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

SCM Repository

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

Annotation of /sml/trunk/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1115 - (view) (download)

1 : dbm 570 This is the HISTORY file for the Yale SML/NJ CVS repository.
2 :    
3 :     An entry should be made for _every_ commit to the repository.
4 :     The entries in this file will be used when creating the README
5 :     for new versions, so keep that in mind when writing the
6 :     description.
7 :    
8 :     The form of an entry should be:
9 :    
10 :     Name:
11 : george 1003 Date: yyyy/mm/dd
12 : dbm 570 Tag: <post-commit CVS tag>
13 :     Description:
14 : leunga 1025
15 : george 1003 ----------------------------------------------------------------------
16 : leunga 1115 Name: Allen Leung
17 :     Date: 2002/03/03 13:20:00 EST
18 :     Tag: leunga-20020303-mlrisc-tools
19 :    
20 :     Added #[ ... ] expressions to mlrisc tools
21 :    
22 :     ----------------------------------------------------------------------
23 : blume 1096 Name: Matthias Blume
24 : blume 1098 Date: 2002/02/27 12:29:00 EST
25 :     Tag: blume-20020227-cdebug
26 :     Description:
27 :    
28 :     - made types in structure C and C_Debug to be equal
29 :     - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
30 :     - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
31 :    
32 :     ----------------------------------------------------------------------
33 :     Name: Matthias Blume
34 : blume 1096 Date: 2002/02/26 12:00:00 EST
35 :     Tag: blume-20020226-ffi
36 :     Description:
37 :    
38 :     1. Fixed a minor bug in CM's "noweb" tool:
39 :     If numbering is turned off, then truly don't number (i.e., do not
40 :     supply the -L option to noweb). The previous behavior was to supply
41 :     -L'' -- which caused noweb to use the "default" line numbering scheme.
42 :     Thanks to Chris Richards for pointing this out (and supplying the fix).
43 :    
44 :     2. Once again, I reworked some aspects of the FFI:
45 :    
46 :     A. The incomplete/complete type business:
47 :    
48 :     - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
49 :     gone!
50 :     - ML types representing an incomplete type are now *equal* to
51 :     ML types representing their corresponding complete types (just like
52 :     in C). This is still safe because ml-nlffigen will not generate
53 :     RTTI for incomplete types, nor will it generate functions that
54 :     require access to such RTTI. But when ML code generated from both
55 :     incomplete and complete versions of the C type meet, the ML types
56 :     are trivially interoperable.
57 :    
58 :     NOTE: These changes restore the full generality of the translation
59 :     (which was previously lost when I eliminated functorization)!
60 :    
61 :     B. Enum types:
62 :    
63 :     - Structure C now has a type constructor "enum" that is similar to
64 :     how the "su" constructor works. However, "enum" is not a phantom
65 :     type because each "T enum" has values (and is isomorphic to
66 :     MLRep.Signed.int).
67 :     - There are generic access operations for enum objects (using
68 :     MLRep.Signed.int).
69 :     - ml-nlffigen will generate a structure E_foo for each "enum foo".
70 :     * The structure contains the definition of type "mlrep" (the ML-side
71 :     representation type of the enum). Normally, mlrep is the same
72 :     as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
73 :     then mlrep will be defined as a datatype -- thus facilitating
74 :     pattern matching on mlrep values.
75 :     ("-ec" will be suppressed if there are duplicate values in an
76 :     enumeration.)
77 :     * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
78 :     will be generated for each C enum constant xxx.
79 :     * Conversion functions m2i and i2m convert between mlrep and
80 :     MLRep.Signed.int. (Without "-ec", these functions are identities.)
81 :     * Coversion functions c and ml convert between mlrep and "tag enum".
82 :     * Access functions (get/set) fetch and store mlrep values.
83 :     - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
84 :     enumerations are merged into one single enumeration represented by
85 :     structure E_'.
86 :    
87 :     ----------------------------------------------------------------------
88 : leunga 1080 Name: Allen Leung
89 : leunga 1094 Date: 2002/02/25 04:45:00 EST
90 :     Tag: leunga-20020225-cps-spill
91 :    
92 :     This is a new implementation of the CPS spill phase.
93 :     The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
94 :     In case of problems, replace it with the old file spill.sml
95 :    
96 :     The current compiler runs into some serious performance problems when
97 :     constructing a large record. This can happen when we try to compile a
98 :     structure with many items. Even a very simple structure like the following
99 :     makes the compiler slow down.
100 :    
101 :     structure Foo = struct
102 :     val x_1 = 0w1 : Word32.int
103 :     val x_2 = 0w2 : Word32.int
104 :     val x_3 = 0w3 : Word32.int
105 :     ...
106 :     val x_N = 0wN : Word32.int
107 :     end
108 :    
109 :     The following table shows the compile time, from N=1000 to N=4000,
110 :     with the old compiler:
111 :    
112 :     N
113 :     1000 CPS 100 spill 0.04u 0.00s 0.00g
114 :     MLRISC ra 0.06u 0.00s 0.05g
115 :     (spills = 0 reloads = 0)
116 :     TOTAL 0.63u 0.07s 0.21g
117 :    
118 :     1100 CPS 100 spill 8.25u 0.32s 0.64g
119 :     MLRISC ra 5.68u 0.59s 3.93g
120 :     (spills = 0 reloads = 0)
121 :     TOTAL 14.71u 0.99s 4.81g
122 :    
123 :     1500 CPS 100 spill 58.55u 2.34s 1.74g
124 :     MLRISC ra 5.54u 0.65s 3.91g
125 :     (spills = 543 reloads = 1082)
126 :     TOTAL 65.40u 3.13s 6.00g
127 :    
128 :     2000 CPS 100 spill 126.69u 4.84s 3.08g
129 :     MLRISC ra 0.80u 0.10s 0.55g
130 :     (spills = 42 reloads = 84)
131 :     TOTAL 129.42u 5.10s 4.13g
132 :    
133 :     3000 CPS 100 spill 675.59u 19.03s 11.64g
134 :     MLRISC ra 2.69u 0.27s 1.38g
135 :     (spills = 62 reloads = 124)
136 :     TOTAL 682.48u 19.61s 13.99g
137 :    
138 :     4000 CPS 100 spill 2362.82u 56.28s 43.60g
139 :     MLRISC ra 4.96u 0.27s 2.72g
140 :     (spills = 85 reloads = 170)
141 :     TOTAL 2375.26u 57.21s 48.00g
142 :    
143 :     As you can see the old cps spill module suffers from some serious
144 : leunga 1115 performance problem. But since I cannot decipher the old code fully,
145 :     instead of patching the problems up, I'm reimplementing it
146 : leunga 1094 with a different algorithm. The new code is more modular,
147 :     smaller when compiled, and substantially faster
148 :     (O(n log n) time and O(n) space). Timing of the new spill module:
149 :    
150 :     4000 CPS 100 spill 0.02u 0.00s 0.00g
151 :     MLRISC ra 0.25u 0.02s 0.15g
152 :     (spills=1 reloads=3)
153 :     TOTAL 7.74u 0.34s 1.62g
154 :    
155 :     Implementation details:
156 :    
157 :     As far as I can tell, the purpose of the CPS spill module is to make sure the
158 :     number of live variables at any program point (the bandwidth)
159 :     does not exceed a certain limit, which is determined by the
160 :     size of the spill area.
161 :    
162 :     When the bandwidth is too large, we decrease the register pressure by
163 :     packing live variables into spill records. How we achieve this is
164 :     completely different than what we did in the old code.
165 :    
166 :     First, there is something about the MLRiscGen code generator
167 :     that we should be aware of:
168 :    
169 :     o MLRiscGen performs code motion!
170 :    
171 :     In particular, it will move floating point computations and
172 :     address computations involving only the heap pointer to
173 :     their use sites (if there is only a single use).
174 :     What this means is that if we have a CPS record construction
175 :     statement
176 :    
177 :     RECORD(k,vl,w,e)
178 :    
179 :     we should never count the new record address w as live if w
180 :     has only one use (which is often the case).
181 :    
182 :     We should do something similar to floating point, but the transformation
183 :     there is much more complex, so I won't deal with that.
184 :    
185 :     Secondly, there are now two new cps primops at our disposal:
186 :    
187 :     1. rawrecord of record_kind option
188 :     This pure operator allocates some uninitialized storage from the heap.
189 :     There are two forms:
190 :    
191 :     rawrecord NONE [INT n] allocates a tagless record of length n
192 :     rawrecord (SOME rk) [INT n] allocates a tagged record of length n
193 :     and initializes the tag.
194 :    
195 :     2. rawupdate of cty
196 :     rawupdate cty (v,i,x)
197 :     Assigns to x to the ith component of record v.
198 :     The storelist is not updated.
199 :    
200 :     We use these new primops for both spilling and increment record construction.
201 :    
202 :     1. Spilling.
203 :    
204 :     This is implemented with a linear scan algorithm (but generalized
205 :     to trees). The algorithm will create a single spill record at the
206 :     beginning of the cps function and use rawupdate to spill to it,
207 :     and SELECT or SELp to reload from it. So both spills and reloads
208 :     are fine-grain operations. In contrast, in the old algorithm
209 :     "spills" have to be bundled together in records.
210 :    
211 :     Ideally, we should sink the spill record construction to where
212 :     it is needed. We can even split the spill record into multiple ones
213 :     at the places where they are needed. But CPS is not a good
214 :     representation for global code motion, so I'll keep it simple and
215 :     am not attempting this.
216 :    
217 :     2. Incremental record construction (aka record splitting).
218 :    
219 :     Long records with many component values which are simulatenously live
220 :     (recall that single use record addresses are not considered to
221 :     be live) are constructed with rawrecord and rawupdate.
222 :     We allocate space on the heap with rawrecord first, then gradually
223 :     fill it in with rawupdate. This is the technique suggested to me
224 :     by Matthias.
225 :    
226 :     Some restrictions on when this is applicable:
227 :     1. It is not a VECTOR record. The code generator currently does not handle
228 :     this case. VECTOR record uses double indirection like arrays.
229 :     2. All the record component values are defined in the same "basic block"
230 :     as the record constructor. This is to prevent speculative
231 :     record construction.
232 :    
233 :     ----------------------------------------------------------------------
234 :     Name: Allen Leung
235 : leunga 1086 Date: 2002/02/22 01:02:00 EST
236 :     Tag: leunga-20020222-mlrisc-tools
237 :    
238 :     Minor bug fixes in the parser and rewriter
239 :    
240 :     ----------------------------------------------------------------------
241 :     Name: Allen Leung
242 : leunga 1085 Date: 2002/02/21 20:20:00 EST
243 :     Tag: leunga-20020221-peephole
244 :    
245 :     Regenerated the peephole files. Some contained typos in the specification
246 :     and some didn't compile because of pretty printing bugs in the old version
247 :     of 'nowhere'.
248 :    
249 :     ----------------------------------------------------------------------
250 :     Name: Allen Leung
251 : leunga 1079 Date: 2002/02/19 20:20:00 EST
252 :     Tag: leunga-20020219-mlrisc-tools
253 :     Description:
254 :    
255 :     Minor bug fixes to the mlrisc-tools library:
256 :    
257 :     1. Fixed up parsing colon suffixed keywords
258 :     2. Added the ability to shut the error messages up
259 :     3. Reimplemented the pretty printer and fixed up/improved
260 :     the pretty printing of handle and -> types.
261 :     4. Fixed up generation of literal symbols in the nowhere tool.
262 :     5. Added some SML keywords to to sml.sty
263 :    
264 :     ----------------------------------------------------------------------
265 :     Name: Matthias Blume
266 : blume 1078 Date: 2002/02/19 16:20:00 EST
267 :     Tag: blume-20020219-cmffi
268 :     Description:
269 :    
270 :     A wild mix of changes, some minor, some major:
271 :    
272 :     * All C FFI-related libraries are now anchored under $c:
273 :     $/c.cm --> $c/c.cm
274 :     $/c-int.cm --> $c/internals/c-int.cm
275 :     $/memory.cm --> $c/memory/memory.cm
276 :    
277 :     * "make" tool (in CM) now treats its argument pathname slightly
278 :     differently:
279 :     1. If the native expansion is an absolute name, then before invoking
280 :     the "make" command on it, CM will apply OS.Path.mkRelative
281 :     (with relativeTo = OS.FileSys.getDir()) to it.
282 :     2. The argument will be passed through to subsequent phases of CM
283 :     processing without "going native". In particular, if the argument
284 :     was an anchored path, then "make" will not lose track of that anchor.
285 :    
286 :     * Compiler backends now "know" their respective C calling conventions
287 :     instead of having to be told about it by ml-nlffigen. This relieves
288 :     ml-nlffigen from one of its burdens.
289 :    
290 :     * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
291 :    
292 :     * Export C_DEBUG and C_Debug from $c/c.cm.
293 :    
294 :     * C type encoding in ml-nlffi-lib has been improved to model the conceptual
295 :     subtyping relationship between incomplete pointers and their complete
296 :     counterparts. For this, ('t, 'c) ptr has been changed to 'o ptr --
297 :     with the convention of instantiating 'o with ('t, 'c) obj whenever
298 :     the pointer target type is complete. In the incomplete case, 'o
299 :     will be instantiated with some "'c iobj" -- a type obtained by
300 :     using one of the functors PointerToIncompleteType or PointerToCompleteType.
301 :    
302 :     Operations that work on both incomplete and complete pointer types are
303 :     typed as taking an 'o ptr while operations that require the target to
304 :     be known are typed as taking some ('t, 'c) obj ptr.
305 :    
306 :     voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
307 :     where void is an eqtype without any values. This makes it possible
308 :     to work on voidptr values using functions meant to operate on light
309 :     incomplete pointers.
310 :    
311 :     * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
312 :     been vastly simplified.
313 :    
314 :     ----------------------------------------------------------------------
315 :     Name: Matthias Blume
316 : blume 1076 Date: 2002/02/19 10:48:00 EST
317 :     Tag: blume-20020219-pqfix
318 :     Description:
319 :    
320 :     Applied Chris Okasaki's bug fix for priority queues.
321 :    
322 :     ----------------------------------------------------------------------
323 :     Name: Matthias Blume
324 : blume 1073 Date: 2002/02/15 17:05:00
325 : blume 1069 Tag: Release_110_39
326 :     Description:
327 :    
328 : blume 1073 Last-minute retagging is becoming a tradition... :-(
329 :    
330 :     This is the working release 110.39.
331 :    
332 :     ----------------------------------------------------------------------
333 :     Name: Matthias Blume
334 :     Date: 2002/02/15 16:00:00 EST
335 :     Tag: Release_110_39-orig
336 :     Description:
337 :    
338 : blume 1069 Working release 110.39. New bootfiles.
339 :    
340 : blume 1073 (Update: There was a small bug in the installer so it wouldn't work
341 :     with all shells. So I retagged. -Matthias)
342 :    
343 : blume 1069 ----------------------------------------------------------------------
344 :     Name: Matthias Blume
345 : blume 1068 Date: 2002/02/15 14:17:00 EST
346 :     Tag: blume-20020215-showbindings
347 :     Description:
348 :    
349 :     Added EnvRef.listBoundSymbols and CM.State.showBindings. Especially
350 :     the latter can be useful for exploring what bindings are available at
351 :     the interactive prompt. (The first function returns only the list
352 :     of symbols that are really bound, the second prints those but also the
353 :     ones that CM's autoloading mechanism knows about.)
354 :    
355 :     ----------------------------------------------------------------------
356 :     Name: Matthias Blume
357 : blume 1067 Date: 2002/02/15 12:08:00 EST
358 :     Tag: blume-20020215-iptrs
359 :     Description:
360 :    
361 :     Two improvements to ml-nlffigen:
362 :    
363 :     1. Write files only if they do not exist or if their current contents
364 :     do not coincide with what's being written. (That is, avoid messing
365 :     with the time stamps unless absolutely necessary.)
366 :    
367 :     2. Implement a "repository" mechanism for generated files related
368 :     to "incomplete pointer types". See the README file for details.
369 :    
370 :     ----------------------------------------------------------------------
371 :     Name: Matthias Blume
372 : blume 1066 Date: 2002/02/14 11:50:00 EST
373 :     Tag: blume-20020214-quote
374 :     Description:
375 :    
376 :     Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm). This is required
377 :     because of the new and improved tag generation scheme. (Thanks to Allen
378 :     Leung for pointing it out.)
379 :    
380 :     ----------------------------------------------------------------------
381 : george 1065 Name: Lal George
382 :     Date: 2002/02/14 09:55:27 EST 2002
383 :     Tag: george-20020214-isabelle-bug
384 :     Description:
385 :    
386 :     Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
387 :     of Isabelle on the x86.
388 :    
389 :     From Allen:
390 :     -----------
391 :     I've found the problem:
392 :    
393 :     in ra-core.sml, I use the counter "blocked" to keep track of the
394 :     true number of elements in the freeze queue. When the counter goes
395 :     to zero, I skip examining the queue. But I've messed up the
396 :     bookkeeping in combine():
397 :    
398 :     else ();
399 :     case !ucol of
400 :     PSEUDO => (if !cntv > 0 then
401 :     (if !cntu > 0 then blocked := !blocked - 1 else ();
402 :     ^^^^^^^^^^^^^^^^^^^^^^^
403 :     moveu := mergeMoveList(!movev, !moveu)
404 :     )
405 :     else ();
406 :    
407 :     combine() is called to coalesce two nodes u and v.
408 :     I think I was thinking that if the move counts of u and v are both
409 :     greater than zero then after they are coalesced then one node is
410 :     removed from the freeze queue. Apparently I was thinking that
411 :     both u and v are of low degree, but that's clearly not necessarily true.
412 :    
413 :    
414 :     02/12/2002:
415 :     Here's the patch. HOL now compiles.
416 :    
417 :     I don't know how this impact on performance (compile
418 :     time or runtime). This bug caused the RA (especially on the x86)
419 :     to go thru the potential spill phase when there are still nodes on the
420 :     freeze queue.
421 :    
422 :    
423 :    
424 :    
425 :     ----------------------------------------------------------------------
426 : blume 1058 Name: Matthias Blume
427 : blume 1064 Date: 2002/02/13 22:40:00 EST
428 :     Tag: blume-20020213-fptr-rtti
429 :     Description:
430 :    
431 :     Fixed a bug in ml-nlffigen that was introduced with one of the previous
432 :     updates.
433 :    
434 :     ----------------------------------------------------------------------
435 :     Name: Matthias Blume
436 : blume 1063 Date: 2002/02/13 16:41:00 EST
437 :     Tag: blume-20020213-cmlpq
438 :     Description:
439 :    
440 :     Added new priority queue export symbols (which have just been added to
441 :     smlnj-lib.cm) to CML's version of smlnj-lib.cm. (Otherwise CML would
442 :     not compile and the installer would choke.)
443 :    
444 :     ----------------------------------------------------------------------
445 :     Name: Matthias Blume
446 : blume 1062 Date: 2002/02/13 16:15:00 EST
447 :     Tag: blume-20020213-various
448 :     Description:
449 :    
450 :     1. More tweaks to ml-nlffigen:
451 :    
452 :     - better internal datastructures (resulting in slight speedup)
453 :     - "-match" option requires exact match
454 :     - "localized" gensym counters (untagged structs/unions nested within
455 :     other structs/unions or within typedefs get a fresh counter; their
456 :     tag will be prefixed by a concatenation of their parents' tags)
457 :     - bug fixes (related to calculation of transitive closure of types
458 :     to be included in the output)
459 :    
460 :     2. Minor Basis updates:
461 :    
462 :     - added implementations for List.collate and Option.app
463 :    
464 :     ----------------------------------------------------------------------
465 :     Name: Matthias Blume
466 : blume 1060 Date: 2002/02/11 15:55:00 EST
467 :     Tag: blume-20020211-gensym
468 :     Description:
469 :    
470 :     Added a "-gensym" option to command line of ml-nlffigen. This can be
471 :     used to specify a "stem" -- a string that is inserted in all "gensym'd"
472 :     names (ML structure names that correspond to unnamed C structs, unions,
473 :     and enums), so that separate runs of ml-nlffigen do not clash.
474 :    
475 :     ----------------------------------------------------------------------
476 :     Name: Matthias Blume
477 : blume 1059 Date: 2002/02/11 12:05:00 EST
478 :     Tag: blume-20020211-gensml
479 :     Description:
480 :    
481 :     A quick fix for a problem with GenSML (in the pgraph-util library):
482 :     Make generation of toplevel "local" optional. (Strictly speaking,
483 :     signature definitions within "local" are not legal SML.)
484 :    
485 :     Other than that: updates to INSTALL and cm/TODO.
486 :    
487 :     ----------------------------------------------------------------------
488 :     Name: Matthias Blume
489 : blume 1058 Date: 2002/02/08 15:00:00 EST
490 :     Tag: blume-20020208-uniquepid
491 :     Description:
492 :    
493 :     0. Version number has been bumped to 110.38.1. NEW BOOTFILES!!!
494 :    
495 :     1. The installer (config/install.sh) has gotten smarter:
496 :    
497 :     - Configuration options are a bit easier to specify now
498 :     (in config/targets).
499 :     - Bug in recognizing .tar.bz2 files fixed.
500 :     - Installer automatically resolves dependencies between
501 :     configuration options (e.g., if you ask for eXene, you will
502 :     also get cml -- regardless whether you asked for it or not).
503 :     - Installer can run in "quieter mode" by setting the environment
504 :     variable INSTALL_QUIETLY to "true". "Quieter" does not mean
505 :     "completely silent", though.
506 :     - Build HashCons library as part of smlnj-lib.
507 :    
508 :     2. A new scheme for assigning persistent identifiers to compilation
509 :     units (and, by extension, to types etc.) has been put into place.
510 :     This fixes a long-standing bug where types and even dynamic values
511 :     can get internally confused, thereby compromising type safety
512 :     (abstraction) and dynamic correctness. See
513 :    
514 :     http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
515 :    
516 :     for an example of how things could go wrong until now.
517 :    
518 :     The downside of the new scheme is that pids are not quite as
519 :     persistent as they used to be: CM will generate a fresh pid
520 :     for every compilation unit that it thinks it sees for the first
521 :     time. That means that if you compile starting from a clean, fresh
522 :     source tree at two different times, you end up with different
523 :     binaries.
524 :    
525 :     Cutoff recompilation, however, has not been compromised because
526 :     CM keeps pid information in special caches between runs.
527 :    
528 :     ----------------------------------------------------------------------
529 : george 1053 Name: Lal George
530 : george 1055 Date: 2002/02/07 15:34:13 EST 2002
531 :     Tag: <none>
532 :     Description:
533 :    
534 :     Compilers that generate assembly code may produce global labels
535 :     whose value is resolved at link time. The various peephole optimization
536 :     modules did not take this in account.
537 :    
538 :     TODO. The Labels.addrOf function should really return an option
539 :     type so that clients are forced to deal with this issue, rather
540 :     than an exception being raised.
541 :    
542 :     ----------------------------------------------------------------------
543 :     Name: Lal George
544 : george 1053 Date: 2002/02/06 13:55:02 EST
545 :     Tag: george-20020206-ra-breakup
546 :     Description:
547 :    
548 :     1. A bug fix from Allen.
549 :    
550 :     A typo causes extra fstp %st(0)'s to be generated at compensation
551 :     edges, which might cause stack underflow traps at runtime. This
552 :     occurs in fft where there are extraneous fstps right before the 'into'
553 :     trap instruction (in this case they are harmless since none of the
554 :     integers overflow.)
555 :    
556 : george 1054 2. Pulled out various utility modules that were embedded in the modules
557 : george 1053 of the register allocator. I need these modules for other purposes, but
558 : george 1054 they are not complete enough to put into a library (just yet).
559 : george 1053 ----------------------------------------------------------------------
560 : blume 1049 Name: Matthias Blume
561 :     Date: 2002/01/31 16:05:00 EST
562 :     Tag: blume-20020131-sparc-ccalls
563 :     Description:
564 :    
565 :     1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
566 :     of extra stack space by mistake. Fixed.
567 :    
568 :     2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
569 :    
570 :     ----------------------------------------------------------------------
571 : leunga 1048 Name: Allen Leung
572 :     Date: 2002/01/30
573 :     Tag: leunga-20020130-nowhere-bug-fix
574 :     Description:
575 :    
576 :     MLRISC bug fixes:
577 :     1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
578 :     2. MachineInt.fromString was negating its value.
579 :    
580 :     ----------------------------------------------------------------------
581 : blume 1039 Name: Matthias Blume
582 : blume 1047 Date: 2002/01/29
583 :     Tag: blume-20020129-INSTALL
584 :     Description:
585 :    
586 :     - Added somewhat detailed installation instructions (file INSTALL).
587 :     - Fixed curl-detection bug in config/install.sh.
588 :     - It is now possible to select the URL getter using the URLGETTER
589 :     environment variable:
590 :    
591 :     not set / "unknown" --> automatic detection (script tries wget,
592 :     curl, and lynx)
593 :     "wget" / "curl" / "lynx" --> use the specified program (script "knows"
594 :     how to properly invoke them)
595 :     other --> use $URLGETTER directly, it must take
596 :     precisely two command-line arguments
597 :     (source URL and destination file name)
598 :    
599 :     ----------------------------------------------------------------------
600 :     Name: Matthias Blume
601 : blume 1044 Date: 2002/01/28
602 : blume 1045 Tag: blume-20020128-sparc-ccalls
603 :     Description:
604 :    
605 :     - Fixed problem with calculation of "used" registers in sparc-c-calls.
606 :     - Make use of the allocParam argument in sparc-c-calls.
607 :    
608 :     ----------------------------------------------------------------------
609 :     Name: Matthias Blume
610 :     Date: 2002/01/28
611 : blume 1044 Tag: blume-20020128-allocParam
612 :     Description:
613 :    
614 :     John Reppy: Changes c-calls API to accept client-callback for
615 :     allocating extra stack space.
616 :     me: Corresponding changes to mlriscGen (using a dummy argument that
617 :     does not change the current behavior).
618 :    
619 :     ----------------------------------------------------------------------
620 :     Name: Matthias Blume
621 : blume 1040 Date: 2002/01/28 12:00:00
622 :     Tag: Release_110_38
623 :     Description:
624 :    
625 :     This time for real!!!
626 :    
627 :     ----------------------------------------------------------------------
628 :     Name: Matthias Blume
629 : blume 1039 Date: 2002/01/28 10:56:00 EST
630 :     Tag: blume-20020128-retraction
631 :     Description:
632 :    
633 :     0. Retracted earlier 110.38. (The Release_110_38 tag has been replaced
634 :     with blume-Release_110_38-retracted.)
635 :    
636 :     1. Fixed a problem with incorrect rounding modes in real64.sml.
637 :     (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
638 :    
639 :     2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
640 :     and enums fixed. The naming of corresponding ML identifiers should
641 :     now be consistent again.
642 :    
643 :     ----------------------------------------------------------------------
644 : leunga 1038 Name: Allen Leung
645 :     Date: 2002/01/27
646 :     Tag: leunga-20020127-nowhere
647 :     Description:
648 :    
649 :     Added a target called nowhere in the configuration scripts.
650 :     Enabling this will build the MLRISC 'nowhere' tool (for translating
651 :     programs with where-clauses into legal SML code) during installation.
652 :    
653 :     ----------------------------------------------------------------------
654 : blume 1036 Name: Matthias Blume
655 : leunga 1038 Date: 2002/01/25 21:27:00 EST
656 : blume 1039 Tag: blume-Release_110_38-retracted
657 : blume 1037 Description:
658 :    
659 :     Call it a (working) release! Version is 110.38. Bootfiles are ready.
660 :    
661 :     README will be added later.
662 :    
663 : blume 1039 !!! NOTE: Re-tagged as blume-Release_110_38-retracted. Original tag
664 :     (Release_110_38) removed. Reason: Last-minute bug fixes.
665 :    
666 : blume 1037 ----------------------------------------------------------------------
667 :     Name: Matthias Blume
668 : blume 1036 Date: 2002/01/25
669 :     Tag: blume-20020125-ffi
670 :     Description:
671 :    
672 :     A large number of tweaks and improvements to ml-nlffi-lib and
673 :     ml-nlffigen:
674 :    
675 :     - ML represenation types have been streamlined
676 :     - getter and setter functions work with concrete values, not abstract
677 :     ones where possible
678 :     - ml-nlffigen command line more flexible (see README file there)
679 :     - some bugs have been fixed (hopefully)
680 :    
681 :     ----------------------------------------------------------------------
682 : george 1034 Name: Lal George
683 :     Date: 2002/01/24
684 :     Tag: george-20020124-risc-ra-interface
685 :     Description:
686 :    
687 :     There is a dramatic simplification in the interface to the
688 :     register allocator for RISC architectures as a result of making
689 :     parallel copy instructions explicit.
690 :    
691 :     ----------------------------------------------------------------------
692 : blume 1028 Name: Matthias Blume
693 : blume 1032 Date: 2002/01/22
694 :     Tag: blume-20020122-x86-ccalls
695 :     Description:
696 :    
697 :     Bug fix for c-calls on x86 (having to do with how char- and
698 :     short-arguments are being handled).
699 :    
700 :     ----------------------------------------------------------------------
701 :     Name: Matthias Blume
702 : blume 1031 Date: 2002/01/21
703 :     Tag: blume-20020121-ff
704 :     Description:
705 :    
706 :     Another day of fiddling with the FFI...
707 :    
708 :     1. Bug fix/workaround: CKIT does not complain about negative array
709 :     dimensions, so ml-nlffigen has to guard itself against this possibility.
710 :     (Otherwise a negative dimension would send it into an infinite loop.)
711 :    
712 :     2. Some of the abstract types (light objects, light pointers, most "base"
713 :     types) in structure C are now eqtypes.
714 :    
715 :     3. Added constructors and test functions for NULL function pointers.
716 :    
717 :     ----------------------------------------------------------------------
718 :     Name: Matthias Blume
719 : blume 1028 Date: 2002/01/18
720 : blume 1030 Tag: blume-20020118-ready-for-new-release
721 :     Description:
722 :    
723 :     Made config/srcarchiveurl point to a new place. (Will provide boot
724 :     files shortly.)
725 :    
726 :     Maybe we christen this to be 110.38?
727 :    
728 :     ----------------------------------------------------------------------
729 :     Name: Matthias Blume
730 :     Date: 2002/01/18
731 : blume 1029 Tag: blume-20020118-more-ffifiddle
732 :     Description:
733 :    
734 :     Today's FFI fiddling:
735 :    
736 :     - Provided a structure CGetSet with "convenient" versions of C.Get.* and
737 :     C.Set.* that use concrete (MLRep.*) arguments and results instead
738 :     of abstract ones.
739 :    
740 :     - Provided word-style bit operations etc. for "int" representation
741 :     types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
742 :     and Long.
743 :    
744 :     ----------------------------------------------------------------------
745 :     Name: Matthias Blume
746 :     Date: 2002/01/18
747 : blume 1028 Tag: blume-20020118-use-x86-fp
748 :     Description:
749 :    
750 :     Now that x86-fast-fp seems to be working, I turned it back on again
751 :     by default. (Seems to work fine now, even with the FFI.)
752 :    
753 :     Other than that, I added some documentation about the FFI to
754 :     src/ml-nlffigen/README and updated the FFI test examples in
755 :     src/ml-nlffi-lib/Tests/*.
756 :    
757 :     ----------------------------------------------------------------------
758 : leunga 1025 Name: Allen Leung
759 :     Date: 2002/01/17
760 :     Tag: leunga-20020117-x86-fast-fp-call
761 :     Description:
762 :    
763 :     1. Fixed a problem with handling return fp values when x86's fast fp
764 :     mode is turned on.
765 :    
766 :     2. Minor pretty printing fix for cellset. Print %st(0) as %st(0) instead
767 :     of %f32.
768 :    
769 :     3. Added a constructor INT32lit to the ast of MLRISC tools.
770 :    
771 :     ----------------------------------------------------------------------
772 : blume 1021 Name: Matthias Blume
773 :     Date: 2002/01/16
774 :     Tag: blume-20020116-ffifiddle
775 :     Description:
776 :    
777 :     More fiddling with the FFI interface:
778 :    
779 :     - Make constness 'c instead of rw wherever possible. This eliminates
780 :     the need for certain explicit coercions. (However, due to ML's
781 :     value polymorphism, there will still be many cases where explicit
782 :     coercions are necessary. Phantom types are not the whole answer
783 :     to modeling a subtyping relationship in ML.)
784 :    
785 :     - ro/rw coersions for pointers added. (Avoids the detour through */&.)
786 :    
787 :     - "printf" test example added to src/ml-nlffi-lib/Tests. (Demonstrates
788 :     clumsy workaround for varargs problem.)
789 :    
790 :     ----------------------------------------------------------------------
791 : george 1016 Name: Lal George
792 :     Date: 2002/01/15
793 :     Tag: <none>
794 :     Description:
795 :    
796 :     1. Since COPY instructions are no longer native to the architecture,
797 :     a generic functor can be used to implement the expandCopies function.
798 :    
799 :     2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
800 :     TEXT segment.
801 :    
802 :     ----------------------------------------------------------------------
803 : blume 1011 Name: Matthias Blume
804 : blume 1015 Date: 2002/01/15
805 :     Tag: blume-20020115-ffiupdates
806 :     Description:
807 :    
808 :     1. Fix for bug resulting in single-precision float values being returned
809 :     incorrectly from FFI calls.
810 :    
811 :     2. Small modifications to C FFI API:
812 :    
813 :     - memory-allocation routines return straight objects (no options)
814 :     and raise an exception in out-of-memory situations
815 :     - unsafe extensions to cast between function pointers and pointers
816 :     from/to ints
817 :     - added structure C_Debug as an alternative to structure C where
818 :     pointer-dereferencing (|*| and |*!) always check for null-pointers
819 :     - added open_lib' to DynLinkage; open_lib' works like open_lib
820 :     but also takes a (possibly empty) list of existing library handles
821 :     that the current library depends on
822 :    
823 :     ----------------------------------------------------------------------
824 :     Name: Matthias Blume
825 : blume 1011 Date: 2002/01/10
826 :     Tag: blume-20020110-newffigen
827 :     Description:
828 :    
829 :     1. Updates to portable graph code.
830 :    
831 :     2. Major update to ml-nlffigen and ml-nlffi-lib. Things are much
832 :     more scalable now so that even huge interfaces such as the one
833 :     for GTK compile in finite time and space. :-)
834 :     See src/ml-nlffigen/README for details on what's new.
835 :    
836 :     ----------------------------------------------------------------------
837 : george 1003 Name: Lal George
838 : george 1009 Date: 2001/01/09 14:31:35 EST 2002
839 :     Tag: george-20011206-rm-native-copy
840 :     Description:
841 :    
842 :     Removed the native COPY and FCOPY instructions
843 :     from all the architectures and replaced it with the
844 :     explicit COPY instruction from the previous commit.
845 :    
846 :     It is now possible to simplify many of the optimizations
847 :     modules that manipulate copies. This has not been
848 :     done in this change.
849 :    
850 :     ----------------------------------------------------------------------
851 :     Name: Lal George
852 : george 1003 Date: 2001/12/06 16:50:13 EST 2001
853 :     Tag: george-20011206-mlrisc-instruction
854 :     Description:
855 : george 824
856 : george 1003 Changed the representation of instructions from being fully abstract
857 :     to being partially concrete. That is to say:
858 :    
859 :     from
860 :     type instruction
861 :    
862 :     to
863 :     type instr (* machine instruction *)
864 :    
865 :     datatype instruction =
866 :     LIVE of {regs: C.cellset, spilled: C.cellset}
867 :     | KILL of {regs: C.cellset, spilled: C.cellset}
868 :     | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
869 :     | ANNOTATION of {i: instruction, a: Annotations.annotation}
870 :     | INSTR of instr
871 :    
872 :     This makes the handling of certain special instructions that appear on
873 :     all architectures easier and uniform.
874 :    
875 :     LIVE and KILL say that a list of registers are live or killed at the
876 :     program point where they appear. No spill code is generated when an
877 :     element of the 'regs' field is spilled, but the register is moved to
878 :     the 'spilled' (which is present, more for debugging than anything else).
879 :    
880 :     LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
881 :     We used to generate:
882 :    
883 :     DEFFREG f1
884 :     f1 := f2 + f3
885 :     trapb
886 :    
887 :     but now generate:
888 :    
889 :     f1 := f2 + f3
890 :     trapb
891 :     LIVE {regs=[f1,f2,f3], spilled=[]}
892 :    
893 :     Furthermore, the DEFFREG (hack) required that all floating point instruction
894 :     use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
895 :     defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
896 :     in a cleaner alpha implementation. (Hopefully, intel will not get rid of
897 :     this architecture).
898 :    
899 :     COPYXXX is intended to replace the parallel COPY and FCOPY available on
900 :     all the architectures. This will result in further simplification of the
901 :     register allocator that must be aware of them for coalescing purposes, and
902 :     will also simplify certain aspects of the machine description that provides
903 :     callbacks related to parallel copies.
904 :    
905 :     ANNOTATION should be obvious, and now INSTR represents the honest to God
906 :     machine instruction set!
907 :    
908 :     The <arch>/instructions/<arch>Instr.sml files define certain utility
909 :     functions for making porting easier -- essentially converting upper case
910 :     to lower case. All machine instructions (of type instr) are in upper case,
911 :     and the lower case form generates an MLRISC instruction. For example on
912 :     the alpha we have:
913 :    
914 :     datatype instr =
915 :     LDA of {r:cell, b:cell, d:operand}
916 :     | ...
917 :    
918 :     val lda : {r:cell, b:cell, d:operand} -> instruction
919 :     ...
920 :    
921 :     where lda is just (INSTR o LDA), etc.
922 :    
923 : leunga 815 ----------------------------------------------------------------------
924 : blume 986 Name: Matthias Blume
925 : blume 993 Date: 2001/11/22 21:40:00 EST
926 :     Tag: Release_110_37
927 :     Description:
928 :    
929 :     Release 110.37. This time for real.
930 :    
931 :     ----------------------------------------------------------------------
932 :     Name: Matthias Blume
933 : blume 988 Date: 2001/11/21 16:35:00 EST
934 :     Tag: blume-20011121-foot-in-mouth
935 :     Description:
936 :    
937 :     Removed the "Release_110_37" tag because of a serious bug.
938 :     This will be re-tagged once the bug is fixed.
939 :    
940 :     ----------------------------------------------------------------------
941 :     Name: Matthias Blume
942 : blume 987 Date: 2001/11/21 16:14:00 EST
943 :     Tag: blume-20011121-forgottenfile
944 :     Description:
945 :    
946 :     Forgot to add a file. (Just a .tex-file -- part of
947 :     the CM manual source.)
948 :    
949 :     ----------------------------------------------------------------------
950 :     Name: Matthias Blume
951 : blume 986 Date: 2001/11/21 16:10:00 EST
952 : blume 988 Tag: blume-20011121-invalid_110_37
953 : blume 986 Description:
954 :    
955 : blume 988 Note: I removed the original tag "Release_110_37" from this commit
956 :     because we found a serious bug in all non-x86 backends.
957 :     - Matthias
958 :    
959 : blume 986 1. Modifications to the SML/NJ code generator and to the runtime system
960 :     so that code object name strings are directly inserted into code
961 :     objects at code generation time. The only business the runtime system
962 :     has with this is now to read the name strings on occasions.
963 :     (The encoding of the name string has also changed somewhat.)
964 :    
965 :     2. CM now implements a simple "set calculus" for specifying export lists.
966 :     In particular, it is now possible to refer to the export lists of
967 :     other libraries/groups/sources and form unions as well as differences.
968 :     See the latest CM manual for details.
969 :    
970 :     3. An separate notion of "proxy" libraries has again be eliminated from
971 :     CM's model. (Proxy libraries are now simply a special case of using
972 :     the export list calculus.)
973 :    
974 :     4. Some of the existing libraries now take advantage of the new set
975 :     calculus.
976 :     (Notice that not all libraries have been converted because some
977 :     of the existing .cm-files are supposed to be backward compatible
978 :     with 110.0.x.)
979 :    
980 :     5. Some cleanup in stand-alone programs. (Don't use "exnMessage" -- use
981 :     "General.exnMessage"! The former relies on a certain hook to be
982 :     initialized, and that often does not happen in the stand-alone case.)
983 :    
984 :     ----------------------------------------------------------------------
985 : george 984 Name: Lal George
986 :     Date: 2001/11/21 13:56:18 EST
987 :     Tag: george-2001121-pseudo-ops
988 :     Description:
989 :    
990 :     Implemented a complete redesign of MLRISC pseudo-ops. Now there
991 :     ought to never be any question of incompatabilities with
992 :     pseudo-op syntax expected by host assemblers.
993 :    
994 :     For now, only modules supporting GAS syntax are implemented
995 :     but more should follow, such as MASM, and vendor assembler
996 :     syntax, e.g. IBM as, Sun as, etc.
997 :    
998 :     ----------------------------------------------------------------------
999 : blume 977 Name: Matthias Blume
1000 :     Date: 2001/11/14 11:52:00 EST
1001 :     Tag: blume-20011114-srcname
1002 :     Description:
1003 :    
1004 :     1. Routed the name of the current source file to mlriscgen where it
1005 :     should be directly emitted into the code object. (This last part
1006 :     is yet to be done.)
1007 :    
1008 :     2. Some cleanup of the pgraph code to make it match the proposal that
1009 :     I put out the other day. (The proposal notwithstanding, things are
1010 :     still in flux here.)
1011 :    
1012 :     ----------------------------------------------------------------------
1013 : george 976 Name: Lal George
1014 :     Date: 2001/11/14 09:44:04 EST
1015 :     Tag:
1016 :     Description:
1017 :    
1018 :     Fix for a backpatching bug reported by Allen.
1019 :    
1020 :     Because the boundary between short and long span-dependent
1021 :     instructions is +/- 128, there are an astounding number of
1022 :     span-dependent instructions whose size is over estimated.
1023 :    
1024 :     Allen came up with the idea of letting the size of span
1025 :     dependent instructions be non-monotonic, for a maxIter
1026 :     number of times, after which the size must be monotonically
1027 :     increasing.
1028 :    
1029 :     This table shows the number of span-dependent instructions
1030 :     whose size was over-estimated as a function of maxIter, for the
1031 :     file Parse/parse/ml.grm.sml:
1032 :    
1033 :     maxIter # of instructions:
1034 :     10 687
1035 :     20 438
1036 :     30 198
1037 :     40 0
1038 :    
1039 :     In compiling the compiler, there is no significant difference in
1040 :     compilation speed between maxIter=10 and maxIter=40. Actually,
1041 :     my measurements showed that maxIter=40 was a tad faster than
1042 :     maxIter=10! Also 96% of the files in the compiler reach a fix
1043 :     point within 13 iterations, so fixing maxIter at 40, while high,
1044 :     is okay.
1045 :    
1046 :     ----------------------------------------------------------------------
1047 : blume 975 Name: Matthias Blume
1048 :     Date: 2001/10/31 15:25:00 EST
1049 :     Tag: blume-20011031-pgraph
1050 :     Description:
1051 :    
1052 :     CKIT:
1053 :     * Changed the "Function" constructor of type Ast.ctype to carry optional
1054 :     argument identifiers.
1055 :     * Changed the return type of TypeUtil.getFunction accordingly.
1056 :     * Type equality ignores the argument names.
1057 :     * TypeUtil.composite tries to preserve argument names but gives up quickly
1058 :     if there is a mismatch.
1059 :    
1060 :     installation script:
1061 :     * attempts to use "curl" if available (unless "wget" is available as well)
1062 :    
1063 :     CM:
1064 :     * has an experimental implementation of "portable graphs" which I will
1065 :     soon propose as an implementation-independent library format
1066 :     * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
1067 :    
1068 :     NLFFI-LIB:
1069 :     * some cleanup (all cosmetic)
1070 :    
1071 :     NLFFIGEN:
1072 :     * temporarily disabled the mechanism that suppresses ML output for
1073 :     C definitions whose identifiers start with an underscore character
1074 :     * generate val bindings for enum constants
1075 :     * user can request that only one style (light or heavy) is being used;
1076 :     default is to use both (command-line arguments: -heavy and -light)
1077 :     * fixed bug in handling of function types involving incomplete pointers
1078 :     * generate ML entry points that take record arguments (i.e., using
1079 :     named arguments) for C functions that have a prototype with named
1080 :     arguments
1081 :     (see changes to CKIT)
1082 :    
1083 :     ----------------------------------------------------------------------
1084 : leunga 972 Name: Allen Leung
1085 :     Date: 2001/10/27 20:34:00 EDT
1086 :     Tag: leunga-20011027-x86-fast-fp-call
1087 :     Description:
1088 :    
1089 :     Fixed the bug described in blume-20010920-slowfp.
1090 :    
1091 :     The fix involves
1092 :     1. generating FCOPYs in FSTP in ia32-svid
1093 :     2. marking a CALL with the appropriate annotation
1094 :    
1095 :     ----------------------------------------------------------------------
1096 : blume 961 Name: Matthias Blume
1097 :     Date: 2001/10/16 11:32:00 EDT
1098 :     Tag: blume-20011016-netbsd
1099 :     Description:
1100 :    
1101 :     Underscore patch from Chris Richards (fixing problem with compiling
1102 :     runtime system under recent NetBSD).
1103 :    
1104 :     ----------------------------------------------------------------------
1105 : george 959 Name: Allen Leung
1106 :     Date: 2001/10/12 17:18:32 EDT 2001
1107 :     Tag: leung-20011012-x86-printflowgraph
1108 :     Description:
1109 :    
1110 :     X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
1111 :    
1112 :     ----------------------------------------------------------------------
1113 : george 952 Name: Lal George
1114 : george 958 Date: 2001/10/11 23:51:34 EDT
1115 :     Tag: george-20011011-too-many-instrs
1116 :     Description:
1117 :    
1118 :     The representation of a program point never expected to see more
1119 :     than 65536 instructions in a basic block!
1120 :    
1121 :     ----------------------------------------------------------------------
1122 :     Name: Lal George
1123 : george 952 Date: 2001/10/09 09:41:37 EDT
1124 :     Tag: george-20011008-mlrisc-labels
1125 :     Description:
1126 :    
1127 :     Changed the machine description files to support printing of
1128 :     local and global labels in assembly code, based on host assembler
1129 :     conventions.
1130 :    
1131 :     ----------------------------------------------------------------------
1132 : blume 916 Name: Matthias Blume
1133 : blume 939 Date: 2001/09/25 15:25:00 EDT
1134 :     Tag: blume-20010925-exninfo
1135 :     Description:
1136 :    
1137 :     I provided a non-hook implementation of exnName (at the toplevel) and
1138 :     made the "dummy" implementation of exnMessage (at the toplevel) more
1139 :     useful: if nothing gets "hooked in", then at least you are going to
1140 :     see the exception name and a message indicating why you don't see more.
1141 :    
1142 :     [For the time being, programs that need exnMessage and want to use
1143 :     ml-build should either use General.exnMessage (strongly recommended) or
1144 :     refer to structure General at some other point so that CM sees a
1145 :     static dependency.]
1146 :    
1147 :     [Similar remarks go for "print" and "use": If you want to use their
1148 :     functionality in stand-alone programs generated by ml-build, then use
1149 :     TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
1150 :    
1151 :     ----------------------------------------------------------------------
1152 :     Name: Matthias Blume
1153 : blume 938 Date: 2001/09/20 17:28:00 EDT
1154 :     Tag: blume-20010920-slowfp
1155 :     Description:
1156 :    
1157 :     Allen says that x86-fast-fp is not safe yet, so I turned it off again...
1158 :    
1159 :     ----------------------------------------------------------------------
1160 :     Name: Matthias Blume
1161 : blume 937 Date: 2001/09/20 17:20:00 EDT
1162 :     Tag: blume-20010920-canonicalpaths
1163 :     Description:
1164 :    
1165 :     0. Updated the BOOT file (something that I forgot to do earlier).
1166 :    
1167 :     1. Small internal change to CM so that it avoids "/../" in filenames
1168 :     as much as possible (but only where it is safe).
1169 :    
1170 :     2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
1171 :     that arguments that contain delimiters are passed through correctly.
1172 :     This change also means that all "special" arguments of the form
1173 :     @SMLxxx... must come first.
1174 :    
1175 :     3. Changed install script to put relative anchor names for tool commands
1176 :     into pathconfig.
1177 :    
1178 :     ----------------------------------------------------------------------
1179 :     Name: Matthias Blume
1180 : blume 932 Date: 2001/09/18 15:35:00 EDT
1181 :     Tag: blume-20010918-readme11036
1182 :     Description:
1183 :    
1184 :     Added README files.
1185 :    
1186 :     ----------------------------------------------------------------------
1187 :     Name: Matthias Blume
1188 : blume 929 Date: 2001/09/18 11:45:00 EDT
1189 :     Tag: Release_110_36 (retag)
1190 :     Description:
1191 :    
1192 :     Fixed mistake in config/preloads. Retagged as 110.36.
1193 :    
1194 :     ----------------------------------------------------------------------
1195 :     Name: Matthias Blume
1196 : blume 928 Date: 2001/09/18 09:40:00 EDT
1197 : blume 929 Tag: Release_110_36_orig (tag changed)
1198 : blume 928 Description:
1199 :    
1200 :     New version (110.36). New bootfiles.
1201 :    
1202 :     ----------------------------------------------------------------------
1203 :     Name: Matthias Blume
1204 : blume 927 Date: 2001/09/14 16:15:00 EDT
1205 :     Tag: blume-20010914-x86fastfp
1206 :     Description:
1207 :    
1208 :     John committed some changes that Allen made, in particular a (hopefully)
1209 :     correctly working version of the x86-fp module.
1210 :    
1211 :     I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
1212 :     flag to "true". Everything seems to compile to a fixpoint ok, and
1213 :     "mandelbrot" speeds up by about 15%.
1214 :    
1215 :     ----------------------------------------------------------------------
1216 :     Name: Matthias Blume
1217 : blume 923 Date: 2001/09/13 11:20:00 EDT
1218 :     Tag: blume-20010913-minimal
1219 :     Description:
1220 :    
1221 :     1. Stefan Monnier's patch to fix a miscompilation problem that
1222 :     was brought to light by John Reppy's work on Moby.
1223 :    
1224 :     2. Implemented a minimal "structure Compiler" that contains just
1225 :     "version" and "architecture". The minimal version will be
1226 :     available when the full version is not. This is for backward-
1227 :     compatibility with code that wants to test Compiler.version.
1228 :    
1229 :     ----------------------------------------------------------------------
1230 :     Name: Matthias Blume
1231 : blume 920 Date: 2001/08/28 14:03:00 EDT
1232 :     Tag: blume-20010828-ml-lex
1233 :     Description:
1234 :    
1235 :     Fix for bug 1581, received from Neophytos Michael.
1236 :    
1237 :     ----------------------------------------------------------------------
1238 :     Name: Matthias Blume
1239 : blume 918 Date: 2001/08/27 11:20:00 EDT
1240 :     Tag: blume-20010827-readme11035
1241 :     Description:
1242 :    
1243 :     Fleshed out the README file for 110.35.
1244 :    
1245 :     ----------------------------------------------------------------------
1246 :     Name: Matthias Blume
1247 : blume 916 Date: 2001/08/24 17:10:00 EDT
1248 :     Tag: Release_110_35
1249 :     Description:
1250 :    
1251 :     New version number (110.35). New bootfiles.
1252 :    
1253 :     ----------------------------------------------------------------------
1254 : george 915 Name: Lal George
1255 : george 909 Date: 2001/08/24 13:47:18 EDT 2001
1256 :     Tag: george-20010824-MLRISC-graphs
1257 :     Description:
1258 :    
1259 :     removed clusters from MLRISC completely and replaced with graphs.
1260 :    
1261 :     ----------------------------------------------------------------------
1262 :     Name: Matthias Blume
1263 : blume 905 Date: 2001/08/23 17:50:00 EDT
1264 :     Tag: blume-20010823-toplevel
1265 :     Description:
1266 :    
1267 :     - some reorganization of the code that implements various kinds of
1268 :     environments in the compiler (static, dynamic, symbolic, combined)
1269 :     - re-implemented the EnvRef module so that evalStream works properly
1270 :     (if the stream contains references to "use", "CM.make", etc.)
1271 :     - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
1272 :    
1273 :     ----------------------------------------------------------------------
1274 :     Name: Matthias Blume
1275 : blume 904 Date: 2001/08/20 15:50 EDT
1276 :     Tag: blume20010820-slipup
1277 :     Description:
1278 :    
1279 :     I forgot to commit a few files. Here they are...
1280 :    
1281 :     ----------------------------------------------------------------------
1282 :     Name: Matthias Blume
1283 : blume 903 Date: 2001/08/20 15:35:00 EDT
1284 :     Tag: blume-20010820-debugprof
1285 :     Description:
1286 :    
1287 :     !!!! NEW BOOTFILES !!!!
1288 :    
1289 :     This is another round of reorganizing the compiler sources. This
1290 :     time the main goal was to factor out all the "instrumentation"
1291 :     passes (for profiling and backtracing) into their own library.
1292 :     The difficulty was to do it in such a way that it does not depend
1293 :     on elaborate.cm but only on elabdata.cm.
1294 :    
1295 :     Therefore there have been further changes to both elaborate.cm and
1296 :     elabdata.cm -- more "generic" things have been moved from the former
1297 :     to the latter. As a result, I was forced to split the assignment
1298 :     of numbers indicating "primtyc"s into two portions: SML-generic and
1299 :     SML/NJ-specific. Since it would have been awkward to maintain,
1300 :     I bit the bullet and actually _changed_ the mapping between these
1301 :     numbers and primtycs. The bottom line of this is that you need
1302 :     a new set of bin- and bootfiles.
1303 :    
1304 :     I have built new bootfiles for all architectures, so doing a fresh
1305 :     checkout and config/install.sh should be all you need.
1306 :    
1307 :     The newly created library's name is
1308 :    
1309 :     $smlnj/viscomp/debugprof.cm
1310 :    
1311 :     and its sources live under
1312 :    
1313 :     src/compiler/DebugProf
1314 :    
1315 :     ----------------------------------------------------------------------
1316 :     Name: Matthias Blume
1317 : blume 902 Date: 2001/08/15 17:15:00 EDT
1318 :     Tag: blume-20010815-compreorg
1319 :     Description:
1320 :    
1321 :     This is a first cut at reorganizing the CM libraries that make up the
1322 :     core of the compiler. The idea is to separate out pieces that could
1323 :     be used independently by tools, e.g., the parser, the typechecker, etc.
1324 :    
1325 :     The current status is a step in this direction, but it is not quite
1326 :     satisfactory yet. Expect more changes in the future.
1327 :    
1328 :     Here is the current (new) organization...
1329 :    
1330 :     What used to be $smlnj/viscomp/core.cm is now divided into
1331 :     six CM libraries:
1332 :    
1333 :     $smlnj/viscomp/basics.cm
1334 :     /parser.cm
1335 :     /elabdata.cm
1336 :     /elaborate.cm
1337 :     /execute.cm
1338 :     /core.cm
1339 :    
1340 :     The CM files for these libraries live under src/system/smlnj/viscomp.
1341 :     All these libraries are proxy libraries that contain precisely
1342 :     one CM library component. Here are the locations of the components
1343 :     (all within the src/compiler tree):
1344 :    
1345 :     Basics/basics.cm
1346 :     Parse/parser.cm
1347 :     ElabData/elabdata.cm
1348 :     Elaborator/elaborate.cm
1349 :     Execution/execute.cm
1350 :     core.cm
1351 :    
1352 :     [This organization is the same that has been used already
1353 :     for a while for the architecture-specific parts of the visible
1354 :     compiler and for the old version of core.cm.]
1355 :    
1356 :     As you will notice, many source files have been moved from their
1357 :     respective original locations to a new home in one of the above
1358 :     subtrees.
1359 :    
1360 :     The division of labor between the new libraries is the following:
1361 :    
1362 :     basics.cm:
1363 :     - Simple, basic definitions that pertain to many (or all) of
1364 :     the other libraries.
1365 :     parser.cm:
1366 :     - The SML parser, producing output of type Ast.dec.
1367 :     - The type family for Ast is also defined and exported here.
1368 :     elabdata.cm:
1369 :     - The datatypes that describe input and output of the elaborator.
1370 :     This includes types, absyn, and static environments.
1371 :     elaborator.cm:
1372 :     - The SML/NJ type checker and elaborator.
1373 :     This maps an Ast.dec (with a given static environment) to
1374 :     an Absyn.dec (with a new static environment).
1375 :     - This libraries implements certain modules that used to be
1376 :     structures as functors (to remove dependencies on FLINT).
1377 :     execute.cm:
1378 :     - Everything having to do with executing binary code objects.
1379 :     - Dynamic environments.
1380 :     core.cm:
1381 :     - SML/NJ-specific instantiations of the elaborator and MLRISC.
1382 :     - Top-level modules.
1383 :     - FLINT (this should eventually become its own library)
1384 :    
1385 :     Notes:
1386 :    
1387 :     I am not 100% happy with the way I separated the elaborator (and its
1388 :     data structures) from FLINT. Two instances of the same problem:
1389 :    
1390 :     1. Data structures contain certain fields that carry FLINT-specific
1391 :     information. I hacked around this using exn and the property list
1392 :     module from smlnj-lib. But the fact that there are middle-end
1393 :     specific fields around at all is a bit annoying.
1394 :    
1395 :     2. The elaborator calculates certain FLINT-related information. I tried
1396 :     to make this as abstract as I could using functorization, but, again,
1397 :     the fact that the elaborator has to perform calculations on behalf
1398 :     of the middle-end at all is not nice.
1399 :    
1400 :     3. Having to used exn and property lists is unfortunate because it
1401 :     weakens type checking. The other alternative (parameterizing
1402 :     nearly *everything*) is not appealing, though.
1403 :    
1404 :     I removed the "rebinding =" warning hack because due to the new organization
1405 :     it was awkward to maintain it. As a result, the compiler now issues some of
1406 :     these warnings when compiling init.cmi during bootstrap compilation. On
1407 :     the plus side, you also get a warning when you do, for example:
1408 :     val op = = Int32.+
1409 :     which was not the case up to now.
1410 :    
1411 :     I placed "assign" and "deref" into the _Core structure so that the
1412 :     code that deals with the "lazy" keyword can find them there. This
1413 :     removes the need for having access to the primitive environment
1414 :     during elaboration.
1415 :    
1416 :     ----------------------------------------------------------------------
1417 :     Name: Matthias Blume
1418 : blume 898 Date: 2001/08/13
1419 :     Tag: blume-20010813-closures
1420 :     Description:
1421 :    
1422 :     This fix was sent to us by Zhong Shao. It is supposed to improve the
1423 :     performance of certain loops by avoiding needless closure allocation.
1424 :    
1425 :     ----------------------------------------------------------------------
1426 : george 890 Name: Lal George
1427 : george 896 Date: 2001/07/31 10:03:23 EDT 2001
1428 :     Tag: george-20010731-x86-fmalloc
1429 :     Description: Fixed bug in x86 calls
1430 :    
1431 :     There was a bug where call instructions would mysteriously
1432 :     vanish. The call instruction had to be one that returned
1433 :     a floating point value.
1434 :    
1435 :     ----------------------------------------------------------------------
1436 :     Name: Lal George
1437 : george 890 Date: 2001/07/19 16:36:29 EDT 2001
1438 :     Tag: george-20010719-simple-cells
1439 :     Description:
1440 :    
1441 :     I have dramatically simplified the interface for CELLS in MLRISC.
1442 :    
1443 :     In summary, the cells interface is broken up into three parts:
1444 :    
1445 :     1. CellsBasis : CELLS_BASIS
1446 :    
1447 :     CellsBasis is a top level structure and common for all
1448 :     architectures. it contains the definitions of basic datatypes
1449 :     and utility functions over these types.
1450 :    
1451 :     2. functor Cells() : CELLS
1452 :    
1453 :     Cells generates an interface for CELLS that incorporates the
1454 :     specific resources on the target architecture, such as the
1455 :     presence of special register classes, their number and size,
1456 :     and various useful substructures.
1457 :    
1458 :     3. <ARCH>CELLS
1459 :    
1460 :     e.g. SparcCells: SPARCCELLS
1461 :    
1462 :     <ARCH>CELLS usually contains additional bindings for special
1463 :     registers on the architecture, such as:
1464 :    
1465 :     val r0 : cell (* register zero *)
1466 :     val y : cell (* Y register *)
1467 :     val psr : cell (* processor status register *)
1468 :     ...
1469 :    
1470 :     The structure returned by applying the Cells functor is opened
1471 :     in this interface.
1472 :    
1473 :     The main implication of all this is that the datatypes for cells is
1474 :     split between CellsBasis and CELLS -- a fairly simple change for user
1475 :     code.
1476 :    
1477 :     In the old scheme the CELLS interface had a definitional binding of
1478 :     the form:
1479 :    
1480 :     signature CELLS = sig
1481 :    
1482 :     structure CellsBasis = CellsBasis
1483 :    
1484 :     ...
1485 :    
1486 :     end
1487 :    
1488 :     With all the sharing constraints that goes on in MLRISC, this old
1489 :     design quickly leads to errors such as:
1490 :    
1491 :     "structure definition spec inside of sharing ... "
1492 :    
1493 :    
1494 :     and appears to require an unacceptable amount of sharing and where
1495 :     constraint hackery.
1496 :    
1497 :     I think this error message (the interaction of definitional specs and
1498 :     sharing) requires more explanation on our web page.
1499 :    
1500 :     ----------------------------------------------------------------------
1501 : blume 826 Name: Matthias Blume
1502 : blume 879 Date: 2001/07/19 15:00:00 EDT
1503 :     Tag: blume-20010719-libreorg
1504 :     Description:
1505 :    
1506 :     This update puts together a fairly extensive but straightforward change
1507 :     to the way the libraries that implement the interactive system are
1508 :     organized:
1509 :    
1510 :     The biggest change is the elimination of structure Compiler. As a
1511 :     replacement for this structure, there is now a CM library
1512 :     (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1513 :     that exports all the substructures of the original structure Compiler
1514 :     directly. So instead of saying Compiler.Foo.bar one now simply
1515 :     says Foo.bar. (The CM libraries actually export a collection of
1516 :     structures that is richer than the collection of substructures of
1517 :     structure Compiler.)
1518 :    
1519 :     To make the transition smooth, there is a separate library called
1520 :     $smlnj/compiler/compiler.cm which puts together and exports the
1521 :     original structure Compiler (or at least something very close to it).
1522 :    
1523 :     There are five members of the original structure Compiler
1524 :     that are not exported directly but which instead became members
1525 :     of a new structure Backend (described by signature BACKEND). These are:
1526 :     structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1527 :     Interact (: INTERACT), structure Machine (: MACHINE), and val
1528 :     architecture (: string).
1529 :    
1530 :     Structure Compiler.Version has become structure CompilerVersion.
1531 :    
1532 :     Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1533 :     by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1534 :     or x86, respectively.
1535 :     Each of these exports the same frontend structures that
1536 :     $smlnj/compiler.cm exports. But they do not have a structure Backend
1537 :     and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1538 :     Hppa, PPC, Sparc, or X86, respectively.
1539 :    
1540 :     Library $smlnj/compiler/all.cm exports the union of the exports of
1541 :     $smlnj/compiler/<arch>.cm
1542 :    
1543 :     There are no structures <Arch>Compiler anymore, use
1544 :     $smlnj/compiler/<arch>.cm instead.
1545 :    
1546 :     Library host-compiler-0.cm is gone. Instead, the internal library
1547 :     that instantiates CM is now called cm0.cm. Selection of the host
1548 :     compiler (backend) is no longer done here but. (Responsibility for it
1549 :     now lies with $smlnj/compiler/current.cm. This seems to be more
1550 :     logical.)
1551 :    
1552 :     Many individual files have been moved or renamed. Some files have
1553 :     been split into multiple files, and some "dead" files have been deleted.
1554 :    
1555 :     Aside from these changes to library organization, there are also changes
1556 :     to the way the code itself is organized:
1557 :    
1558 :     Structure Binfile has been re-implemented in such a way that it no
1559 :     longer needs any knowledge of the compiler. It exclusively deals
1560 :     with the details of binfile layout. It no longer invokes the
1561 :     compiler (for the purpose of creating new prospective binfile
1562 :     content), and it no longer has any knowledge of how to interpret
1563 :     pickles.
1564 :    
1565 :     Structure Compile (: COMPILE) has been stripped down to the bare
1566 :     essentials of compilation. It no longer deals with linking/execution.
1567 :     The interface has been cleaned up considerably.
1568 :    
1569 :     Utility routines for dealing with linking and execution have been
1570 :     moved into their own substructures.
1571 :    
1572 :     (The ultimate goal of these changes is to provide a light-weight
1573 :     binfile loader/linker (at least for, e.g., stable libraries) that
1574 :     does not require CM or the compiler to be present.)
1575 :    
1576 :     CM documentation has been updated to reflect the changes to library
1577 :     organization.
1578 :    
1579 :     ----------------------------------------------------------------------
1580 :     Name: Matthias Blume
1581 : blume 875 Date: 2001/07/10 17:30:00 EDT
1582 :     Tag: Release_110_34
1583 :     Description:
1584 :    
1585 : blume 876 Minor tweak to 110.34 (re-tagged):
1586 :    
1587 :     - README.html file added to CVS repository
1588 :     - runtime compiles properly under FreeBSD 3.X and 4.X
1589 :    
1590 :     ----------------------------------------------------------------------
1591 :     Name: Matthias Blume
1592 :     Date: 2001/07/10 17:30:00 EDT
1593 :     Tag: Release_110_34
1594 :     Description:
1595 :    
1596 : blume 875 New version number (110.34). New bootfiles.
1597 :    
1598 :     ----------------------------------------------------------------------
1599 :     Name: Matthias Blume
1600 : blume 874 Date: 2001/07/09 16:00:00 EDT
1601 :     Tag: blume-20010709-more-varargs
1602 :     Description:
1603 :    
1604 :     I changed the handling of varargs in ml-nlffigen again:
1605 :     The ellipsis ... will now simply be ignored (with an accompanying warning).
1606 :    
1607 :     The immediate effect is that you can actually call a varargs function
1608 :     from ML -- but you can't actually supply any arguments beyond the ones
1609 :     specified explicitly. (For example, you can call printf with its format
1610 :     string, but you cannot pass additional arguments.)
1611 :    
1612 :     This behavior is only marginally more useful than the one before, but
1613 :     it has the advantage that a function or, more importantly, a function
1614 :     type never gets dropped on the floor, thus avoiding follow-up problems with
1615 :     other types that refer to the offending one.
1616 :    
1617 :     ----------------------------------------------------------------------
1618 :     Name: Matthias Blume
1619 : blume 873 Date: 2001/07/09 11:25:00 EDT
1620 :     Tag: blume-20010709-varargs
1621 :     Description:
1622 :    
1623 :     1. ckit-lib.cm now exports structure Error
1624 :     2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1625 :     with a warning accompanied by a source location. Moreover, it
1626 :     merely skips the offending function or type and proceeds with the
1627 :     rest of its work.u As a result, one can safely feed C code containing
1628 :     "..." to ml-nlffigen.
1629 :     3. There are some internal improvements to CM, providing slightly
1630 :     more general string substitutions in the tools subsystem.
1631 :    
1632 :     ----------------------------------------------------------------------
1633 :     Name: Matthias Blume
1634 : blume 854 Date: 2001/06/27 15:10:00 EDT
1635 :     Tag: blume-20010627-concur
1636 :     Description:
1637 :    
1638 :     Fixed a small bug in CM's handling of parallel compilation.
1639 :     (You could observe the bug by Control-C-interrupting an ordinary
1640 :     CMB.make or CM.stabilize and then attaching some compile servers.
1641 :     The result was that all of a sudden the previously interrupted
1642 :     compilation would continue on its own. This was because of
1643 :     an over-optimization: CM did not bother to clean out certain queues
1644 :     when no servers were attached "anyway", resulting in the contents
1645 :     of these queues to grab control when new servers did get attached.)
1646 :    
1647 :     There is also another minor update to the CM manual.
1648 :    
1649 :     ----------------------------------------------------------------------
1650 :     Name: Matthias Blume
1651 : blume 853 Date: 2001/06/26 16:15:00 EDT
1652 :     Tag: blume-20010626-cmdoc
1653 :     Description:
1654 :    
1655 :     Minor typo fixed in CM manual (syntax diagram for libraries).
1656 :    
1657 :     ----------------------------------------------------------------------
1658 :     Name: Matthias Blume
1659 : blume 852 Date: 2001/06/25 22:55:00 EDT
1660 :     Tag: blume-20010625-x86pc
1661 :     Description:
1662 :    
1663 :     Fixed a nasty bug in the X86 assembly code that caused signal
1664 :     handlers to fail (crash) randomly.
1665 :    
1666 :     ----------------------------------------------------------------------
1667 :     Name: Matthias Blume
1668 : blume 846 Date: 2001/06/25 12:05:00 EDT
1669 :     Tag: blume-20010625-nlffigen
1670 :     Description:
1671 :    
1672 :     This update fixes a number of minor bugs in ml-nlffigen as reported by
1673 :     Nick Carter <nbc@andrew.cmu.edu>.
1674 :    
1675 :     1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1676 :     2. Default names for generated files are now derived from the name of
1677 :     the C file *without its directory*. In particular, this causes generated
1678 :     files to be placed locally even if the C file is in some system directory.
1679 :     3. Default names for generated signatures and structures are also derived
1680 :     from the C file name without its directory. This avoids silly things
1681 :     like "structure GL/GL".
1682 :     (Other silly names are still possible because ml-nlffigen does not do
1683 : blume 847 a thorough check of whether generated names are legal ML identifiers.
1684 :     When in doubt, use command line arguments to force particular names.)
1685 : blume 846
1686 :     ----------------------------------------------------------------------
1687 :     Name: Matthias Blume
1688 : blume 845 Date: 2001/06/21 12:25:00 EDT
1689 :     Tag: blume-20010621-eXene
1690 :     Description:
1691 :    
1692 :     eXene now compiles and (sort of) works again.
1693 :    
1694 :     The library name (for version > 110.33) is $/eXene.cm.
1695 :    
1696 :     I also added an new example in src/eXene/examples/nbody. See the
1697 :     README file there for details.
1698 :    
1699 :     ----------------------------------------------------------------------
1700 :     Name: Matthias Blume
1701 : blume 844 Date: 2001/06/20 16:40:00 EDT
1702 :     Tag: blume-20010620-cml
1703 :     Description:
1704 :    
1705 :     CML now compiles and works again.
1706 :    
1707 :     Libraries (for version > 110.33):
1708 :    
1709 :     $cml/cml.cm Main CML library.
1710 :     $cml/basis.cm CML's version of $/basis.cm.
1711 :     $cml/cml-internal.cm Internal helper library.
1712 :     $cml/core-cml.cm Internal helper library.
1713 :     $cml-lib/trace-cml.cm Tracing facility.
1714 :     $cml-lib/smlnj-lib.cm CML's version of $/smlnj-lib.cm
1715 :    
1716 :     The installer (config/install.sh) has been taught how to properly
1717 :     install this stuff.
1718 :    
1719 :     ----------------------------------------------------------------------
1720 :     Name: Matthias Blume
1721 : blume 843 Date: 2001/06/19 17:55:00 EDT
1722 :     Tag: blume-20010619-instantiate
1723 :     Description:
1724 :    
1725 :     This un-breaks the fix for bug 1432.
1726 :     (The bug was originally fixed in 110.9 but I broke it again some
1727 :     time after that.)
1728 :    
1729 :     ----------------------------------------------------------------------
1730 :     Name: Matthias Blume
1731 : blume 842 Date: 2001/06/19 17:25:00 EDT
1732 :     Tag: blume-20010619-signals
1733 :     Description:
1734 :    
1735 :     This should (hopefully) fix the long-standing signal handling bug.
1736 :     (The runtime system was constructing a continuation record with an
1737 :     incorrect descriptor which would cause the GC to drop data on the floor...)
1738 :    
1739 :     ----------------------------------------------------------------------
1740 :     Name: Matthias Blume
1741 : blume 840 Date: 2001/06/15 15:05:00 EDT
1742 : blume 841 Tag: blume-20010615-moresparc
1743 :     Description:
1744 :    
1745 :     Here is a short late-hour update related to Sparc c-calls:
1746 :    
1747 :     -- made handling of double-word arguments a bit smarter
1748 :    
1749 :     -- instruction selection phase tries to collapse certain clumsily
1750 :     constructed ML-Trees; typical example:
1751 :    
1752 :     ADD(ty,ADD(_,e,LI d1),LI d2) -> ADD(ty,e,LI(d1+d2))
1753 :    
1754 :     This currently has no further impact on SML/NJ since mlriscGen does
1755 :     not seem to generate such patterns in the first place, and c-calls
1756 :     (which did generate them in the beginning) has meanwhile been fixed
1757 :     so as to avoid them as well.
1758 :    
1759 :     ----------------------------------------------------------------------
1760 :     Name: Matthias Blume
1761 :     Date: 2001/06/15 15:05:00 EDT
1762 : blume 840 Tag: blume-20010615-sparc
1763 :     Description:
1764 :    
1765 :     The purpose of this update is to provide an implementation of NLFFI
1766 :     on Sparc machines.
1767 :    
1768 :     Here are the changes in detail:
1769 :    
1770 :     * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1771 :     the Sparc implementation of the c-calls API.
1772 :     * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1773 :     for accessing the ML frame. Thus, we have a real frame pointer and
1774 :     can freely modify %sp without need for an omit-frame-ptr phase.
1775 :     The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1776 :     this case.
1777 :     * ml-nlffigen has been taught to produce code for different architectures
1778 :     and calling conventions.
1779 :     * In a way similar to what was done in the x86 case, the Sparc
1780 :     backend uses its own specific extension to mltree. (For example,
1781 :     it needs to be able to generate UNIMP instructions which are part
1782 :     of the calling convention.)
1783 :     * ml-nlffi-lib was reorganized to make it more modular (in particular,
1784 :     to make it easier to plug in new machine- and os-dependent parts).
1785 :    
1786 :     There are some other fairly unrelated bug fixes and cleanups as well:
1787 :    
1788 :     * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1789 :     that they properly share their libraries with existing SML/NJ libraries.
1790 :     * I fixed a minor cosmetic bug in CM, supressing certain spurious
1791 :     follow-up error messages.
1792 :     * Updates to CM/CMB documentation.
1793 :    
1794 :     TODO items:
1795 :    
1796 :     * MLRISC should use a different register as its asmTemp on the Sparc.
1797 :     (The current %o2 is a really bad choice because it is part of the
1798 :     calling conventions, so things might interfere in unexpected ways.)
1799 :    
1800 :     ----------------------------------------------------------------------
1801 :     Name: Matthias Blume
1802 : blume 839 Date: 2001/06/07
1803 :     Tag: blume-20010607-calls
1804 :     Description:
1805 :    
1806 :     A number of internal changes related to C calls and calling conventions:
1807 :    
1808 :     1. ML-Tree CALL statements now carry a "pops" field. It indicates the
1809 :     number of bytes popped implicitly (by the callee). In most cases
1810 :     this field is 0 but on x86/win32 it is some non-zero value. This
1811 :     is information provided for the benefit of the "omit-frameptr" pass.
1812 :     2. The CALL instruction on the x86 carries a similar "pops" field.
1813 :     The instruction selection phase copies its value from the ML-Tree
1814 :     CALL statement.
1815 :     3. On all other architectures, the instruction selection phase checks
1816 :     whether "pops=0" and complains if not.
1817 :     4. The c-calls implementation for x86 now accepts two calling conventions:
1818 :     "ccall" and "stdcall". When "ccall" is selected, the caller cleans
1819 :     up after the call and pops is set to 0. For "stdcall", the caller
1820 :     does nothing, leaving the cleanup to the callee; pops is set to
1821 :     the number of bytes that were pushed onto the stack.
1822 :     5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1823 :     distinguish between "ccall" and "stdcall".
1824 :     6. The UNIMP instruction has been added to the supported Sparc instruction
1825 :     set. (This is needed for implementing the official C calling convention
1826 :     on this architecture.)
1827 :     7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1828 :     work with the latest CM.
1829 :    
1830 :     ----------------------------------------------------------------------
1831 :     Name: Matthias Blume
1832 : blume 838 Date: 2001/06/05 15:10:00 EDT
1833 :     Tag: blume-20010605-cm-index
1834 :     Description:
1835 :    
1836 :     0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1837 :    
1838 :     1. CM can now generate "index files". These are human-readable files
1839 :     that list on a per-.cm-file basis each toplevel symbol defined or
1840 :     imported. The location of the index file for
1841 :     <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1842 :     To enable index-file generation, set CM.Control.generate_index to true
1843 :     or export an environment-symbol: export CM_GENERATE_INDEX=true.
1844 :    
1845 :     The CM manual has been updated accordingly.
1846 :    
1847 :     2. I made some slight modifications to the c-calls API in MLRISC.
1848 :    
1849 :     a) There is now a callback to support saving/restoring of
1850 :     dedicated but caller-save registers around the actual call
1851 :     instruction.
1852 :     b) One can optionally specify a comment-annotation for the
1853 :     call instruction.
1854 :    
1855 :     3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1856 :     (For example, the comment annotation shows the C prototype of
1857 :     the function being called.)
1858 :    
1859 :     ----------------------------------------------------------------------
1860 :     Name: Matthias Blume
1861 : blume 837 Date: 2001/06/01 13:30:00 EDT
1862 :     Tag: blume-20010601-nlffi-cleanup
1863 :     Description:
1864 :    
1865 :     This is mostly a cleanup of MLFFI stuff:
1866 :    
1867 :     - some signature files have been put into a more exposed place
1868 :     - the ugly 'f type parameter is gone (simplifies types tremendously!)
1869 :     - ml-nlffigen changed accordingly
1870 :     - tutorial updated
1871 :    
1872 :     Other changes:
1873 :    
1874 :     - author's affiliation in CM manual(s) updated
1875 :     - some more recognized keywords added to Allen's sml.sty
1876 :    
1877 :     ----------------------------------------------------------------------
1878 :     Name: Matthias Blume
1879 : blume 836 Date: 2001/05/25 15:30:00 EDT
1880 :     Tag: blume-20010525-iptr
1881 :     Description:
1882 :    
1883 :     - put the official 110.33-README (as it appears on the ftp server) under
1884 :     CVS
1885 :     - fixed a small bug related to incomplete pointer types in
1886 :     ml-nlffigen
1887 :     - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1888 :     (it does not need the 'f type parameter)
1889 :    
1890 :     ----------------------------------------------------------------------
1891 :     Name: Matthias Blume
1892 : blume 834 Date: 2001/05/23 14:30:00 EDT
1893 :     Tag: Release_110_33
1894 :     Description:
1895 :    
1896 :     New version number (110.33). New bootfiles.
1897 :    
1898 :     ----------------------------------------------------------------------
1899 :     Name: Matthias Blume
1900 : blume 833 Date: 2001/05/22 18:06:00 EDT
1901 :     Tag: blume-20010522-targets
1902 :     Description:
1903 :    
1904 :     Made install.sh use file config/targets.customized if it exists, falling
1905 :     back to config/targets if it doesn't. This way one can have a customized
1906 :     version of the targets file without touching the "real thing", thus
1907 :     eliminating the constant fear of accidentally checking something bogus
1908 :     back into the CVS repository... (File config/targets.customized must
1909 :     not be added to the repository!)
1910 :    
1911 :     ----------------------------------------------------------------------
1912 :     Name: Matthias Blume
1913 : blume 832 Date: 2001/05/22 16:30:00 EDT
1914 :     Tag: blume-20010522-minitut
1915 :     Description:
1916 :    
1917 :     1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1918 :     struct returns.
1919 :     2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt. This is some very
1920 :     incomplete, preliminary documentation for NLFFI.
1921 :    
1922 :     ----------------------------------------------------------------------
1923 :     Name: Matthias Blume
1924 : blume 829 Date: 2001/05/14 11:30:00 EDT
1925 :     Tag: blume-20010514-script
1926 :     Description:
1927 :    
1928 :     Some bugs in install script fixed.
1929 :    
1930 :     In addition to that I also made a slight change to the NLFFI API:
1931 :     Functors generated by ml-nlffigen now take the dynamic library as a
1932 :     straight functor argument, not as a suspended one. (The original
1933 :     functor code used to force the suspension right away anyway, so there
1934 :     was nothing gained by this complication of the interface.)
1935 :    
1936 :     ----------------------------------------------------------------------
1937 :     Name: Matthias Blume
1938 : blume 828 Date: 2001/05/11 14:35:00 EDT
1939 :     Tag: blume-20010511-ml-nlffi
1940 :     Description:
1941 :    
1942 :     I finally took the plunge and added my new FFI code to the main
1943 :     repository. For x86-linux it is now ready for prime-time.
1944 :    
1945 :     There are two new subdirectories of "src":
1946 :    
1947 :     - ml-nlffi-lib:
1948 :     The utility library for programs using the FFI interface.
1949 :     Here is the implementation of $/c.cm and its associated low-level
1950 :     partners $/c-int.cm and $/memory.cm.
1951 :     - ml-nlffigen:
1952 :     A stand-alone program for generating ML glue code from C source
1953 :     code.
1954 :    
1955 :     Building ml-nlffigen requires $/ckit-lib.cm.
1956 :    
1957 :     The config/install.sh script has been updates to do the Right Thing
1958 :     (hopefully).
1959 :    
1960 :     Notice that the source tree for the C-Kit will not be put under "src"
1961 :     but directly under the installation root directory. (This is the
1962 :     structure that currently exists on the CVS server when you check out
1963 :     module "sml".) Fortunately, config/install.sh knows about this oddity.
1964 :    
1965 :     Bugs: No documentation yet.
1966 :    
1967 :     ----------------------------------------------------------------------
1968 :     Name: Matthias Blume
1969 : blume 826 Date: 2001/05/09 16:35:00 EDT
1970 :     Tag: blume-20010509-cpscontract
1971 :     Description:
1972 :    
1973 :     Fixed a bug in the accounting code in cpsopt/contract.sml. (The
1974 :     wrapper/unwrapper elimination did not decrement usage counts and some
1975 :     dead variables got overlooked by the dead-up logic.)
1976 :    
1977 :     ----------------------------------------------------------------------
1978 : george 824 Name: Lal George
1979 :     Date: 2001/05/08 17:26:09 EDT
1980 :     Tag: george-20010508-omit-frameptr
1981 :     Description:
1982 :    
1983 :     Changes to implement the omit-frame-pointer optimization to support
1984 :     raw C calls. For now, there is only support on the Intel x86, but
1985 :     other architectures will follow as more experience is gained with this.
1986 :    
1987 :    
1988 :     ----------------------------------------------------------------------
1989 : blume 818 Name: Matthias Blume
1990 : blume 819 Date: 2001/05/07 14:40:00 EDT
1991 :     Tag: blume-20010507-proxies
1992 :     Description:
1993 :    
1994 :     I made into "proxy libraries" all libraries that qualify for such a
1995 :     change. (A qualifying library is a library that has another library or
1996 :     groups as its sole member and repeats that member's export list
1997 :     verbatim. A proxy library avoids this repetition by omitting its export
1998 :     list, effectively inheriting the list that its (only) member exports.
1999 :     See the CM manual for more explanation.)
2000 :     The main effect is that explicit export lists for these libraries
2001 :     do not have to be kepts in sync, making maintenance a bit easier.
2002 :    
2003 :     I also added copyright notices to many .cm-files.
2004 :    
2005 :     Last but not least, I made a new set of bootfiles.
2006 :    
2007 :     ----------------------------------------------------------------------
2008 :     Name: Matthias Blume
2009 : blume 818 Date: 2001/05/04 17:00:00 EDT
2010 :     Tag: blume-20010504-cm-lsplit
2011 :     Description:
2012 :    
2013 :     0. John merged pending changes to $/smlnj-lib.cm
2014 :    
2015 :     1. Allen's previous change accidentally backed out of one of Lal's
2016 :     earlier changes. I undid this mistake (re-introducing Lal's change).
2017 :    
2018 :     2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
2019 :     within the compiler where applicable. There is some code simplification
2020 :     because of that.
2021 :    
2022 :     3. The "split" phase (in FLINT) is now part of the default list of phases.
2023 :     Compiler.Control.LambdaSplitting.* can be used to globally control the
2024 :     lambda-splitting (cross-module-inlining) engine. In addition to that,
2025 :     it can now also be controlled on a per-source basis: CM has been taught
2026 :     a new tool parameter applicable to ML source files.
2027 :    
2028 :     - To turn lambda-splitting off completely:
2029 :     local open Compiler.Control.LambdaSplitting in
2030 :     val _ = set Off
2031 :     end
2032 :     - To make "no lambda-splitting" the global default (but allow per-source
2033 :     overriding); this is the initial setting:
2034 :     local open Compiler.Control.LambdaSplitting in
2035 :     val _ = set (Default NONE)
2036 :     end
2037 :     - To make "lambda-splitting with aggressiveness a" the global default
2038 :     (and allow per-source overriding):
2039 :     local open Compiler.Control.LambdaSplitting in
2040 :     val _ = set (Default (SOME a))
2041 :     end
2042 :    
2043 :     - To turn lambda-splitting off for a given ML souce file (say: a.sml)
2044 :     write (in the respective .cm-file):
2045 :     a.sml (lambdasplitting:off)
2046 :     - To turn lambda-splitting for a.sml on with minimal aggressiveness:
2047 :     a.sml (lambdasplitting:on)
2048 :     - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
2049 :     <a> is a decimal non-negative integer):
2050 :     a.sml (lambdasplitting:<a>)
2051 :     - To turn lambda-splitting for a.sml on with maximal aggressiveness:
2052 :     a.sml (lambdasplitting:infinity)
2053 :     - To use the global default for a.sml:
2054 :     a.sml (lambdasplitting:default)
2055 :     or simply
2056 :     a.sml
2057 :    
2058 :     ----------------------------------------------------------------------
2059 : leunga 815 Name: Allen Leung
2060 :     Date: 2001/05/04 01:57:00 EDT
2061 :     Tag: leunga-20010504-sync
2062 :     Description:
2063 : leunga 788
2064 : leunga 815 MLRISC features.
2065 :    
2066 :     1. Fix to CMPXCHG instructions.
2067 :     2. Changed RA interface to allow annotations in callbacks.
2068 :     3. Added a new method to the stream interface to allow annotations updates.
2069 :    
2070 : george 749 ----------------------------------------------------------------------
2071 : blume 812 Name: Matthias Blume
2072 : blume 814 Date: 2001/05/01 11:45:00 EDT
2073 :     Tag: blume-20010501-pcedittmp
2074 :     Description:
2075 :    
2076 :     Changed install.sh to use the current working directory instead of
2077 :     /usr/tmp for a temporary file (pcedittmp). The previous choice
2078 :     of /usr/tmp caused trouble with MacOS X because of file premission
2079 :     problems.
2080 :    
2081 :     ----------------------------------------------------------------------
2082 :     Name: Matthias Blume
2083 : blume 812 Date: 2001/04/20 11:10:00 EDT
2084 :     Tag: blume-20010420-inMLflag
2085 :     Description:
2086 :    
2087 :     - added vp_limitPtrMask to vproc-state.h
2088 :     (for use by the raw-C-calls mechanism to implement proper interrupt
2089 :     handling)
2090 :     - made the ML compiler aware of various data-structure offsets so it
2091 :     can generate code for accessing the vp_inML flag and vp_limitPtrMask
2092 :     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
2093 :     raw C-calls
2094 :    
2095 :     ----------------------------------------------------------------------
2096 : george 810 Name: Lal George
2097 :     Date: 2001/04/20 09:15:28 EDT
2098 :     Tag: george-20010420-macosX
2099 :     Description:
2100 :    
2101 :     - Changes to port to Mac OS X; Darwin.
2102 :    
2103 :     - In the process I found that sqrt was broken on the PPC, because the
2104 :     fsqrt instruction is not implemented.
2105 :    
2106 :     ----------------------------------------------------------------------
2107 : blume 799 Name: Matthias Blume
2108 : blume 808 Date: 2001/04/18 12:45:00 EDT
2109 :     Tag: blume-20010418-ccalls
2110 :     Description:
2111 :    
2112 :     - fixed two off-by-4 errors in the x86-specific c-calls implementation
2113 :     (this bug prevented structure arguments containing pointers from being
2114 :     passed correctly)
2115 :     - changed the raw-C-call code in mlriscGen.sml in such a way that
2116 :     structure arguments are represented as a pointer to the beginning
2117 :     of the structure (instead of having a series of synthesized arguments,
2118 :     one for each structure member)
2119 :    
2120 :     - made makeml script's verbosity level configurable via environment
2121 :     variable (MAKEML_VERBOSITY)
2122 :    
2123 : blume 813 - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
2124 : blume 808 in rawmem-x86.sml; we are now using the real thing
2125 :    
2126 :     ----------------------------------------------------------------------
2127 :     Name: Matthias Blume
2128 : blume 806 Date: 2001/03/22 16:25:00 EST
2129 :     Tag: blume-20010322-bootfiles
2130 :     Description:
2131 :    
2132 :     Created a new set of bootfiles (for your automatic installation convenience).
2133 :    
2134 :     ----------------------------------------------------------------------
2135 :     Name: Matthias Blume
2136 : blume 805 Date: 2001/03/22 15:10:00 EST
2137 :     Tag: blume-20010322-rawmem-parcm
2138 :     Description:
2139 :    
2140 :     1. All "raw memory access" primitives for the new FFI are implemented now
2141 :     (at least on the x86).
2142 :     2. Some further cleanup of CM's parallel make mechanism.
2143 :    
2144 :     ----------------------------------------------------------------------
2145 :     Name: Matthias Blume
2146 : blume 801 Date: 2001/03/19 17:53:00 EST
2147 :     Tag: blume-20010319-parallel
2148 :     Description:
2149 :    
2150 :     Parallel make (using compile servers) now works again.
2151 :    
2152 :     To this end, CM.stabilize and CMB.make have been modified to work in
2153 :     two passes when compile servers are attached:
2154 :     1. Compile everything, do not perform stabilization; this pass
2155 :     uses compile servers
2156 :     2. Stabilize everything; this pass does not use compile servers
2157 :     If there are no compile servers, the two passes are combined into one
2158 :     (as before). Splitting the passes increases the inherent parallelism
2159 :     in the dependency graph because the entire graph including all
2160 :     libraries is available at the same time. This, in turn, improves
2161 :     server utilization. The downside is that the master process will
2162 :     have to do some extra work after compilation is done (because for
2163 :     technical reasons it must re-read all the binfiles during stabilization).
2164 :    
2165 :     ----------------------------------------------------------------------
2166 :     Name: Matthias Blume
2167 : blume 800 Date: 2001/03/16 12:22:00 EST
2168 :     Tag: blume-20010316-bootfiles
2169 :     Description:
2170 :    
2171 :     Created a new set of bootfiles (for your automatic installation convenience).
2172 :    
2173 :     ----------------------------------------------------------------------
2174 :     Name: Matthias Blume
2175 : blume 799 Date: 2001/03/16 11:00:00 EST
2176 :     Tag: blume-20010316-MLTREE-fixup
2177 :     Description:
2178 :    
2179 :     This is a minor fixup for an (untagged) earlier commit by Allen.
2180 :     (A file was missing).
2181 :    
2182 :     ----------------------------------------------------------------------
2183 : leunga 788 Name: Allen Leung
2184 : leunga 796 Date: Mon Mar 5 18:54:57 EST 2001
2185 :     Tag: leunga-20010305-cut-support
2186 :    
2187 :     1. New support for alternative control-flow in MLTREE.
2188 :     Currently we support
2189 :    
2190 :     FLOW_TO(CALL ...., [k1,...,kn])
2191 :    
2192 :     This is needed for 'cuts to' in C-- and try/handle-like constructs
2193 :     in Moby
2194 :    
2195 :     New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
2196 :    
2197 :     2. Register Allocator
2198 :    
2199 :     Changes in interface [from Fermin, John]
2200 :    
2201 :     3. Alpha 8-bit SLL support [Fermin]
2202 :    
2203 :     4. All architectures
2204 :    
2205 :     A new module (ClusterExpandCopies) for expanding parallel copies.
2206 :    
2207 :     ----------------------------------------------------------------------
2208 :     Name: Allen Leung
2209 : leunga 788 Date: 2001/02/27 23:07:00 EST
2210 :     Tag: leunga-20010227-minor-stuff
2211 :    
2212 :     1. Alpha bug fix for CMOVNE
2213 :     2. Handle mltree COND(..,FCMP ...,...)
2214 :     3. Bug fix in simplifier
2215 :    
2216 :     ----------------------------------------------------------------------
2217 : blume 777 Name: Matthias Blume
2218 : blume 781 Date: 2001/01/30 17:50:00 EST
2219 :     Tag: blume-20010130-sync
2220 :     Description:
2221 :    
2222 :     This is just a minor update to sync my devel branch with the main brach.
2223 :     The only visible change is the addition of some README files.
2224 :    
2225 :     ----------------------------------------------------------------------
2226 :     Name: Matthias Blume
2227 : blume 778 Date: 2001/01/12 23:30:00 JST
2228 :     Tag: blume-20010112-bootfiles
2229 :     Description:
2230 :    
2231 :     Made a new set of bootfiles that goes with the current state of the
2232 :     repository.
2233 :    
2234 :     ----------------------------------------------------------------------
2235 :     Name: Matthias Blume
2236 : blume 777 Date: 2001/01/12 21:20:00 JST
2237 :     Tag: blume-20010112-sync
2238 :     Description:
2239 :    
2240 :     I am just flushing out some minor changes that had accumulated in
2241 :     my private branch in order to sync with the main tree. (This is
2242 :     mainly because I had CVS trouble when trying to merge _into_ my
2243 :     private branch.)
2244 :    
2245 :     Most people should be completely unaffected by this.
2246 :    
2247 :     ----------------------------------------------------------------------
2248 : leunga 775 Name: Allen Leung
2249 :     Date: Thu Jan 11 21:03:00 EST 2001
2250 :     Tag: leunga-20010111-labexp=mltree
2251 :     Description:
2252 :    
2253 :     1. Removed the type LabelExp and replace it by MLTree.
2254 :     2. Rewritten mltree-simplify with the pattern matcher tool.
2255 :     3. There were some bugs in alpha code generator which would break
2256 :     64-bit code generation.
2257 :     4. Redo the tools to generate code with the
2258 :     5. The CM files in MLRISC (and in src/system/smlnj/MLRISC)
2259 :     are now generated by perl scripts.
2260 :    
2261 :     ----------------------------------------------------------------------
2262 : blume 771 Name: Matthias Blume
2263 : blume 774 Date: 2001/01/10 21:55:00 JST
2264 :     Tag: blume-20010110-rcc
2265 :     Description:
2266 :    
2267 :     The RCC stuff now seems to work (but only on the x86).
2268 :     This required hacking of the c-calls interface (and -implementation) in
2269 :     MLRISC.
2270 :    
2271 :     Normal compiler users should be unaffected.
2272 :    
2273 :     ----------------------------------------------------------------------
2274 :     Name: Matthias Blume
2275 : blume 773 Date: 2001/01/09 01:20:00 JST
2276 :     Tag: blume-20010109-rcc
2277 :     Description:
2278 :    
2279 :     This is a fairly big patch, flushing out a large number of pending
2280 :     changes that I made to my development copy over the last couple of days.
2281 :    
2282 :     Of practical relevance at this moment is a workaround for a pickling
2283 :     bug that Allen ran into the other day. The cause of the bug itself is
2284 :     still unknown and it might be hard to fix it properly, but the
2285 :     workaround has some merits of its own (namely somewhat reducing pickling
2286 :     overhead for certain libraries). Therefore, I think this solution should
2287 :     be satisfactory at this time.
2288 :    
2289 :     The rest of the changes (i.e., the vast majority) has to do with my
2290 :     ongoing efforts of providing direct support for C function calls from
2291 :     ML. At the moment there is a new primop "RAW_CCALL", typing magic
2292 :     in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
2293 :     case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
2294 :     translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
2295 :     handle RCC.
2296 :    
2297 :     The last part (the changes to mlriscGen.sml) are still known to be
2298 :     wrong on the x86 and not implemented on all other architectures. But
2299 :     the infrastructure is in place. I had to change a few functor
2300 :     signatures in the backend to be able to route the CCalls interface
2301 :     from MLRISC there, and I had to specialize the mltree type (on the
2302 :     x86) to include the necessary extensions. (The extensions themselves
2303 :     were already there and redy to go in MLRISC/x86).
2304 :    
2305 :     Everything should be very happy as soon as someone helps me with
2306 :     mlriscGen.sml...
2307 :    
2308 :     In any case, nothing of this should matter to anyone as long as the
2309 :     new primop is not being used (which is going to be the case unless you
2310 :     find it where I hid it :). The rest of the compiler is completely
2311 :     unaffected.
2312 :    
2313 :     ----------------------------------------------------------------------
2314 :     Name: Matthias Blume
2315 : blume 772 Date: 2001/01/05 00:30:00 JST
2316 :     Tag: blume-20010105-primops
2317 :     Description:
2318 :    
2319 :     Added some experimental support for work that I am doing right now.
2320 :     These changes mostly concern added primops, but there is also a new
2321 :     experimental C library in the runtime system (but currently not enabled
2322 :     anywhere except on Linux/X86).
2323 :    
2324 :     In the course of adding primops (and playing with them), I discovered that
2325 :     Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
2326 :     broken. (Zhong was very right he labeled this stuff as "major gross hack".)
2327 :     To recover, I made type information in INL_PRIM mandatory and changed
2328 :     prim.sml as well as built-in.sml accordingly. The InLine structure now
2329 :     has complete, correct type information (i.e., no bottom types).
2330 :    
2331 :     Since all these changes mean that we need new binfiles, I also bumped the
2332 :     version number to 110.32.1.
2333 :    
2334 :     ----------------------------------------------------------------------
2335 :     Name: Matthias Blume
2336 : blume 771 Date: 2000/12/30 22:10:00 JST
2337 :     Tag: blume-20001230-various
2338 :     Description:
2339 :    
2340 :     Added proxy libraries for MLRISC and let MLRISC libraries refer
2341 :     to each other using path anchors. (See CM manual for explanation.)
2342 :    
2343 :     Updated CM documentation.
2344 :    
2345 :     Fixed some bugs in CM.
2346 :    
2347 :     Implemented "proxy" libraries (= syntactic sugar for CM).
2348 :    
2349 :     Added "-quiet" option to makeml and changed runtime system accordingly.
2350 :    
2351 :     Added cleanup handler for exportML to reset timers and compiler stats.
2352 :    
2353 :     ----------------------------------------------------------------------
2354 : george 761 Name: Lal George
2355 :     Date: 2000/12/22 22:22:58 EST 2000
2356 :     Tag: Release_110_32
2357 :     Description:
2358 :    
2359 :     Infinite precision used throughout MLRISC.
2360 :     see MLRISC/mltree/machine-int.sig
2361 :    
2362 :     ----------------------------------------------------------------------
2363 : blume 756 Name: Matthias Blume
2364 : blume 760 Date: 2000/12/22 23:16:00 JST
2365 :     Tag: blume-20001222-warn
2366 :     Description:
2367 : blume 759
2368 : blume 760 Corrected wording and formatting of some CM warning message which I
2369 :     broke in my previous patch.
2370 : blume 759
2371 :     ----------------------------------------------------------------------
2372 :     Name: Matthias Blume
2373 : blume 763 Date: 2000/12/22 21:20:00 JST
2374 :     Tag: blume-20001222-anchorenv
2375 :     Description:
2376 :    
2377 :     Fixed CM's handling of anchor environments in connection with CMB.make.
2378 :    
2379 :     ----------------------------------------------------------------------
2380 :     Name: Matthias Blume
2381 : blume 758 Date: 2000/12/22 13:15:00 JST
2382 :     Tag: blume-20001222-cleanup
2383 :     Description:
2384 :    
2385 :     Removed src/cm/ffi which does not (and did not) belong here.
2386 :    
2387 :     ----------------------------------------------------------------------
2388 :     Name: Matthias Blume
2389 : blume 757 Date: 2000/12/21 23:55:00 JST
2390 :     Tag: blume-20001221-exn
2391 :     Description:
2392 :    
2393 :     Probably most important: CM no longer silently swallows all exceptions
2394 :     in the compiler.
2395 :     Plus: some other minor CM changes. For example, CM now reports some
2396 :     sizes for generated binfiles (code, data, envpickle, lambdapickle).
2397 :    
2398 :     ----------------------------------------------------------------------
2399 :     Name: Matthias Blume
2400 : blume 756 Date: 2000/12/15 00:01:05 JST
2401 :     Tag: blume-20001215-dirtool
2402 :     Description:
2403 :    
2404 :     - "dir" tool added.
2405 :     - improvements and cleanup to Tools structure
2406 :     - documentation updates
2407 :    
2408 :     ----------------------------------------------------------------------
2409 : leunga 752 Name: Allen Leung
2410 : leunga 755 Date: Thu Dec 14 03:45:24 EST 2000
2411 :     Description:
2412 :     Tag: leunga-20001214-int-inf
2413 :     Description:
2414 :    
2415 :     In IntInf, added these standard functions, which are missing from our
2416 :     implementation:
2417 :    
2418 :     andb : int * int -> int
2419 :     xorb : int * int -> int
2420 :     orb : int * int -> int
2421 :     notb : int -> int
2422 :     << : int * word -> int
2423 :     ~>> : int * word -> int
2424 :    
2425 :     Not tested, I hope they are correct.
2426 :    
2427 :     ----------------------------------------------------------------------
2428 :     Name: Allen Leung
2429 : leunga 752 Date: Fri Dec 8 19:23:26 EST 2000
2430 :     Description:
2431 :     Tag: leunga-20001208-nowhere
2432 :     Description:
2433 :    
2434 :     Slight improvements to the 'nowhere' tool to handle OR-patterns,
2435 :     to generate better error messages etc. Plus a brief manual.
2436 :    
2437 :     ----------------------------------------------------------------------
2438 : george 749 Name: Lal George
2439 :     Date: 2000/12/08 09:54:02 EST 2000
2440 :     Tag: Release_110_31
2441 :     Description:
2442 : george 705
2443 : george 749 - Version 110.31
2444 : leunga 591 ----------------------------------------------------------------------
2445 : leunga 744 Name: Allen Leung
2446 :     Date: Thu Dec 7 22:01:04 EST 2000
2447 :     Tag: leunga-20001207-cell-monster-hack
2448 :     Description:
2449 :    
2450 :     Major MLRISC internal changes. Affect all clients.
2451 :     Summary:
2452 :    
2453 :     1. Type CELLS.cell = int is now replaced by a datatype.
2454 :     As a result, the old regmap is now gone. Almost all interfaces
2455 :     in MLRISC change as a consequence.
2456 :    
2457 :     2. A new brand version of machine description tool (v3.0) that generates
2458 :     modules expecting the new interface. The old version is removed.
2459 :    
2460 :     3. The RA interface has been further abstracted into two new functors.
2461 :     RISC_RA and X86RA. These functors have much simpler interfaces.
2462 :     [See also directory MLRISC/demo.]
2463 :    
2464 :     4. Some other new source->source code generation tools are available:
2465 :    
2466 :     a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
2467 :     b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
2468 :     I use this tool to generate the peephole optimizers---with the new
2469 :     cell type changes, peephole rules are becoming difficult to write
2470 :     without conditional pattern matching.
2471 :    
2472 :     5. More Intmap -> IntHashTable change. Previous changes by Matthias didn't
2473 :     cover the entire MLRISC source tree so many things broke.
2474 :    
2475 :     6. CM files have been moved to the subdirectory MLRISC/cm.
2476 :     They are moved because there are a lot of them and they clutter up the
2477 :     root dir.
2478 :    
2479 :     7. More detailed documentation to come...
2480 :    
2481 :     NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
2482 :     a makeml -rebuild first. This is because of other other
2483 :     changes that Matthias has made (see below).
2484 :    
2485 :    
2486 :     ----------------------------------------------------------------------
2487 : blume 733 Name: Matthias Blume
2488 : blume 742 Date: 2000/11/30 23:12:00 JST
2489 :     Tag: blume-20001130-filereorg
2490 :     Description:
2491 :    
2492 :     Some manual updates and some file reorganizations in CM.
2493 :    
2494 :     ----------------------------------------------------------------------
2495 :     Name: Matthias Blume
2496 : blume 737 Date: 2000/11/24 17:45:00 JST
2497 :     Tag: blume-20001124-link
2498 :     Description:
2499 :    
2500 :     Drastically improved link traversal code for the case that the dynamic
2501 :     value was already loaded at bootstrap time. As a result, CM and CMB
2502 :     now both load blazingly fast -- even on a very slow machine. Also,
2503 :     memory consumption has been further reduced by this.
2504 :    
2505 :     Warning: The format of the PIDMAP file has changed. THerefore, to
2506 :     bootstrap you have to do this:
2507 :    
2508 :     1. Run CMB.make
2509 :     2. Make a symbolic link for the boot directory:
2510 :     ln -s sml.boot.ARCH-OS xxx
2511 :     3. "Rebuild" the boot directory:
2512 :     ./makeml -boot xxx -rebuild sml ; rm xxx
2513 :     4. Boot normally:
2514 :     ./makeml
2515 :    
2516 :     ----------------------------------------------------------------------
2517 :     Name: Matthias Blume
2518 : blume 735 Date: 2000/11/21 21:20:00 JST
2519 :     Tag: blume-20001121-tools
2520 :     Description:
2521 :    
2522 :     Continued hacking on autoloading problem -- with success this time.
2523 :     Also changed tool-plugin mechanism. See new CM manual.
2524 :    
2525 :     ----------------------------------------------------------------------
2526 :     Name: Matthias Blume
2527 : blume 734 Date: 2000/11/19 14:30:00 JST
2528 :     Tag: blume-20001119-autoload
2529 :     Description:
2530 :    
2531 :     Some hacking to make autoloading faster. Success for CMB, no success
2532 :     so far for CM. There is a reduced structure CM' that autoloads faster.
2533 :     (This is a temporary, non-documented hack to be eliminated again when
2534 :     the general problem is solved.)
2535 :    
2536 :     ----------------------------------------------------------------------
2537 :     Name: Matthias Blume
2538 : blume 733 Date: 2000/11/17 14:10:00 JST
2539 :     Tag: blume-20001117-pickle-lib
2540 :     Description:
2541 :    
2542 :     1. Eliminated comp-lib.cm
2543 :     2. Made pickle-lib.cm
2544 :     3. Eliminated all uses of intset.sml (from comp-lib.cm)
2545 :     4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
2546 :     equivalent constructs from smlnj-lib.cm (INtHashTable).
2547 :     5. Point 4. also goes for those uses of intmap.* in MLRISC.
2548 :     Duplicated intmap modules thrown out.
2549 :     6. Hunted down all duplicated SCC code and replaced it with
2550 :     equivalent stuff (GraphSCCFn from smlnj-lib.cm).
2551 :     7. Rewrote Feedback module.
2552 :     8. Moved sortedlist.sml into viscomp-lib.cm. Eventually it
2553 :     should be thrown out and equivalent modules from smlnj-lib.cm
2554 :     should be used (IntRedBlackSet, IntListSet, ...).
2555 :    
2556 :     Confirmed that compiler compiles to fixpoint.
2557 :    
2558 :     ----------------------------------------------------------------------
2559 : leunga 731 Name: Allen Leung
2560 :     Date: 2000/11/10 18:00:00
2561 :     Tag: leunga-20001110-new-x86-fp
2562 :    
2563 :     A new x86 floating point code generator has been added.
2564 :     By default this is turned off. To turn this on, do:
2565 :    
2566 :     CM.autoload "$smlnj/compiler.cm";
2567 :     Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
2568 :    
2569 :     Changes:
2570 :    
2571 :     1. Changed FTAN to FPTAN so that the assembly output is correct.
2572 :     2. Changed the extension callback for FTANGENT to generate:
2573 :    
2574 :     fptan
2575 :     fstp %st(0)
2576 :     instead of
2577 :     fptan
2578 :     fstpl ftempmem
2579 :    
2580 :     3. Numerous assembly fixes for x86.
2581 :    
2582 :     5. Cleaned up the machine code output module x86/x86MC.sml and added
2583 :     support for a whole bunch of instructions and addressing modes:
2584 :    
2585 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st, %st(n)
2586 :     faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp %st, %st(n)
2587 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st(n), %st
2588 :     fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
2589 :     fxch %st(n)
2590 :     fld %st(n)
2591 :     fst %st(n)
2592 :     fst mem
2593 :     fstp %st(n)
2594 :     fucom %st(n)
2595 :     fucomp %st(n)
2596 :    
2597 :     All these are now generated when the fast fp mode is turned on.
2598 :    
2599 :     6. Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
2600 :    
2601 :     ----------------------------------------------------------------------
2602 : blume 729 Name: Matthias Blume
2603 :     Date: 2000/11/09 11:20:00 JST
2604 :     Tag: blume-20001109-scc
2605 :     Description:
2606 :    
2607 :     Eliminated some code duplication:
2608 :    
2609 :     1. Added "where" clause to GraphSCCFn in SML/NJ Library.
2610 :     (Otherwise the functor is useless.)
2611 :     2. Used GraphSCCFn where SCCUtilFun was used previously.
2612 :     3. Got rid of SCCUtilFun (in comp-lib.cm).
2613 :    
2614 :     ----------------------------------------------------------------------
2615 : george 721 Name: Lal George
2616 :     Date: 2000/11/06 09:02:21 EST 2000
2617 :     Tag: Release_110_30
2618 :     Description:
2619 :    
2620 :     - Version 110.30
2621 :     ----------------------------------------------------------------------
2622 : blume 715 Name: Matthias Blume
2623 : blume 716 Date: 2000/11/04 14:45:00
2624 :     Tag: blume-20001104-mlbuild
2625 :     Description:
2626 :    
2627 :     - Made ml-build faster on startup.
2628 :     - Documentation fixes.
2629 :    
2630 :     ----------------------------------------------------------------------
2631 :     Name: Matthias Blume
2632 : blume 715 Date: 2000/11/02 17:00:00 JST
2633 :     Tag: blume-20001102-condcomp
2634 :     Description:
2635 :    
2636 :     - Small tweaks to pickler -- new BOOTFILES!
2637 :     - Version bumped to 110.29.2.
2638 :     - Added conditional compilation facility to init.cmi (see comment there).
2639 :     ----------------------------------------------------------------------
2640 : leunga 713 Name: Allen Leung
2641 :     Date: 2000/10/23 19:31:00
2642 :     Tag: leunga-20001023-demo-ra
2643 :    
2644 :     1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
2645 :     2. Test programs for the graph library updated
2646 :     3. Some new MLRISC demo programs added
2647 :    
2648 :     ----------------------------------------------------------------------
2649 : blume 710 Name: Matthias Blume
2650 : blume 711 Date: 2000/08/31 22:15:00 JST
2651 :     Tag: blume-20001017-errmsg
2652 :     Description:
2653 :    
2654 :     More error message grief: Where there used to be no messages, there
2655 :     now were some that had bogus error regions. Fixed.
2656 :    
2657 :     ----------------------------------------------------------------------
2658 :     Name: Matthias Blume
2659 : blume 710 Date: 2000/08/31 17:30:00 JST
2660 :     Tag: blume-20001017-v110p29p1
2661 :     Description:
2662 :    
2663 :     I made a version 110.29.1 with new bootfiles.
2664 :    
2665 :     Changes: Modified pickler/unpickler for faster and leaner unpickling.
2666 :     CM documentation changes and a small bugfix in CM's error reporting.
2667 :    
2668 :     ----------------------------------------------------------------------
2669 : george 705 Name: Lal George
2670 :     Date: 2000/09/27 14:42:35 EDT
2671 :     Tag: george-20000927-nodestatus
2672 :     Description:
2673 :    
2674 :     Changed the type of the nodestatus, so that:
2675 :    
2676 :     SPILLED(~1) is now SPILLED
2677 :     SPILLED(m) where m>=0 is now MEMREG(m)
2678 :     SPILLED(s) where s<~1 is now SPILL_LOC(~s)
2679 :    
2680 :     ----------------------------------------------------------------------
2681 : blume 697 Name: Matthias Blume
2682 : blume 703 Date: 2000/09/07 14:45:00 JST
2683 :     Tag: blume-20000907-cmerrmsg
2684 :     Description:
2685 :    
2686 :     Small tweak to CM to avoid getting ML syntax error messages twice.
2687 :    
2688 :     ----------------------------------------------------------------------
2689 :     Name: Matthias Blume
2690 : blume 701 Date: 2000/08/31 18:00:00 JST
2691 :     Tag: blume-20000831-cvsbootfiles
2692 :     Description:
2693 :    
2694 :     New URL for boot files (because the 110.29 files on the BL server do
2695 :     now work correctly with my updated install scripts for yacc and lex).
2696 :    
2697 :     ----------------------------------------------------------------------
2698 :     Name: Matthias Blume
2699 : blume 697 Date: 2000/08/08 12:33:00 JST
2700 :     Tag: blume-20000808-manual
2701 :     Description:
2702 :    
2703 :     Tiny update to CM manual.
2704 :    
2705 :     ----------------------------------------------------------------------
2706 : leunga 695 Name: Allen Leung
2707 :     Date: 2000/08/7 19:31:00
2708 :     Tag: leunga-20000807-a-whole-bunch-of-stuff
2709 :    
2710 :     Moby, C--, SSA, x86, machine descriptions etc. Should only affect C--
2711 :     and Mobdy.
2712 :    
2713 :     1. x86
2714 :    
2715 :     a. Fixes to peephole module by John and Dan.
2716 :     b. Assembly fix to SETcc by Allen.
2717 :     c. Fix to c-call by John.
2718 :     d. Fix to spilling by John. (This one deals with the missing FSTPT case)
2719 :     e. Instruction selection optimization to SETcc as suggested by John.
2720 :    
2721 :     For example,
2722 :    
2723 :     MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
2724 :    
2725 :     should generate:
2726 :    
2727 :     MOVL a, x
2728 :     SUBL b, x
2729 :     SHRL 31, x
2730 :    
2731 :     2. IR stuff
2732 :    
2733 :     A bunch of new DJ-graph related algorithms added. These
2734 :     speed up SSA construction.
2735 :    
2736 :     3. SSA + Scheduling
2737 :    
2738 :     Added code for SSA and scheduling to the repository
2739 :    
2740 :     ----------------------------------------------------------------------
2741 : george 694 Name: Lal George
2742 :     Date: 2000/07/27 11:53:14 EDT
2743 :    
2744 :     Tag: lal-20000727-linux-ppc
2745 :     Description:
2746 :    
2747 :     Made changes to support Linux PPC.
2748 :     p.s. I have confirmation that the 110.29 boot files work fine.
2749 :    
2750 :     ----------------------------------------------------------------------
2751 : blume 684 Name: Matthias Blume
2752 : blume 692 Date: 2000/07/27 17:40:00 JST
2753 :     Tag: blume-20000727-scripts
2754 :     Description:
2755 :    
2756 :     !!!! WARNING !!!!
2757 :     You must recompile the runtime system!
2758 :     !!!! WARNING !!!!
2759 :    
2760 :     This is basically another round of script-enhancements:
2761 :    
2762 :     1. sml, ml-build, and ml-makedepend accept options -D and -U to define
2763 :     and undefine CM preprocessor symbols.
2764 :    
2765 :     2. ml-build avoids generating a new heap image if it finds that the
2766 :     existing one is still ok. (The condition is that no ML file had to
2767 :     be recompiled and all ML files are found to be older that the heap
2768 :     file.)
2769 :    
2770 :     To make this work smoothly, I also hacked the runtime system as
2771 :     well as SMLofNJ.SysInfo to get access to the heap image suffix
2772 :     (.sparc-solaris, ...) that is currently being used.
2773 :    
2774 :     Moreover, the signature of CM.mk_standalone has changed. See the
2775 :     CM manual.
2776 :    
2777 :     3. ml-makedepend accepts additional options -n, -a, and -o. (See the
2778 :     CM manual for details.)
2779 :    
2780 :     4. More CM manual updates:
2781 :     - all of the above has been documented.
2782 :     - there is now a section describing the (CM-related) command line
2783 :     arguments that are accepted by the "sml" command
2784 :    
2785 :     ----------------------------------------------------------------------
2786 :     Name: Matthias Blume
2787 : blume 691 Date: 2000/07/25 16:20:00 JST
2788 :     Tag: blume-20000725-makedepend
2789 :     Description:
2790 :    
2791 :     Added a script called ml-makedepend. This can be used in makefiles
2792 :     for Unix' make in a way very similar to the "makedepend" command for
2793 :     C.
2794 :    
2795 :     The script internally uses function CM.sources.
2796 :    
2797 :     Synopsis:
2798 :    
2799 :     ml-makedepend [-f makefile] cmfile targetname
2800 :    
2801 :     The default for the makefile is "makefile" (or "Makefile" should
2802 :     "makefile" not exist).
2803 :    
2804 :     ml-makedepend adds a cmfile/targetname-specific section to this
2805 :     makefile (after removing the previous version of this section). The
2806 :     section contains a single dependency specification with targetname on
2807 :     the LHS (targetname is an arbitrary name), and a list of files derived
2808 :     from the cmfile on the RHS. Some of the files on the RHS are
2809 :     ARCH/OPSYS-specific. Therefore, ml-makedepend inserts references to
2810 :     "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
2811 :     path names. The makefile writer is responsible for making sure that
2812 :     these variables have correct at the time "make" is invoked.
2813 :    
2814 :     ----------------------------------------------------------------------
2815 :     Name: Matthias Blume
2816 : blume 690 Date: 2000/07/22 23:30:00 JST
2817 :     Tag: blume-20000722-urlupdate
2818 :     Description:
2819 :    
2820 :     Changed BOOT and config/srcarchiveurl to point to BL server:
2821 :    
2822 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
2823 :    
2824 :     ----------------------------------------------------------------------
2825 :     Name: Matthias Blume
2826 : blume 686 Date: 2000/07/18 18:00:00 JST
2827 :     Tag: blume-20000718-Version_110_29
2828 :     Description:
2829 :    
2830 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
2831 :    
2832 :     2. Updated config/version to 110.29
2833 :    
2834 :     3. Updated config/srcarchiveurl
2835 :    
2836 :     3. New boot files!
2837 :     ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
2838 :    
2839 :     ----------------------------------------------------------------------
2840 :     Name: Matthias Blume
2841 : blume 684 Date: 2000/07/11 13:58:00 JST
2842 :     Tag: blume-20000711-doctypo
2843 :     Description:
2844 :    
2845 :     Fixed a few typos in CM manual.
2846 :    
2847 :     ----------------------------------------------------------------------
2848 : leunga 683 Name: Allen Leung
2849 :     Date: 2000/06/15 00:38:00
2850 :     Tag: leunga-20000704-sparc-x86
2851 :    
2852 :     1. x86 peephole improvement sp += k; sp -= k => nop [from John]
2853 :     2. fix to x86 RET bug [found by Dan Grossman]
2854 :     3. sparc assembly bug fix for ticc instructions [found by Fermin]
2855 :    
2856 :     Affects c-- and moby only
2857 :    
2858 :     ----------------------------------------------------------------------
2859 : blume 680 Name: Matthias Blume
2860 : blume 682 Date: 2000/07/04 15:26:00
2861 :     Tag: blume-20000704-trigger
2862 :     Description:
2863 :    
2864 :     1. Improvements to CM manual.
2865 :     2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
2866 :     of getting a back-trace. The function, when called, raises an
2867 :     internal exception which explicitly carries the full back-trace history,
2868 :     so it is unaffected by any intervening handle-raise pairs ("trivial"
2869 :     or not). The interactive loop will print that history once it arrives
2870 :     at top level.
2871 :     Short of having all exceptions implicitly carry the full history, the
2872 :     recommended way of using this facility is:
2873 :     - compile your program with instrumentation "on"
2874 :     - run it, when it raises an exception, look at the history
2875 :     - if the history is "cut off" because of some handler, go and modify
2876 :     your program so that it explicitly calls BTrace.trigger
2877 :     - recompile (still instrumented), and rerun; look at the full history
2878 :    
2879 :     ----------------------------------------------------------------------
2880 :     Name: Matthias Blume
2881 :     Date: 2000/07/03 15:36:00 JST
2882 : blume 680 Tag: blume-20000702-manual
2883 :     Description:
2884 :    
2885 :     Small corrections and updates to CM manual.
2886 :    
2887 :     ----------------------------------------------------------------------
2888 : blume 682 Name: Matthias Blume
2889 : blume 679 Date: 2000/06/29 16:04:00 JST
2890 :     Tag: blume-20000629-yacctool
2891 :     Description:
2892 :    
2893 :     Changes:
2894 :    
2895 :     1. Class "mlyacc" now takes separate arguments to pass options to
2896 :     generated .sml- and .sig-files independently.
2897 :     2. Corresponding CM manual updates.
2898 :     3. BTrace module now also reports call sites. (However, for loop clusters
2899 :     it only shows from where the cluster was entered.) There are associated
2900 :     modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
2901 :    
2902 :     ----------------------------------------------------------------------
2903 : blume 675 Name: Matthias Blume
2904 : blume 678 Date: 2000/06/27 16:51:00 JST
2905 :     Tag: blume-20000627-noweb
2906 :     Description:
2907 :    
2908 :     Changes:
2909 :    
2910 :     1. Implemented "subdir" and "witness" options for noweb tool.
2911 :     This caused some slight internal changes in CM's tool implementation.
2912 :     2. Fixed bug in "tool plugin" mechanism. This is essentially cleaning
2913 :     some remaining issues from earlier path anchor changes.
2914 :     3. Updated CM manual accordingly.
2915 :    
2916 :     4. Changed implementation of back-tracing so that I now consider it
2917 :     ready for prime-time.
2918 :    
2919 :     In particular, you don't have to explicitly trigger the back-trace
2920 :     anymore. Instead, if you are running BTrace-instrumented code and
2921 :     there is an uncaught exception (regardless of whether or not it was
2922 :     raised in instrumented code), the top-level evalloop will print
2923 :     the back-trace.
2924 :    
2925 :     Features:
2926 :    
2927 :     - Instrumented and uninstrumented code work together seemlessly.
2928 :     (Of course, uninstrumented code is never mentioned in actual
2929 :     back-traces.)
2930 :    
2931 :     - Asymptotic time- and space-complexity of instrumented code is
2932 :     equal to that of uninstrumented code. (This means that
2933 :     tail-recursion is preserved by the instrumentation phase.)
2934 :    
2935 :     - Modules whose code has been instrumented in different sessions
2936 :     work together without problem.
2937 :    
2938 :     - There is no penalty whatsoever on uninstrumented code.
2939 :    
2940 :     - There is no penalty on "raise" expressions, even in
2941 :     instrumented code.
2942 :    
2943 :     A potential bug (or perhaps it is a feature, too):
2944 :    
2945 :     A back-trace reaches no further than the outermost instrumented
2946 :     non-trivial "raise". Here, a "trivial" raise is one that is the
2947 :     sole RHS of a "handle" rule. Thus, back-traces reach trough
2948 :    
2949 :     <exp> handle e => raise e
2950 :    
2951 :     and even
2952 :    
2953 :     <exp> handle Foo => raise Bar
2954 :    
2955 :     and, of course, through
2956 :    
2957 :     <exp> handle Foo => ...
2958 :    
2959 :     if the exception was not Foo.
2960 :    
2961 :     Back-traces always reach right through any un-instrumented code
2962 :     including any of its "handle" expressions, trivial or not.
2963 :    
2964 :     To try this out, do the following:
2965 :    
2966 :     - Erase all existing binfiles for your program.
2967 :     (You may keep binfiles for those modules where you think you
2968 :     definitely don't need back-tracing.)
2969 :     - Turn on back-trace instrumentation:
2970 :     SMLofNJ.Internals.BTrace.mode (SOME true);
2971 :     - Recompile your program. (I.e., run "CM.make" or "use".)
2972 :     - You may now turn instrumentation off again (if you want):
2973 :     SMLofNJ.Internals.BTrace.mode (SOME false);
2974 :     - Run your program as usual. If it raises an exception that
2975 :     reaches the interactive toplevel, then a back-trace will
2976 :     automatically be printed. After that, the toplevel loop
2977 :     will print the exception history as usual.
2978 :    
2979 :     ----------------------------------------------------------------------
2980 :     Name: Matthias Blume
2981 : blume 677 Date: 2000/06/26 09:56:46 JST
2982 :     Tag: blume-20000626-setup
2983 :     Description:
2984 :    
2985 :     CM: - setup-parameter to "sml" added; this can be used to run arbitrary
2986 :     ML code before and after compiling a file (e.g., to set compiler
2987 :     flags)
2988 :    
2989 :     Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
2990 :     - associated changes to btrace.sml (BTrace instrumentation pass)
2991 :     - cleaner implementation of btimp.sml (BTrace tracing and report
2992 :     module)
2993 :    
2994 :     CM manual: * new path encoding documented
2995 :     * description of setup-parameter to "sml" added
2996 :    
2997 :     The biggest user-visible change to back-tracing is that it is no
2998 :     longer necessary to compile all traced modules within the same
2999 :     session. (This was a real limitation.)
3000 :    
3001 :     ----------------------------------------------------------------------
3002 :     Name: Matthias Blume
3003 : blume 676 Date: 2000/06/24 12:40:00 JST
3004 :     Tag: blume-20000624-startup
3005 :     Description:
3006 :    
3007 :     Fixes startup slowdown problem. (I was calling SrcPath.sync a _tad_
3008 :     bit too often -- to put it mildly. :)
3009 :    
3010 :     ----------------------------------------------------------------------
3011 :     Name: Matthias Blume
3012 : blume 675 Date: 2000/06/23 18:20:00 JST
3013 :     Tag: blume-20000623-btrace
3014 :     Description:
3015 :    
3016 :     This updates adds a backtrace facility to aid programmers in debugging
3017 :     their programs. This involves the following changes:
3018 :    
3019 :     1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
3020 :     keeping track of the current call stack. When programs are compiled
3021 :     in a special mode, the compiler will insert calls to these hooks
3022 :     into the user program.
3023 :     "Hook" means that it is possible for different implementations of
3024 :     back-tracing to register themselves (at different times).
3025 :    
3026 :     2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
3027 :     as an Absyn.dec->Absyn.dec rewrite. Normally this phase is turned off.
3028 :     It can be turned on using this call:
3029 :     SMLofNJ.Internals.BTrace.mode (SOME true);
3030 :     Turning it off again:
3031 :     SMLofNJ.Internals.BTrace.mode (SOME false);
3032 :     Querying the current status:
3033 :     SMLofNJ.Internals.BTrace.mode NONE;
3034 :     Annotated programs are about twice as big as normal ones, and they
3035 :     run a factor of 2 to 4 slower with a dummy back-trace plugin (one
3036 :     where all hooks do nothing). The slowdown with a plugin that is
3037 :     actually useful (such as the one supplied by default) is even greater,
3038 :     but in the case of the default plugin it is still only an constant
3039 :     factor (amortized).
3040 :    
3041 :     3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
3042 :     with a sub-structure BTrace for controlling back-tracing. In particular,
3043 :     the above-mentioned function "mode" controls whether the annotation
3044 :     phase is invoked by the compiler. Another important function is
3045 :     "trigger": when called it aborts the current execution and causes
3046 :     the top-level loop to print a full back-trace.
3047 :    
3048 :     4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
3049 :     for back-tracing. It keeps track of the dynamic call stack and in
3050 :     addition to that it keeps a partial history at each "level" of that
3051 :     stack. For example, if a tail-calls b, b tail-calls c, and c tail-calls
3052 :     d and b (at separate times, dynamically), then the report will show:
3053 :    
3054 :     GOTO d
3055 :     /c
3056 :     GOTO \b
3057 :     CALL a
3058 :    
3059 :     This shows that there was an initial non-tail call of a, then a
3060 :     tail-call to b or c, looping behavior in a cluster of functions that
3061 :     consist of b and c, and then a goto from that cluster (i.e., either from
3062 :     b or from c) to d.
3063 :    
3064 :     Note that (depending on the user program) the amount of information
3065 :     that the back-trace module has to keep track of at each level is bounded
3066 :     by a constant. Thus, the whole implementation has the same asymptotical
3067 :     complexity as the original program (both in space and in time).
3068 :    
3069 :     5. compiler/TopLevel/interact/evalloop.sml has been modified to
3070 :     handle the special exception SMLofNJ.Internals.BTrace.BTrace
3071 :     which is raised by the "trigger" function mentioned above.
3072 :    
3073 :     Notes on usage:
3074 :    
3075 :     - Annotated code works well together with unannotated code:
3076 :     Unannotated calls simply do not show up at all in the backtrace.
3077 :    
3078 :     - It is not a good idea to let modules that were annotated during
3079 :     different sessions run at the same time. This is because the compiler
3080 :     chooses small integers to identify individual functions, and there
3081 :     will be clashes if different modules were compiled in separate sessions.
3082 :     (Nothing will crash, and you will even be told about the clashes, but
3083 :     back-trace information will in general not be useful.)
3084 :    
3085 :     - Back-tracing can be confused by callcc and capture.
3086 :    
3087 :     - The only way of getting a back-trace right now is to explicitly
3088 :     invoke the "trigger" function from your user program. Eventually, we
3089 :     should make every exception carry back-trace information (if
3090 :     available). But since this creates more overhead at "raise"-time
3091 :     (similar to the current exnHistory overhead), I have not yet
3092 :     implemented this. (The implementation will be rather easy.) With
3093 :     exceptions carrying back-trace information, this facility will be even
3094 :     more useful because users don't need to modify their programs...
3095 :    
3096 :     - While it is possible to compile the compiler with back-trace
3097 :     annotations turned on (I did it to get some confidence in
3098 :     correctness), you must make absolutely sure that core.sml and
3099 :     btimp.sml are compiled WITHOUT annotation! (core.sml cannot actually
3100 :     be compiled with annotation because there is no core access yet, but
3101 :     if you compile btimp.sml with annotation, then the system will go into
3102 :     an infinite recursion and crash.)
3103 :     Since CM currently does not know about BTrace, the only way to turn
3104 :     annotations on and off for different modules of the compiler is to
3105 :     interrupt CMB.make, change the settings, and re-invoke it. Of course,
3106 :     this is awkward and clumsy.
3107 :    
3108 :     Sample sessions:
3109 :    
3110 :     Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
3111 :     - SMLofNJ.Internals.BTrace.mode (SOME true);
3112 :     [autoloading]
3113 :     [autoloading done]
3114 :     val it = false : bool
3115 :     - structure X = struct
3116 :     - fun main n = let
3117 :     - fun a (x, 0) = d x
3118 :     - | a (x, n) = b (x, n - 1)
3119 :     - and b (x, n) = c (x, n)
3120 :     - and c (x, n) = a (x, n)
3121 :     - and d x = e (x, 3)
3122 :     - and e (x, 0) = f x
3123 :     - | e (x, n) = e (x, n - 1)
3124 :     - and f 0 = SMLofNJ.Internals.BTrace.trigger ()
3125 :     - | f n = n * g (n - 1)
3126 :     - and g n = a (n, 3)
3127 :     - in
3128 :     - f n
3129 :     - end
3130 :     - end;
3131 :     structure X : sig val main : int -> int end
3132 :     - X.main 3;
3133 :     *** BACK-TRACE ***
3134 :     GOTO stdIn:4.2-13.20: X.main[2].f
3135 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
3136 :     GOTO stdIn:4.2-13.20: X.main[2].d
3137 :     / stdIn:4.2-13.20: X.main[2].a
3138 :     | stdIn:4.2-13.20: X.main[2].b
3139 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
3140 :     CALL stdIn:4.2-13.20: X.main[2].g
3141 :     GOTO stdIn:4.2-13.20: X.main[2].f
3142 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
3143 :     GOTO stdIn:4.2-13.20: X.main[2].d
3144 :     / stdIn:4.2-13.20: X.main[2].a
3145 :     | stdIn:4.2-13.20: X.main[2].b
3146 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
3147 :     CALL stdIn:4.2-13.20: X.main[2].g
3148 :     GOTO stdIn:4.2-13.20: X.main[2].f
3149 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
3150 :     GOTO stdIn:4.2-13.20: X.main[2].d
3151 :     / stdIn:4.2-13.20: X.main[2].a
3152 :     | stdIn:4.2-13.20: X.main[2].b
3153 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
3154 :     CALL stdIn:4.2-13.20: X.main[2].g
3155 :     GOTO stdIn:4.2-13.20: X.main[2].f
3156 :     CALL stdIn:2.15-17.4: X.main[2]
3157 :     -
3158 :    
3159 :     (Note that because of a FLINt bug the above code currently does not
3160 :     compile without BTrace turned on.)
3161 :    
3162 :     Here is another example, using my modified Tiger compiler:
3163 :    
3164 :     Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
3165 :     - SMLofNJ.Internals.BTrace.mode (SOME true);
3166 :     [autoloading]
3167 :     [autoloading done]
3168 :     val it = false : bool
3169 :     - CM.make "sources.cm";
3170 :     [autoloading]
3171 :     ...
3172 :     [autoloading done]
3173 :     [scanning sources.cm]
3174 :     [parsing (sources.cm):parse.sml]
3175 :     [creating directory CM/SKEL ...]
3176 :     [parsing (sources.cm):tiger.lex.sml]
3177 :     ...
3178 :     [wrote CM/sparc-unix/semant.sml]
3179 :     [compiling (sources.cm):main.sml]
3180 :     [wrote CM/sparc-unix/main.sml]
3181 :     [New bindings added.]
3182 :     val it = true : bool
3183 :     - Main.compile ("../testcases/merge.tig", "foo.out");
3184 :     *** BACK-TRACE ***
3185 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
3186 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3187 :     CALL lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
3188 :     GOTO lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
3189 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3190 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3191 :     CALL lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
3192 :     / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
3193 :     CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
3194 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3195 :     CALL lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
3196 :     CALL main.sml:1.18-118.4: Main.compile[2]
3197 :     -
3198 :    
3199 :     ----------------------------------------------------------------------
3200 : blume 673 Name: Matthias Blumen
3201 :     Date: 2000/06/21 18:00:00 JST
3202 :     Tag: blume-20000621-manual
3203 :     Description:
3204 :    
3205 :     CM manual update: Path environments documented.
3206 :    
3207 :     ----------------------------------------------------------------------
3208 : blume 666 Name: Matthias Blume
3209 : blume 672 Date: 2000/06/19 13:40:00
3210 :     Tag: blume-20000619-manual
3211 :     Description:
3212 :    
3213 :     CM manual and system/README update. This only covers the fact that
3214 :     there are no more implicit anchors. (Path environments and the "bind"
3215 :     option to "cm" have yet to be documented.)
3216 :    
3217 :     ----------------------------------------------------------------------
3218 :     Name: Matthias Blume
3219 : blume 671 Date: 2000/06/19 11:05:00 JST
3220 :     Tag: blume-20000619-chdir-bugfix
3221 :     Description:
3222 :    
3223 :     Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
3224 :    
3225 :     ----------------------------------------------------------------------
3226 :     Name: Matthias Blume
3227 : blume 670 Date: 2000/06/18 22:00:10 JST
3228 :     Tag: blume-20000618-implicit-anchors-really-gone
3229 :     Description:
3230 :    
3231 :     I updates the previous HISTORY entry where I forgot to mention that
3232 :     implicit anchors are no longer with us.
3233 :    
3234 :     The current update also gets rid of the (now useless) controller
3235 :     CM.Control.implicit_anchors.
3236 :    
3237 :     ----------------------------------------------------------------------
3238 :     Name: Matthias Blume
3239 : blume 666 Date: 2000/06/16 17:30:00 JST
3240 :     Tag: blume-20000616-anchorenv
3241 :     Description:
3242 :    
3243 :     This patch implements the long anticipated (just kidding :) "anchor
3244 :     environment" mechanism. In the course of doing this, I also
3245 :     re-implemented CM's internal "SrcPath" module from scratch. The new
3246 :     one should be more robust in certain boundary cases. In any case, it
3247 :     is a lot cleaner than its predecessor (IMHO).
3248 :    
3249 :     This time, although there is yet another boot file format change, I
3250 :     kept the unpickler backward-compatible. As a result, no new bootfiles
3251 :     are necessary and bootstrapping is straightforward. (You cannot read
3252 :     new bootfiles into an old system, but the other way around is no
3253 :     problem.)
3254 :    
3255 :     Visible changes:
3256 :    
3257 : blume 670 ** 0. Implicit path anchors (without the leading $-symbol) are no
3258 :     longer recognized at all. This means that such path names are not
3259 :     illegal either. For example, the name basis.cm simply refers to a
3260 :     local file called "basis.cm" (i.e, the name is an ordinary path
3261 :     relative to .cm-files directory). Or, to put it differently, only
3262 :     names that start with $ are anchored paths.
3263 : blume 666
3264 : blume 670 ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
3265 :     vanished.
3266 :    
3267 :     John (Reppy) had critizised this as soon as I originally proposed and
3268 : blume 666 implemented it, but at that time I did not really deeply believe
3269 :     him. :) Now I came full-circle because I need the $<singlearc> syntax
3270 :     in another place where it cannot be seen as an abbreviation for
3271 :     $/<singlearc>. To avoid the confusion, $<singlearc> now means what it
3272 :     seems to mean (i.e., it "expands" into the corresponding anchor
3273 :     value).
3274 :    
3275 :     However, when paths are used as members in CM description files, it
3276 :     continues to be true that there must be at least another arc after the
3277 :     anchor. This is now enforced separately during semantic analysis
3278 : blume 670 (i.e., from a lexical/syntactical point of view, the notation is ok.)
3279 : blume 666
3280 :     ** 2. The "cm" class now accepts an option "bind". The option's value
3281 :     is a sub-option list of precisely two items -- one labeled "anchor"
3282 :     and the other one labeled "value". As you might expect, "anchor" is
3283 :     used to specify an anchor name to be bound, and "value" specifies what
3284 :     the anchor is being bound to.
3285 :    
3286 :     The value must be a directory name and can be given in either standard
3287 :     syntax (including the possibility that it is itself an anchored path)
3288 :     or native syntax.
3289 :    
3290 :     Examples:
3291 :    
3292 :     foo.cm (bind:(anchor:bar value:$mystuff/bar))
3293 :     lib.cm (bind:(anchor:a value:"H:\\x\\y\\z")) (* only works under windows *)
3294 :    
3295 :     and so on.
3296 :    
3297 :     The meaning of this is that the .cm-file will be processed with an
3298 :     augmented anchor environment where the given anchor(s) is/are bound to
3299 :     the given values(s).
3300 :    
3301 :     The rationale for having this feature is this: Suppose you are trying
3302 :     to use two different (already stable) libraries a.cm and b.cm (that
3303 :     you perhaps didn't write yourself). Further, suppose each of these
3304 :     two libraries internally uses its own auxiliary library $aux/lib.cm.
3305 :     Normally you would now have a problem because the anchor "lib" can not
3306 :     be bound to more than one value globally. Therefore, the project that
3307 :     uses both a.cm and b.cm must locally redirect the anchor to some other
3308 :     place:
3309 :    
3310 :     a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
3311 :     b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
3312 :    
3313 :     This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
3314 :     /usr/lib/smlnj/b-stuff/aux.cm, respectively.
3315 :    
3316 :     Hard-wiring path names is a bit inflexible (and CM will verbosely warn
3317 :     you when you do so at the time of CM.stabilize). Therefore, you can
3318 :     also use an anchored path as the value:
3319 :    
3320 :     a.cm (bind:(anchor:lib value:$a-lib))
3321 :     b.cm (bind:(anchor:lib value:$b-lib))
3322 :    
3323 :     Now you can globally configure (using the usual CM.Anchor.anchor or
3324 :     pathconfig machinery) bindings for "a-lib" and "b-lib". Since "lib"
3325 :     itself is always locally bound, setting it globally is no longer
3326 :     meaningful or necessary (but it does not hurt either). In fact, "lib"
3327 :     can still be used as a global anchor for separate purposes. As a
3328 :     matter of fact, one can locally define "lib" in terms of a global
3329 :     "lib":
3330 :    
3331 :     a.cm (bind:(anchor:lib value:$lib/a))
3332 :     b.cm (bind:(anchor:lib value:$lib/b))
3333 :    
3334 :     ** 3: The encoding of path names has changed. This affects the way
3335 :     path names are shown in CM's progress report and also the internal
3336 :     protocol encoding used for parallel make.
3337 :    
3338 :     The encoding now uses one or more ':'-separated segments. Each
3339 :     segments corresponds to a file that has been specified relative to the
3340 :     file given by its preceding segment. The first segment is either
3341 :     relative to the CWD, absolute, or anchored. Each segment itself is
3342 :     basically a Unix pathname; all segments but the first are relative.
3343 :    
3344 :     Example:
3345 :    
3346 :     $foo/bar/baz.cm:a/b/c.sml
3347 :    
3348 :     This path denotes the file bar/a/b/c.sml relative to the directory
3349 :     denoted by anchor "foo". Notice that the encoding also includes
3350 :     baz.cm which is the .cm-file that listed a/b/c.sml. As usual, such
3351 :     paths are resolved relative to the .cm-files directory, so baz.cm must
3352 :     be ignored to get the "real" pathname.
3353 :    
3354 :     To make this fact more obvious, CM puts the names of such "virtual
3355 :     arcs" into parentheses when they appear in progress reports. (No
3356 :     parentheses will appear in the internal protocol encoding.) Thus,
3357 :     what you really see is:
3358 :    
3359 :     $foo/bar/(baz.cm):a/b/c.sml
3360 :    
3361 :     I find this notation to be much more informative than before.
3362 :    
3363 :     Another new feature of the encoding is that special characters
3364 :     including parentheses, colons, (back)slashes, and white space are
3365 :     written as \ddd (where ddd is the decimal encoding of the character).
3366 :    
3367 : blume 670 *** The CM manual still needs to be updated.
3368 :    
3369 : blume 666 ----------------------------------------------------------------------
3370 : leunga 660 Name: Allen Leung
3371 :     Date: 2000/06/15 00:38:00
3372 :     Tag: leunga-20000615-x86-peephole
3373 :    
3374 :     x86 Peephole fix by Fermin. Affects c-- and moby only.
3375 :    
3376 :     ----------------------------------------------------------------------
3377 : blume 658 Name: Matthias Blume
3378 :     Date: 2000/06/12 11:40:00
3379 :     Tag: blume-20000612-parmakefix
3380 :     Description:
3381 :    
3382 :     More cleanup after changing the file naming scheme: This time I
3383 :     repaired the parallel make mechanism for CMB.make which I broke earlier.
3384 :    
3385 :     ----------------------------------------------------------------------
3386 : leunga 657 Name: Allen Leung
3387 :     Date: 2000/06/09 01:25:00
3388 :     Tag: leunga-20000609-various
3389 :    
3390 :     None of these things should affect normal SML/NJ operations
3391 :    
3392 :     1. Peephole improvements provided by Fermin (c--)
3393 :     2. New annotation DEFUSE for adding extra dependence (moby)
3394 :     3. New X86 LOCK instructions (moby)
3395 :     4. New machine description language for reservation tables (scheduling)
3396 :     5. Fixes to various optimization/analysis modules (branch chaining, dominator
3397 :     trees etc.)
3398 :     6. I've changed the CM files so that they can work with versions
3399 :     110.0.6, 110.25 and 110.28
3400 :    
3401 :     ----------------------------------------------------------------------
3402 : blume 652 Name: Matthias Blume
3403 : blume 656 Date: 2000/06/09 12:40:00
3404 :     Tag: blume-20000609-log
3405 :     Description:
3406 :    
3407 :     - Removed all(?) remaining RCS Log entries from sources.
3408 :    
3409 :     - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
3410 :     anchored paths).
3411 :    
3412 :     ----------------------------------------------------------------------
3413 :     Name: Matthias Blume
3414 : blume 653 Date: 2000/06/07 17:00:00 JST
3415 :     Tag: blume-20000607-no-implicit-anchors
3416 :     Description:
3417 :    
3418 :     1. This update changes the default setting for
3419 :     CM.Control.implicit_anchors from true to false. This means that
3420 :     implicit anchors are no longer permitted by default. I also tried to
3421 :     make sure that nothing else still relies on implicit anchors.
3422 :     (This is the next step on the schedule towards a CM that does not even
3423 :     have the notion of implicit anchors anymore.)
3424 :    
3425 :     2. More CM manual updates.
3426 :    
3427 :     3. I managed to track down and fix the pickling bug I mentioned last
3428 :     time. Because of the previously existing workaround, this entails no
3429 :     immediate practical changes.
3430 :    
3431 :     ----------------------------------------------------------------------
3432 :     Name: Matthias Blume
3433 : blume 652 Date: 2000/06/06 11:15:00 JST
3434 :     Tag: blume-20000606-lazierpickle
3435 :     Description:
3436 :    
3437 :     !!!! NEW BOOT FILES !!!!
3438 :    
3439 :     * The main purpose of this update is to make library pickles lazier in
3440 :     order to reduce the initial space penalty for autoloading a library.
3441 :     As a result, it is now possible to have $smlnj/compiler.cm
3442 :     pre-registered. This should take care of the many complaints or
3443 :     inquiries about missing structure Compiler. This required changes to
3444 :     CM's internal data structures and small tweaks to some algorithms.
3445 :    
3446 :     As a neat additional effect, it is no longer necessary (for the sake
3447 :     of lean heap image files) to distinguish between a "minimal" CM and a
3448 :     "full" CM. Now, there is only one CM (i.e., the "full" version:
3449 :     $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
3450 :     interactive top level. ($smlnj/cm/minimal.cm is gone.)
3451 :    
3452 :     To make the life of compiler-hackers easier, "makeml" now also
3453 :     pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm). In other
3454 :     words, after you bootstrap a new sml for the first time, you will not
3455 :     have to autoload $smlnj/cmb.cm again afterwards. (The first time
3456 :     around you will still have to do it, though.)
3457 :    
3458 :     * A second change consists of major updates to the CM manual. There
3459 :     are now several appendices with summary information and also a full
3460 :     specification of the CM description file syntax.
3461 :    
3462 :     * In directory src/system I added the script "allcross". This script
3463 :     invokes sml and cross-compiles the compiler for all supported
3464 :     architectures. (Useful when providing a new set of boot files.)
3465 :    
3466 :     * There seems to be a latent bug in my "lazy pickles" mechanism. I
3467 :     added a small tweak to pickle-util.sml to work around this problem,
3468 :     but it is not a proper fix yet. I will investigate further. (The
3469 :     effect of the bug was an inflation of library pickle size.)
3470 :    
3471 :     * Version number increased to 110.28.1 (to avoid compatibility problems).
3472 :    
3473 :     ----------------------------------------------------------------------
3474 : leunga 646 Name: Allen Leung
3475 : leunga 648 Date: 2000/05/25 17:28 EDT
3476 :     Tag: leunga-20000525-ra
3477 :     Description:
3478 :    
3479 :     Fixed a bug in freezing phase of the register allocator.
3480 :    
3481 :     ----------------------------------------------------------------------
3482 :     Name: Allen Leung
3483 : leunga 647 Date: 2000/05/15 22:53 EDT
3484 : leunga 646 Tag: leunga-20000515-alpha-x86-ra
3485 :     Description:
3486 :    
3487 :     1. Alpha
3488 :    
3489 :     Slight cleanup. Removed the instruction SGNXL
3490 :    
3491 :     2. X86
3492 :    
3493 :     Added the following instructions to the instruction set:
3494 :    
3495 :     ROLx, RORx,
3496 :     BTx, BTSx, BTLx, BTRx,
3497 :     XCHGx, and variants with the LOCK prefix
3498 :    
3499 :     3. Register Allocation
3500 :    
3501 :     The module ra-rewrite-with-renaming has been improved.
3502 :    
3503 : leunga 647 These have no effect on SML/NJ.
3504 :    
3505 : leunga 646 ----------------------------------------------------------------------
3506 : blume 642 Name: Matthias Blume
3507 : blume 645 Date: 2000/05/15 16:20:00 JST
3508 :     Tag: blume-20000515-lightrebuild
3509 :     Description:
3510 :    
3511 :     1. I added an alternative to "-rebuild" to "makeml". The difference is
3512 :     that prior to calling CMB.make' the CM-variable "LIGHT" will be
3513 :     defined. In effect, the command will not build any cross-compiler
3514 :     backends and therefore finish more quickly.
3515 :    
3516 :     The "fixpt" script also takes a "-light" switch to be able to use
3517 :     this new facility while compiling for a fixpoint.
3518 :    
3519 :     2. I replaced all mentions of anchored paths in group owner specifications
3520 :     with simple relative paths (usually starting with "..").
3521 :     The rationale is that a library's internal workings should not be
3522 :     compromised by the lack of some anchor. (An anchor is necessary
3523 :     for someone who wants to refer to the library by an anchored path,
3524 :     but it should not be necessary to build the same library in the first
3525 :     place.)
3526 :    
3527 :     3. I changed the way CM's tool mechanism determines the shell command
3528 :     string used for things like ml-yacc etc. so that it does not break
3529 :     when CM.Control.implicit_anchors is turned off.
3530 :    
3531 :     ----------------------------------------------------------------------
3532 :     Name: Matthias Blume
3533 : blume 644 Date: 2000/05/12 18:20:00 JST
3534 :     Tag: blume-20000512-ml-build
3535 :     Description:
3536 :    
3537 :     Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
3538 :     getting installed properly (by config/install.sh).
3539 :    
3540 :     ----------------------------------------------------------------------
3541 :     Name: Matthias Blume
3542 : blume 643 Date: 2000/05/12 17:30:00 JST
3543 :     Tag: blume-20000512-anchors
3544 :     Description:
3545 :    
3546 :     !!! NEW BOOT FILES !!!
3547 :    
3548 :     This change is in preparation of fading out support for "implicitly
3549 :     anchored path names". I went through all sources and used the
3550 :     explicit (and relatively new) $-notation. See system/README and the
3551 :     CM manual for more info on this.
3552 :    
3553 :     I also modified the anchoring scheme for some things such as "smlnj",
3554 :     "MLRISC", "cm", etc. to take advantage of the fact that explicit
3555 :     anchors are more expressive: anchor name and first arc do not have to
3556 :     coincide. This entails the following user-visible change:
3557 :    
3558 :     You have to write $smlnj/foo/bar instead of smlnj/foo/bar. In
3559 :     particular, when you fire up sml with a command-line argument, say,
3560 :     e.g.:
3561 :    
3562 :     sml '$smlnj/cmb.cm'
3563 :    
3564 :     At the ML toplevel prompt:
3565 :    
3566 :     CM.autoload "$smlnj/cmb.cm";
3567 :    
3568 :     There is also a new controller in CM.Control that can be used to turn
3569 :     off all remaining support for implicit anchors by saying:
3570 :    
3571 :     CM.autoload "$smlnj/
3572 :     #set CM.Control.implicit_anchors false;
3573 :    
3574 :     This causes CM to reject implicitly anchored paths. This is (for the
3575 :     time being) less permissive than the "final" version where there will
3576 :     be no more such implicit anchors and relative paths will be just that:
3577 :     relative.
3578 :    
3579 :     The next step (version after next version?) will be to make the
3580 :     default for CM.Control.implicit_anchors false. After the dust has
3581 :     settled, I can then produce the "final" version of this...
3582 :    
3583 :     Note: Since bootstrapping is a bit tricky, I provided new boot files.
3584 :    
3585 :     ----------------------------------------------------------------------
3586 :     Name: Matthias Blume
3587 : blume 642 Date: 2000/05/11 16:30:00 JST
3588 :     Tag: blume-20000511-sources
3589 :     Description:
3590 :    
3591 :     The main change is that I added function CM.sources as a generalized
3592 :     version of the earlier CM.makedepend. This entails the following
3593 :     additional changes:
3594 :    
3595 :     - CM.makedepend has been dropped.
3596 :    
3597 :     - CM manual has been updated.
3598 :    
3599 :     - TOOLS signature and API have been changed.
3600 :    
3601 :     ----------------------------------------------------------------------
3602 : leunga 640 Name: Allen Leung
3603 : leunga 641 Date: 2000/05/10 21:17 EDT
3604 :     Tag: leunga-20000510-moby-c--ssa
3605 :     Description:
3606 :    
3607 :     Various bug fixes and new features for C--, Moby and MLRISC optimizations.
3608 :     None of these affect SML/NJ.
3609 :    
3610 :     1. Register Allocation
3611 :    
3612 :     a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
3613 :     This module tries to remove local (i.e. basic block level) redundancies
3614 :     during spilling.
3615 :    
3616 :     b. A new framework for performing region based register allocation.
3617 :     Not yet entirely functional.
3618 :    
3619 :     2. X86
3620 :    
3621 :     a. DefUse for POP was missing the stack pointer [found by Lal]
3622 :     b. Reload for CALL was incorrect in X86Spill [found by John]
3623 :     c. Various fixes in X86Spill so that it can be used correctly for
3624 :     the new spilling module.
3625 :    
3626 :     3. SSA/IR
3627 :    
3628 :     a. New module ir/dj-dataflow.sml implements elimination based
3629 :     data flow analysis.
3630 :    
3631 :     4. MLRiscGen
3632 :    
3633 :     a. Fix for gc type annotation
3634 :    
3635 :     5. MDGen
3636 :    
3637 :     Various fixes for machine description -> ml code translation. For ssa
3638 :     only.
3639 :    
3640 :     ----------------------------------------------------------------------
3641 :     Name: Allen Leung
3642 : leunga 640 Date: 2000/05/08 22:17 EDT
3643 :     Tag: leunga-20000508-labexp
3644 :     Description:
3645 :    
3646 :     Fermin has found a few assembly problems with constant expressions
3647 :     generated in LabelExp. Mostly, the problems involve extra parentheses,
3648 :     which choke on dumb assemblers. This is his fix.
3649 :    
3650 :     ----------------------------------------------------------------------
3651 : dbm 635 Name: Dave MacQueen
3652 :     Date: 2000/04/09 14:00 EDT
3653 :     Tag: dbm-20000502-Version_110_28
3654 :     Description:
3655 :    
3656 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
3657 :    
3658 :     2. Updated config/version to 110.28
3659 :    
3660 :     3. Updated config/srcarchiveurl
3661 :    
3662 :     3. New boot files!
3663 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
3664 :    
3665 :     ----------------------------------------------------------------------
3666 : blume 634 Name: Matthias Blume
3667 :     Date: 2000/05/01 19:05:00 JST
3668 :     Tag: blume-20000501-noweb
3669 :     Description:
3670 :    
3671 :     A new noweb tool has been added. The existing system is entirely
3672 :     unaffected by this, but some CM users have asked for renewed noweb
3673 :     support. Everything is documented in the CM manual.
3674 :    
3675 :     New (plugin) libraries:
3676 :    
3677 :     noweb-tool.cm
3678 :     nw-ext.cm
3679 :    
3680 :     ----------------------------------------------------------------------
3681 : dbm 633 Name: Dave MacQueen
3682 :     Date: 2000/04/30 12:40PM EDT
3683 :     Tag: dbm-20000430-bug_fixes
3684 :     Description:
3685 :    
3686 :     1. Fix for bug 1498
3687 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sig
3688 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sml
3689 :     added toRealArray function
3690 :     smlnj/src/compiler/MiscUtil/print/ppobj.sml
3691 :     added check for tag Obj.RealArray to array printing case in ppObj
3692 :    
3693 :     2. Fix for bug 1510
3694 :     smlnj/src/compiler/Semant/types/typesutil.sml
3695 :     fixed definition of dummyargs (used by equalTycon) so that
3696 :     dummy args are distinct types
3697 :    
3698 :     ----------------------------------------------------------------------
3699 : blume 629 Name: Matthias Blume
3700 : blume 632 Date: 2000/04/30 01:00:00 JST
3701 :     Tag: blume-20000430-versions
3702 :     Description:
3703 :    
3704 :     1. CM version numbering added. This is an implementation of Lal's
3705 :     proposal for adding version numbers and version checking to .cm
3706 :     files. Lal said that his proposal was just that -- a proposal.
3707 :     For the time being I went ahead and implemented it so that people
3708 :     can comment on it. Everything is completely backward-compatible
3709 :     (except for the stable library format, i.e., new bootfiles!).
3710 :    
3711 :     As usual, see the CM manual for details.
3712 :    
3713 :     2. An alternative syntax for anchored paths has been implemented.
3714 :     Dave has recently voiced the same concerns that I had when I did
3715 :     this, so there should be some support. My take is that eventually
3716 :     I will let support for the current syntax (where anchors are
3717 :     "implicit") fade out in favor of the new, explicit syntax.
3718 :     In order to be backward-compatible, both old and new syntax are
3719 :     currently supported.
3720 :    
3721 :     Again, see the CM manual for details.
3722 :    
3723 :     3. Parallel make is trying to be slightly smarter: When the master
3724 :     process finds a "bottleneck", i.e., when there is only one
3725 :     compilation unit that can be compiled and everybody else is
3726 :     waiting on it, then it will simply compile it directly instead
3727 :     of clumsily telling one of the slaves to do it.
3728 :    
3729 :     4. Support for "unsharing" added. This is necessary in order to be
3730 :     able to have two different versions of the same library running
3731 :     at the same time (e.g., for trying out a new MLRISC while still
3732 :     having the old MLRISC linked into the current compiler, etc.)
3733 :     See the CM manual.
3734 :    
3735 :     5. Simple "makedepend" functionality added for generating Makefile
3736 :     dependency information. (This is rather crude at the moment.
3737 :     Expect some changes here in the future.)
3738 :    
3739 :     6. ".fun" added as a recognized suffix for ML files. Also documented
3740 :     explicitly in the manual that the fallback behavior (unknown suffix
3741 :     -> ML file) is not an official feature!
3742 :    
3743 :     7. Small changes to the pickler for stable libraries.
3744 :    
3745 :     8. Several internal changes to CM (for cleanup/improvement).
3746 :    
3747 :    
3748 :     !!!! NEW BINFILES !!!!
3749 :    
3750 :     ----------------------------------------------------------------------
3751 :     Name: Matthias Blume
3752 : blume 631 Date: 2000/04/28 17:30:00 JST
3753 :     Tag: blume-20000428-pathconfig
3754 :     Description:
3755 :    
3756 :     1. I changed config/install.sh to remove duplicate entries from the
3757 :     lib/pathconfig file at the end. Moreover, the final version of
3758 :     lib/pathconfig is sorted alphabetically. The same (sorting) is done
3759 :     in src/system/installml.
3760 :    
3761 :     2. The config/install.sh script now consistently uses relative
3762 :     pathnames in lib/pathconfig whenever the anchor is in the lib
3763 :     directory. (So far this was true for the libraries that come
3764 :     pre-compiled and bundled as part of the bootfiles but not for
3765 :     libraries that are compiled by the script itself.)
3766 :    
3767 :     ----------------------------------------------------------------------
3768 :     Name: Matthias Blume
3769 : blume 629 Date: 2000/04/26 13:10:00 JST
3770 :     Tag: blume-20000426-fun_suffix
3771 :     Description:
3772 :    
3773 :     Added ".fun" as a recognized file name suffix (for ML code).
3774 :    
3775 :     ----------------------------------------------------------------------
3776 : leunga 628 Name: Allen Leung
3777 :     Date: 2000/04/25 17:00:00 EST
3778 :     Tag: leunga-20000425-alpha-ra
3779 :     Description:
3780 :    
3781 :     1. Alpha
3782 :    
3783 :     PSEUDOARITH was missing in AlphaRewrite. This causes an endless loop
3784 :     in C--.
3785 :    
3786 :     2. RA
3787 :    
3788 :     Added a flag "ra-dump-size" to print out the size of the flowgraph
3789 :     and the interference graph.
3790 :    
3791 :     ----------------------------------------------------------------------
3792 : dbm 627 Name: Dave MacQueen
3793 :     Date: 2000/04/25/
3794 :     Tag: dbm-20000425-mlyacc_doc_examples
3795 :     Description:
3796 :     Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
3797 :     Updated all three examples in src/ml-yacc/examples to run
3798 :     under 110.* using CM.make.
3799 :    
3800 :     ----------------------------------------------------------------------
3801 : leunga 624 Name: Allen Leung
3802 :     Date: 2000/04/20 23:04:00 EST
3803 :     Tag: leunga-20000420-ssa-c---stuff
3804 :     Description:
3805 :    
3806 :     This update synchronizes my repository with Yale's. Most of these
3807 :     changes, however, do not affect SML/NJ at all (the RA is an exception).
3808 :    
3809 :     1. Register Allocator
3810 :    
3811 :     a. An improvement in the interference graph construction:
3812 :     Given a copy
3813 :    
3814 :     s <- t
3815 :    
3816 : leunga 628 no interference edge between s and t is added for this definition of s.
3817 : leunga 624
3818 :     b. I've added two new spill heuristic modules that Fermin and I developed
3819 :     (in the new library RA.cm). These are unused in SML/NJ but maybe
3820 :     useful for others (Moby?)
3821 :    
3822 :     2. X86
3823 :    
3824 :     a. Various fixes in the backend provided by Fermin [C--] and Lal.
3825 :    
3826 :     3. Alpha
3827 :    
3828 :     a. Added the BSR instruction and code generation that goes with it [C--]
3829 :     b. Other fixes too numerous to recount provided by Fermin [C--]
3830 :    
3831 :     4. Regmaps
3832 :    
3833 :     a. The regmaps are not initialized with the identity physical bindings
3834 :     at creation time. This is unneeded.
3835 :    
3836 :     5. MLRISC Optimizations
3837 :    
3838 :     a. The DJ-Graph module can now compute the iterated dominance frontiers
3839 :     intersects with liveness incrementally in linear time! Woohoo!
3840 :     This is now used in my new SSA construction algorithm.
3841 :    
3842 :     b. THe branch reorganization module is now smarter about linear chains of
3843 :     basic blocks.
3844 :    
3845 :    
3846 :     ----------------------------------------------------------------------
3847 : blume 615 Name: Matthias Blume
3848 :     Date: 2000/04/12 13:52:00 JST
3849 :     Tag: blume_main_v110p27_1
3850 :     Description:
3851 :    
3852 :     Changed install.sh script to handle archive files without version number
3853 :     and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
3854 :     name of the boot file archive.
3855 :    
3856 :     ----------------------------------------------------------------------
3857 : dbm 607 Name: Dave MacQueen
3858 :     Date: 2000/04/09 14:00 EDT
3859 :     Tag: dbm-20000410-Version_110_27
3860 :     Description:
3861 :    
3862 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
3863 :    
3864 :     2. Updated src/config/version to 110.27
3865 :    
3866 :     3. New boot files!
3867 :    
3868 :     ----------------------------------------------------------------------
3869 : leunga 606 Name: Allen Leung
3870 :     Date: 2000/04/09 19:09:00 EST
3871 :     Tag: leunga-20000409-misc
3872 :     Description:
3873 :    
3874 :     1. Yet another fix for x86 assembly for idivl, imull, mull and friends.
3875 :    
3876 :     2. Miscellaneous improvements to MLRISC (unused in sml/nj)
3877 :    
3878 :     ----------------------------------------------------------------------
3879 : monnier 605 Name: Stefan
3880 :     Date: 2000/04/07 10:00:00 EDT
3881 :     Tag: monnier-20000406-branch-handling
3882 :     Description:
3883 :    
3884 :     Improved handling of branches (mostly those generated from
3885 :     polymorphic equality), removed switchoff and changed the
3886 :     default optimization settings (more cpsopt and less flintopt).
3887 :    
3888 :     ----------------------------------------------------------------------
3889 : leunga 601 Name: Allen Leung
3890 : leunga 602 Date: 2000/04/06 01:30:00 EST
3891 :     Tag: leunga-20000406-peephole-x86-SSA-2
3892 :     Description:
3893 :    
3894 :     Forgot a few files.
3895 :    
3896 :     ----------------------------------------------------------------------
3897 :     Name: Allen Leung
3898 : leunga 601 Date: 2000/04/06 00:36:00 EST
3899 :     Tag: leunga-20000406-peephole-x86-SSA
3900 :     Description:
3901 :    
3902 :     1. New Peephole code
3903 :    
3904 :     2. Minor improvement to X86 instruction selection
3905 :    
3906 :     3. Various fixes to SSA and machine description -> code translator
3907 :    
3908 :     ----------------------------------------------------------------------
3909 : blume 595 Name: Matthias Blume
3910 :     Date: 2000/04/05 12:30:00 JST
3911 :     Tag: blume_main_v110p26p2_3
3912 :     Description:
3913 :    
3914 :     This update just merges three minor cosmetic updates to CM's sources
3915 :     to get ready for the 110.27 code freeze on Friday. No functionality
3916 :     has changed.
3917 :    
3918 :     ----------------------------------------------------------------------
3919 : leunga 593 Name: Allen Leung
3920 : leunga 594 Date: 2000/04/04 19:39:00 EST
3921 :     Tag: leunga-20000404-x86-asm
3922 :     Description:
3923 :    
3924 :     1. Fixed a problem in X86 assembly.
3925 :    
3926 :     Things like
3927 :    
3928 :     jmp %eax
3929 :     jmp (%eax)
3930 :    
3931 :     should be output as
3932 :    
3933 :     jmp *%eax
3934 :     jmp *(%eax)
3935 :    
3936 :     2. Assembly output
3937 :    
3938 :     Added a new flag
3939 :    
3940 :     "asm-indent-copies" (default to false)
3941 :    
3942 :     When this flag is on, parallel copies will be indented an extra level.
3943 :    
3944 :     ----------------------------------------------------------------------
3945 :     Name: Allen Leung
3946 : leunga 593 Date: 2000/04/04 03:18:00 EST
3947 :     Tag: leunga-20000404-C--Moby
3948 :     Description:
3949 :    
3950 :     All of these fixes are related to C--, Moby, and my own optimization
3951 :     stuff; so they shouldn't affect SML/NJ.
3952 :    
3953 :     1. X86
3954 :    
3955 :     Various fixes related floating point, and extensions.
3956 :    
3957 :     2. Alpha
3958 :    
3959 :     Some extra patterns related to loads with signed/zero extension
3960 :     provided by Fermin.
3961 :    
3962 :     3. Assembly
3963 :    
3964 : monnier 659 When generating assembly, resolve the value of client defined constants,
3965 : leunga 593 instead of generating symbolic values. This is controlled by the
3966 :     new flag "asm-resolve-constants", which is default to true.
3967 :    
3968 :     4. Machine Descriptions
3969 :    
3970 :     a. The precedence parser was slightly broken when parsing infixr symbols.
3971 :     b. The type generalizing code had the bound variables reversed, resulting
3972 :     in a problem during arity raising.
3973 :     c. Various fixes in machine descriptions.
3974 :    
3975 :     ----------------------------------------------------------------------
3976 : blume 592 Name: Matthias Blume
3977 :     Date: 2000/04/03 16:05:00 JST
3978 :     Tag: blume_main_v110p26p2_2
3979 :     Description:
3980 :    
3981 :     I eliminated coreEnv from compInfo. Access to the "Core" structure is
3982 :     now done via the ordinary static environment that is context to each
3983 :     compilation unit.
3984 :    
3985 :     To this end, I arranged that instead of "structure Core" as "structure
3986 :     _Core" is bound in the pervasive environment. Core access is done via
3987 : monnier 659 _Core (which can never be accidentally rebound because _Core is not a
3988 : blume 592 legal surface-syntax symbol).
3989 :    
3990 :     The current solution is much cleaner because the core environment is
3991 :     now simply part of the pervasive environment which is part of every
3992 :     compilation unit's context anyway. In particular, this eliminates all
3993 :     special-case handling that was necessary until now in order to deal
3994 :     with dynamic and symbolic parts of the core environment.
3995 :    
3996 :     Remaining hackery (to bind the "magic" symbol _Core) is localized in the
3997 : monnier 659 compilation manager's bootstrap compiler (actually: in the "init group"
3998 : blume 592 handling). See the comments in src/system/smlnj/init/init.cmi for
3999 :     more details.
4000 :    
4001 :     I also tried to track down all mentions of "Core" (as string argument
4002 :     to Symbol.strSymbol) in the compiler and replaced them with a
4003 :     reference to the new CoreSym.coreSym. Seems cleaner since the actual
4004 :     name appears in one place only.
4005 :    
4006 :     Binfile and bootfile format have not changed, but the switchover from
4007 :     the old "init.cmi" to the new one is a bit tricky, so I supplied new
4008 :     bootfiles anyway.
4009 :    
4010 :     ----------------------------------------------------------------------
4011 : leunga 591 Name: Allen Leung
4012 :     Date: 2000/04/02 21:17:00 EST
4013 :     Tag: leunga-20000402-mltree
4014 :     Description:
4015 : leunga 585
4016 : leunga 591 1. Renamed the constructor CALL in MLTREE by popular demand.
4017 :     2. Added a bunch of files from my repository. These are currently
4018 :     used by other non-SMLNJ backends.
4019 :    
4020 : leunga 576 ----------------------------------------------------------------------
4021 : leunga 590 Name: Allen Leung
4022 :     Date: 2000/03/31 21:15:00 EST
4023 :     Tag: leunga-20000331-aliasing
4024 :     Description:
4025 :    
4026 :     This update contains a rewritten (and hopefully more correct) module
4027 :     for extracting aliasing information from CPS.
4028 :    
4029 :     To turn on this feature:
4030 :    
4031 :     Compiler.Control.CG.memDisambiguate := true
4032 :    
4033 :     To pretty print the region information with assembly
4034 :    
4035 :     Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
4036 :    
4037 :     To control how many levels of aliasing information are printed, use:
4038 :    
4039 :     Compiler.Control.MLRISC.getInt "points-to-show-level" := n
4040 :    
4041 :     The default of n is 3.
4042 :    
4043 :     ----------------------------------------------------------------------
4044 : dbm 589 Name: David MacQueen
4045 :     Date: 2000/03/31 11:15:00 EST
4046 :     Tag: dbm-20000331-runtime_fix
4047 :     Description:
4048 :    
4049 :     This update contains:
4050 :    
4051 :     1. runtime/c-lib/c-libraries.c
4052 :     includes added in revision 1.2 caused compilation errors on hppa-hpux
4053 :    
4054 :     2. fix for bug 1556
4055 :     system/Basis/Implementation/NJ/internal-signals.sml
4056 :    
4057 :     ----------------------------------------------------------------------
4058 : blume 587 Name: Matthias Blume
4059 : blume 588 Date: 2000/03/31 18:00:00 JST
4060 :     Tag: blume_main_v110p26p2_1
4061 :     Description:
4062 :    
4063 :     This update contains:
4064 :    
4065 :     1. A small change to CM's handling of stable libraries:
4066 :     CM now maintains one "global" modmap that is used for all stable
4067 :     libraries. The use of such a global modmap maximizes sharing and
4068 :     minimizes the need for re-traversing parts of environments during
4069 :     modmap construction. (However, this has minor impact since modmap
4070 :     construction seems to account for just one percent or less of total
4071 :     compile time.)
4072 :    
4073 :     2. I added a "genmap" phase to the statistics. This is where I got the
4074 :     "one percent" number (see above).
4075 :    
4076 :     3. CM's new tool parameter mechanism just became _even_ better. :)
4077 :     - The parser understands named parameters and recursive options.
4078 :     - The "make" and "shell" tools use these new features.
4079 :     (This makes it a lot easier to cascade these tools.)
4080 :     - There is a small syntax change: named parameters use a
4081 :    
4082 :     <name> : ( <option> ... ) or
4083 :     <name> : <string>
4084 :    
4085 :     syntax. Previously, named parameters were implemented in an
4086 :     ad-hoc fashion by each tool individually (by parsing strings)
4087 :     and had the form
4088 :    
4089 :     <name>=<string>
4090 :    
4091 :     See the CM manual for a full description of these issues.
4092 :    
4093 :     ----------------------------------------------------------------------
4094 :     Name: Matthias Blume
4095 : blume 587 Date: 2000/03/30 18:00:00 JST
4096 :     Tag: blume_main_v110p26p2_0
4097 :     Description:
4098 :    
4099 :     !!!!! WARNING !!!!!!
4100 :     !! New binfiles !!
4101 :     !!!!!!!!!!!!!!!!!!!!
4102 :    
4103 :     This update contains:
4104 :    
4105 :     1. Moderate changes to CM:
4106 :    
4107 :     - Changes to CM's tools mechanism. In particular, it is now possible
4108 :     to have tools that accept additional "command line" parameters
4109 :     (specified in the .cm file at each instance where the tool's class is
4110 :     used).
4111 :    
4112 : monnier 659 This was done to accommodate the new "make" and "shell" tools which
4113 :     facilitate fairly seamless hookup to portions of code managed using
4114 : blume 587 Makefiles or Shell scripts.
4115 :    
4116 :     There are no classes "shared" or "private" anymore. Instead, the
4117 :     sharing annotation is now a parameter to the "sml" class.
4118 :    
4119 :     There is a bit of generic machinery for implementing one's own
4120 :     tools that accept command-line parameters. However, I am not yet fully
4121 :     satisfied with that part, so expect changes here in the future.
4122 :    
4123 :     All existing tools are described in the CM manual.
4124 :    
4125 : monnier 659 - Slightly better error handling. (CM now suppresses many followup
4126 : blume 587 error messages that tended to be more annoying than helpful.)
4127 :    
4128 :     2. Major changes to the compiler's static environment data structures.
4129 :    
4130 :     - no CMStaticEnv anymore.
4131 :     - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
4132 :     but it is called Environment), no conversions between different
4133 :     kinds of static environments
4134 :    
4135 :     - There is still a notion of a "modmap", but such modmaps are generated
4136 :     on demand at the time when they are needed. This sounds slow, but I
4137 :     sped up the code that generates modmaps enough for this not to lead to
4138 :     a slowdown of the compiler (at least I didn't detect any).
4139 :    
4140 :     - To facilitate rapid modmap generation, static environments now
4141 :     contain an (optional) "modtree" structure. Modtree annotations are
4142 :     constructed by the unpickler during unpickling. (This means that
4143 :     the elaborator does not have to worry about modtrees at all.)
4144 :     Modtrees have the advantage that they are compositional in the same
4145 :     way as the environment data structure itself is compositional.
4146 :     As a result, modtrees never hang on to parts of an environment that
4147 :     has already been rendered "stale" by filtering or rebinding.
4148 :    
4149 :     - I went through many, many trials and errors before arriving at the
4150 :     current solution. (The initial idea of "linkpaths" did not work.)
4151 :     But the result of all this is that I have touched a lot of files that
4152 :     depend on the "modules" and "types" data structures (most of the
4153 :     elaborator). There were a lot of changes during my "linkpath" trials
4154 :     that could have been reverted to their original state but weren't.
4155 :     Please, don't be too harsh on me for messing with this code a bit more
4156 : monnier 921 than what was strictly necessary... (I _did_ resist the temptation
4157 : blume 587 of doing any "global reformatting" to avoid an untimely death at
4158 :     Dave's hands. :)
4159 :    
4160 :     - One positive aspect of the previous point: At least I made sure that
4161 :     all files that I touched now compile without warnings (other than
4162 :     "polyEqual").
4163 :    
4164 :     - compiler now tends to run "leaner" (i.e., ties up less memory in
4165 :     redundant modmaps)
4166 :    
4167 :     ----------------------------------------------------------------------
4168 : leunga 580 Name: Allen Leung
4169 : leunga 585 Date: 2000/03/29 18:00:00
4170 :     Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
4171 : leunga 586 Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
4172 : leunga 585 Description:
4173 :    
4174 :     This update contains *MAJOR* changes to the way code is generated from CPS
4175 :     in the module mlriscGen, and in various backend modules.
4176 :    
4177 :     CHANGES
4178 :     =======
4179 :    
4180 :     1. MLRiscGen: forward propagation fix.
4181 :    
4182 :     There was a bug in forward propagation introduced at about the same time
4183 :     as the MLRISC x86 backend, which prohibits coalescing to be
4184 :     performed effectively in loops.
4185 :    
4186 :     Effect: speed up of loops in RISC architectures.
4187 :     By itself, this actually slowed down certain benchmarks on the x86.
4188 :    
4189 :     2. MLRiscGen: forward propagating addresses from consing.
4190 :    
4191 :     I've changed the way consing code is generated. Basically I separated
4192 :     out the initialization part:
4193 :    
4194 :     store tag, offset(allocptr)
4195 :     store elem1, offset+4(allocptr)
4196 :     store elem2, offset+8(allocptr)
4197 :     ...
4198 :     store elemn, offset+4n(allocptr)
4199 :    
4200 :     and the address computation part:
4201 :    
4202 :     celladdr <- offset+4+alloctpr
4203 :    
4204 :     and move the address computation part
4205 :    
4206 :     Effect: register pressure is generally lower as a result. This
4207 :     makes compilation of certain expressions much faster, such as
4208 :     long lists with non-trivial elements.
4209 :    
4210 :     [(0,0), (0,0), .... (0,0)]
4211 :    
4212 :     3. MLRiscGen: base pointer elimination.
4213 :    
4214 :     As part of the linkage mechanism, we generate the sequence:
4215 :    
4216 :     L: ... <- start of the code fragment
4217 :    
4218 :     L1:
4219 :     base pointer <- linkreg - L1 + L
4220 :    
4221 :     The base pointer was then used for computing relocatable addresses
4222 :     in the code fragment. Frequently (such as in lots of continuations)
4223 :     this is not needed. We now eliminate this sequence whenever possible.
4224 :    
4225 :     For compile time efficiency, I'm using a very stupid local heuristic.
4226 :     But in general, this should be done as a control flow analysis.
4227 :    
4228 :     Effect: Smaller code size. Speed up of most programs.
4229 :    
4230 :     4. Hppa back end
4231 :    
4232 :     Long jumps in span dependence resolution used to depend on the existence
4233 :     of the base pointer.
4234 :    
4235 :     A jump to a long label L was expanded into the following sequence:
4236 :    
4237 :     LDIL %hi(L-8192), %r29
4238 :     LDO %lo(L-8192)(%r29), %r29
4239 :     ADD %r29, baseptr, %r29
4240 :     BV,n %r0(%r29)
4241 :    
4242 :     In the presence of change (3) above, this will not work. I've changed
4243 :     it so that the following sequence of instructions are generated, which
4244 :     doesn't mention the base pointer at all:
4245 :    
4246 :     BL,n L', %r29 /* branch and link, L' + 4 -> %r29 */
4247 :     L': ADDIL L-(L'+4), %r29 /* Compute address of L */
4248 :     BV,n %r0(%r29) /* Jump */
4249 :    
4250 :     5. Alpha back end
4251 :    
4252 :     New alpha instructions LDB/LDW have been added, as per Fermin's
4253 :     suggestions. This is unrelated to all other changes.
4254 :    
4255 :     6. X86 back end
4256 :    
4257 :     I've changed andl to testl in the floating point test sequence
4258 :     whenever appropriate. The Intel optimization guide states that
4259 : monnier 659 testl is preferable to andl.
4260 : leunga 585
4261 :     7. RA (x86 only)
4262 :    
4263 :     I've improved the spill propagation algorithm, using an approximation
4264 :     of maximal weighted independent sets. This seems to be necessary to
4265 :     alleviate the negative effect in light of the slow down in (1).
4266 :    
4267 :     I'll write down the algorithm one of these days.
4268 :    
4269 :     8. MLRiscGen: frequencies
4270 :    
4271 :     I've added an annotation that states that all call gc blocks have zero
4272 :     execution frequencies. This improves register allocation on the x86.
4273 :    
4274 :     BENCHMARKS
4275 :     ==========
4276 :    
4277 :     I've only perform the comparison on 110.25.
4278 :    
4279 :     The platforms are:
4280 :    
4281 :     HPPA A four processor HP machine (E9000) with 5G of memory.
4282 :     X86 A 300Hhz Pentium II with 128M of memory, and
4283 :     SPARC An Ultra sparc 2 with 512M of memory.
4284 :    
4285 :     I used the following parameters for the SML benchmarks:
4286 :    
4287 :     @SMLalloc
4288 :     HPPA 256k
4289 :     SPARC 512k
4290 :     X86 256k
4291 :    
4292 :     COMPILATION TIME
4293 :     ----------------
4294 :     Here are the numbers comparing the compilation times of the compilers.
4295 :     I've only compared 110.25 compiling the new sources versus
4296 :     a fixpoint version of the new compiler compiling the same.
4297 :    
4298 :     110.25 New
4299 :     Total Time in RA Spill+Reload Total Time In RA Spill+Reload
4300 :     HPPA 627s 116s 2684+3584 599s 95s 1003+1879
4301 :     SPARC 892s 173s 2891+3870 708s 116s 1004+1880
4302 :     X86 999s 315s 94006+130691 987s 296s 108877+141957
4303 :    
4304 :     110.25 New
4305 :     Code Size Code Size
4306 :     HPPA 8596736 8561421
4307 :     SPARC 8974299 8785143
4308 :     X86 9029180 8716783
4309 :    
4310 :     So in summary, things are at least as good as before. Dramatic
4311 :     reduction in compilation is obtained on the Sparc; I can't explain it,
4312 :     but it is reproducible. Perhaps someone should try to reproduce this
4313 :     on their own machines.
4314 :    
4315 :     SML BENCHMARKS
4316 :     --------------
4317 :    
4318 :     On the average, all benchmarks perform at least as well as before.
4319 :    
4320 :     HPPA Compilation Time Spill+Reload Run Time
4321 :     110.25 New 110.25 New 110.25 New
4322 :    
4323 :     barnesHut 3.158 3.015 4.75% 1+1 0+0 2.980 2.922 2.00%
4324 :     boyer 6.152 5.708 7.77% 0+0 0+0 0.218 0.213 2.34%
4325 :     count-graphs 1.168 1.120 4.32% 0+0 0+0 22.705 23.073 -1.60%
4326 :     fft 0.877 0.792 10.74% 1+3 1+3 0.602 0.587 2.56%
4327 :     knuthBendix 3.180 2.857 11.32% 0+0 0+0 0.675 0.662 2.02%
4328 :     lexgen 6.190 5.290 17.01% 0+0 0+0 0.913 0.788 15.86%
4329 :     life 0.803 0.703 14.22% 25+25 0+0 0.153 0.140 9.52%
4330 :     logic 2.048 2.007 2.08% 6+6 1+1 4.133 4.008 3.12%
4331 :     mandelbrot 0.077 0.080 -4.17% 0+0 0+0 0.765 0.712 7.49%
4332 :     mlyacc 22.932 20.937 9.53% 154+181 32+57 0.468 0.430 8.91%
4333 :     nucleic 5.183 5.060 2.44% 2+2 0+0 0.125 0.120 4.17%
4334 :     ratio-regions 3.357 3.142 6.84% 0+0 0+0 116.225 113.173 2.70%
4335 :     ray 1.283 1.290 -0.52% 0+0 0+0 2.887 2.855 1.11%
4336 :     simple 6.307 6.032 4.56% 28+30 5+7 3.705 3.658 1.28%
4337 :     tsp 0.888 0.862 3.09% 0+0 0+0 7.040 6.893 2.13%
4338 :     vliw 24.378 23.455 3.94% 106+127 25+45 2.758 2.707 1.91%
4339 :     --------------------------------------------------------------------------
4340 :     Average 6.12% 4.09%
4341 :    
4342 :     SPARC Compilation Time Spill+Reload Run Time
4343 :     110.25 New 110.25 New 110.25 New
4344 :    
4345 :     barnesHut 3.778 3.592 5.20% 2+2 0+0 3.648 3.453 5.65%
4346 :     boyer 6.632 6.110 8.54% 0+0 0+0 0.258 0.242 6.90%
4347 :     count-graphs 1.435 1.325 8.30% 0+0 0+0 33.672 34.737 -3.07%
4348 :     fft 0.980 0.940 4.26% 3+9 2+6 0.838 0.827 1.41%
4349 :     knuthBendix 3.590 3.138 14.39% 0+0 0+0 0.962 0.967 -0.52%
4350 :     lexgen 6.593 6.072 8.59% 1+1 0+0 1.077 1.078 -0.15%