-readable files
1519 :     that list on a per-.cm-file basis each toplevel symbol defined or
1520 :     imported. The location of the index file for
1521 :     <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1522 :     To enable index-file generation, set CM.Control.generate_index to true
1523 :     or export an environment-symbol: export CM_GENERATE_INDEX=true.
1524 :    
1525 :     The CM manual has been updated accordingly.
1526 :    
1527 :     2. I made some slight modifications to the c-calls API in MLRISC.
1528 :    
1529 :     a) There is now a callback to support saving/restoring of
1530 :     dedicated but caller-save registers around the actual call
1531 :     instruction.
1532 :     b) One can optionally specify a comment-annotation for the
1533 :     call instruction.
1534 :    
1535 :     3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1536 :     (For example, the comment annotation shows the C prototype of
1537 :     the function being called.)
1538 :    
1539 :     ----------------------------------------------------------------------
1540 :     Name: Matthias Blume
1541 : blume 837 Date: 2001/06/01 13:30:00 EDT
1542 :     Tag: blume-20010601-nlffi-cleanup
1543 :     Description:
1544 :    
1545 :     This is mostly a cleanup of MLFFI stuff:
1546 :    
1547 :     - some signature files have been put into a more exposed place
1548 :     - the ugly 'f type parameter is gone (simplifies types tremendously!)
1549 :     - ml-nlffigen changed accordingly
1550 :     - tutorial updated
1551 :    
1552 :     Other changes:
1553 :    
1554 :     - author's affiliation in CM manual(s) updated
1555 :     - some more recognized keywords added to Allen's sml.sty
1556 :    
1557 :     ----------------------------------------------------------------------
1558 :     Name: Matthias Blume
1559 : blume 836 Date: 2001/05/25 15:30:00 EDT
1560 :     Tag: blume-20010525-iptr
1561 :     Description:
1562 :    
1563 :     - put the official 110.33-README (as it appears on the ftp server) under
1564 :     CVS
1565 :     - fixed a small bug related to incomplete pointer types in
1566 :     ml-nlffigen
1567 :     - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1568 :     (it does not need the 'f type parameter)
1569 :    
1570 :     ----------------------------------------------------------------------
1571 :     Name: Matthias Blume
1572 : blume 834 Date: 2001/05/23 14:30:00 EDT
1573 :     Tag: Release_110_33
1574 :     Description:
1575 :    
1576 :     New version number (110.33). New bootfiles.
1577 :    
1578 :     ----------------------------------------------------------------------
1579 :     Name: Matthias Blume
1580 : blume 833 Date: 2001/05/22 18:06:00 EDT
1581 :     Tag: blume-20010522-targets
1582 :     Description:
1583 :    
1584 :     Made install.sh use file config/targets.customized if it exists, falling
1585 :     back to config/targets if it doesn't. This way one can have a customized
1586 :     version of the targets file without touching the "real thing", thus
1587 :     eliminating the constant fear of accidentally checking something bogus
1588 :     back into the CVS repository... (File config/targets.customized must
1589 :     not be added to the repository!)
1590 :    
1591 :     ----------------------------------------------------------------------
1592 :     Name: Matthias Blume
1593 : blume 832 Date: 2001/05/22 16:30:00 EDT
1594 :     Tag: blume-20010522-minitut
1595 :     Description:
1596 :    
1597 :     1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1598 :     struct returns.
1599 :     2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt. This is some very
1600 :     incomplete, preliminary documentation for NLFFI.
1601 :    
1602 :     ----------------------------------------------------------------------
1603 :     Name: Matthias Blume
1604 : blume 829 Date: 2001/05/14 11:30:00 EDT
1605 :     Tag: blume-20010514-script
1606 :     Description:
1607 :    
1608 :     Some bugs in install script fixed.
1609 :    
1610 :     In addition to that I also made a slight change to the NLFFI API:
1611 :     Functors generated by ml-nlffigen now take the dynamic library as a
1612 :     straight functor argument, not as a suspended one. (The original
1613 :     functor code used to force the suspension right away anyway, so there
1614 :     was nothing gained by this complication of the interface.)
1615 :    
1616 :     ----------------------------------------------------------------------
1617 :     Name: Matthias Blume
1618 : blume 828 Date: 2001/05/11 14:35:00 EDT
1619 :     Tag: blume-20010511-ml-nlffi
1620 :     Description:
1621 :    
1622 :     I finally took the plunge and added my new FFI code to the main
1623 :     repository. For x86-linux it is now ready for prime-time.
1624 :    
1625 :     There are two new subdirectories of "src":
1626 :    
1627 :     - ml-nlffi-lib:
1628 :     The utility library for programs using the FFI interface.
1629 :     Here is the implementation of $/c.cm and its associated low-level
1630 :     partners $/c-int.cm and $/memory.cm.
1631 :     - ml-nlffigen:
1632 :     A stand-alone program for generating ML glue code from C source
1633 :     code.
1634 :    
1635 :     Building ml-nlffigen requires $/ckit-lib.cm.
1636 :    
1637 :     The config/install.sh script has been updates to do the Right Thing
1638 :     (hopefully).
1639 :    
1640 :     Notice that the source tree for the C-Kit will not be put under "src"
1641 :     but directly under the installation root directory. (This is the
1642 :     structure that currently exists on the CVS server when you check out
1643 :     module "sml".) Fortunately, config/install.sh knows about this oddity.
1644 :    
1645 :     Bugs: No documentation yet.
1646 :    
1647 :     ----------------------------------------------------------------------
1648 :     Name: Matthias Blume
1649 : blume 826 Date: 2001/05/09 16:35:00 EDT
1650 :     Tag: blume-20010509-cpscontract
1651 :     Description:
1652 :    
1653 :     Fixed a bug in the accounting code in cpsopt/contract.sml. (The
1654 :     wrapper/unwrapper elimination did not decrement usage counts and some
1655 :     dead variables got overlooked by the dead-up logic.)
1656 :    
1657 :     ----------------------------------------------------------------------
1658 : george 824 Name: Lal George
1659 :     Date: 2001/05/08 17:26:09 EDT
1660 :     Tag: george-20010508-omit-frameptr
1661 :     Description:
1662 :    
1663 :     Changes to implement the omit-frame-pointer optimization to support
1664 :     raw C calls. For now, there is only support on the Intel x86, but
1665 :     other architectures will follow as more experience is gained with this.
1666 :    
1667 :    
1668 :     ----------------------------------------------------------------------
1669 : blume 818 Name: Matthias Blume
1670 : blume 819 Date: 2001/05/07 14:40:00 EDT
1671 :     Tag: blume-20010507-proxies
1672 :     Description:
1673 :    
1674 :     I made into "proxy libraries" all libraries that qualify for such a
1675 :     change. (A qualifying library is a library that has another library or
1676 :     groups as its sole member and repeats that member's export list
1677 :     verbatim. A proxy library avoids this repetition by omitting its export
1678 :     list, effectively inheriting the list that its (only) member exports.
1679 :     See the CM manual for more explanation.)
1680 :     The main effect is that explicit export lists for these libraries
1681 :     do not have to be kepts in sync, making maintenance a bit easier.
1682 :    
1683 :     I also added copyright notices to many .cm-files.
1684 :    
1685 :     Last but not least, I made a new set of bootfiles.
1686 :    
1687 :     ----------------------------------------------------------------------
1688 :     Name: Matthias Blume
1689 : blume 818 Date: 2001/05/04 17:00:00 EDT
1690 :     Tag: blume-20010504-cm-lsplit
1691 :     Description:
1692 :    
1693 :     0. John merged pending changes to $/smlnj-lib.cm
1694 :    
1695 :     1. Allen's previous change accidentally backed out of one of Lal's
1696 :     earlier changes. I undid this mistake (re-introducing Lal's change).
1697 :    
1698 :     2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1699 :     within the compiler where applicable. There is some code simplification
1700 :     because of that.
1701 :    
1702 :     3. The "split" phase (in FLINT) is now part of the default list of phases.
1703 :     Compiler.Control.LambdaSplitting.* can be used to globally control the
1704 :     lambda-splitting (cross-module-inlining) engine. In addition to that,
1705 :     it can now also be controlled on a per-source basis: CM has been taught
1706 :     a new tool parameter applicable to ML source files.
1707 :    
1708 :     - To turn lambda-splitting off completely:
1709 :     local open Compiler.Control.LambdaSplitting in
1710 :     val _ = set Off
1711 :     end
1712 :     - To make "no lambda-splitting" the global default (but allow per-source
1713 :     overriding); this is the initial setting:
1714 :     local open Compiler.Control.LambdaSplitting in
1715 :     val _ = set (Default NONE)
1716 :     end
1717 :     - To make "lambda-splitting with aggressiveness a" the global default
1718 :     (and allow per-source overriding):
1719 :     local open Compiler.Control.LambdaSplitting in
1720 :     val _ = set (Default (SOME a))
1721 :     end
1722 :    
1723 :     - To turn lambda-splitting off for a given ML souce file (say: a.sml)
1724 :     write (in the respective .cm-file):
1725 :     a.sml (lambdasplitting:off)
1726 :     - To turn lambda-splitting for a.sml on with minimal aggressiveness:
1727 :     a.sml (lambdasplitting:on)
1728 :     - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1729 :     <a> is a decimal non-negative integer):
1730 :     a.sml (lambdasplitting:<a>)
1731 :     - To turn lambda-splitting for a.sml on with maximal aggressiveness:
1732 :     a.sml (lambdasplitting:infinity)
1733 :     - To use the global default for a.sml:
1734 :     a.sml (lambdasplitting:default)
1735 :     or simply
1736 :     a.sml
1737 :    
1738 :     ----------------------------------------------------------------------
1739 : leunga 815 Name: Allen Leung
1740 :     Date: 2001/05/04 01:57:00 EDT
1741 :     Tag: leunga-20010504-sync
1742 :     Description:
1743 : leunga 788
1744 : leunga 815 MLRISC features.
1745 :    
1746 :     1. Fix to CMPXCHG instructions.
1747 :     2. Changed RA interface to allow annotations in callbacks.
1748 :     3. Added a new method to the stream interface to allow annotations updates.
1749 :    
1750 : george 749 ----------------------------------------------------------------------
1751 : blume 812 Name: Matthias Blume
1752 : blume 814 Date: 2001/05/01 11:45:00 EDT
1753 :     Tag: blume-20010501-pcedittmp
1754 :     Description:
1755 :    
1756 :     Changed install.sh to use the current working directory instead of
1757 :     /usr/tmp for a temporary file (pcedittmp). The previous choice
1758 :     of /usr/tmp caused trouble with MacOS X because of file premission
1759 :     problems.
1760 :    
1761 :     ----------------------------------------------------------------------
1762 :     Name: Matthias Blume
1763 : blume 812 Date: 2001/04/20 11:10:00 EDT
1764 :     Tag: blume-20010420-inMLflag
1765 :     Description:
1766 :    
1767 :     - added vp_limitPtrMask to vproc-state.h
1768 :     (for use by the raw-C-calls mechanism to implement proper interrupt
1769 :     handling)
1770 :     - made the ML compiler aware of various data-structure offsets so it
1771 :     can generate code for accessing the vp_inML flag and vp_limitPtrMask
1772 :     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1773 :     raw C-calls
1774 :    
1775 :     ----------------------------------------------------------------------
1776 : george 810 Name: Lal George
1777 :     Date: 2001/04/20 09:15:28 EDT
1778 :     Tag: george-20010420-macosX
1779 :     Description:
1780 :    
1781 :     - Changes to port to Mac OS X; Darwin.
1782 :    
1783 :     - In the process I found that sqrt was broken on the PPC, because the
1784 :     fsqrt instruction is not implemented.
1785 :    
1786 :     ----------------------------------------------------------------------
1787 : blume 799 Name: Matthias Blume
1788 : blume 808 Date: 2001/04/18 12:45:00 EDT
1789 :     Tag: blume-20010418-ccalls
1790 :     Description:
1791 :    
1792 :     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1793 :     (this bug prevented structure arguments containing pointers from being
1794 :     passed correctly)
1795 :     - changed the raw-C-call code in mlriscGen.sml in such a way that
1796 :     structure arguments are represented as a pointer to the beginning
1797 :     of the structure (instead of having a series of synthesized arguments,
1798 :     one for each structure member)
1799 :    
1800 :     - made makeml script's verbosity level configurable via environment
1801 :     variable (MAKEML_VERBOSITY)
1802 :    
1803 : blume 813 - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1804 : blume 808 in rawmem-x86.sml; we are now using the real thing
1805 :    
1806 :     ----------------------------------------------------------------------
1807 :     Name: Matthias Blume
1808 : blume 806 Date: 2001/03/22 16:25:00 EST
1809 :     Tag: blume-20010322-bootfiles
1810 :     Description:
1811 :    
1812 :     Created a new set of bootfiles (for your automatic installation convenience).
1813 :    
1814 :     ----------------------------------------------------------------------
1815 :     Name: Matthias Blume
1816 : blume 805 Date: 2001/03/22 15:10:00 EST
1817 :     Tag: blume-20010322-rawmem-parcm
1818 :     Description:
1819 :    
1820 :     1. All "raw memory access" primitives for the new FFI are implemented now
1821 :     (at least on the x86).
1822 :     2. Some further cleanup of CM's parallel make mechanism.
1823 :    
1824 :     ----------------------------------------------------------------------
1825 :     Name: Matthias Blume
1826 : blume 801 Date: 2001/03/19 17:53:00 EST
1827 :     Tag: blume-20010319-parallel
1828 :     Description:
1829 :    
1830 :     Parallel make (using compile servers) now works again.
1831 :    
1832 :     To this end, CM.stabilize and CMB.make have been modified to work in
1833 :     two passes when compile servers are attached:
1834 :     1. Compile everything, do not perform stabilization; this pass
1835 :     uses compile servers
1836 :     2. Stabilize everything; this pass does not use compile servers
1837 :     If there are no compile servers, the two passes are combined into one
1838 :     (as before). Splitting the passes increases the inherent parallelism
1839 :     in the dependency graph because the entire graph including all
1840 :     libraries is available at the same time. This, in turn, improves
1841 :     server utilization. The downside is that the master process will
1842 :     have to do some extra work after compilation is done (because for
1843 :     technical reasons it must re-read all the binfiles during stabilization).
1844 :    
1845 :     ----------------------------------------------------------------------
1846 :     Name: Matthias Blume
1847 : blume 800 Date: 2001/03/16 12:22:00 EST
1848 :     Tag: blume-20010316-bootfiles
1849 :     Description:
1850 :    
1851 :     Created a new set of bootfiles (for your automatic installation convenience).
1852 :    
1853 :     ----------------------------------------------------------------------
1854 :     Name: Matthias Blume
1855 : blume 799 Date: 2001/03/16 11:00:00 EST
1856 :     Tag: blume-20010316-MLTREE-fixup
1857 :     Description:
1858 :    
1859 :     This is a minor fixup for an (untagged) earlier commit by Allen.
1860 :     (A file was missing).
1861 :    
1862 :     ----------------------------------------------------------------------
1863 : leunga 788 Name: Allen Leung
1864 : leunga 796 Date: Mon Mar 5 18:54:57 EST 2001
1865 :     Tag: leunga-20010305-cut-support
1866 :    
1867 :     1. New support for alternative control-flow in MLTREE.
1868 :     Currently we support
1869 :    
1870 :     FLOW_TO(CALL ...., [k1,...,kn])
1871 :    
1872 :     This is needed for 'cuts to' in C-- and try/handle-like constructs
1873 :     in Moby
1874 :    
1875 :     New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1876 :    
1877 :     2. Register Allocator
1878 :    
1879 :     Changes in interface [from Fermin, John]
1880 :    
1881 :     3. Alpha 8-bit SLL support [Fermin]
1882 :    
1883 :     4. All architectures
1884 :    
1885 :     A new module (ClusterExpandCopies) for expanding parallel copies.
1886 :    
1887 :     ----------------------------------------------------------------------
1888 :     Name: Allen Leung
1889 : leunga 788 Date: 2001/02/27 23:07:00 EST
1890 :     Tag: leunga-20010227-minor-stuff
1891 :    
1892 :     1. Alpha bug fix for CMOVNE
1893 :     2. Handle mltree COND(..,FCMP ...,...)
1894 :     3. Bug fix in simplifier
1895 :    
1896 :     ----------------------------------------------------------------------
1897 : blume 777 Name: Matthias Blume
1898 : blume 781 Date: 2001/01/30 17:50:00 EST
1899 :     Tag: blume-20010130-sync
1900 :     Description:
1901 :    
1902 :     This is just a minor update to sync my devel branch with the main brach.
1903 :     The only visible change is the addition of some README files.
1904 :    
1905 :     ----------------------------------------------------------------------
1906 :     Name: Matthias Blume
1907 : blume 778 Date: 2001/01/12 23:30:00 JST
1908 :     Tag: blume-20010112-bootfiles
1909 :     Description:
1910 :    
1911 :     Made a new set of bootfiles that goes with the current state of the
1912 :     repository.
1913 :    
1914 :     ----------------------------------------------------------------------
1915 :     Name: Matthias Blume
1916 : blume 777 Date: 2001/01/12 21:20:00 JST
1917 :     Tag: blume-20010112-sync
1918 :     Description:
1919 :    
1920 :     I am just flushing out some minor changes that had accumulated in
1921 :     my private branch in order to sync with the main tree. (This is
1922 :     mainly because I had CVS trouble when trying to merge _into_ my
1923 :     private branch.)
1924 :    
1925 :     Most people should be completely unaffected by this.
1926 :    
1927 :     ----------------------------------------------------------------------
1928 : leunga 775 Name: Allen Leung
1929 :     Date: Thu Jan 11 21:03:00 EST 2001
1930 :     Tag: leunga-20010111-labexp=mltree
1931 :     Description:
1932 :    
1933 :     1. Removed the type LabelExp and replace it by MLTree.
1934 :     2. Rewritten mltree-simplify with the pattern matcher tool.
1935 :     3. There were some bugs in alpha code generator which would break
1936 :     64-bit code generation.
1937 :     4. Redo the tools to generate code with the
1938 :     5. The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1939 :     are now generated by perl scripts.
1940 :    
1941 :     ----------------------------------------------------------------------
1942 : blume 771 Name: Matthias Blume
1943 : blume 774 Date: 2001/01/10 21:55:00 JST
1944 :     Tag: blume-20010110-rcc
1945 :     Description:
1946 :    
1947 :     The RCC stuff now seems to work (but only on the x86).
1948 :     This required hacking of the c-calls interface (and -implementation) in
1949 :     MLRISC.
1950 :    
1951 :     Normal compiler users should be unaffected.
1952 :    
1953 :     ----------------------------------------------------------------------
1954 :     Name: Matthias Blume
1955 : blume 773 Date: 2001/01/09 01:20:00 JST
1956 :     Tag: blume-20010109-rcc
1957 :     Description:
1958 :    
1959 :     This is a fairly big patch, flushing out a large number of pending
1960 :     changes that I made to my development copy over the last couple of days.
1961 :    
1962 :     Of practical relevance at this moment is a workaround for a pickling
1963 :     bug that Allen ran into the other day. The cause of the bug itself is
1964 :     still unknown and it might be hard to fix it properly, but the
1965 :     workaround has some merits of its own (namely somewhat reducing pickling
1966 :     overhead for certain libraries). Therefore, I think this solution should
1967 :     be satisfactory at this time.
1968 :    
1969 :     The rest of the changes (i.e., the vast majority) has to do with my
1970 :     ongoing efforts of providing direct support for C function calls from
1971 :     ML. At the moment there is a new primop "RAW_CCALL", typing magic
1972 :     in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1973 :     case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1974 :     translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1975 :     handle RCC.
1976 :    
1977 :     The last part (the changes to mlriscGen.sml) are still known to be
1978 :     wrong on the x86 and not implemented on all other architectures. But
1979 :     the infrastructure is in place. I had to change a few functor
1980 :     signatures in the backend to be able to route the CCalls interface
1981 :     from MLRISC there, and I had to specialize the mltree type (on the
1982 :     x86) to include the necessary extensions. (The extensions themselves
1983 :     were already there and redy to go in MLRISC/x86).
1984 :    
1985 :     Everything should be very happy as soon as someone helps me with
1986 :     mlriscGen.sml...
1987 :    
1988 :     In any case, nothing of this should matter to anyone as long as the
1989 :     new primop is not being used (which is going to be the case unless you
1990 :     find it where I hid it :). The rest of the compiler is completely
1991 :     unaffected.
1992 :    
1993 :     ----------------------------------------------------------------------
1994 :     Name: Matthias Blume
1995 : blume 772 Date: 2001/01/05 00:30:00 JST
1996 :     Tag: blume-20010105-primops
1997 :     Description:
1998 :    
1999 :     Added some experimental support for work that I am doing right now.
2000 :     These changes mostly concern added primops, but there is also a new
2001 :     experimental C library in the runtime system (but currently not enabled
2002 :     anywhere except on Linux/X86).
2003 :    
2004 :     In the course of adding primops (and playing with them), I discovered that
2005 :     Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
2006 :     broken. (Zhong was very right he labeled this stuff as "major gross hack".)
2007 :     To recover, I made type information in INL_PRIM mandatory and changed
2008 :     prim.sml as well as built-in.sml accordingly. The InLine structure now
2009 :     has complete, correct type information (i.e., no bottom types).
2010 :    
2011 :     Since all these changes mean that we need new binfiles, I also bumped the
2012 :     version number to 110.32.1.
2013 :    
2014 :     ----------------------------------------------------------------------
2015 :     Name: Matthias Blume
2016 : blume 771 Date: 2000/12/30 22:10:00 JST
2017 :     Tag: blume-20001230-various
2018 :     Description:
2019 :    
2020 :     Added proxy libraries for MLRISC and let MLRISC libraries refer
2021 :     to each other using path anchors. (See CM manual for explanation.)
2022 :    
2023 :     Updated CM documentation.
2024 :    
2025 :     Fixed some bugs in CM.
2026 :    
2027 :     Implemented "proxy" libraries (= syntactic sugar for CM).
2028 :    
2029 :     Added "-quiet" option to makeml and changed runtime system accordingly.
2030 :    
2031 :     Added cleanup handler for exportML to reset timers and compiler stats.
2032 :    
2033 :     ----------------------------------------------------------------------
2034 : george 761 Name: Lal George
2035 :     Date: 2000/12/22 22:22:58 EST 2000
2036 :     Tag: Release_110_32
2037 :     Description:
2038 :    
2039 :     Infinite precision used throughout MLRISC.
2040 :     see MLRISC/mltree/machine-int.sig
2041 :    
2042 :     ----------------------------------------------------------------------
2043 : blume 756 Name: Matthias Blume
2044 : blume 760 Date: 2000/12/22 23:16:00 JST
2045 :     Tag: blume-20001222-warn
2046 :     Description:
2047 : blume 759
2048 : blume 760 Corrected wording and formatting of some CM warning message which I
2049 :     broke in my previous patch.
2050 : blume 759
2051 :     ----------------------------------------------------------------------
2052 :     Name: Matthias Blume
2053 : blume 763 Date: 2000/12/22 21:20:00 JST
2054 :     Tag: blume-20001222-anchorenv
2055 :     Description:
2056 :    
2057 :     Fixed CM's handling of anchor environments in connection with CMB.make.
2058 :    
2059 :     ----------------------------------------------------------------------
2060 :     Name: Matthias Blume
2061 : blume 758 Date: 2000/12/22 13:15:00 JST
2062 :     Tag: blume-20001222-cleanup
2063 :     Description:
2064 :    
2065 :     Removed src/cm/ffi which does not (and did not) belong here.
2066 :    
2067 :     ----------------------------------------------------------------------
2068 :     Name: Matthias Blume
2069 : blume 757 Date: 2000/12/21 23:55:00 JST
2070 :     Tag: blume-20001221-exn
2071 :     Description:
2072 :    
2073 :     Probably most important: CM no longer silently swallows all exceptions
2074 :     in the compiler.
2075 :     Plus: some other minor CM changes. For example, CM now reports some
2076 :     sizes for generated binfiles (code, data, envpickle, lambdapickle).
2077 :    
2078 :     ----------------------------------------------------------------------
2079 :     Name: Matthias Blume
2080 : blume 756 Date: 2000/12/15 00:01:05 JST
2081 :     Tag: blume-20001215-dirtool
2082 :     Description:
2083 :    
2084 :     - "dir" tool added.
2085 :     - improvements and cleanup to Tools structure
2086 :     - documentation updates
2087 :    
2088 :     ----------------------------------------------------------------------
2089 : leunga 752 Name: Allen Leung
2090 : leunga 755 Date: Thu Dec 14 03:45:24 EST 2000
2091 :     Description:
2092 :     Tag: leunga-20001214-int-inf
2093 :     Description:
2094 :    
2095 :     In IntInf, added these standard functions, which are missing from our
2096 :     implementation:
2097 :    
2098 :     andb : int * int -> int
2099 :     xorb : int * int -> int
2100 :     orb : int * int -> int
2101 :     notb : int -> int
2102 :     << : int * word -> int
2103 :     ~>> : int * word -> int
2104 :    
2105 :     Not tested, I hope they are correct.
2106 :    
2107 :     ----------------------------------------------------------------------
2108 :     Name: Allen Leung
2109 : leunga 752 Date: Fri Dec 8 19:23:26 EST 2000
2110 :     Description:
2111 :     Tag: leunga-20001208-nowhere
2112 :     Description:
2113 :    
2114 :     Slight improvements to the 'nowhere' tool to handle OR-patterns,
2115 :     to generate better error messages etc. Plus a brief manual.
2116 :    
2117 :     ----------------------------------------------------------------------
2118 : george 749 Name: Lal George
2119 :     Date: 2000/12/08 09:54:02 EST 2000
2120 :     Tag: Release_110_31
2121 :     Description:
2122 : george 705
2123 : george 749 - Version 110.31
2124 : leunga 591 ----------------------------------------------------------------------
2125 : leunga 744 Name: Allen Leung
2126 :     Date: Thu Dec 7 22:01:04 EST 2000
2127 :     Tag: leunga-20001207-cell-monster-hack
2128 :     Description:
2129 :    
2130 :     Major MLRISC internal changes. Affect all clients.
2131 :     Summary:
2132 :    
2133 :     1. Type CELLS.cell = int is now replaced by a datatype.
2134 :     As a result, the old regmap is now gone. Almost all interfaces
2135 :     in MLRISC change as a consequence.
2136 :    
2137 :     2. A new brand version of machine description tool (v3.0) that generates
2138 :     modules expecting the new interface. The old version is removed.
2139 :    
2140 :     3. The RA interface has been further abstracted into two new functors.
2141 :     RISC_RA and X86RA. These functors have much simpler interfaces.
2142 :     [See also directory MLRISC/demo.]
2143 :    
2144 :     4. Some other new source->source code generation tools are available:
2145 :    
2146 :     a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
2147 :     b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
2148 :     I use this tool to generate the peephole optimizers---with the new
2149 :     cell type changes, peephole rules are becoming difficult to write
2150 :     without conditional pattern matching.
2151 :    
2152 :     5. More Intmap -> IntHashTable change. Previous changes by Matthias didn't
2153 :     cover the entire MLRISC source tree so many things broke.
2154 :    
2155 :     6. CM files have been moved to the subdirectory MLRISC/cm.
2156 :     They are moved because there are a lot of them and they clutter up the
2157 :     root dir.
2158 :    
2159 :     7. More detailed documentation to come...
2160 :    
2161 :     NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
2162 :     a makeml -rebuild first. This is because of other other
2163 :     changes that Matthias has made (see below).
2164 :    
2165 :    
2166 :     ----------------------------------------------------------------------
2167 : blume 733 Name: Matthias Blume
2168 : blume 742 Date: 2000/11/30 23:12:00 JST
2169 :     Tag: blume-20001130-filereorg
2170 :     Description:
2171 :    
2172 :     Some manual updates and some file reorganizations in CM.
2173 :    
2174 :     ----------------------------------------------------------------------
2175 :     Name: Matthias Blume
2176 : blume 737 Date: 2000/11/24 17:45:00 JST
2177 :     Tag: blume-20001124-link
2178 :     Description:
2179 :    
2180 :     Drastically improved link traversal code for the case that the dynamic
2181 :     value was already loaded at bootstrap time. As a result, CM and CMB
2182 :     now both load blazingly fast -- even on a very slow machine. Also,
2183 :     memory consumption has been further reduced by this.
2184 :    
2185 :     Warning: The format of the PIDMAP file has changed. THerefore, to
2186 :     bootstrap you have to do this:
2187 :    
2188 :     1. Run CMB.make
2189 :     2. Make a symbolic link for the boot directory:
2190 :     ln -s sml.boot.ARCH-OS xxx
2191 :     3. "Rebuild" the boot directory:
2192 :     ./makeml -boot xxx -rebuild sml ; rm xxx
2193 :     4. Boot normally:
2194 :     ./makeml
2195 :    
2196 :     ----------------------------------------------------------------------
2197 :     Name: Matthias Blume
2198 : blume 735 Date: 2000/11/21 21:20:00 JST
2199 :     Tag: blume-20001121-tools
2200 :     Description:
2201 :    
2202 :     Continued hacking on autoloading problem -- with success this time.
2203 :     Also changed tool-plugin mechanism. See new CM manual.
2204 :    
2205 :     ----------------------------------------------------------------------
2206 :     Name: Matthias Blume
2207 : blume 734 Date: 2000/11/19 14:30:00 JST
2208 :     Tag: blume-20001119-autoload
2209 :     Description:
2210 :    
2211 :     Some hacking to make autoloading faster. Success for CMB, no success
2212 :     so far for CM. There is a reduced structure CM' that autoloads faster.
2213 :     (This is a temporary, non-documented hack to be eliminated again when
2214 :     the general problem is solved.)
2215 :    
2216 :     ----------------------------------------------------------------------
2217 :     Name: Matthias Blume
2218 : blume 733 Date: 2000/11/17 14:10:00 JST
2219 :     Tag: blume-20001117-pickle-lib
2220 :     Description:
2221 :    
2222 :     1. Eliminated comp-lib.cm
2223 :     2. Made pickle-lib.cm
2224 :     3. Eliminated all uses of intset.sml (from comp-lib.cm)
2225 :     4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
2226 :     equivalent constructs from smlnj-lib.cm (INtHashTable).
2227 :     5. Point 4. also goes for those uses of intmap.* in MLRISC.
2228 :     Duplicated intmap modules thrown out.
2229 :     6. Hunted down all duplicated SCC code and replaced it with
2230 :     equivalent stuff (GraphSCCFn from smlnj-lib.cm).
2231 :     7. Rewrote Feedback module.
2232 :     8. Moved sortedlist.sml into viscomp-lib.cm. Eventually it
2233 :     should be thrown out and equivalent modules from smlnj-lib.cm
2234 :     should be used (IntRedBlackSet, IntListSet, ...).
2235 :    
2236 :     Confirmed that compiler compiles to fixpoint.
2237 :    
2238 :     ----------------------------------------------------------------------
2239 : leunga 731 Name: Allen Leung
2240 :     Date: 2000/11/10 18:00:00
2241 :     Tag: leunga-20001110-new-x86-fp
2242 :    
2243 :     A new x86 floating point code generator has been added.
2244 :     By default this is turned off. To turn this on, do:
2245 :    
2246 :     CM.autoload "$smlnj/compiler.cm";
2247 :     Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
2248 :    
2249 :     Changes:
2250 :    
2251 :     1. Changed FTAN to FPTAN so that the assembly output is correct.
2252 :     2. Changed the extension callback for FTANGENT to generate:
2253 :    
2254 :     fptan
2255 :     fstp %st(0)
2256 :     instead of
2257 :     fptan
2258 :     fstpl ftempmem
2259 :    
2260 :     3. Numerous assembly fixes for x86.
2261 :    
2262 :     5. Cleaned up the machine code output module x86/x86MC.sml and added
2263 :     support for a whole bunch of instructions and addressing modes:
2264 :    
2265 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st, %st(n)
2266 :     faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp %st, %st(n)
2267 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st(n), %st
2268 :     fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
2269 :     fxch %st(n)
2270 :     fld %st(n)
2271 :     fst %st(n)
2272 :     fst mem
2273 :     fstp %st(n)
2274 :     fucom %st(n)
2275 :     fucomp %st(n)
2276 :    
2277 :     All these are now generated when the fast fp mode is turned on.
2278 :    
2279 :     6. Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
2280 :    
2281 :     ----------------------------------------------------------------------
2282 : blume 729 Name: Matthias Blume
2283 :     Date: 2000/11/09 11:20:00 JST
2284 :     Tag: blume-20001109-scc
2285 :     Description:
2286 :    
2287 :     Eliminated some code duplication:
2288 :    
2289 :     1. Added "where" clause to GraphSCCFn in SML/NJ Library.
2290 :     (Otherwise the functor is useless.)
2291 :     2. Used GraphSCCFn where SCCUtilFun was used previously.
2292 :     3. Got rid of SCCUtilFun (in comp-lib.cm).
2293 :    
2294 :     ----------------------------------------------------------------------
2295 : george 721 Name: Lal George
2296 :     Date: 2000/11/06 09:02:21 EST 2000
2297 :     Tag: Release_110_30
2298 :     Description:
2299 :    
2300 :     - Version 110.30
2301 :     ----------------------------------------------------------------------
2302 : blume 715 Name: Matthias Blume
2303 : blume 716 Date: 2000/11/04 14:45:00
2304 :     Tag: blume-20001104-mlbuild
2305 :     Description:
2306 :    
2307 :     - Made ml-build faster on startup.
2308 :     - Documentation fixes.
2309 :    
2310 :     ----------------------------------------------------------------------
2311 :     Name: Matthias Blume
2312 : blume 715 Date: 2000/11/02 17:00:00 JST
2313 :     Tag: blume-20001102-condcomp
2314 :     Description:
2315 :    
2316 :     - Small tweaks to pickler -- new BOOTFILES!
2317 :     - Version bumped to 110.29.2.
2318 :     - Added conditional compilation facility to init.cmi (see comment there).
2319 :     ----------------------------------------------------------------------
2320 : leunga 713 Name: Allen Leung
2321 :     Date: 2000/10/23 19:31:00
2322 :     Tag: leunga-20001023-demo-ra
2323 :    
2324 :     1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
2325 :     2. Test programs for the graph library updated
2326 :     3. Some new MLRISC demo programs added
2327 :    
2328 :     ----------------------------------------------------------------------
2329 : blume 710 Name: Matthias Blume
2330 : blume 711 Date: 2000/08/31 22:15:00 JST
2331 :     Tag: blume-20001017-errmsg
2332 :     Description:
2333 :    
2334 :     More error message grief: Where there used to be no messages, there
2335 :     now were some that had bogus error regions. Fixed.
2336 :    
2337 :     ----------------------------------------------------------------------
2338 :     Name: Matthias Blume
2339 : blume 710 Date: 2000/08/31 17:30:00 JST
2340 :     Tag: blume-20001017-v110p29p1
2341 :     Description:
2342 :    
2343 :     I made a version 110.29.1 with new bootfiles.
2344 :    
2345 :     Changes: Modified pickler/unpickler for faster and leaner unpickling.
2346 :     CM documentation changes and a small bugfix in CM's error reporting.
2347 :    
2348 :     ----------------------------------------------------------------------
2349 : george 705 Name: Lal George
2350 :     Date: 2000/09/27 14:42:35 EDT
2351 :     Tag: george-20000927-nodestatus
2352 :     Description:
2353 :    
2354 :     Changed the type of the nodestatus, so that:
2355 :    
2356 :     SPILLED(~1) is now SPILLED
2357 :     SPILLED(m) where m>=0 is now MEMREG(m)
2358 :     SPILLED(s) where s<~1 is now SPILL_LOC(~s)
2359 :    
2360 :     ----------------------------------------------------------------------
2361 : blume 697 Name: Matthias Blume
2362 : blume 703 Date: 2000/09/07 14:45:00 JST
2363 :     Tag: blume-20000907-cmerrmsg
2364 :     Description:
2365 :    
2366 :     Small tweak to CM to avoid getting ML syntax error messages twice.
2367 :    
2368 :     ----------------------------------------------------------------------
2369 :     Name: Matthias Blume
2370 : blume 701 Date: 2000/08/31 18:00:00 JST
2371 :     Tag: blume-20000831-cvsbootfiles
2372 :     Description:
2373 :    
2374 :     New URL for boot files (because the 110.29 files on the BL server do
2375 :     now work correctly with my updated install scripts for yacc and lex).
2376 :    
2377 :     ----------------------------------------------------------------------
2378 :     Name: Matthias Blume
2379 : blume 697 Date: 2000/08/08 12:33:00 JST
2380 :     Tag: blume-20000808-manual
2381 :     Description:
2382 :    
2383 :     Tiny update to CM manual.
2384 :    
2385 :     ----------------------------------------------------------------------
2386 : leunga 695 Name: Allen Leung
2387 :     Date: 2000/08/7 19:31:00
2388 :     Tag: leunga-20000807-a-whole-bunch-of-stuff
2389 :    
2390 :     Moby, C--, SSA, x86, machine descriptions etc. Should only affect C--
2391 :     and Mobdy.
2392 :    
2393 :     1. x86
2394 :    
2395 :     a. Fixes to peephole module by John and Dan.
2396 :     b. Assembly fix to SETcc by Allen.
2397 :     c. Fix to c-call by John.
2398 :     d. Fix to spilling by John. (This one deals with the missing FSTPT case)
2399 :     e. Instruction selection optimization to SETcc as suggested by John.
2400 :    
2401 :     For example,
2402 :    
2403 :     MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
2404 :    
2405 :     should generate:
2406 :    
2407 :     MOVL a, x
2408 :     SUBL b, x
2409 :     SHRL 31, x
2410 :    
2411 :     2. IR stuff
2412 :    
2413 :     A bunch of new DJ-graph related algorithms added. These
2414 :     speed up SSA construction.
2415 :    
2416 :     3. SSA + Scheduling
2417 :    
2418 :     Added code for SSA and scheduling to the repository
2419 :    
2420 :     ----------------------------------------------------------------------
2421 : george 694 Name: Lal George
2422 :     Date: 2000/07/27 11:53:14 EDT
2423 :    
2424 :     Tag: lal-20000727-linux-ppc
2425 :     Description:
2426 :    
2427 :     Made changes to support Linux PPC.
2428 :     p.s. I have confirmation that the 110.29 boot files work fine.
2429 :    
2430 :     ----------------------------------------------------------------------
2431 : blume 684 Name: Matthias Blume
2432 : blume 692 Date: 2000/07/27 17:40:00 JST
2433 :     Tag: blume-20000727-scripts
2434 :     Description:
2435 :    
2436 :     !!!! WARNING !!!!
2437 :     You must recompile the runtime system!
2438 :     !!!! WARNING !!!!
2439 :    
2440 :     This is basically another round of script-enhancements:
2441 :    
2442 :     1. sml, ml-build, and ml-makedepend accept options -D and -U to define
2443 :     and undefine CM preprocessor symbols.
2444 :    
2445 :     2. ml-build avoids generating a new heap image if it finds that the
2446 :     existing one is still ok. (The condition is that no ML file had to
2447 :     be recompiled and all ML files are found to be older that the heap
2448 :     file.)
2449 :    
2450 :     To make this work smoothly, I also hacked the runtime system as
2451 :     well as SMLofNJ.SysInfo to get access to the heap image suffix
2452 :     (.sparc-solaris, ...) that is currently being used.
2453 :    
2454 :     Moreover, the signature of CM.mk_standalone has changed. See the
2455 :     CM manual.
2456 :    
2457 :     3. ml-makedepend accepts additional options -n, -a, and -o. (See the
2458 :     CM manual for details.)
2459 :    
2460 :     4. More CM manual updates:
2461 :     - all of the above has been documented.
2462 :     - there is now a section describing the (CM-related) command line
2463 :     arguments that are accepted by the "sml" command
2464 :    
2465 :     ----------------------------------------------------------------------
2466 :     Name: Matthias Blume
2467 : blume 691 Date: 2000/07/25 16:20:00 JST
2468 :     Tag: blume-20000725-makedepend
2469 :     Description:
2470 :    
2471 :     Added a script called ml-makedepend. This can be used in makefiles
2472 :     for Unix' make in a way very similar to the "makedepend" command for
2473 :     C.
2474 :    
2475 :     The script internally uses function CM.sources.
2476 :    
2477 :     Synopsis:
2478 :    
2479 :     ml-makedepend [-f makefile] cmfile targetname
2480 :    
2481 :     The default for the makefile is "makefile" (or "Makefile" should
2482 :     "makefile" not exist).
2483 :    
2484 :     ml-makedepend adds a cmfile/targetname-specific section to this
2485 :     makefile (after removing the previous version of this section). The
2486 :     section contains a single dependency specification with targetname on
2487 :     the LHS (targetname is an arbitrary name), and a list of files derived
2488 :     from the cmfile on the RHS. Some of the files on the RHS are
2489 :     ARCH/OPSYS-specific. Therefore, ml-makedepend inserts references to
2490 :     "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
2491 :     path names. The makefile writer is responsible for making sure that
2492 :     these variables have correct at the time "make" is invoked.
2493 :    
2494 :     ----------------------------------------------------------------------
2495 :     Name: Matthias Blume
2496 : blume 690 Date: 2000/07/22 23:30:00 JST
2497 :     Tag: blume-20000722-urlupdate
2498 :     Description:
2499 :    
2500 :     Changed BOOT and config/srcarchiveurl to point to BL server:
2501 :    
2502 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
2503 :    
2504 :     ----------------------------------------------------------------------
2505 :     Name: Matthias Blume
2506 : blume 686 Date: 2000/07/18 18:00:00 JST
2507 :     Tag: blume-20000718-Version_110_29
2508 :     Description:
2509 :    
2510 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
2511 :    
2512 :     2. Updated config/version to 110.29
2513 :    
2514 :     3. Updated config/srcarchiveurl
2515 :    
2516 :     3. New boot files!
2517 :     ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
2518 :    
2519 :     ----------------------------------------------------------------------
2520 :     Name: Matthias Blume
2521 : blume 684 Date: 2000/07/11 13:58:00 JST
2522 :     Tag: blume-20000711-doctypo
2523 :     Description:
2524 :    
2525 :     Fixed a few typos in CM manual.
2526 :    
2527 :     ----------------------------------------------------------------------
2528 : leunga 683 Name: Allen Leung
2529 :     Date: 2000/06/15 00:38:00
2530 :     Tag: leunga-20000704-sparc-x86
2531 :    
2532 :     1. x86 peephole improvement sp += k; sp -= k => nop [from John]
2533 :     2. fix to x86 RET bug [found by Dan Grossman]
2534 :     3. sparc assembly bug fix for ticc instructions [found by Fermin]
2535 :    
2536 :     Affects c-- and moby only
2537 :    
2538 :     ----------------------------------------------------------------------
2539 : blume 680 Name: Matthias Blume
2540 : blume 682 Date: 2000/07/04 15:26:00
2541 :     Tag: blume-20000704-trigger
2542 :     Description:
2543 :    
2544 :     1. Improvements to CM manual.
2545 :     2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
2546 :     of getting a back-trace. The function, when called, raises an
2547 :     internal exception which explicitly carries the full back-trace history,
2548 :     so it is unaffected by any intervening handle-raise pairs ("trivial"
2549 :     or not). The interactive loop will print that history once it arrives
2550 :     at top level.
2551 :     Short of having all exceptions implicitly carry the full history, the
2552 :     recommended way of using this facility is:
2553 :     - compile your program with instrumentation "on"
2554 :     - run it, when it raises an exception, look at the history
2555 :     - if the history is "cut off" because of some handler, go and modify
2556 :     your program so that it explicitly calls BTrace.trigger
2557 :     - recompile (still instrumented), and rerun; look at the full history
2558 :    
2559 :     ----------------------------------------------------------------------
2560 :     Name: Matthias Blume
2561 :     Date: 2000/07/03 15:36:00 JST
2562 : blume 680 Tag: blume-20000702-manual
2563 :     Description:
2564 :    
2565 :     Small corrections and updates to CM manual.
2566 :    
2567 :     ----------------------------------------------------------------------
2568 : blume 682 Name: Matthias Blume
2569 : blume 679 Date: 2000/06/29 16:04:00 JST
2570 :     Tag: blume-20000629-yacctool
2571 :     Description:
2572 :    
2573 :     Changes:
2574 :    
2575 :     1. Class "mlyacc" now takes separate arguments to pass options to
2576 :     generated .sml- and .sig-files independently.
2577 :     2. Corresponding CM manual updates.
2578 :     3. BTrace module now also reports call sites. (However, for loop clusters
2579 :     it only shows from where the cluster was entered.) There are associated
2580 :     modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
2581 :    
2582 :     ----------------------------------------------------------------------
2583 : blume 675 Name: Matthias Blume
2584 : blume 678 Date: 2000/06/27 16:51:00 JST
2585 :     Tag: blume-20000627-noweb
2586 :     Description:
2587 :    
2588 :     Changes:
2589 :    
2590 :     1. Implemented "subdir" and "witness" options for noweb tool.
2591 :     This caused some slight internal changes in CM's tool implementation.
2592 :     2. Fixed bug in "tool plugin" mechanism. This is essentially cleaning
2593 :     some remaining issues from earlier path anchor changes.
2594 :     3. Updated CM manual accordingly.
2595 :    
2596 :     4. Changed implementation of back-tracing so that I now consider it
2597 :     ready for prime-time.
2598 :    
2599 :     In particular, you don't have to explicitly trigger the back-trace
2600 :     anymore. Instead, if you are running BTrace-instrumented code and
2601 :     there is an uncaught exception (regardless of whether or not it was
2602 :     raised in instrumented code), the top-level evalloop will print
2603 :     the back-trace.
2604 :    
2605 :     Features:
2606 :    
2607 :     - Instrumented and uninstrumented code work together seemlessly.
2608 :     (Of course, uninstrumented code is never mentioned in actual
2609 :     back-traces.)
2610 :    
2611 :     - Asymptotic time- and space-complexity of instrumented code is
2612 :     equal to that of uninstrumented code. (This means that
2613 :     tail-recursion is preserved by the instrumentation phase.)
2614 :    
2615 :     - Modules whose code has been instrumented in different sessions
2616 :     work together without problem.
2617 :    
2618 :     - There is no penalty whatsoever on uninstrumented code.
2619 :    
2620 :     - There is no penalty on "raise" expressions, even in
2621 :     instrumented code.
2622 :    
2623 :     A potential bug (or perhaps it is a feature, too):
2624 :    
2625 :     A back-trace reaches no further than the outermost instrumented
2626 :     non-trivial "raise". Here, a "trivial" raise is one that is the
2627 :     sole RHS of a "handle" rule. Thus, back-traces reach trough
2628 :    
2629 :     <exp> handle e => raise e
2630 :    
2631 :     and even
2632 :    
2633 :     <exp> handle Foo => raise Bar
2634 :    
2635 :     and, of course, through
2636 :    
2637 :     <exp> handle Foo => ...
2638 :    
2639 :     if the exception was not Foo.
2640 :    
2641 :     Back-traces always reach right through any un-instrumented code
2642 :     including any of its "handle" expressions, trivial or not.
2643 :    
2644 :     To try this out, do the following:
2645 :    
2646 :     - Erase all existing binfiles for your program.
2647 :     (You may keep binfiles for those modules where you think you
2648 :     definitely don't need back-tracing.)
2649 :     - Turn on back-trace instrumentation:
2650 :     SMLofNJ.Internals.BTrace.mode (SOME true);
2651 :     - Recompile your program. (I.e., run "CM.make" or "use".)
2652 :     - You may now turn instrumentation off again (if you want):
2653 :     SMLofNJ.Internals.BTrace.mode (SOME false);
2654 :     - Run your program as usual. If it raises an exception that
2655 :     reaches the interactive toplevel, then a back-trace will
2656 :     automatically be printed. After that, the toplevel loop
2657 :     will print the exception history as usual.
2658 :    
2659 :     ----------------------------------------------------------------------
2660 :     Name: Matthias Blume
2661 : blume 677 Date: 2000/06/26 09:56:46 JST
2662 :     Tag: blume-20000626-setup
2663 :     Description:
2664 :    
2665 :     CM: - setup-parameter to "sml" added; this can be used to run arbitrary
2666 :     ML code before and after compiling a file (e.g., to set compiler
2667 :     flags)
2668 :    
2669 :     Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
2670 :     - associated changes to btrace.sml (BTrace instrumentation pass)
2671 :     - cleaner implementation of btimp.sml (BTrace tracing and report
2672 :     module)
2673 :    
2674 :     CM manual: * new path encoding documented
2675 :     * description of setup-parameter to "sml" added
2676 :    
2677 :     The biggest user-visible change to back-tracing is that it is no
2678 :     longer necessary to compile all traced modules within the same
2679 :     session. (This was a real limitation.)
2680 :    
2681 :     ----------------------------------------------------------------------
2682 :     Name: Matthias Blume
2683 : blume 676 Date: 2000/06/24 12:40:00 JST
2684 :     Tag: blume-20000624-startup
2685 :     Description:
2686 :    
2687 :     Fixes startup slowdown problem. (I was calling SrcPath.sync a _tad_
2688 :     bit too often -- to put it mildly. :)
2689 :    
2690 :     ----------------------------------------------------------------------
2691 :     Name: Matthias Blume
2692 : blume 675 Date: 2000/06/23 18:20:00 JST
2693 :     Tag: blume-20000623-btrace
2694 :     Description:
2695 :    
2696 :     This updates adds a backtrace facility to aid programmers in debugging
2697 :     their programs. This involves the following changes:
2698 :    
2699 :     1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
2700 :     keeping track of the current call stack. When programs are compiled
2701 :     in a special mode, the compiler will insert calls to these hooks
2702 :     into the user program.
2703 :     "Hook" means that it is possible for different implementations of
2704 :     back-tracing to register themselves (at different times).
2705 :    
2706 :     2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
2707 :     as an Absyn.dec->Absyn.dec rewrite. Normally this phase is turned off.
2708 :     It can be turned on using this call:
2709 :     SMLofNJ.Internals.BTrace.mode (SOME true);
2710 :     Turning it off again:
2711 :     SMLofNJ.Internals.BTrace.mode (SOME false);
2712 :     Querying the current status:
2713 :     SMLofNJ.Internals.BTrace.mode NONE;
2714 :     Annotated programs are about twice as big as normal ones, and they
2715 :     run a factor of 2 to 4 slower with a dummy back-trace plugin (one
2716 :     where all hooks do nothing). The slowdown with a plugin that is
2717 :     actually useful (such as the one supplied by default) is even greater,
2718 :     but in the case of the default plugin it is still only an constant
2719 :     factor (amortized).
2720 :    
2721 :     3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
2722 :     with a sub-structure BTrace for controlling back-tracing. In particular,
2723 :     the above-mentioned function "mode" controls whether the annotation
2724 :     phase is invoked by the compiler. Another important function is
2725 :     "trigger": when called it aborts the current execution and causes
2726 :     the top-level loop to print a full back-trace.
2727 :    
2728 :     4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
2729 :     for back-tracing. It keeps track of the dynamic call stack and in
2730 :     addition to that it keeps a partial history at each "level" of that
2731 :     stack. For example, if a tail-calls b, b tail-calls c, and c tail-calls
2732 :     d and b (at separate times, dynamically), then the report will show:
2733 :    
2734 :     GOTO d
2735 :     /c
2736 :     GOTO \b
2737 :     CALL a
2738 :    
2739 :     This shows that there was an initial non-tail call of a, then a
2740 :     tail-call to b or c, looping behavior in a cluster of functions that
2741 :     consist of b and c, and then a goto from that cluster (i.e., either from
2742 :     b or from c) to d.
2743 :    
2744 :     Note that (depending on the user program) the amount of information
2745 :     that the back-trace module has to keep track of at each level is bounded
2746 :     by a constant. Thus, the whole implementation has the same asymptotical
2747 :     complexity as the original program (both in space and in time).
2748 :    
2749 :     5. compiler/TopLevel/interact/evalloop.sml has been modified to
2750 :     handle the special exception SMLofNJ.Internals.BTrace.BTrace
2751 :     which is raised by the "trigger" function mentioned above.
2752 :    
2753 :     Notes on usage:
2754 :    
2755 :     - Annotated code works well together with unannotated code:
2756 :     Unannotated calls simply do not show up at all in the backtrace.
2757 :    
2758 :     - It is not a good idea to let modules that were annotated during
2759 :     different sessions run at the same time. This is because the compiler
2760 :     chooses small integers to identify individual functions, and there
2761 :     will be clashes if different modules were compiled in separate sessions.
2762 :     (Nothing will crash, and you will even be told about the clashes, but
2763 :     back-trace information will in general not be useful.)
2764 :    
2765 :     - Back-tracing can be confused by callcc and capture.
2766 :    
2767 :     - The only way of getting a back-trace right now is to explicitly
2768 :     invoke the "trigger" function from your user program. Eventually, we
2769 :     should make every exception carry back-trace information (if
2770 :     available). But since this creates more overhead at "raise"-time
2771 :     (similar to the current exnHistory overhead), I have not yet
2772 :     implemented this. (The implementation will be rather easy.) With
2773 :     exceptions carrying back-trace information, this facility will be even
2774 :     more useful because users don't need to modify their programs...
2775 :    
2776 :     - While it is possible to compile the compiler with back-trace
2777 :     annotations turned on (I did it to get some confidence in
2778 :     correctness), you must make absolutely sure that core.sml and
2779 :     btimp.sml are compiled WITHOUT annotation! (core.sml cannot actually
2780 :     be compiled with annotation because there is no core access yet, but
2781 :     if you compile btimp.sml with annotation, then the system will go into
2782 :     an infinite recursion and crash.)
2783 :     Since CM currently does not know about BTrace, the only way to turn
2784 :     annotations on and off for different modules of the compiler is to
2785 :     interrupt CMB.make, change the settings, and re-invoke it. Of course,
2786 :     this is awkward and clumsy.
2787 :    
2788 :     Sample sessions:
2789 :    
2790 :     Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2791 :     - SMLofNJ.Internals.BTrace.mode (SOME true);
2792 :     [autoloading]
2793 :     [autoloading done]
2794 :     val it = false : bool
2795 :     - structure X = struct
2796 :     - fun main n = let
2797 :     - fun a (x, 0) = d x
2798 :     - | a (x, n) = b (x, n - 1)
2799 :     - and b (x, n) = c (x, n)
2800 :     - and c (x, n) = a (x, n)
2801 :     - and d x = e (x, 3)
2802 :     - and e (x, 0) = f x
2803 :     - | e (x, n) = e (x, n - 1)
2804 :     - and f 0 = SMLofNJ.Internals.BTrace.trigger ()
2805 :     - | f n = n * g (n - 1)
2806 :     - and g n = a (n, 3)
2807 :     - in
2808 :     - f n
2809 :     - end
2810 :     - end;
2811 :     structure X : sig val main : int -> int end
2812 :     - X.main 3;
2813 :     *** BACK-TRACE ***
2814 :     GOTO stdIn:4.2-13.20: X.main[2].f
2815 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
2816 :     GOTO stdIn:4.2-13.20: X.main[2].d
2817 :     / stdIn:4.2-13.20: X.main[2].a
2818 :     | stdIn:4.2-13.20: X.main[2].b
2819 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
2820 :     CALL stdIn:4.2-13.20: X.main[2].g
2821 :     GOTO stdIn:4.2-13.20: X.main[2].f
2822 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
2823 :     GOTO stdIn:4.2-13.20: X.main[2].d
2824 :     / stdIn:4.2-13.20: X.main[2].a
2825 :     | stdIn:4.2-13.20: X.main[2].b
2826 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
2827 :     CALL stdIn:4.2-13.20: X.main[2].g
2828 :     GOTO stdIn:4.2-13.20: X.main[2].f
2829 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
2830 :     GOTO stdIn:4.2-13.20: X.main[2].d
2831 :     / stdIn:4.2-13.20: X.main[2].a
2832 :     | stdIn:4.2-13.20: X.main[2].b
2833 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
2834 :     CALL stdIn:4.2-13.20: X.main[2].g
2835 :     GOTO stdIn:4.2-13.20: X.main[2].f
2836 :     CALL stdIn:2.15-17.4: X.main[2]
2837 :     -
2838 :    
2839 :     (Note that because of a FLINt bug the above code currently does not
2840 :     compile without BTrace turned on.)
2841 :    
2842 :     Here is another example, using my modified Tiger compiler:
2843 :    
2844 :     Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
2845 :     - SMLofNJ.Internals.BTrace.mode (SOME true);
2846 :     [autoloading]
2847 :     [autoloading done]
2848 :     val it = false : bool
2849 :     - CM.make "sources.cm";
2850 :     [autoloading]
2851 :     ...
2852 :     [autoloading done]
2853 :     [scanning sources.cm]
2854 :     [parsing (sources.cm):parse.sml]
2855 :     [creating directory CM/SKEL ...]
2856 :     [parsing (sources.cm):tiger.lex.sml]
2857 :     ...
2858 :     [wrote CM/sparc-unix/semant.sml]
2859 :     [compiling (sources.cm):main.sml]
2860 :     [wrote CM/sparc-unix/main.sml]
2861 :     [New bindings added.]
2862 :     val it = true : bool
2863 :     - Main.compile ("../testcases/merge.tig", "foo.out");
2864 :     *** BACK-TRACE ***
2865 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
2866 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2867 :     CALL lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2868 :     GOTO lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
2869 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2870 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2871 :     CALL lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
2872 :     / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
2873 :     CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
2874 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
2875 :     CALL lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
2876 :     CALL main.sml:1.18-118.4: Main.compile[2]
2877 :     -
2878 :    
2879 :     ----------------------------------------------------------------------
2880 : blume 673 Name: Matthias Blumen
2881 :     Date: 2000/06/21 18:00:00 JST
2882 :     Tag: blume-20000621-manual
2883 :     Description:
2884 :    
2885 :     CM manual update: Path environments documented.
2886 :    
2887 :     ----------------------------------------------------------------------
2888 : blume 666 Name: Matthias Blume
2889 : blume 672 Date: 2000/06/19 13:40:00
2890 :     Tag: blume-20000619-manual
2891 :     Description:
2892 :    
2893 :     CM manual and system/README update. This only covers the fact that
2894 :     there are no more implicit anchors. (Path environments and the "bind"
2895 :     option to "cm" have yet to be documented.)
2896 :    
2897 :     ----------------------------------------------------------------------
2898 :     Name: Matthias Blume
2899 : blume 671 Date: 2000/06/19 11:05:00 JST
2900 :     Tag: blume-20000619-chdir-bugfix
2901 :     Description:
2902 :    
2903 :     Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
2904 :    
2905 :     ----------------------------------------------------------------------
2906 :     Name: Matthias Blume
2907 : blume 670 Date: 2000/06/18 22:00:10 JST
2908 :     Tag: blume-20000618-implicit-anchors-really-gone
2909 :     Description:
2910 :    
2911 :     I updates the previous HISTORY entry where I forgot to mention that
2912 :     implicit anchors are no longer with us.
2913 :    
2914 :     The current update also gets rid of the (now useless) controller
2915 :     CM.Control.implicit_anchors.
2916 :    
2917 :     ----------------------------------------------------------------------
2918 :     Name: Matthias Blume
2919 : blume 666 Date: 2000/06/16 17:30:00 JST
2920 :     Tag: blume-20000616-anchorenv
2921 :     Description:
2922 :    
2923 :     This patch implements the long anticipated (just kidding :) "anchor
2924 :     environment" mechanism. In the course of doing this, I also
2925 :     re-implemented CM's internal "SrcPath" module from scratch. The new
2926 :     one should be more robust in certain boundary cases. In any case, it
2927 :     is a lot cleaner than its predecessor (IMHO).
2928 :    
2929 :     This time, although there is yet another boot file format change, I
2930 :     kept the unpickler backward-compatible. As a result, no new bootfiles
2931 :     are necessary and bootstrapping is straightforward. (You cannot read
2932 :     new bootfiles into an old system, but the other way around is no
2933 :     problem.)
2934 :    
2935 :     Visible changes:
2936 :    
2937 : blume 670 ** 0. Implicit path anchors (without the leading $-symbol) are no
2938 :     longer recognized at all. This means that such path names are not
2939 :     illegal either. For example, the name basis.cm simply refers to a
2940 :     local file called "basis.cm" (i.e, the name is an ordinary path
2941 :     relative to .cm-files directory). Or, to put it differently, only
2942 :     names that start with $ are anchored paths.
2943 : blume 666
2944 : blume 670 ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
2945 :     vanished.
2946 :    
2947 :     John (Reppy) had critizised this as soon as I originally proposed and
2948 : blume 666 implemented it, but at that time I did not really deeply believe
2949 :     him. :) Now I came full-circle because I need the $<singlearc> syntax
2950 :     in another place where it cannot be seen as an abbreviation for
2951 :     $/<singlearc>. To avoid the confusion, $<singlearc> now means what it
2952 :     seems to mean (i.e., it "expands" into the corresponding anchor
2953 :     value).
2954 :    
2955 :     However, when paths are used as members in CM description files, it
2956 :     continues to be true that there must be at least another arc after the
2957 :     anchor. This is now enforced separately during semantic analysis
2958 : blume 670 (i.e., from a lexical/syntactical point of view, the notation is ok.)
2959 : blume 666
2960 :     ** 2. The "cm" class now accepts an option "bind". The option's value
2961 :     is a sub-option list of precisely two items -- one labeled "anchor"
2962 :     and the other one labeled "value". As you might expect, "anchor" is
2963 :     used to specify an anchor name to be bound, and "value" specifies what
2964 :     the anchor is being bound to.
2965 :    
2966 :     The value must be a directory name and can be given in either standard
2967 :     syntax (including the possibility that it is itself an anchored path)
2968 :     or native syntax.
2969 :    
2970 :     Examples:
2971 :    
2972 :     foo.cm (bind:(anchor:bar value:$mystuff/bar))
2973 :     lib.cm (bind:(anchor:a value:"H:\\x\\y\\z")) (* only works under windows *)
2974 :    
2975 :     and so on.
2976 :    
2977 :     The meaning of this is that the .cm-file will be processed with an
2978 :     augmented anchor environment where the given anchor(s) is/are bound to
2979 :     the given values(s).
2980 :    
2981 :     The rationale for having this feature is this: Suppose you are trying
2982 :     to use two different (already stable) libraries a.cm and b.cm (that
2983 :     you perhaps didn't write yourself). Further, suppose each of these
2984 :     two libraries internally uses its own auxiliary library $aux/lib.cm.
2985 :     Normally you would now have a problem because the anchor "lib" can not
2986 :     be bound to more than one value globally. Therefore, the project that
2987 :     uses both a.cm and b.cm must locally redirect the anchor to some other
2988 :     place:
2989 :    
2990 :     a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
2991 :     b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
2992 :    
2993 :     This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
2994 :     /usr/lib/smlnj/b-stuff/aux.cm, respectively.
2995 :    
2996 :     Hard-wiring path names is a bit inflexible (and CM will verbosely warn
2997 :     you when you do so at the time of CM.stabilize). Therefore, you can
2998 :     also use an anchored path as the value:
2999 :    
3000 :     a.cm (bind:(anchor:lib value:$a-lib))
3001 :     b.cm (bind:(anchor:lib value:$b-lib))
3002 :    
3003 :     Now you can globally configure (using the usual CM.Anchor.anchor or
3004 :     pathconfig machinery) bindings for "a-lib" and "b-lib". Since "lib"
3005 :     itself is always locally bound, setting it globally is no longer
3006 :     meaningful or necessary (but it does not hurt either). In fact, "lib"
3007 :     can still be used as a global anchor for separate purposes. As a
3008 :     matter of fact, one can locally define "lib" in terms of a global
3009 :     "lib":
3010 :    
3011 :     a.cm (bind:(anchor:lib value:$lib/a))
3012 :     b.cm (bind:(anchor:lib value:$lib/b))
3013 :    
3014 :     ** 3: The encoding of path names has changed. This affects the way
3015 :     path names are shown in CM's progress report and also the internal
3016 :     protocol encoding used for parallel make.
3017 :    
3018 :     The encoding now uses one or more ':'-separated segments. Each
3019 :     segments corresponds to a file that has been specified relative to the
3020 :     file given by its preceding segment. The first segment is either
3021 :     relative to the CWD, absolute, or anchored. Each segment itself is
3022 :     basically a Unix pathname; all segments but the first are relative.
3023 :    
3024 :     Example:
3025 :    
3026 :     $foo/bar/baz.cm:a/b/c.sml
3027 :    
3028 :     This path denotes the file bar/a/b/c.sml relative to the directory
3029 :     denoted by anchor "foo". Notice that the encoding also includes
3030 :     baz.cm which is the .cm-file that listed a/b/c.sml. As usual, such
3031 :     paths are resolved relative to the .cm-files directory, so baz.cm must
3032 :     be ignored to get the "real" pathname.
3033 :    
3034 :     To make this fact more obvious, CM puts the names of such "virtual
3035 :     arcs" into parentheses when they appear in progress reports. (No
3036 :     parentheses will appear in the internal protocol encoding.) Thus,
3037 :     what you really see is:
3038 :    
3039 :     $foo/bar/(baz.cm):a/b/c.sml
3040 :    
3041 :     I find this notation to be much more informative than before.
3042 :    
3043 :     Another new feature of the encoding is that special characters
3044 :     including parentheses, colons, (back)slashes, and white space are
3045 :     written as \ddd (where ddd is the decimal encoding of the character).
3046 :    
3047 : blume 670 *** The CM manual still needs to be updated.
3048 :    
3049 : blume 666 ----------------------------------------------------------------------
3050 : leunga 660 Name: Allen Leung
3051 :     Date: 2000/06/15 00:38:00
3052 :     Tag: leunga-20000615-x86-peephole
3053 :    
3054 :     x86 Peephole fix by Fermin. Affects c-- and moby only.
3055 :    
3056 :     ----------------------------------------------------------------------
3057 : blume 658 Name: Matthias Blume
3058 :     Date: 2000/06/12 11:40:00
3059 :     Tag: blume-20000612-parmakefix
3060 :     Description:
3061 :    
3062 :     More cleanup after changing the file naming scheme: This time I
3063 :     repaired the parallel make mechanism for CMB.make which I broke earlier.
3064 :    
3065 :     ----------------------------------------------------------------------
3066 : leunga 657 Name: Allen Leung
3067 :     Date: 2000/06/09 01:25:00
3068 :     Tag: leunga-20000609-various
3069 :    
3070 :     None of these things should affect normal SML/NJ operations
3071 :    
3072 :     1. Peephole improvements provided by Fermin (c--)
3073 :     2. New annotation DEFUSE for adding extra dependence (moby)
3074 :     3. New X86 LOCK instructions (moby)
3075 :     4. New machine description language for reservation tables (scheduling)
3076 :     5. Fixes to various optimization/analysis modules (branch chaining, dominator
3077 :     trees etc.)
3078 :     6. I've changed the CM files so that they can work with versions
3079 :     110.0.6, 110.25 and 110.28
3080 :    
3081 :     ----------------------------------------------------------------------
3082 : blume 652 Name: Matthias Blume
3083 : blume 656 Date: 2000/06/09 12:40:00
3084 :     Tag: blume-20000609-log
3085 :     Description:
3086 :    
3087 :     - Removed all(?) remaining RCS Log entries from sources.
3088 :    
3089 :     - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
3090 :     anchored paths).
3091 :    
3092 :     ----------------------------------------------------------------------
3093 :     Name: Matthias Blume
3094 : blume 653 Date: 2000/06/07 17:00:00 JST
3095 :     Tag: blume-20000607-no-implicit-anchors
3096 :     Description:
3097 :    
3098 :     1. This update changes the default setting for
3099 :     CM.Control.implicit_anchors from true to false. This means that
3100 :     implicit anchors are no longer permitted by default. I also tried to
3101 :     make sure that nothing else still relies on implicit anchors.
3102 :     (This is the next step on the schedule towards a CM that does not even
3103 :     have the notion of implicit anchors anymore.)
3104 :    
3105 :     2. More CM manual updates.
3106 :    
3107 :     3. I managed to track down and fix the pickling bug I mentioned last
3108 :     time. Because of the previously existing workaround, this entails no
3109 :     immediate practical changes.
3110 :    
3111 :     ----------------------------------------------------------------------
3112 :     Name: Matthias Blume
3113 : blume 652 Date: 2000/06/06 11:15:00 JST
3114 :     Tag: blume-20000606-lazierpickle
3115 :     Description:
3116 :    
3117 :     !!!! NEW BOOT FILES !!!!
3118 :    
3119 :     * The main purpose of this update is to make library pickles lazier in
3120 :     order to reduce the initial space penalty for autoloading a library.
3121 :     As a result, it is now possible to have $smlnj/compiler.cm
3122 :     pre-registered. This should take care of the many complaints or
3123 :     inquiries about missing structure Compiler. This required changes to
3124 :     CM's internal data structures and small tweaks to some algorithms.
3125 :    
3126 :     As a neat additional effect, it is no longer necessary (for the sake
3127 :     of lean heap image files) to distinguish between a "minimal" CM and a
3128 :     "full" CM. Now, there is only one CM (i.e., the "full" version:
3129 :     $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
3130 :     interactive top level. ($smlnj/cm/minimal.cm is gone.)
3131 :    
3132 :     To make the life of compiler-hackers easier, "makeml" now also
3133 :     pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm). In other
3134 :     words, after you bootstrap a new sml for the first time, you will not
3135 :     have to autoload $smlnj/cmb.cm again afterwards. (The first time
3136 :     around you will still have to do it, though.)
3137 :    
3138 :     * A second change consists of major updates to the CM manual. There
3139 :     are now several appendices with summary information and also a full
3140 :     specification of the CM description file syntax.
3141 :    
3142 :     * In directory src/system I added the script "allcross". This script
3143 :     invokes sml and cross-compiles the compiler for all supported
3144 :     architectures. (Useful when providing a new set of boot files.)
3145 :    
3146 :     * There seems to be a latent bug in my "lazy pickles" mechanism. I
3147 :     added a small tweak to pickle-util.sml to work around this problem,
3148 :     but it is not a proper fix yet. I will investigate further. (The
3149 :     effect of the bug was an inflation of library pickle size.)
3150 :    
3151 :     * Version number increased to 110.28.1 (to avoid compatibility problems).
3152 :    
3153 :     ----------------------------------------------------------------------
3154 : leunga 646 Name: Allen Leung
3155 : leunga 648 Date: 2000/05/25 17:28 EDT
3156 :     Tag: leunga-20000525-ra
3157 :     Description:
3158 :    
3159 :     Fixed a bug in freezing phase of the register allocator.
3160 :    
3161 :     ----------------------------------------------------------------------
3162 :     Name: Allen Leung
3163 : leunga 647 Date: 2000/05/15 22:53 EDT
3164 : leunga 646 Tag: leunga-20000515-alpha-x86-ra
3165 :     Description:
3166 :    
3167 :     1. Alpha
3168 :    
3169 :     Slight cleanup. Removed the instruction SGNXL
3170 :    
3171 :     2. X86
3172 :    
3173 :     Added the following instructions to the instruction set:
3174 :    
3175 :     ROLx, RORx,
3176 :     BTx, BTSx, BTLx, BTRx,
3177 :     XCHGx, and variants with the LOCK prefix
3178 :    
3179 :     3. Register Allocation
3180 :    
3181 :     The module ra-rewrite-with-renaming has been improved.
3182 :    
3183 : leunga 647 These have no effect on SML/NJ.
3184 :    
3185 : leunga 646 ----------------------------------------------------------------------
3186 : blume 642 Name: Matthias Blume
3187 : blume 645 Date: 2000/05/15 16:20:00 JST
3188 :     Tag: blume-20000515-lightrebuild
3189 :     Description:
3190 :    
3191 :     1. I added an alternative to "-rebuild" to "makeml". The difference is
3192 :     that prior to calling CMB.make' the CM-variable "LIGHT" will be
3193 :     defined. In effect, the command will not build any cross-compiler
3194 :     backends and therefore finish more quickly.
3195 :    
3196 :     The "fixpt" script also takes a "-light" switch to be able to use
3197 :     this new facility while compiling for a fixpoint.
3198 :    
3199 :     2. I replaced all mentions of anchored paths in group owner specifications
3200 :     with simple relative paths (usually starting with "..").
3201 :     The rationale is that a library's internal workings should not be
3202 :     compromised by the lack of some anchor. (An anchor is necessary
3203 :     for someone who wants to refer to the library by an anchored path,
3204 :     but it should not be necessary to build the same library in the first
3205 :     place.)
3206 :    
3207 :     3. I changed the way CM's tool mechanism determines the shell command
3208 :     string used for things like ml-yacc etc. so that it does not break
3209 :     when CM.Control.implicit_anchors is turned off.
3210 :    
3211 :     ----------------------------------------------------------------------
3212 :     Name: Matthias Blume
3213 : blume 644 Date: 2000/05/12 18:20:00 JST
3214 :     Tag: blume-20000512-ml-build
3215 :     Description:
3216 :    
3217 :     Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
3218 :     getting installed properly (by config/install.sh).
3219 :    
3220 :     ----------------------------------------------------------------------
3221 :     Name: Matthias Blume
3222 : blume 643 Date: 2000/05/12 17:30:00 JST
3223 :     Tag: blume-20000512-anchors
3224 :     Description:
3225 :    
3226 :     !!! NEW BOOT FILES !!!
3227 :    
3228 :     This change is in preparation of fading out support for "implicitly
3229 :     anchored path names". I went through all sources and used the
3230 :     explicit (and relatively new) $-notation. See system/README and the
3231 :     CM manual for more info on this.
3232 :    
3233 :     I also modified the anchoring scheme for some things such as "smlnj",
3234 :     "MLRISC", "cm", etc. to take advantage of the fact that explicit
3235 :     anchors are more expressive: anchor name and first arc do not have to
3236 :     coincide. This entails the following user-visible change:
3237 :    
3238 :     You have to write $smlnj/foo/bar instead of smlnj/foo/bar. In
3239 :     particular, when you fire up sml with a command-line argument, say,
3240 :     e.g.:
3241 :    
3242 :     sml '$smlnj/cmb.cm'
3243 :    
3244 :     At the ML toplevel prompt:
3245 :    
3246 :     CM.autoload "$smlnj/cmb.cm";
3247 :    
3248 :     There is also a new controller in CM.Control that can be used to turn
3249 :     off all remaining support for implicit anchors by saying:
3250 :    
3251 :     CM.autoload "$smlnj/
3252 :     #set CM.Control.implicit_anchors false;
3253 :    
3254 :     This causes CM to reject implicitly anchored paths. This is (for the
3255 :     time being) less permissive than the "final" version where there will
3256 :     be no more such implicit anchors and relative paths will be just that:
3257 :     relative.
3258 :    
3259 :     The next step (version after next version?) will be to make the
3260 :     default for CM.Control.implicit_anchors false. After the dust has
3261 :     settled, I can then produce the "final" version of this...
3262 :    
3263 :     Note: Since bootstrapping is a bit tricky, I provided new boot files.
3264 :    
3265 :     ----------------------------------------------------------------------
3266 :     Name: Matthias Blume
3267 : blume 642 Date: 2000/05/11 16:30:00 JST
3268 :     Tag: blume-20000511-sources
3269 :     Description:
3270 :    
3271 :     The main change is that I added function CM.sources as a generalized
3272 :     version of the earlier CM.makedepend. This entails the following
3273 :     additional changes:
3274 :    
3275 :     - CM.makedepend has been dropped.
3276 :    
3277 :     - CM manual has been updated.
3278 :    
3279 :     - TOOLS signature and API have been changed.
3280 :    
3281 :     ----------------------------------------------------------------------
3282 : leunga 640 Name: Allen Leung
3283 : leunga 641 Date: 2000/05/10 21:17 EDT
3284 :     Tag: leunga-20000510-moby-c--ssa
3285 :     Description:
3286 :    
3287 :     Various bug fixes and new features for C--, Moby and MLRISC optimizations.
3288 :     None of these affect SML/NJ.
3289 :    
3290 :     1. Register Allocation
3291 :    
3292 :     a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
3293 :     This module tries to remove local (i.e. basic block level) redundancies
3294 :     during spilling.
3295 :    
3296 :     b. A new framework for performing region based register allocation.
3297 :     Not yet entirely functional.
3298 :    
3299 :     2. X86
3300 :    
3301 :     a. DefUse for POP was missing the stack pointer [found by Lal]
3302 :     b. Reload for CALL was incorrect in X86Spill [found by John]
3303 :     c. Various fixes in X86Spill so that it can be used correctly for
3304 :     the new spilling module.
3305 :    
3306 :     3. SSA/IR
3307 :    
3308 :     a. New module ir/dj-dataflow.sml implements elimination based
3309 :     data flow analysis.
3310 :    
3311 :     4. MLRiscGen
3312 :    
3313 :     a. Fix for gc type annotation
3314 :    
3315 :     5. MDGen
3316 :    
3317 :     Various fixes for machine description -> ml code translation. For ssa
3318 :     only.
3319 :    
3320 :     ----------------------------------------------------------------------
3321 :     Name: Allen Leung
3322 : leunga 640 Date: 2000/05/08 22:17 EDT
3323 :     Tag: leunga-20000508-labexp
3324 :     Description:
3325 :    
3326 :     Fermin has found a few assembly problems with constant expressions
3327 :     generated in LabelExp. Mostly, the problems involve extra parentheses,
3328 :     which choke on dumb assemblers. This is his fix.
3329 :    
3330 :     ----------------------------------------------------------------------
3331 : dbm 635 Name: Dave MacQueen
3332 :     Date: 2000/04/09 14:00 EDT
3333 :     Tag: dbm-20000502-Version_110_28
3334 :     Description:
3335 :    
3336 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
3337 :    
3338 :     2. Updated config/version to 110.28
3339 :    
3340 :     3. Updated config/srcarchiveurl
3341 :    
3342 :     3. New boot files!
3343 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
3344 :    
3345 :     ----------------------------------------------------------------------
3346 : blume 634 Name: Matthias Blume
3347 :     Date: 2000/05/01 19:05:00 JST
3348 :     Tag: blume-20000501-noweb
3349 :     Description:
3350 :    
3351 :     A new noweb tool has been added. The existing system is entirely
3352 :     unaffected by this, but some CM users have asked for renewed noweb
3353 :     support. Everything is documented in the CM manual.
3354 :    
3355 :     New (plugin) libraries:
3356 :    
3357 :     noweb-tool.cm
3358 :     nw-ext.cm
3359 :    
3360 :     ----------------------------------------------------------------------
3361 : dbm 633 Name: Dave MacQueen
3362 :     Date: 2000/04/30 12:40PM EDT
3363 :     Tag: dbm-20000430-bug_fixes
3364 :     Description:
3365 :    
3366 :     1. Fix for bug 1498
3367 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sig
3368 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sml
3369 :     added toRealArray function
3370 :     smlnj/src/compiler/MiscUtil/print/ppobj.sml
3371 :     added check for tag Obj.RealArray to array printing case in ppObj
3372 :    
3373 :     2. Fix for bug 1510
3374 :     smlnj/src/compiler/Semant/types/typesutil.sml
3375 :     fixed definition of dummyargs (used by equalTycon) so that
3376 :     dummy args are distinct types
3377 :    
3378 :     ----------------------------------------------------------------------
3379 : blume 629 Name: Matthias Blume
3380 : blume 632 Date: 2000/04/30 01:00:00 JST
3381 :     Tag: blume-20000430-versions
3382 :     Description:
3383 :    
3384 :     1. CM version numbering added. This is an implementation of Lal's
3385 :     proposal for adding version numbers and version checking to .cm
3386 :     files. Lal said that his proposal was just that -- a proposal.
3387 :     For the time being I went ahead and implemented it so that people
3388 :     can comment on it. Everything is completely backward-compatible
3389 :     (except for the stable library format, i.e., new bootfiles!).
3390 :    
3391 :     As usual, see the CM manual for details.
3392 :    
3393 :     2. An alternative syntax for anchored paths has been implemented.
3394 :     Dave has recently voiced the same concerns that I had when I did
3395 :     this, so there should be some support. My take is that eventually
3396 :     I will let support for the current syntax (where anchors are
3397 :     "implicit") fade out in favor of the new, explicit syntax.
3398 :     In order to be backward-compatible, both old and new syntax are
3399 :     currently supported.
3400 :    
3401 :     Again, see the CM manual for details.
3402 :    
3403 :     3. Parallel make is trying to be slightly smarter: When the master
3404 :     process finds a "bottleneck", i.e., when there is only one
3405 :     compilation unit that can be compiled and everybody else is
3406 :     waiting on it, then it will simply compile it directly instead
3407 :     of clumsily telling one of the slaves to do it.
3408 :    
3409 :     4. Support for "unsharing" added. This is necessary in order to be
3410 :     able to have two different versions of the same library running
3411 :     at the same time (e.g., for trying out a new MLRISC while still
3412 :     having the old MLRISC linked into the current compiler, etc.)
3413 :     See the CM manual.
3414 :    
3415 :     5. Simple "makedepend" functionality added for generating Makefile
3416 :     dependency information. (This is rather crude at the moment.
3417 :     Expect some changes here in the future.)
3418 :    
3419 :     6. ".fun" added as a recognized suffix for ML files. Also documented
3420 :     explicitly in the manual that the fallback behavior (unknown suffix
3421 :     -> ML file) is not an official feature!
3422 :    
3423 :     7. Small changes to the pickler for stable libraries.
3424 :    
3425 :     8. Several internal changes to CM (for cleanup/improvement).
3426 :    
3427 :    
3428 :     !!!! NEW BINFILES !!!!
3429 :    
3430 :     ----------------------------------------------------------------------
3431 :     Name: Matthias Blume
3432 : blume 631 Date: 2000/04/28 17:30:00 JST
3433 :     Tag: blume-20000428-pathconfig
3434 :     Description:
3435 :    
3436 :     1. I changed config/install.sh to remove duplicate entries from the
3437 :     lib/pathconfig file at the end. Moreover, the final version of
3438 :     lib/pathconfig is sorted alphabetically. The same (sorting) is done
3439 :     in src/system/installml.
3440 :    
3441 :     2. The config/install.sh script now consistently uses relative
3442 :     pathnames in lib/pathconfig whenever the anchor is in the lib
3443 :     directory. (So far this was true for the libraries that come
3444 :     pre-compiled and bundled as part of the bootfiles but not for
3445 :     libraries that are compiled by the script itself.)
3446 :    
3447 :     ----------------------------------------------------------------------
3448 :     Name: Matthias Blume
3449 : blume 629 Date: 2000/04/26 13:10:00 JST
3450 :     Tag: blume-20000426-fun_suffix
3451 :     Description:
3452 :    
3453 :     Added ".fun" as a recognized file name suffix (for ML code).
3454 :    
3455 :     ----------------------------------------------------------------------
3456 : leunga 628 Name: Allen Leung
3457 :     Date: 2000/04/25 17:00:00 EST
3458 :     Tag: leunga-20000425-alpha-ra
3459 :     Description:
3460 :    
3461 :     1. Alpha
3462 :    
3463 :     PSEUDOARITH was missing in AlphaRewrite. This causes an endless loop
3464 :     in C--.
3465 :    
3466 :     2. RA
3467 :    
3468 :     Added a flag "ra-dump-size" to print out the size of the flowgraph
3469 :     and the interference graph.
3470 :    
3471 :     ----------------------------------------------------------------------
3472 : dbm 627 Name: Dave MacQueen
3473 :     Date: 2000/04/25/
3474 :     Tag: dbm-20000425-mlyacc_doc_examples
3475 :     Description:
3476 :     Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
3477 :     Updated all three examples in src/ml-yacc/examples to run
3478 :     under 110.* using CM.make.
3479 :    
3480 :     ----------------------------------------------------------------------
3481 : leunga 624 Name: Allen Leung
3482 :     Date: 2000/04/20 23:04:00 EST
3483 :     Tag: leunga-20000420-ssa-c---stuff
3484 :     Description:
3485 :    
3486 :     This update synchronizes my repository with Yale's. Most of these
3487 :     changes, however, do not affect SML/NJ at all (the RA is an exception).
3488 :    
3489 :     1. Register Allocator
3490 :    
3491 :     a. An improvement in the interference graph construction:
3492 :     Given a copy
3493 :    
3494 :     s <- t
3495 :    
3496 : leunga 628 no interference edge between s and t is added for this definition of s.
3497 : leunga 624
3498 :     b. I've added two new spill heuristic modules that Fermin and I developed
3499 :     (in the new library RA.cm). These are unused in SML/NJ but maybe
3500 :     useful for others (Moby?)
3501 :    
3502 :     2. X86
3503 :    
3504 :     a. Various fixes in the backend provided by Fermin [C--] and Lal.
3505 :    
3506 :     3. Alpha
3507 :    
3508 :     a. Added the BSR instruction and code generation that goes with it [C--]
3509 :     b. Other fixes too numerous to recount provided by Fermin [C--]
3510 :    
3511 :     4. Regmaps
3512 :    
3513 :     a. The regmaps are not initialized with the identity physical bindings
3514 :     at creation time. This is unneeded.
3515 :    
3516 :     5. MLRISC Optimizations
3517 :    
3518 :     a. The DJ-Graph module can now compute the iterated dominance frontiers
3519 :     intersects with liveness incrementally in linear time! Woohoo!
3520 :     This is now used in my new SSA construction algorithm.
3521 :    
3522 :     b. THe branch reorganization module is now smarter about linear chains of
3523 :     basic blocks.
3524 :    
3525 :    
3526 :     ----------------------------------------------------------------------
3527 : blume 615 Name: Matthias Blume
3528 :     Date: 2000/04/12 13:52:00 JST
3529 :     Tag: blume_main_v110p27_1
3530 :     Description:
3531 :    
3532 :     Changed install.sh script to handle archive files without version number
3533 :     and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
3534 :     name of the boot file archive.
3535 :    
3536 :     ----------------------------------------------------------------------
3537 : dbm 607 Name: Dave MacQueen
3538 :     Date: 2000/04/09 14:00 EDT
3539 :     Tag: dbm-20000410-Version_110_27
3540 :     Description:
3541 :    
3542 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
3543 :    
3544 :     2. Updated src/config/version to 110.27
3545 :    
3546 :     3. New boot files!
3547 :    
3548 :     ----------------------------------------------------------------------
3549 : leunga 606 Name: Allen Leung
3550 :     Date: 2000/04/09 19:09:00 EST
3551 :     Tag: leunga-20000409-misc
3552 :     Description:
3553 :    
3554 :     1. Yet another fix for x86 assembly for idivl, imull, mull and friends.
3555 :    
3556 :     2. Miscellaneous improvements to MLRISC (unused in sml/nj)
3557 :    
3558 :     ----------------------------------------------------------------------
3559 : monnier 605 Name: Stefan
3560 :     Date: 2000/04/07 10:00:00 EDT
3561 :     Tag: monnier-20000406-branch-handling
3562 :     Description:
3563 :    
3564 :     Improved handling of branches (mostly those generated from
3565 :     polymorphic equality), removed switchoff and changed the
3566 :     default optimization settings (more cpsopt and less flintopt).
3567 :    
3568 :     ----------------------------------------------------------------------
3569 : leunga 601 Name: Allen Leung
3570 : leunga 602 Date: 2000/04/06 01:30:00 EST
3571 :     Tag: leunga-20000406-peephole-x86-SSA-2
3572 :     Description:
3573 :    
3574 :     Forgot a few files.
3575 :    
3576 :     ----------------------------------------------------------------------
3577 :     Name: Allen Leung
3578 : leunga 601 Date: 2000/04/06 00:36:00 EST
3579 :     Tag: leunga-20000406-peephole-x86-SSA
3580 :     Description:
3581 :    
3582 :     1. New Peephole code
3583 :    
3584 :     2. Minor improvement to X86 instruction selection
3585 :    
3586 :     3. Various fixes to SSA and machine description -> code translator
3587 :    
3588 :     ----------------------------------------------------------------------
3589 : blume 595 Name: Matthias Blume
3590 :     Date: 2000/04/05 12:30:00 JST
3591 :     Tag: blume_main_v110p26p2_3
3592 :     Description:
3593 :    
3594 :     This update just merges three minor cosmetic updates to CM's sources
3595 :     to get ready for the 110.27 code freeze on Friday. No functionality
3596 :     has changed.
3597 :    
3598 :     ----------------------------------------------------------------------
3599 : leunga 593 Name: Allen Leung
3600 : leunga 594 Date: 2000/04/04 19:39:00 EST
3601 :     Tag: leunga-20000404-x86-asm
3602 :     Description:
3603 :    
3604 :     1. Fixed a problem in X86 assembly.
3605 :    
3606 :     Things like
3607 :    
3608 :     jmp %eax
3609 :     jmp (%eax)
3610 :    
3611 :     should be output as
3612 :    
3613 :     jmp *%eax
3614 :     jmp *(%eax)
3615 :    
3616 :     2. Assembly output
3617 :    
3618 :     Added a new flag
3619 :    
3620 :     "asm-indent-copies" (default to false)
3621 :    
3622 :     When this flag is on, parallel copies will be indented an extra level.
3623 :    
3624 :     ----------------------------------------------------------------------
3625 :     Name: Allen Leung
3626 : leunga 593 Date: 2000/04/04 03:18:00 EST
3627 :     Tag: leunga-20000404-C--Moby
3628 :     Description:
3629 :    
3630 :     All of these fixes are related to C--, Moby, and my own optimization
3631 :     stuff; so they shouldn't affect SML/NJ.
3632 :    
3633 :     1. X86
3634 :    
3635 :     Various fixes related floating point, and extensions.
3636 :    
3637 :     2. Alpha
3638 :    
3639 :     Some extra patterns related to loads with signed/zero extension
3640 :     provided by Fermin.
3641 :    
3642 :     3. Assembly
3643 :    
3644 : monnier 659 When generating assembly, resolve the value of client defined constants,
3645 : leunga 593 instead of generating symbolic values. This is controlled by the
3646 :     new flag "asm-resolve-constants", which is default to true.
3647 :    
3648 :     4. Machine Descriptions
3649 :    
3650 :     a. The precedence parser was slightly broken when parsing infixr symbols.
3651 :     b. The type generalizing code had the bound variables reversed, resulting
3652 :     in a problem during arity raising.
3653 :     c. Various fixes in machine descriptions.
3654 :    
3655 :     ----------------------------------------------------------------------
3656 : blume 592 Name: Matthias Blume
3657 :     Date: 2000/04/03 16:05:00 JST
3658 :     Tag: blume_main_v110p26p2_2
3659 :     Description:
3660 :    
3661 :     I eliminated coreEnv from compInfo. Access to the "Core" structure is
3662 :     now done via the ordinary static environment that is context to each
3663 :     compilation unit.
3664 :    
3665 :     To this end, I arranged that instead of "structure Core" as "structure
3666 :     _Core" is bound in the pervasive environment. Core access is done via
3667 : monnier 659 _Core (which can never be accidentally rebound because _Core is not a
3668 : blume 592 legal surface-syntax symbol).
3669 :    
3670 :     The current solution is much cleaner because the core environment is
3671 :     now simply part of the pervasive environment which is part of every
3672 :     compilation unit's context anyway. In particular, this eliminates all
3673 :     special-case handling that was necessary until now in order to deal
3674 :     with dynamic and symbolic parts of the core environment.
3675 :    
3676 :     Remaining hackery (to bind the "magic" symbol _Core) is localized in the
3677 : monnier 659 compilation manager's bootstrap compiler (actually: in the "init group"
3678 : blume 592 handling). See the comments in src/system/smlnj/init/init.cmi for
3679 :     more details.
3680 :    
3681 :     I also tried to track down all mentions of "Core" (as string argument
3682 :     to Symbol.strSymbol) in the compiler and replaced them with a
3683 :     reference to the new CoreSym.coreSym. Seems cleaner since the actual
3684 :     name appears in one place only.
3685 :    
3686 :     Binfile and bootfile format have not changed, but the switchover from
3687 :     the old "init.cmi" to the new one is a bit tricky, so I supplied new
3688 :     bootfiles anyway.
3689 :    
3690 :     ----------------------------------------------------------------------
3691 : leunga 591 Name: Allen Leung
3692 :     Date: 2000/04/02 21:17:00 EST
3693 :     Tag: leunga-20000402-mltree
3694 :     Description:
3695 : leunga 585
3696 : leunga 591 1. Renamed the constructor CALL in MLTREE by popular demand.
3697 :     2. Added a bunch of files from my repository. These are currently
3698 :     used by other non-SMLNJ backends.
3699 :    
3700 : leunga 576 ----------------------------------------------------------------------
3701 : leunga 590 Name: Allen Leung
3702 :     Date: 2000/03/31 21:15:00 EST
3703 :     Tag: leunga-20000331-aliasing
3704 :     Description:
3705 :    
3706 :     This update contains a rewritten (and hopefully more correct) module
3707 :     for extracting aliasing information from CPS.
3708 :    
3709 :     To turn on this feature:
3710 :    
3711 :     Compiler.Control.CG.memDisambiguate := true
3712 :    
3713 :     To pretty print the region information with assembly
3714 :    
3715 :     Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
3716 :    
3717 :     To control how many levels of aliasing information are printed, use:
3718 :    
3719 :     Compiler.Control.MLRISC.getInt "points-to-show-level" := n
3720 :    
3721 :     The default of n is 3.
3722 :    
3723 :     ----------------------------------------------------------------------
3724 : dbm 589 Name: David MacQueen
3725 :     Date: 2000/03/31 11:15:00 EST
3726 :     Tag: dbm-20000331-runtime_fix
3727 :     Description:
3728 :    
3729 :     This update contains:
3730 :    
3731 :     1. runtime/c-lib/c-libraries.c
3732 :     includes added in revision 1.2 caused compilation errors on hppa-hpux
3733 :    
3734 :     2. fix for bug 1556
3735 :     system/Basis/Implementation/NJ/internal-signals.sml
3736 :    
3737 :     ----------------------------------------------------------------------
3738 : blume 587 Name: Matthias Blume
3739 : blume 588 Date: 2000/03/31 18:00:00 JST
3740 :     Tag: blume_main_v110p26p2_1
3741 :     Description:
3742 :    
3743 :     This update contains:
3744 :    
3745 :     1. A small change to CM's handling of stable libraries:
3746 :     CM now maintains one "global" modmap that is used for all stable
3747 :     libraries. The use of such a global modmap maximizes sharing and
3748 :     minimizes the need for re-traversing parts of environments during
3749 :     modmap construction. (However, this has minor impact since modmap
3750 :     construction seems to account for just one percent or less of total
3751 :     compile time.)
3752 :    
3753 :     2. I added a "genmap" phase to the statistics. This is where I got the
3754 :     "one percent" number (see above).
3755 :    
3756 :     3. CM's new tool parameter mechanism just became _even_ better. :)
3757 :     - The parser understands named parameters and recursive options.
3758 :     - The "make" and "shell" tools use these new features.
3759 :     (This makes it a lot easier to cascade these tools.)
3760 :     - There is a small syntax change: named parameters use a
3761 :    
3762 :     <name> : ( <option> ... ) or
3763 :     <name> : <string>
3764 :    
3765 :     syntax. Previously, named parameters were implemented in an
3766 :     ad-hoc fashion by each tool individually (by parsing strings)
3767 :     and had the form
3768 :    
3769 :     <name>=<string>
3770 :    
3771 :     See the CM manual for a full description of these issues.
3772 :    
3773 :     ----------------------------------------------------------------------
3774 :     Name: Matthias Blume
3775 : blume 587 Date: 2000/03/30 18:00:00 JST
3776 :     Tag: blume_main_v110p26p2_0
3777 :     Description:
3778 :    
3779 :     !!!!! WARNING !!!!!!
3780 :     !! New binfiles !!
3781 :     !!!!!!!!!!!!!!!!!!!!
3782 :    
3783 :     This update contains:
3784 :    
3785 :     1. Moderate changes to CM:
3786 :    
3787 :     - Changes to CM's tools mechanism. In particular, it is now possible
3788 :     to have tools that accept additional "command line" parameters
3789 :     (specified in the .cm file at each instance where the tool's class is
3790 :     used).
3791 :    
3792 : monnier 659 This was done to accommodate the new "make" and "shell" tools which
3793 :     facilitate fairly seamless hookup to portions of code managed using
3794 : blume 587 Makefiles or Shell scripts.
3795 :    
3796 :     There are no classes "shared" or "private" anymore. Instead, the
3797 :     sharing annotation is now a parameter to the "sml" class.
3798 :    
3799 :     There is a bit of generic machinery for implementing one's own
3800 :     tools that accept command-line parameters. However, I am not yet fully
3801 :     satisfied with that part, so expect changes here in the future.
3802 :    
3803 :     All existing tools are described in the CM manual.
3804 :    
3805 : monnier 659 - Slightly better error handling. (CM now suppresses many followup
3806 : blume 587 error messages that tended to be more annoying than helpful.)
3807 :    
3808 :     2. Major changes to the compiler's static environment data structures.
3809 :    
3810 :     - no CMStaticEnv anymore.
3811 :     - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
3812 :     but it is called Environment), no conversions between different
3813 :     kinds of static environments
3814 :    
3815 :     - There is still a notion of a "modmap", but such modmaps are generated
3816 :     on demand at the time when they are needed. This sounds slow, but I
3817 :     sped up the code that generates modmaps enough for this not to lead to
3818 :     a slowdown of the compiler (at least I didn't detect any).
3819 :    
3820 :     - To facilitate rapid modmap generation, static environments now
3821 :     contain an (optional) "modtree" structure. Modtree annotations are
3822 :     constructed by the unpickler during unpickling. (This means that
3823 :     the elaborator does not have to worry about modtrees at all.)
3824 :     Modtrees have the advantage that they are compositional in the same
3825 :     way as the environment data structure itself is compositional.
3826 :     As a result, modtrees never hang on to parts of an environment that
3827 :     has already been rendered "stale" by filtering or rebinding.
3828 :    
3829 :     - I went through many, many trials and errors before arriving at the
3830 :     current solution. (The initial idea of "linkpaths" did not work.)
3831 :     But the result of all this is that I have touched a lot of files that
3832 :     depend on the "modules" and "types" data structures (most of the
3833 :     elaborator). There were a lot of changes during my "linkpath" trials
3834 :     that could have been reverted to their original state but weren't.
3835 :     Please, don't be too harsh on me for messing with this code a bit more
3836 : monnier 921 than what was strictly necessary... (I _did_ resist the temptation
3837 : blume 587 of doing any "global reformatting" to avoid an untimely death at
3838 :     Dave's hands. :)
3839 :    
3840 :     - One positive aspect of the previous point: At least I made sure that
3841 :     all files that I touched now compile without warnings (other than
3842 :     "polyEqual").
3843 :    
3844 :     - compiler now tends to run "leaner" (i.e., ties up less memory in
3845 :     redundant modmaps)
3846 :    
3847 :     ----------------------------------------------------------------------
3848 : leunga 580 Name: Allen Leung
3849 : leunga 585 Date: 2000/03/29 18:00:00
3850 :     Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
3851 : leunga 586 Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
3852 : leunga 585 Description:
3853 :    
3854 :     This update contains *MAJOR* changes to the way code is generated from CPS
3855 :     in the module mlriscGen, and in various backend modules.
3856 :    
3857 :     CHANGES
3858 :     =======
3859 :    
3860 :     1. MLRiscGen: forward propagation fix.
3861 :    
3862 :     There was a bug in forward propagation introduced at about the same time
3863 :     as the MLRISC x86 backend, which prohibits coalescing to be
3864 :     performed effectively in loops.
3865 :    
3866 :     Effect: speed up of loops in RISC architectures.
3867 :     By itself, this actually slowed down certain benchmarks on the x86.
3868 :    
3869 :     2. MLRiscGen: forward propagating addresses from consing.
3870 :    
3871 :     I've changed the way consing code is generated. Basically I separated
3872 :     out the initialization part:
3873 :    
3874 :     store tag, offset(allocptr)
3875 :     store elem1, offset+4(allocptr)
3876 :     store elem2, offset+8(allocptr)
3877 :     ...
3878 :     store elemn, offset+4n(allocptr)
3879 :    
3880 :     and the address computation part:
3881 :    
3882 :     celladdr <- offset+4+alloctpr
3883 :    
3884 :     and move the address computation part
3885 :    
3886 :     Effect: register pressure is generally lower as a result. This
3887 :     makes compilation of certain expressions much faster, such as
3888 :     long lists with non-trivial elements.
3889 :    
3890 :     [(0,0), (0,0), .... (0,0)]
3891 :    
3892 :     3. MLRiscGen: base pointer elimination.
3893 :    
3894 :     As part of the linkage mechanism, we generate the sequence:
3895 :    
3896 :     L: ... <- start of the code fragment
3897 :    
3898 :     L1:
3899 :     base pointer <- linkreg - L1 + L
3900 :    
3901 :     The base pointer was then used for computing relocatable addresses
3902 :     in the code fragment. Frequently (such as in lots of continuations)
3903 :     this is not needed. We now eliminate this sequence whenever possible.
3904 :    
3905 :     For compile time efficiency, I'm using a very stupid local heuristic.
3906 :     But in general, this should be done as a control flow analysis.
3907 :    
3908 :     Effect: Smaller code size. Speed up of most programs.
3909 :    
3910 :     4. Hppa back end
3911 :    
3912 :     Long jumps in span dependence resolution used to depend on the existence
3913 :     of the base pointer.
3914 :    
3915 :     A jump to a long label L was expanded into the following sequence:
3916 :    
3917 :     LDIL %hi(L-8192), %r29
3918 :     LDO %lo(L-8192)(%r29), %r29
3919 :     ADD %r29, baseptr, %r29
3920 :     BV,n %r0(%r29)
3921 :    
3922 :     In the presence of change (3) above, this will not work. I've changed
3923 :     it so that the following sequence of instructions are generated, which
3924 :     doesn't mention the base pointer at all:
3925 :    
3926 :     BL,n L', %r29 /* branch and link, L' + 4 -> %r29 */
3927 :     L': ADDIL L-(L'+4), %r29 /* Compute address of L */
3928 :     BV,n %r0(%r29) /* Jump */
3929 :    
3930 :     5. Alpha back end
3931 :    
3932 :     New alpha instructions LDB/LDW have been added, as per Fermin's
3933 :     suggestions. This is unrelated to all other changes.
3934 :    
3935 :     6. X86 back end
3936 :    
3937 :     I've changed andl to testl in the floating point test sequence
3938 :     whenever appropriate. The Intel optimization guide states that
3939 : monnier 659 testl is preferable to andl.
3940 : leunga 585
3941 :     7. RA (x86 only)
3942 :    
3943 :     I've improved the spill propagation algorithm, using an approximation
3944 :     of maximal weighted independent sets. This seems to be necessary to
3945 :     alleviate the negative effect in light of the slow down in (1).
3946 :    
3947 :     I'll write down the algorithm one of these days.
3948 :    
3949 :     8. MLRiscGen: frequencies
3950 :    
3951 :     I've added an annotation that states that all call gc blocks have zero
3952 :     execution frequencies. This improves register allocation on the x86.
3953 :    
3954 :     BENCHMARKS
3955 :     ==========
3956 :    
3957 :     I've only perform the comparison on 110.25.
3958 :    
3959 :     The platforms are:
3960 :    
3961 :     HPPA A four processor HP machine (E9000) with 5G of memory.
3962 :     X86 A 300Hhz Pentium II with 128M of memory, and
3963 :     SPARC An Ultra sparc 2 with 512M of memory.
3964 :    
3965 :     I used the following parameters for the SML benchmarks:
3966 :    
3967 :     @SMLalloc
3968 :     HPPA 256k
3969 :     SPARC 512k
3970 :     X86 256k
3971 :    
3972 :     COMPILATION TIME
3973 :     ----------------
3974 :     Here are the numbers comparing the compilation times of the compilers.
3975 :     I've only compared 110.25 compiling the new sources versus
3976 :     a fixpoint version of the new compiler compiling the same.
3977 :    
3978 :     110.25 New
3979 :     Total Time in RA Spill+Reload Total Time In RA Spill+Reload
3980 :     HPPA 627s 116s 2684+3584 599s 95s 1003+1879
3981 :     SPARC 892s 173s 2891+3870 708s 116s 1004+1880
3982 :     X86 999s 315s 94006+130691 987s 296s 108877+141957
3983 :    
3984 :     110.25 New
3985 :     Code Size Code Size
3986 :     HPPA 8596736 8561421
3987 :     SPARC 8974299 8785143
3988 :     X86 9029180 8716783
3989 :    
3990 :     So in summary, things are at least as good as before. Dramatic
3991 :     reduction in compilation is obtained on the Sparc; I can't explain it,
3992 :     but it is reproducible. Perhaps someone should try to reproduce this
3993 :     on their own machines.
3994 :    
3995 :     SML BENCHMARKS
3996 :     --------------
3997 :    
3998 :     On the average, all benchmarks perform at least as well as before.
3999 :    
4000 :     HPPA Compilation Time Spill+Reload Run Time
4001 :     110.25 New 110.25 New 110.25 New
4002 :    
4003 :     barnesHut 3.158 3.015 4.75% 1+1 0+0 2.980 2.922 2.00%
4004 :     boyer 6.152 5.708 7.77% 0+0 0+0 0.218 0.213 2.34%
4005 :     count-graphs 1.168 1.120 4.32% 0+0 0+0 22.705 23.073 -1.60%
4006 :     fft 0.877 0.792 10.74% 1+3 1+3 0.602 0.587 2.56%
4007 :     knuthBendix 3.180 2.857 11.32% 0+0 0+0 0.675 0.662 2.02%
4008 :     lexgen 6.190 5.290 17.01% 0+0 0+0 0.913 0.788 15.86%
4009 :     life 0.803 0.703 14.22% 25+25 0+0 0.153 0.140 9.52%
4010 :     logic 2.048 2.007 2.08% 6+6 1+1 4.133 4.008 3.12%
4011 :     mandelbrot 0.077 0.080 -4.17% 0+0 0+0 0.765 0.712 7.49%
4012 :     mlyacc 22.932 20.937 9.53% 154+181 32+57 0.468 0.430 8.91%
4013 :     nucleic 5.183 5.060 2.44% 2+2 0+0 0.125 0.120 4.17%
4014 :     ratio-regions 3.357 3.142 6.84% 0+0 0+0 116.225 113.173 2.70%
4015 :     ray 1.283 1.290 -0.52% 0+0 0+0 2.887 2.855 1.11%
4016 :     simple 6.307 6.032 4.56% 28+30 5+7 3.705 3.658 1.28%
4017 :     tsp 0.888 0.862 3.09% 0+0 0+0 7.040 6.893 2.13%
4018 :     vliw 24.378 23.455 3.94% 106+127 25+45 2.758 2.707 1.91%
4019 :     --------------------------------------------------------------------------
4020 :     Average 6.12% 4.09%
4021 :    
4022 :     SPARC Compilation Time Spill+Reload Run Time
4023 :     110.25 New 110.25 New 110.25 New
4024 :    
4025 :     barnesHut 3.778 3.592 5.20% 2+2 0+0 3.648 3.453 5.65%
4026 :     boyer 6.632 6.110 8.54% 0+0 0+0 0.258 0.242 6.90%
4027 :     count-graphs 1.435 1.325 8.30% 0+0 0+0 33.672 34.737 -3.07%
4028 :     fft 0.980 0.940 4.26% 3+9 2+6 0.838 0.827 1.41%
4029 :     knuthBendix 3.590 3.138 14.39% 0+0 0+0 0.962 0.967 -0.52%
4030 :     lexgen 6.593 6.072 8.59% 1+1 0+0 1.077 1.078 -0.15%
4031 :     life 0.972 0.868 11.90% 26+26 0+0 0.143 0.140 2.38%
4032 :     logic 2.525 2.387 5.80% 7+7 1+1 5.625 5.158 9.05%
4033 :     mandelbrot 0.090 0.093 -3.57% 0+0 0+0 0.855 0.728 17.39%
4034 :     mlyacc 26.732 23.827 12.19% 162+189 32+57 0.550 0.560 -1.79%
4035 :     nucleic 6.233 6.197 0.59% 3+3 0+0 0.163 0.173 -5.77%
4036 :     ratio-regions 3.780 3.507 7.79% 0+0 0+0 133.993 131.035 2.26%
4037 :     ray 1.595 1.550 2.90% 1+1 0+0 3.440 3.418 0.63%
4038 :     simple 6.972 6.487 7.48% 29+32 5+7 3.523 3.525 -0.05%
4039 :     tsp 1.115 1.063 4.86% 0+0 0+0 7.393 7.265 1.77%
4040 :     vliw 27.765 24.818 11.87% 110+135 25+45 2.265 2.135 6.09%
4041 :     ----------------------------------------------------------------------------
4042 :     Average 6.94% 2.64%
4043 :    
4044 :     X86 Compilation Time Spill+Reload Run Time
4045 :     110.25 New 110.25 New 110.25 New
4046 :    
4047 :     barnesHut 5.530 5.420 2.03% 593+893 597+915 3.532 3.440 2.66%
4048 :     boyer 8.768 7.747 13.19% 493+199 301+289 0.327 0.297 10.11%
4049 :     count-graphs 2.040 2.010 1.49% 298+394 315+457 26.578 28.660 -7.26%
4050 :     fft 1.327 1.302 1.92% 112+209 115+210 1.055 0.962 9.71%
4051 :     knuthBendix 5.218 5.475 -4.69% 451+598 510+650 0.928 0.932 -0.36%
4052 :     lexgen 9.970 9.623 3.60% 1014+841 1157+885 0.947 0.928 1.97%
4053 :     life 1.183 1.183 0.00% 162+182 145+148 0.127 0.103 22.58%
4054 :     logic 3.285 3.512 -6.45% 514+684 591+836 5.682 5.577 1.88%
4055 :     mandelbrot 0.147 0.143 2.33% 38+41 33+54 0.703 0.690 1.93%
4056 :     mlyacc 35.457 32.763 8.22% 3496+4564 3611+4860 0.552 0.550 0.30%
4057 :     nucleic 7.100 6.888 3.07% 239+168 201+158 0.175 0.173 0.96%
4058 :     ratio-regions 6.388 6.843 -6.65% 1182+257 981+300 120.142 120.345 -0.17%
4059 :     ray 2.332 2.338 -0.29% 346+398 402+494 3.593 3.540 1.51%
4060 :     simple 9.912 9.903 0.08% 1475+941 1579+1168 3.057 3.178 -3.83%
4061 :     tsp 1.623 1.532 5.98% 266+200 250+211 8.045 7.878 2.12%
4062 :     vliw 33.947 35.470 -4.29% 2629+2774 2877+3171 2.072 1.890 9.61%
4063 :     ----------------------------------------------------------------------------
4064 :     Average 1.22% 3.36%
4065 :    
4066 :     ----------------------------------------------------------------------
4067 :     Name: Allen Leung
4068 : leunga 583 Date: 2000/03/23 16:25:00
4069 :     Tag: leunga-20000323-fix_x86_alpha
4070 :     Description:
4071 :    
4072 :     1. X86 fixes/changes
4073 :    
4074 :     a. The old code generated for SETcc was completely wrong.
4075 :     The Intel optimization guide is VERY misleading.
4076 :    
4077 :     2. ALPHA fixes/changes
4078 :    
4079 :     a. Added the instructions LDBU, LDWU, STB, STW as per Fermin's suggestion.
4080 :     b. Added a new mode byteWordLoadStores to the functor parameter to Alpha()
4081 :     c. Added reassociation code for address computation.
4082 :    
4083 :     ----------------------------------------------------------------------
4084 :     Name: Allen Leung
4085 : leunga 580 Date: 2000/03/22 01:23:00
4086 :     Tag: leunga-20000322-fix_x86_hppa_ra
4087 :     Description:
4088 :    
4089 :     1. X86 fixes/changes
4090 :    
4091 :     a. x86Rewrite bug with MUL3 (found by Lal)
4092 :     b. Added the instructions FSTS, FSTL
4093 :    
4094 :     2. PA-RISC fixes/changes
4095 :    
4096 :     a. B label should not be a delay slot candidate! Why did this work?
4097 :     b. ADDT(32, REG(32, r), LI n) now generates one instruction instead of two,
4098 :     as it should be.
4099 :     c. The assembly syntax for fstds and fstdd was wrong.
4100 :     d. Added the composite instruction COMICLR/LDO, which is the immediate
4101 :     operand variant of COMCLR/LDO.
4102 :    
4103 :     3. Generic MLRISC
4104 :    
4105 :     a. shuffle.sml rewritten to be slightly more efficient
4106 :     b. DIV bug in mltree-simplify fixed (found by Fermin)
4107 :    
4108 :     4. Register Allocator
4109 :    
4110 :     a. I now release the interference graph earlier during spilling.
4111 :     May improve memory usage.
4112 :    
4113 :     ----------------------------------------------------------------------
4114 : blume 577 Name: Matthias Blume
4115 : blume 578 Date: 2000/03/14 14:15:32
4116 :     Tag: blume_main_v110p26p1_2
4117 :     Description:
4118 :    
4119 :     1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
4120 :     additional argument called "template" which is an optional string that
4121 : monnier 659 specifies the layout of the tool command line. See the CM manual for
4122 : blume 578 explanation.
4123 :    
4124 : monnier 659 2. A special-purpose tool can be "registered" by simply dropping the
4125 : blume 578 corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
4126 :     directory where the .cm file lives that uses this tool. (The
4127 :     behavior/misfeature until now was to look for the tool description
4128 :     files in the current working directory.) As before, tool description
4129 :     files could also be anchored -- in which case they can live anywhere
4130 :     they like. Following the recent e-mail discussion, this change should
4131 :     make it easier to have special-purpose tools that are shipped together
4132 :     with the sources of the program that uses them.
4133 :    
4134 :     ----------------------------------------------------------------------
4135 :     Name: Matthias Blume
4136 : blume 577 Date: 2000/03/10 07:48:34
4137 :     Tag: blume_main_v110p26p1_1
4138 :     Description:
4139 :    
4140 :     I added a re-written version of Dave's fixpt script to src/system.
4141 :     Changes relative to the original version:
4142 :     - sh-ified (not everybody has ksh)
4143 :     - automatically figures out which architecture it runs on
4144 :     - uses ./makeml a bit more cleverly
4145 :     - never invokes ./installml (and, thus, does not clobber your
4146 :     good and working installation of sml in case something goes wrong)
4147 :     - accepts max iteration count using option "-iter <n>"
4148 :     - accepts a "base" name using option "-base <base>"
4149 :    
4150 :     It does not build any extraneous heap images but directly rebuilds
4151 :     bin- and boot-hierarchies using makeml's "-rebuild" switch. Finally,
4152 :     it can incorporate existing bin- and boot- hierarchies. For example,
4153 :     suppose the base is set to "sml" (which is the default). Then it
4154 :     successively builds
4155 :    
4156 :     sml.bin.<arch>-unix and sml.boot.<arch>-unix
4157 :     then sml1.bin.<arch>-unix and sml1.boot.<arch>-unix
4158 :     then sml2.bin.<arch>-unix and sml2.boot.<arch>-unix
4159 :     ...
4160 :     then sml<n>.bin.<arch>-unix and sml<n>.boot.<arch>-unix
4161 :    
4162 :     and so on. If any of these already exist, it will just use what's
4163 :     there. In particular, many people will have the initial set of bin
4164 :     and boot files around, so this saves time for at least one full
4165 :     rebuild. Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
4166 :     <k>=1,2,... is normally not a good idea when invoking fixpt. However,
4167 :     they might be the result of an earlier partial run of fixpt (which
4168 : monnier 659 perhaps got accidentally killed). In this case, fixpt will quickly
4169 : blume 577 move through what exists before continuing where it left off earlier,
4170 :     and, thus, saves a lot of time.
4171 :    
4172 :     ----------------------------------------------------------------------
4173 : leunga 576 Name: Allen Leung
4174 :     Date: 00/03/10 02:20:00
4175 :     Tag: leunga-20000310-fix_x86_asm_ra
4176 :     Description:
4177 : dbm 570
4178 : leunga 576 More assembly output problems involving the indexed addressing mode
4179 :     on the x86 have been found and corrected. Thanks to Fermin Reig for the
4180 :     fix.
4181 :    
4182 :     The interface and implementation of the register allocator have been changed
4183 :     slightly to accommodate the possibility to skip the register allocation
4184 :     phases completely and go directly to memory allocation. This is needed
4185 :     for C-- use.
4186 :    
4187 : dbm 570 ----------------------------------------------------------------------
4188 : blume 572 Name: Matthias Blume
4189 : blume 575 Date: 00/03/09 10:23:53
4190 :     Tag: blume_main_v110p26p1_0
4191 :     Description:
4192 :    
4193 :     * Complete re-organization of library names. Many libraries have been
4194 :     consolidated so that they share the same path anchor. For example,
4195 :     all MLRISC-related libraries are anchored at MLRISC, most libraries that
4196 :     are SML/NJ-specific are under "smlnj". Notice that names like
4197 :     host-cmb.cm or host-compiler.cm no longer exist. See system/README
4198 :     for a complete description of the new naming scheme. Quick reference:
4199 :    
4200 :     host-cmb.cm -> smlnj/cmb.cm
4201 :     host-compiler.cm -> smlnj/compiler.cm
4202 :     full-cm.cm -> smlnj/cm.cm
4203 :     <arch>-<os>.cm -> smlnj/cmb/<arch>-<os>.cm
4204 :     <arch>-compiler.cm -> smlnj/compiler/<arch>.cm
4205 :    
4206 :     * Bug fixes in CM.
4207 :     - exceptions in user code are being passed through (i.e., reach top level)
4208 :     - more bugs in paranoia mode fixed
4209 :     - bug related to checking group owners fixed
4210 :    
4211 :     * New install.sh script that automagically fetches archive files:
4212 :     The new file config/srcarchiveurl must contain the URL of the
4213 :     (remote) directory that contains bin files (or other source archives).
4214 :     If install.sh does not find the archive locally, it tries to get
4215 :     it from that remote directory.
4216 :     This should simplify installation further: For machines that have
4217 :     access to the internet, just fetch <version>-config.tgz, unpack it,
4218 : monnier 659 edit config/targets, and go (run config/install.sh). The script will
4219 : blume 575 fetch everything else that it might need all by itself.
4220 :    
4221 :     For CVS users, this mechanism is not relevant for source archives, but
4222 :     it is convenient for getting new sets of binfiles.
4223 :    
4224 :     Archives should be tar files compressed with either gzip, compress, or
4225 :     bzip2. The script recognizes .tgz, .tar, tar.gz, tz, .tar.Z, and .tar.bz2.
4226 :    
4227 :     ----------------------------------------------------------------------
4228 :     Name: Matthias Blume
4229 : blume 572 Date: 2000/03/07 04:01:04
4230 :     Tag: blume_main_v110_26_2
4231 : dbm 570 Description:
4232 : blume 572 - size info in BOOTLIST
4233 :     * no fixed upper limits for number of bootfiles or length of
4234 :     bootfile names in runtime
4235 :     * falling back to old behavior if no BOOTLIST size info found
4236 :     - allocation size heuristics in .run-sml
4237 :     * tries to read cache size from /proc/cpuinfo (this is important for
4238 :     small-cache Celeron systems!)
4239 :     - install.sh robustified
4240 :     - CM manual updates
4241 :     - paranoid mode
4242 :     * no more CMB.deliver() (i.e., all done by CMB.make())
4243 :     * can re-use existing sml.boot.* files
4244 :     * init.cmi now treated as library
4245 :     * library stamps for consistency checks
4246 :     - sml.boot.<arch>-<os>/PIDMAP file
4247 :     * This file is read by the CM startup code. This is used to minimize
4248 :     the amount of dynamic state that needs to be stowed away for the
4249 :     purpose of sharing between interactive system and user code.
4250 :     - CM.Anchor.anchor instead of CM.Anchor.{set,cancel}
4251 :     * Upon request by Elsa. Anchors now controlled by get-set-pair
4252 :     like most other CM state variables.
4253 :     - Compiler.CMSA eliminated
4254 :     * No longer supported by CM anyway.
4255 :     - fixed bugs in pickler that kept biting Stefan
4256 :     * past refs to past refs (was caused by the possibility that
4257 :     ad-hoc sharing is more discriminating than hash-cons sharing)
4258 :     * integer overflow on LargeInt.minInt
4259 :     - ml-{lex,yacc} build scripts now use new mechanism
4260 :     for building standalone programs
4261 :     - fixed several gcc -Wall warnings that were caused by missing header
4262 :     files, missing initializations, etc., in runtime (not all warnings
4263 :     eliminated, though)

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