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 1128 - (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 : blume 1128 Name: Matthias Blume
17 :     Date: 2002/03/08 16:05:00
18 :     Tag: blume-20020308-entrypoints
19 :     Description:
20 :    
21 :     Version number bumped to 110.39.1. NEW BOOTFILES!
22 :    
23 :     Entrypoints: non-zero offset into a code object where execution should begin.
24 :    
25 :     - Added the notion of an entrypoint to CodeObj.
26 :     - Added reading/writing of entrypoint info to Binfile.
27 :     - Made runtime system bootloader aware of entrypoints.
28 :     - Use the address of the label of the first function given to mlriscGen
29 :     as the entrypoint. This address is currently always 0, but it will
30 :     not be 0 once we turn on block placement.
31 :     - Removed the linkage cluster code (which was The Other Way(tm) of dealing
32 :     with entry points) from mlriscGen.
33 :    
34 :     ----------------------------------------------------------------------
35 : leunga 1127 Name: Allen Leung
36 :     Date: 2002/03/07 20:45:00 EST
37 :     Tag: leunga-20020307-x86-cmov
38 :     Description:
39 :    
40 :     Bug fixes for CMOVcc on x86.
41 :    
42 :     1. Added machine code generation for CMOVcc
43 :     2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
44 :     3. CMOVcc cannot have an immediate operand as argument.
45 :    
46 :     ----------------------------------------------------------------------
47 : blume 1126 Name: Matthias Blume
48 :     Date: 2002/03/07 16:15:00 EST
49 :     Tag: blume-20020307-controls
50 :     Description:
51 :    
52 :     This is a very large but mostly boring patch which makes (almost)
53 :     every tuneable compiler knob (i.e., pretty much everything under
54 :     Control.* plus a few other things) configurable via both the command
55 :     line and environment variables in the style CM did its configuration
56 :     until now.
57 :    
58 :     Try starting sml with '-h' (or, if you are brave, '-H')
59 :    
60 :     To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
61 :     implements the underlying generic mechanism.
62 :    
63 :     The interface to some of the existing such facilities has changed somewhat.
64 :     For example, the MLRiscControl module now provides mkFoo instead of getFoo.
65 :     (The getFoo interface is still there for backward-compatibility, but its
66 :     use is deprecated.)
67 :    
68 :     The ml-build script passes -Cxxx=yyy command-line arguments through so
69 :     that one can now twiddle the compiler settings when using this "batch"
70 :     compiler.
71 :    
72 :     TODO items:
73 :    
74 :     We should go through and throw out all controls that are no longer
75 :     connected to anything. Moreover, we should go through and provide
76 :     meaningful (and correct!) documentation strings for those controls
77 :     that still are connected.
78 :    
79 :     Currently, multiple calls to Controls.new are accepted (only the first
80 :     has any effect). Eventually we should make sure that every control
81 :     is being made (via Controls.new) exactly once. Future access can then
82 :     be done using Controls.acc.
83 :    
84 :     Finally, it would probably be a good idea to use the getter-setter
85 :     interface to controls rather than ref cells. For the time being, both
86 :     styles are provided by the Controls module, but getter-setter pairs are
87 :     better if thread-safety is of any concern because they can be wrapped.
88 :    
89 :     *****************************************
90 :    
91 :     One bug fix: The function blockPlacement in three of the MLRISC
92 :     backpatch files used to be hard-wired to one of two possibilities at
93 :     link time (according to the value of the placementFlag). But (I
94 :     think) it should rather sense the flag every time.
95 :    
96 :     *****************************************
97 :    
98 :     Other assorted changes (by other people who did not supply a HISTORY entry):
99 :    
100 :     1. the cross-module inliner now works much better (Monnier)
101 :     2. representation of weights, frequencies, and probabilities in MLRISC
102 :     changed in preparation of using those for weighted block placement
103 :     (Reppy, George)
104 :    
105 :     ----------------------------------------------------------------------
106 : george 1116 Name: Lal George
107 : george 1124 Date: 2002/03/07 14:44:24 EST 2002
108 :     Tag: george-20020307-weighted-block-placement
109 :    
110 :     Tested the weighted block placement optimization on all architectures
111 :     (except the hppa) using AMPL to generate the block and edge frequencies.
112 :     Changes were required in the machine properties to correctly
113 :     categorize trap instructions. There is an MLRISC flag
114 :     "weighted-block-placement" that can be used to enable weighted block
115 :     placement, but this will be ineffective without block/edge
116 :     frequencies (coming soon).
117 :    
118 :    
119 :     ----------------------------------------------------------------------
120 :     Name: Lal George
121 : george 1116 Date: 2002/03/05 17:24:48 EST
122 :     Tag: george-20020305-linkage-cluster
123 :    
124 : george 1124 In order to support the block placement optimization, a new cluster
125 :     is generated as the very first cluster (called the linkage cluster).
126 :     It contains a single jump to the 'real' entry point for the compilation
127 :     unit. Block placement has no effect on the linkage cluster itself, but
128 :     all the other clusters have full freedom in the manner in which they
129 :     reorder blocks or functions.
130 : george 1116
131 :     On the x86 the typical linkage code that is generated is:
132 :     ----------------------
133 :     .align 2
134 :     L0:
135 :     addl $L1-L0, 72(%esp)
136 : george 1124 jmp L1
137 : george 1116
138 :    
139 :     .align 2
140 :     L1:
141 :     ----------------------
142 :    
143 :     72(%esp) is the memory location for the stdlink register. This
144 :     must contain the address of the CPS function being called. In the
145 : george 1124 above example, it contains the address of L0; before
146 : george 1116 calling L1 (the real entry point for the compilation unit), it
147 :     must contain the address for L1, and hence
148 :    
149 :     addl $L1-L0, 72(%esp)
150 :    
151 : george 1124 I have tested this on all architectures except the hppa.The increase
152 :     in code size is of course negligible
153 : george 1116
154 :     ----------------------------------------------------------------------
155 : leunga 1115 Name: Allen Leung
156 :     Date: 2002/03/03 13:20:00 EST
157 :     Tag: leunga-20020303-mlrisc-tools
158 :    
159 :     Added #[ ... ] expressions to mlrisc tools
160 :    
161 :     ----------------------------------------------------------------------
162 : blume 1096 Name: Matthias Blume
163 : blume 1098 Date: 2002/02/27 12:29:00 EST
164 :     Tag: blume-20020227-cdebug
165 :     Description:
166 :    
167 :     - made types in structure C and C_Debug to be equal
168 :     - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
169 :     - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
170 :    
171 :     ----------------------------------------------------------------------
172 :     Name: Matthias Blume
173 : blume 1096 Date: 2002/02/26 12:00:00 EST
174 :     Tag: blume-20020226-ffi
175 :     Description:
176 :    
177 :     1. Fixed a minor bug in CM's "noweb" tool:
178 :     If numbering is turned off, then truly don't number (i.e., do not
179 :     supply the -L option to noweb). The previous behavior was to supply
180 :     -L'' -- which caused noweb to use the "default" line numbering scheme.
181 :     Thanks to Chris Richards for pointing this out (and supplying the fix).
182 :    
183 :     2. Once again, I reworked some aspects of the FFI:
184 :    
185 :     A. The incomplete/complete type business:
186 :    
187 :     - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
188 :     gone!
189 :     - ML types representing an incomplete type are now *equal* to
190 :     ML types representing their corresponding complete types (just like
191 :     in C). This is still safe because ml-nlffigen will not generate
192 :     RTTI for incomplete types, nor will it generate functions that
193 :     require access to such RTTI. But when ML code generated from both
194 :     incomplete and complete versions of the C type meet, the ML types
195 :     are trivially interoperable.
196 :    
197 :     NOTE: These changes restore the full generality of the translation
198 :     (which was previously lost when I eliminated functorization)!
199 :    
200 :     B. Enum types:
201 :    
202 :     - Structure C now has a type constructor "enum" that is similar to
203 :     how the "su" constructor works. However, "enum" is not a phantom
204 :     type because each "T enum" has values (and is isomorphic to
205 :     MLRep.Signed.int).
206 :     - There are generic access operations for enum objects (using
207 :     MLRep.Signed.int).
208 :     - ml-nlffigen will generate a structure E_foo for each "enum foo".
209 :     * The structure contains the definition of type "mlrep" (the ML-side
210 :     representation type of the enum). Normally, mlrep is the same
211 :     as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
212 :     then mlrep will be defined as a datatype -- thus facilitating
213 :     pattern matching on mlrep values.
214 :     ("-ec" will be suppressed if there are duplicate values in an
215 :     enumeration.)
216 :     * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
217 :     will be generated for each C enum constant xxx.
218 :     * Conversion functions m2i and i2m convert between mlrep and
219 :     MLRep.Signed.int. (Without "-ec", these functions are identities.)
220 :     * Coversion functions c and ml convert between mlrep and "tag enum".
221 :     * Access functions (get/set) fetch and store mlrep values.
222 :     - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
223 :     enumerations are merged into one single enumeration represented by
224 :     structure E_'.
225 :    
226 :     ----------------------------------------------------------------------
227 : leunga 1080 Name: Allen Leung
228 : leunga 1094 Date: 2002/02/25 04:45:00 EST
229 :     Tag: leunga-20020225-cps-spill
230 :    
231 :     This is a new implementation of the CPS spill phase.
232 :     The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
233 :     In case of problems, replace it with the old file spill.sml
234 :    
235 :     The current compiler runs into some serious performance problems when
236 :     constructing a large record. This can happen when we try to compile a
237 :     structure with many items. Even a very simple structure like the following
238 :     makes the compiler slow down.
239 :    
240 :     structure Foo = struct
241 :     val x_1 = 0w1 : Word32.int
242 :     val x_2 = 0w2 : Word32.int
243 :     val x_3 = 0w3 : Word32.int
244 :     ...
245 :     val x_N = 0wN : Word32.int
246 :     end
247 :    
248 :     The following table shows the compile time, from N=1000 to N=4000,
249 :     with the old compiler:
250 :    
251 :     N
252 :     1000 CPS 100 spill 0.04u 0.00s 0.00g
253 :     MLRISC ra 0.06u 0.00s 0.05g
254 :     (spills = 0 reloads = 0)
255 :     TOTAL 0.63u 0.07s 0.21g
256 :    
257 :     1100 CPS 100 spill 8.25u 0.32s 0.64g
258 :     MLRISC ra 5.68u 0.59s 3.93g
259 :     (spills = 0 reloads = 0)
260 :     TOTAL 14.71u 0.99s 4.81g
261 :    
262 :     1500 CPS 100 spill 58.55u 2.34s 1.74g
263 :     MLRISC ra 5.54u 0.65s 3.91g
264 :     (spills = 543 reloads = 1082)
265 :     TOTAL 65.40u 3.13s 6.00g
266 :    
267 :     2000 CPS 100 spill 126.69u 4.84s 3.08g
268 :     MLRISC ra 0.80u 0.10s 0.55g
269 :     (spills = 42 reloads = 84)
270 :     TOTAL 129.42u 5.10s 4.13g
271 :    
272 :     3000 CPS 100 spill 675.59u 19.03s 11.64g
273 :     MLRISC ra 2.69u 0.27s 1.38g
274 :     (spills = 62 reloads = 124)
275 :     TOTAL 682.48u 19.61s 13.99g
276 :    
277 :     4000 CPS 100 spill 2362.82u 56.28s 43.60g
278 :     MLRISC ra 4.96u 0.27s 2.72g
279 :     (spills = 85 reloads = 170)
280 :     TOTAL 2375.26u 57.21s 48.00g
281 :    
282 :     As you can see the old cps spill module suffers from some serious
283 : leunga 1115 performance problem. But since I cannot decipher the old code fully,
284 :     instead of patching the problems up, I'm reimplementing it
285 : leunga 1094 with a different algorithm. The new code is more modular,
286 :     smaller when compiled, and substantially faster
287 :     (O(n log n) time and O(n) space). Timing of the new spill module:
288 :    
289 :     4000 CPS 100 spill 0.02u 0.00s 0.00g
290 :     MLRISC ra 0.25u 0.02s 0.15g
291 :     (spills=1 reloads=3)
292 :     TOTAL 7.74u 0.34s 1.62g
293 :    
294 :     Implementation details:
295 :    
296 :     As far as I can tell, the purpose of the CPS spill module is to make sure the
297 :     number of live variables at any program point (the bandwidth)
298 :     does not exceed a certain limit, which is determined by the
299 :     size of the spill area.
300 :    
301 :     When the bandwidth is too large, we decrease the register pressure by
302 :     packing live variables into spill records. How we achieve this is
303 :     completely different than what we did in the old code.
304 :    
305 :     First, there is something about the MLRiscGen code generator
306 :     that we should be aware of:
307 :    
308 :     o MLRiscGen performs code motion!
309 :    
310 :     In particular, it will move floating point computations and
311 :     address computations involving only the heap pointer to
312 :     their use sites (if there is only a single use).
313 :     What this means is that if we have a CPS record construction
314 :     statement
315 :    
316 :     RECORD(k,vl,w,e)
317 :    
318 :     we should never count the new record address w as live if w
319 :     has only one use (which is often the case).
320 :    
321 :     We should do something similar to floating point, but the transformation
322 :     there is much more complex, so I won't deal with that.
323 :    
324 :     Secondly, there are now two new cps primops at our disposal:
325 :    
326 :     1. rawrecord of record_kind option
327 :     This pure operator allocates some uninitialized storage from the heap.
328 :     There are two forms:
329 :    
330 :     rawrecord NONE [INT n] allocates a tagless record of length n
331 :     rawrecord (SOME rk) [INT n] allocates a tagged record of length n
332 :     and initializes the tag.
333 :    
334 :     2. rawupdate of cty
335 :     rawupdate cty (v,i,x)
336 :     Assigns to x to the ith component of record v.
337 :     The storelist is not updated.
338 :    
339 :     We use these new primops for both spilling and increment record construction.
340 :    
341 :     1. Spilling.
342 :    
343 :     This is implemented with a linear scan algorithm (but generalized
344 :     to trees). The algorithm will create a single spill record at the
345 :     beginning of the cps function and use rawupdate to spill to it,
346 :     and SELECT or SELp to reload from it. So both spills and reloads
347 :     are fine-grain operations. In contrast, in the old algorithm
348 :     "spills" have to be bundled together in records.
349 :    
350 :     Ideally, we should sink the spill record construction to where
351 :     it is needed. We can even split the spill record into multiple ones
352 :     at the places where they are needed. But CPS is not a good
353 :     representation for global code motion, so I'll keep it simple and
354 :     am not attempting this.
355 :    
356 :     2. Incremental record construction (aka record splitting).
357 :    
358 :     Long records with many component values which are simulatenously live
359 :     (recall that single use record addresses are not considered to
360 :     be live) are constructed with rawrecord and rawupdate.
361 :     We allocate space on the heap with rawrecord first, then gradually
362 :     fill it in with rawupdate. This is the technique suggested to me
363 :     by Matthias.
364 :    
365 :     Some restrictions on when this is applicable:
366 :     1. It is not a VECTOR record. The code generator currently does not handle
367 :     this case. VECTOR record uses double indirection like arrays.
368 :     2. All the record component values are defined in the same "basic block"
369 :     as the record constructor. This is to prevent speculative
370 :     record construction.
371 :    
372 :     ----------------------------------------------------------------------
373 :     Name: Allen Leung
374 : leunga 1086 Date: 2002/02/22 01:02:00 EST
375 :     Tag: leunga-20020222-mlrisc-tools
376 :    
377 :     Minor bug fixes in the parser and rewriter
378 :    
379 :     ----------------------------------------------------------------------
380 :     Name: Allen Leung
381 : leunga 1085 Date: 2002/02/21 20:20:00 EST
382 :     Tag: leunga-20020221-peephole
383 :    
384 :     Regenerated the peephole files. Some contained typos in the specification
385 :     and some didn't compile because of pretty printing bugs in the old version
386 :     of 'nowhere'.
387 :    
388 :     ----------------------------------------------------------------------
389 :     Name: Allen Leung
390 : leunga 1079 Date: 2002/02/19 20:20:00 EST
391 :     Tag: leunga-20020219-mlrisc-tools
392 :     Description:
393 :    
394 :     Minor bug fixes to the mlrisc-tools library:
395 :    
396 :     1. Fixed up parsing colon suffixed keywords
397 :     2. Added the ability to shut the error messages up
398 :     3. Reimplemented the pretty printer and fixed up/improved
399 :     the pretty printing of handle and -> types.
400 :     4. Fixed up generation of literal symbols in the nowhere tool.
401 :     5. Added some SML keywords to to sml.sty
402 :    
403 :     ----------------------------------------------------------------------
404 :     Name: Matthias Blume
405 : blume 1078 Date: 2002/02/19 16:20:00 EST
406 :     Tag: blume-20020219-cmffi
407 :     Description:
408 :    
409 :     A wild mix of changes, some minor, some major:
410 :    
411 :     * All C FFI-related libraries are now anchored under $c:
412 :     $/c.cm --> $c/c.cm
413 :     $/c-int.cm --> $c/internals/c-int.cm
414 :     $/memory.cm --> $c/memory/memory.cm
415 :    
416 :     * "make" tool (in CM) now treats its argument pathname slightly
417 :     differently:
418 :     1. If the native expansion is an absolute name, then before invoking
419 :     the "make" command on it, CM will apply OS.Path.mkRelative
420 :     (with relativeTo = OS.FileSys.getDir()) to it.
421 :     2. The argument will be passed through to subsequent phases of CM
422 :     processing without "going native". In particular, if the argument
423 :     was an anchored path, then "make" will not lose track of that anchor.
424 :    
425 :     * Compiler backends now "know" their respective C calling conventions
426 :     instead of having to be told about it by ml-nlffigen. This relieves
427 :     ml-nlffigen from one of its burdens.
428 :    
429 :     * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
430 :    
431 :     * Export C_DEBUG and C_Debug from $c/c.cm.
432 :    
433 :     * C type encoding in ml-nlffi-lib has been improved to model the conceptual
434 :     subtyping relationship between incomplete pointers and their complete
435 :     counterparts. For this, ('t, 'c) ptr has been changed to 'o ptr --
436 :     with the convention of instantiating 'o with ('t, 'c) obj whenever
437 :     the pointer target type is complete. In the incomplete case, 'o
438 :     will be instantiated with some "'c iobj" -- a type obtained by
439 :     using one of the functors PointerToIncompleteType or PointerToCompleteType.
440 :    
441 :     Operations that work on both incomplete and complete pointer types are
442 :     typed as taking an 'o ptr while operations that require the target to
443 :     be known are typed as taking some ('t, 'c) obj ptr.
444 :    
445 :     voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
446 :     where void is an eqtype without any values. This makes it possible
447 :     to work on voidptr values using functions meant to operate on light
448 :     incomplete pointers.
449 :    
450 :     * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
451 :     been vastly simplified.
452 :    
453 :     ----------------------------------------------------------------------
454 :     Name: Matthias Blume
455 : blume 1076 Date: 2002/02/19 10:48:00 EST
456 :     Tag: blume-20020219-pqfix
457 :     Description:
458 :    
459 :     Applied Chris Okasaki's bug fix for priority queues.
460 :    
461 :     ----------------------------------------------------------------------
462 :     Name: Matthias Blume
463 : blume 1073 Date: 2002/02/15 17:05:00
464 : blume 1069 Tag: Release_110_39
465 :     Description:
466 :    
467 : blume 1073 Last-minute retagging is becoming a tradition... :-(
468 :    
469 :     This is the working release 110.39.
470 :    
471 :     ----------------------------------------------------------------------
472 :     Name: Matthias Blume
473 :     Date: 2002/02/15 16:00:00 EST
474 :     Tag: Release_110_39-orig
475 :     Description:
476 :    
477 : blume 1069 Working release 110.39. New bootfiles.
478 :    
479 : blume 1073 (Update: There was a small bug in the installer so it wouldn't work
480 :     with all shells. So I retagged. -Matthias)
481 :    
482 : blume 1069 ----------------------------------------------------------------------
483 :     Name: Matthias Blume
484 : blume 1068 Date: 2002/02/15 14:17:00 EST
485 :     Tag: blume-20020215-showbindings
486 :     Description:
487 :    
488 :     Added EnvRef.listBoundSymbols and CM.State.showBindings. Especially
489 :     the latter can be useful for exploring what bindings are available at
490 :     the interactive prompt. (The first function returns only the list
491 :     of symbols that are really bound, the second prints those but also the
492 :     ones that CM's autoloading mechanism knows about.)
493 :    
494 :     ----------------------------------------------------------------------
495 :     Name: Matthias Blume
496 : blume 1067 Date: 2002/02/15 12:08:00 EST
497 :     Tag: blume-20020215-iptrs
498 :     Description:
499 :    
500 :     Two improvements to ml-nlffigen:
501 :    
502 :     1. Write files only if they do not exist or if their current contents
503 :     do not coincide with what's being written. (That is, avoid messing
504 :     with the time stamps unless absolutely necessary.)
505 :    
506 :     2. Implement a "repository" mechanism for generated files related
507 :     to "incomplete pointer types". See the README file for details.
508 :    
509 :     ----------------------------------------------------------------------
510 :     Name: Matthias Blume
511 : blume 1066 Date: 2002/02/14 11:50:00 EST
512 :     Tag: blume-20020214-quote
513 :     Description:
514 :    
515 :     Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm). This is required
516 :     because of the new and improved tag generation scheme. (Thanks to Allen
517 :     Leung for pointing it out.)
518 :    
519 :     ----------------------------------------------------------------------
520 : george 1065 Name: Lal George
521 :     Date: 2002/02/14 09:55:27 EST 2002
522 :     Tag: george-20020214-isabelle-bug
523 :     Description:
524 :    
525 :     Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
526 :     of Isabelle on the x86.
527 :    
528 :     From Allen:
529 :     -----------
530 :     I've found the problem:
531 :    
532 :     in ra-core.sml, I use the counter "blocked" to keep track of the
533 :     true number of elements in the freeze queue. When the counter goes
534 :     to zero, I skip examining the queue. But I've messed up the
535 :     bookkeeping in combine():
536 :    
537 :     else ();
538 :     case !ucol of
539 :     PSEUDO => (if !cntv > 0 then
540 :     (if !cntu > 0 then blocked := !blocked - 1 else ();
541 :     ^^^^^^^^^^^^^^^^^^^^^^^
542 :     moveu := mergeMoveList(!movev, !moveu)
543 :     )
544 :     else ();
545 :    
546 :     combine() is called to coalesce two nodes u and v.
547 :     I think I was thinking that if the move counts of u and v are both
548 :     greater than zero then after they are coalesced then one node is
549 :     removed from the freeze queue. Apparently I was thinking that
550 :     both u and v are of low degree, but that's clearly not necessarily true.
551 :    
552 :    
553 :     02/12/2002:
554 :     Here's the patch. HOL now compiles.
555 :    
556 :     I don't know how this impact on performance (compile
557 :     time or runtime). This bug caused the RA (especially on the x86)
558 :     to go thru the potential spill phase when there are still nodes on the
559 :     freeze queue.
560 :    
561 :    
562 :    
563 :    
564 :     ----------------------------------------------------------------------
565 : blume 1058 Name: Matthias Blume
566 : blume 1064 Date: 2002/02/13 22:40:00 EST
567 :     Tag: blume-20020213-fptr-rtti
568 :     Description:
569 :    
570 :     Fixed a bug in ml-nlffigen that was introduced with one of the previous
571 :     updates.
572 :    
573 :     ----------------------------------------------------------------------
574 :     Name: Matthias Blume
575 : blume 1063 Date: 2002/02/13 16:41:00 EST
576 :     Tag: blume-20020213-cmlpq
577 :     Description:
578 :    
579 :     Added new priority queue export symbols (which have just been added to
580 :     smlnj-lib.cm) to CML's version of smlnj-lib.cm. (Otherwise CML would
581 :     not compile and the installer would choke.)
582 :    
583 :     ----------------------------------------------------------------------
584 :     Name: Matthias Blume
585 : blume 1062 Date: 2002/02/13 16:15:00 EST
586 :     Tag: blume-20020213-various
587 :     Description:
588 :    
589 :     1. More tweaks to ml-nlffigen:
590 :    
591 :     - better internal datastructures (resulting in slight speedup)
592 :     - "-match" option requires exact match
593 :     - "localized" gensym counters (untagged structs/unions nested within
594 :     other structs/unions or within typedefs get a fresh counter; their
595 :     tag will be prefixed by a concatenation of their parents' tags)
596 :     - bug fixes (related to calculation of transitive closure of types
597 :     to be included in the output)
598 :    
599 :     2. Minor Basis updates:
600 :    
601 :     - added implementations for List.collate and Option.app
602 :    
603 :     ----------------------------------------------------------------------
604 :     Name: Matthias Blume
605 : blume 1060 Date: 2002/02/11 15:55:00 EST
606 :     Tag: blume-20020211-gensym
607 :     Description:
608 :    
609 :     Added a "-gensym" option to command line of ml-nlffigen. This can be
610 :     used to specify a "stem" -- a string that is inserted in all "gensym'd"
611 :     names (ML structure names that correspond to unnamed C structs, unions,
612 :     and enums), so that separate runs of ml-nlffigen do not clash.
613 :    
614 :     ----------------------------------------------------------------------
615 :     Name: Matthias Blume
616 : blume 1059 Date: 2002/02/11 12:05:00 EST
617 :     Tag: blume-20020211-gensml
618 :     Description:
619 :    
620 :     A quick fix for a problem with GenSML (in the pgraph-util library):
621 :     Make generation of toplevel "local" optional. (Strictly speaking,
622 :     signature definitions within "local" are not legal SML.)
623 :    
624 :     Other than that: updates to INSTALL and cm/TODO.
625 :    
626 :     ----------------------------------------------------------------------
627 :     Name: Matthias Blume
628 : blume 1058 Date: 2002/02/08 15:00:00 EST
629 :     Tag: blume-20020208-uniquepid
630 :     Description:
631 :    
632 :     0. Version number has been bumped to 110.38.1. NEW BOOTFILES!!!
633 :    
634 :     1. The installer (config/install.sh) has gotten smarter:
635 :    
636 :     - Configuration options are a bit easier to specify now
637 :     (in config/targets).
638 :     - Bug in recognizing .tar.bz2 files fixed.
639 :     - Installer automatically resolves dependencies between
640 :     configuration options (e.g., if you ask for eXene, you will
641 :     also get cml -- regardless whether you asked for it or not).
642 :     - Installer can run in "quieter mode" by setting the environment
643 :     variable INSTALL_QUIETLY to "true". "Quieter" does not mean
644 :     "completely silent", though.
645 :     - Build HashCons library as part of smlnj-lib.
646 :    
647 :     2. A new scheme for assigning persistent identifiers to compilation
648 :     units (and, by extension, to types etc.) has been put into place.
649 :     This fixes a long-standing bug where types and even dynamic values
650 :     can get internally confused, thereby compromising type safety
651 :     (abstraction) and dynamic correctness. See
652 :    
653 :     http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
654 :    
655 :     for an example of how things could go wrong until now.
656 :    
657 :     The downside of the new scheme is that pids are not quite as
658 :     persistent as they used to be: CM will generate a fresh pid
659 :     for every compilation unit that it thinks it sees for the first
660 :     time. That means that if you compile starting from a clean, fresh
661 :     source tree at two different times, you end up with different
662 :     binaries.
663 :    
664 :     Cutoff recompilation, however, has not been compromised because
665 :     CM keeps pid information in special caches between runs.
666 :    
667 :     ----------------------------------------------------------------------
668 : george 1053 Name: Lal George
669 : george 1055 Date: 2002/02/07 15:34:13 EST 2002
670 :     Tag: <none>
671 :     Description:
672 :    
673 :     Compilers that generate assembly code may produce global labels
674 :     whose value is resolved at link time. The various peephole optimization
675 :     modules did not take this in account.
676 :    
677 :     TODO. The Labels.addrOf function should really return an option
678 :     type so that clients are forced to deal with this issue, rather
679 :     than an exception being raised.
680 :    
681 :     ----------------------------------------------------------------------
682 :     Name: Lal George
683 : george 1053 Date: 2002/02/06 13:55:02 EST
684 :     Tag: george-20020206-ra-breakup
685 :     Description:
686 :    
687 :     1. A bug fix from Allen.
688 :    
689 :     A typo causes extra fstp %st(0)'s to be generated at compensation
690 :     edges, which might cause stack underflow traps at runtime. This
691 :     occurs in fft where there are extraneous fstps right before the 'into'
692 :     trap instruction (in this case they are harmless since none of the
693 :     integers overflow.)
694 :    
695 : george 1054 2. Pulled out various utility modules that were embedded in the modules
696 : george 1053 of the register allocator. I need these modules for other purposes, but
697 : george 1054 they are not complete enough to put into a library (just yet).
698 : george 1053 ----------------------------------------------------------------------
699 : blume 1049 Name: Matthias Blume
700 :     Date: 2002/01/31 16:05:00 EST
701 :     Tag: blume-20020131-sparc-ccalls
702 :     Description:
703 :    
704 :     1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
705 :     of extra stack space by mistake. Fixed.
706 :    
707 :     2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
708 :    
709 :     ----------------------------------------------------------------------
710 : leunga 1048 Name: Allen Leung
711 :     Date: 2002/01/30
712 :     Tag: leunga-20020130-nowhere-bug-fix
713 :     Description:
714 :    
715 :     MLRISC bug fixes:
716 :     1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
717 :     2. MachineInt.fromString was negating its value.
718 :    
719 :     ----------------------------------------------------------------------
720 : blume 1039 Name: Matthias Blume
721 : blume 1047 Date: 2002/01/29
722 :     Tag: blume-20020129-INSTALL
723 :     Description:
724 :    
725 :     - Added somewhat detailed installation instructions (file INSTALL).
726 :     - Fixed curl-detection bug in config/install.sh.
727 :     - It is now possible to select the URL getter using the URLGETTER
728 :     environment variable:
729 :    
730 :     not set / "unknown" --> automatic detection (script tries wget,
731 :     curl, and lynx)
732 :     "wget" / "curl" / "lynx" --> use the specified program (script "knows"
733 :     how to properly invoke them)
734 :     other --> use $URLGETTER directly, it must take
735 :     precisely two command-line arguments
736 :     (source URL and destination file name)
737 :    
738 :     ----------------------------------------------------------------------
739 :     Name: Matthias Blume
740 : blume 1044 Date: 2002/01/28
741 : blume 1045 Tag: blume-20020128-sparc-ccalls
742 :     Description:
743 :    
744 :     - Fixed problem with calculation of "used" registers in sparc-c-calls.
745 :     - Make use of the allocParam argument in sparc-c-calls.
746 :    
747 :     ----------------------------------------------------------------------
748 :     Name: Matthias Blume
749 :     Date: 2002/01/28
750 : blume 1044 Tag: blume-20020128-allocParam
751 :     Description:
752 :    
753 :     John Reppy: Changes c-calls API to accept client-callback for
754 :     allocating extra stack space.
755 :     me: Corresponding changes to mlriscGen (using a dummy argument that
756 :     does not change the current behavior).
757 :    
758 :     ----------------------------------------------------------------------
759 :     Name: Matthias Blume
760 : blume 1040 Date: 2002/01/28 12:00:00
761 :     Tag: Release_110_38
762 :     Description:
763 :    
764 :     This time for real!!!
765 :    
766 :     ----------------------------------------------------------------------
767 :     Name: Matthias Blume
768 : blume 1039 Date: 2002/01/28 10:56:00 EST
769 :     Tag: blume-20020128-retraction
770 :     Description:
771 :    
772 :     0. Retracted earlier 110.38. (The Release_110_38 tag has been replaced
773 :     with blume-Release_110_38-retracted.)
774 :    
775 :     1. Fixed a problem with incorrect rounding modes in real64.sml.
776 :     (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
777 :    
778 :     2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
779 :     and enums fixed. The naming of corresponding ML identifiers should
780 :     now be consistent again.
781 :    
782 :     ----------------------------------------------------------------------
783 : leunga 1038 Name: Allen Leung
784 :     Date: 2002/01/27
785 :     Tag: leunga-20020127-nowhere
786 :     Description:
787 :    
788 :     Added a target called nowhere in the configuration scripts.
789 :     Enabling this will build the MLRISC 'nowhere' tool (for translating
790 :     programs with where-clauses into legal SML code) during installation.
791 :    
792 :     ----------------------------------------------------------------------
793 : blume 1036 Name: Matthias Blume
794 : leunga 1038 Date: 2002/01/25 21:27:00 EST
795 : blume 1039 Tag: blume-Release_110_38-retracted
796 : blume 1037 Description:
797 :    
798 :     Call it a (working) release! Version is 110.38. Bootfiles are ready.
799 :    
800 :     README will be added later.
801 :    
802 : blume 1039 !!! NOTE: Re-tagged as blume-Release_110_38-retracted. Original tag
803 :     (Release_110_38) removed. Reason: Last-minute bug fixes.
804 :    
805 : blume 1037 ----------------------------------------------------------------------
806 :     Name: Matthias Blume
807 : blume 1036 Date: 2002/01/25
808 :     Tag: blume-20020125-ffi
809 :     Description:
810 :    
811 :     A large number of tweaks and improvements to ml-nlffi-lib and
812 :     ml-nlffigen:
813 :    
814 :     - ML represenation types have been streamlined
815 :     - getter and setter functions work with concrete values, not abstract
816 :     ones where possible
817 :     - ml-nlffigen command line more flexible (see README file there)
818 :     - some bugs have been fixed (hopefully)
819 :    
820 :     ----------------------------------------------------------------------
821 : george 1034 Name: Lal George
822 :     Date: 2002/01/24
823 :     Tag: george-20020124-risc-ra-interface
824 :     Description:
825 :    
826 :     There is a dramatic simplification in the interface to the
827 :     register allocator for RISC architectures as a result of making
828 :     parallel copy instructions explicit.
829 :    
830 :     ----------------------------------------------------------------------
831 : blume 1028 Name: Matthias Blume
832 : blume 1032 Date: 2002/01/22
833 :     Tag: blume-20020122-x86-ccalls
834 :     Description:
835 :    
836 :     Bug fix for c-calls on x86 (having to do with how char- and
837 :     short-arguments are being handled).
838 :    
839 :     ----------------------------------------------------------------------
840 :     Name: Matthias Blume
841 : blume 1031 Date: 2002/01/21
842 :     Tag: blume-20020121-ff
843 :     Description:
844 :    
845 :     Another day of fiddling with the FFI...
846 :    
847 :     1. Bug fix/workaround: CKIT does not complain about negative array
848 :     dimensions, so ml-nlffigen has to guard itself against this possibility.
849 :     (Otherwise a negative dimension would send it into an infinite loop.)
850 :    
851 :     2. Some of the abstract types (light objects, light pointers, most "base"
852 :     types) in structure C are now eqtypes.
853 :    
854 :     3. Added constructors and test functions for NULL function pointers.
855 :    
856 :     ----------------------------------------------------------------------
857 :     Name: Matthias Blume
858 : blume 1028 Date: 2002/01/18
859 : blume 1030 Tag: blume-20020118-ready-for-new-release
860 :     Description:
861 :    
862 :     Made config/srcarchiveurl point to a new place. (Will provide boot
863 :     files shortly.)
864 :    
865 :     Maybe we christen this to be 110.38?
866 :    
867 :     ----------------------------------------------------------------------
868 :     Name: Matthias Blume
869 :     Date: 2002/01/18
870 : blume 1029 Tag: blume-20020118-more-ffifiddle
871 :     Description:
872 :    
873 :     Today's FFI fiddling:
874 :    
875 :     - Provided a structure CGetSet with "convenient" versions of C.Get.* and
876 :     C.Set.* that use concrete (MLRep.*) arguments and results instead
877 :     of abstract ones.
878 :    
879 :     - Provided word-style bit operations etc. for "int" representation
880 :     types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
881 :     and Long.
882 :    
883 :     ----------------------------------------------------------------------
884 :     Name: Matthias Blume
885 :     Date: 2002/01/18
886 : blume 1028 Tag: blume-20020118-use-x86-fp
887 :     Description:
888 :    
889 :     Now that x86-fast-fp seems to be working, I turned it back on again
890 :     by default. (Seems to work fine now, even with the FFI.)
891 :    
892 :     Other than that, I added some documentation about the FFI to
893 :     src/ml-nlffigen/README and updated the FFI test examples in
894 :     src/ml-nlffi-lib/Tests/*.
895 :    
896 :     ----------------------------------------------------------------------
897 : leunga 1025 Name: Allen Leung
898 :     Date: 2002/01/17
899 :     Tag: leunga-20020117-x86-fast-fp-call
900 :     Description:
901 :    
902 :     1. Fixed a problem with handling return fp values when x86's fast fp
903 :     mode is turned on.
904 :    
905 :     2. Minor pretty printing fix for cellset. Print %st(0) as %st(0) instead
906 :     of %f32.
907 :    
908 :     3. Added a constructor INT32lit to the ast of MLRISC tools.
909 :    
910 :     ----------------------------------------------------------------------
911 : blume 1021 Name: Matthias Blume
912 :     Date: 2002/01/16
913 :     Tag: blume-20020116-ffifiddle
914 :     Description:
915 :    
916 :     More fiddling with the FFI interface:
917 :    
918 :     - Make constness 'c instead of rw wherever possible. This eliminates
919 :     the need for certain explicit coercions. (However, due to ML's
920 :     value polymorphism, there will still be many cases where explicit
921 :     coercions are necessary. Phantom types are not the whole answer
922 :     to modeling a subtyping relationship in ML.)
923 :    
924 :     - ro/rw coersions for pointers added. (Avoids the detour through */&.)
925 :    
926 :     - "printf" test example added to src/ml-nlffi-lib/Tests. (Demonstrates
927 :     clumsy workaround for varargs problem.)
928 :    
929 :     ----------------------------------------------------------------------
930 : george 1016 Name: Lal George
931 :     Date: 2002/01/15
932 :     Tag: <none>
933 :     Description:
934 :    
935 :     1. Since COPY instructions are no longer native to the architecture,
936 :     a generic functor can be used to implement the expandCopies function.
937 :    
938 :     2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
939 :     TEXT segment.
940 :    
941 :     ----------------------------------------------------------------------
942 : blume 1011 Name: Matthias Blume
943 : blume 1015 Date: 2002/01/15
944 :     Tag: blume-20020115-ffiupdates
945 :     Description:
946 :    
947 :     1. Fix for bug resulting in single-precision float values being returned
948 :     incorrectly from FFI calls.
949 :    
950 :     2. Small modifications to C FFI API:
951 :    
952 :     - memory-allocation routines return straight objects (no options)
953 :     and raise an exception in out-of-memory situations
954 :     - unsafe extensions to cast between function pointers and pointers
955 :     from/to ints
956 :     - added structure C_Debug as an alternative to structure C where
957 :     pointer-dereferencing (|*| and |*!) always check for null-pointers
958 :     - added open_lib' to DynLinkage; open_lib' works like open_lib
959 :     but also takes a (possibly empty) list of existing library handles
960 :     that the current library depends on
961 :    
962 :     ----------------------------------------------------------------------
963 :     Name: Matthias Blume
964 : blume 1011 Date: 2002/01/10
965 :     Tag: blume-20020110-newffigen
966 :     Description:
967 :    
968 :     1. Updates to portable graph code.
969 :    
970 :     2. Major update to ml-nlffigen and ml-nlffi-lib. Things are much
971 :     more scalable now so that even huge interfaces such as the one
972 :     for GTK compile in finite time and space. :-)
973 :     See src/ml-nlffigen/README for details on what's new.
974 :    
975 :     ----------------------------------------------------------------------
976 : george 1003 Name: Lal George
977 : george 1009 Date: 2001/01/09 14:31:35 EST 2002
978 :     Tag: george-20011206-rm-native-copy
979 :     Description:
980 :    
981 :     Removed the native COPY and FCOPY instructions
982 :     from all the architectures and replaced it with the
983 :     explicit COPY instruction from the previous commit.
984 :    
985 :     It is now possible to simplify many of the optimizations
986 :     modules that manipulate copies. This has not been
987 :     done in this change.
988 :    
989 :     ----------------------------------------------------------------------
990 :     Name: Lal George
991 : george 1003 Date: 2001/12/06 16:50:13 EST 2001
992 :     Tag: george-20011206-mlrisc-instruction
993 :     Description:
994 : george 824
995 : george 1003 Changed the representation of instructions from being fully abstract
996 :     to being partially concrete. That is to say:
997 :    
998 :     from
999 :     type instruction
1000 :    
1001 :     to
1002 :     type instr (* machine instruction *)
1003 :    
1004 :     datatype instruction =
1005 :     LIVE of {regs: C.cellset, spilled: C.cellset}
1006 :     | KILL of {regs: C.cellset, spilled: C.cellset}
1007 :     | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
1008 :     | ANNOTATION of {i: instruction, a: Annotations.annotation}
1009 :     | INSTR of instr
1010 :    
1011 :     This makes the handling of certain special instructions that appear on
1012 :     all architectures easier and uniform.
1013 :    
1014 :     LIVE and KILL say that a list of registers are live or killed at the
1015 :     program point where they appear. No spill code is generated when an
1016 :     element of the 'regs' field is spilled, but the register is moved to
1017 :     the 'spilled' (which is present, more for debugging than anything else).
1018 :    
1019 :     LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
1020 :     We used to generate:
1021 :    
1022 :     DEFFREG f1
1023 :     f1 := f2 + f3
1024 :     trapb
1025 :    
1026 :     but now generate:
1027 :    
1028 :     f1 := f2 + f3
1029 :     trapb
1030 :     LIVE {regs=[f1,f2,f3], spilled=[]}
1031 :    
1032 :     Furthermore, the DEFFREG (hack) required that all floating point instruction
1033 :     use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
1034 :     defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
1035 :     in a cleaner alpha implementation. (Hopefully, intel will not get rid of
1036 :     this architecture).
1037 :    
1038 :     COPYXXX is intended to replace the parallel COPY and FCOPY available on
1039 :     all the architectures. This will result in further simplification of the
1040 :     register allocator that must be aware of them for coalescing purposes, and
1041 :     will also simplify certain aspects of the machine description that provides
1042 :     callbacks related to parallel copies.
1043 :    
1044 :     ANNOTATION should be obvious, and now INSTR represents the honest to God
1045 :     machine instruction set!
1046 :    
1047 :     The <arch>/instructions/<arch>Instr.sml files define certain utility
1048 :     functions for making porting easier -- essentially converting upper case
1049 :     to lower case. All machine instructions (of type instr) are in upper case,
1050 :     and the lower case form generates an MLRISC instruction. For example on
1051 :     the alpha we have:
1052 :    
1053 :     datatype instr =
1054 :     LDA of {r:cell, b:cell, d:operand}
1055 :     | ...
1056 :    
1057 :     val lda : {r:cell, b:cell, d:operand} -> instruction
1058 :     ...
1059 :    
1060 :     where lda is just (INSTR o LDA), etc.
1061 :    
1062 : leunga 815 ----------------------------------------------------------------------
1063 : blume 986 Name: Matthias Blume
1064 : blume 993 Date: 2001/11/22 21:40:00 EST
1065 :     Tag: Release_110_37
1066 :     Description:
1067 :    
1068 :     Release 110.37. This time for real.
1069 :    
1070 :     ----------------------------------------------------------------------
1071 :     Name: Matthias Blume
1072 : blume 988 Date: 2001/11/21 16:35:00 EST
1073 :     Tag: blume-20011121-foot-in-mouth
1074 :     Description:
1075 :    
1076 :     Removed the "Release_110_37" tag because of a serious bug.
1077 :     This will be re-tagged once the bug is fixed.
1078 :    
1079 :     ----------------------------------------------------------------------
1080 :     Name: Matthias Blume
1081 : blume 987 Date: 2001/11/21 16:14:00 EST
1082 :     Tag: blume-20011121-forgottenfile
1083 :     Description:
1084 :    
1085 :     Forgot to add a file. (Just a .tex-file -- part of
1086 :     the CM manual source.)
1087 :    
1088 :     ----------------------------------------------------------------------
1089 :     Name: Matthias Blume
1090 : blume 986 Date: 2001/11/21 16:10:00 EST
1091 : blume 988 Tag: blume-20011121-invalid_110_37
1092 : blume 986 Description:
1093 :    
1094 : blume 988 Note: I removed the original tag "Release_110_37" from this commit
1095 :     because we found a serious bug in all non-x86 backends.
1096 :     - Matthias
1097 :    
1098 : blume 986 1. Modifications to the SML/NJ code generator and to the runtime system
1099 :     so that code object name strings are directly inserted into code
1100 :     objects at code generation time. The only business the runtime system
1101 :     has with this is now to read the name strings on occasions.
1102 :     (The encoding of the name string has also changed somewhat.)
1103 :    
1104 :     2. CM now implements a simple "set calculus" for specifying export lists.
1105 :     In particular, it is now possible to refer to the export lists of
1106 :     other libraries/groups/sources and form unions as well as differences.
1107 :     See the latest CM manual for details.
1108 :    
1109 :     3. An separate notion of "proxy" libraries has again be eliminated from
1110 :     CM's model. (Proxy libraries are now simply a special case of using
1111 :     the export list calculus.)
1112 :    
1113 :     4. Some of the existing libraries now take advantage of the new set
1114 :     calculus.
1115 :     (Notice that not all libraries have been converted because some
1116 :     of the existing .cm-files are supposed to be backward compatible
1117 :     with 110.0.x.)
1118 :    
1119 :     5. Some cleanup in stand-alone programs. (Don't use "exnMessage" -- use
1120 :     "General.exnMessage"! The former relies on a certain hook to be
1121 :     initialized, and that often does not happen in the stand-alone case.)
1122 :    
1123 :     ----------------------------------------------------------------------
1124 : george 984 Name: Lal George
1125 :     Date: 2001/11/21 13:56:18 EST
1126 :     Tag: george-2001121-pseudo-ops
1127 :     Description:
1128 :    
1129 :     Implemented a complete redesign of MLRISC pseudo-ops. Now there
1130 :     ought to never be any question of incompatabilities with
1131 :     pseudo-op syntax expected by host assemblers.
1132 :    
1133 :     For now, only modules supporting GAS syntax are implemented
1134 :     but more should follow, such as MASM, and vendor assembler
1135 :     syntax, e.g. IBM as, Sun as, etc.
1136 :    
1137 :     ----------------------------------------------------------------------
1138 : blume 977 Name: Matthias Blume
1139 :     Date: 2001/11/14 11:52:00 EST
1140 :     Tag: blume-20011114-srcname
1141 :     Description:
1142 :    
1143 :     1. Routed the name of the current source file to mlriscgen where it
1144 :     should be directly emitted into the code object. (This last part
1145 :     is yet to be done.)
1146 :    
1147 :     2. Some cleanup of the pgraph code to make it match the proposal that
1148 :     I put out the other day. (The proposal notwithstanding, things are
1149 :     still in flux here.)
1150 :    
1151 :     ----------------------------------------------------------------------
1152 : george 976 Name: Lal George
1153 :     Date: 2001/11/14 09:44:04 EST
1154 :     Tag:
1155 :     Description:
1156 :    
1157 :     Fix for a backpatching bug reported by Allen.
1158 :    
1159 :     Because the boundary between short and long span-dependent
1160 :     instructions is +/- 128, there are an astounding number of
1161 :     span-dependent instructions whose size is over estimated.
1162 :    
1163 :     Allen came up with the idea of letting the size of span
1164 :     dependent instructions be non-monotonic, for a maxIter
1165 :     number of times, after which the size must be monotonically
1166 :     increasing.
1167 :    
1168 :     This table shows the number of span-dependent instructions
1169 :     whose size was over-estimated as a function of maxIter, for the
1170 :     file Parse/parse/ml.grm.sml:
1171 :    
1172 :     maxIter # of instructions:
1173 :     10 687
1174 :     20 438
1175 :     30 198
1176 :     40 0
1177 :    
1178 :     In compiling the compiler, there is no significant difference in
1179 :     compilation speed between maxIter=10 and maxIter=40. Actually,
1180 :     my measurements showed that maxIter=40 was a tad faster than
1181 :     maxIter=10! Also 96% of the files in the compiler reach a fix
1182 :     point within 13 iterations, so fixing maxIter at 40, while high,
1183 :     is okay.
1184 :    
1185 :     ----------------------------------------------------------------------
1186 : blume 975 Name: Matthias Blume
1187 :     Date: 2001/10/31 15:25:00 EST
1188 :     Tag: blume-20011031-pgraph
1189 :     Description:
1190 :    
1191 :     CKIT:
1192 :     * Changed the "Function" constructor of type Ast.ctype to carry optional
1193 :     argument identifiers.
1194 :     * Changed the return type of TypeUtil.getFunction accordingly.
1195 :     * Type equality ignores the argument names.
1196 :     * TypeUtil.composite tries to preserve argument names but gives up quickly
1197 :     if there is a mismatch.
1198 :    
1199 :     installation script:
1200 :     * attempts to use "curl" if available (unless "wget" is available as well)
1201 :    
1202 :     CM:
1203 :     * has an experimental implementation of "portable graphs" which I will
1204 :     soon propose as an implementation-independent library format
1205 :     * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
1206 :    
1207 :     NLFFI-LIB:
1208 :     * some cleanup (all cosmetic)
1209 :    
1210 :     NLFFIGEN:
1211 :     * temporarily disabled the mechanism that suppresses ML output for
1212 :     C definitions whose identifiers start with an underscore character
1213 :     * generate val bindings for enum constants
1214 :     * user can request that only one style (light or heavy) is being used;
1215 :     default is to use both (command-line arguments: -heavy and -light)
1216 :     * fixed bug in handling of function types involving incomplete pointers
1217 :     * generate ML entry points that take record arguments (i.e., using
1218 :     named arguments) for C functions that have a prototype with named
1219 :     arguments
1220 :     (see changes to CKIT)
1221 :    
1222 :     ----------------------------------------------------------------------
1223 : leunga 972 Name: Allen Leung
1224 :     Date: 2001/10/27 20:34:00 EDT
1225 :     Tag: leunga-20011027-x86-fast-fp-call
1226 :     Description:
1227 :    
1228 :     Fixed the bug described in blume-20010920-slowfp.
1229 :    
1230 :     The fix involves
1231 :     1. generating FCOPYs in FSTP in ia32-svid
1232 :     2. marking a CALL with the appropriate annotation
1233 :    
1234 :     ----------------------------------------------------------------------
1235 : blume 961 Name: Matthias Blume
1236 :     Date: 2001/10/16 11:32:00 EDT
1237 :     Tag: blume-20011016-netbsd
1238 :     Description:
1239 :    
1240 :     Underscore patch from Chris Richards (fixing problem with compiling
1241 :     runtime system under recent NetBSD).
1242 :    
1243 :     ----------------------------------------------------------------------
1244 : george 959 Name: Allen Leung
1245 :     Date: 2001/10/12 17:18:32 EDT 2001
1246 :     Tag: leung-20011012-x86-printflowgraph
1247 :     Description:
1248 :    
1249 :     X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
1250 :    
1251 :     ----------------------------------------------------------------------
1252 : george 952 Name: Lal George
1253 : george 958 Date: 2001/10/11 23:51:34 EDT
1254 :     Tag: george-20011011-too-many-instrs
1255 :     Description:
1256 :    
1257 :     The representation of a program point never expected to see more
1258 :     than 65536 instructions in a basic block!
1259 :    
1260 :     ----------------------------------------------------------------------
1261 :     Name: Lal George
1262 : george 952 Date: 2001/10/09 09:41:37 EDT
1263 :     Tag: george-20011008-mlrisc-labels
1264 :     Description:
1265 :    
1266 :     Changed the machine description files to support printing of
1267 :     local and global labels in assembly code, based on host assembler
1268 :     conventions.
1269 :    
1270 :     ----------------------------------------------------------------------
1271 : blume 916 Name: Matthias Blume
1272 : blume 939 Date: 2001/09/25 15:25:00 EDT
1273 :     Tag: blume-20010925-exninfo
1274 :     Description:
1275 :    
1276 :     I provided a non-hook implementation of exnName (at the toplevel) and
1277 :     made the "dummy" implementation of exnMessage (at the toplevel) more
1278 :     useful: if nothing gets "hooked in", then at least you are going to
1279 :     see the exception name and a message indicating why you don't see more.
1280 :    
1281 :     [For the time being, programs that need exnMessage and want to use
1282 :     ml-build should either use General.exnMessage (strongly recommended) or
1283 :     refer to structure General at some other point so that CM sees a
1284 :     static dependency.]
1285 :    
1286 :     [Similar remarks go for "print" and "use": If you want to use their
1287 :     functionality in stand-alone programs generated by ml-build, then use
1288 :     TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
1289 :    
1290 :     ----------------------------------------------------------------------
1291 :     Name: Matthias Blume
1292 : blume 938 Date: 2001/09/20 17:28:00 EDT
1293 :     Tag: blume-20010920-slowfp
1294 :     Description:
1295 :    
1296 :     Allen says that x86-fast-fp is not safe yet, so I turned it off again...
1297 :    
1298 :     ----------------------------------------------------------------------
1299 :     Name: Matthias Blume
1300 : blume 937 Date: 2001/09/20 17:20:00 EDT
1301 :     Tag: blume-20010920-canonicalpaths
1302 :     Description:
1303 :    
1304 :     0. Updated the BOOT file (something that I forgot to do earlier).
1305 :    
1306 :     1. Small internal change to CM so that it avoids "/../" in filenames
1307 :     as much as possible (but only where it is safe).
1308 :    
1309 :     2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
1310 :     that arguments that contain delimiters are passed through correctly.
1311 :     This change also means that all "special" arguments of the form
1312 :     @SMLxxx... must come first.
1313 :    
1314 :     3. Changed install script to put relative anchor names for tool commands
1315 :     into pathconfig.
1316 :    
1317 :     ----------------------------------------------------------------------
1318 :     Name: Matthias Blume
1319 : blume 932 Date: 2001/09/18 15:35:00 EDT
1320 :     Tag: blume-20010918-readme11036
1321 :     Description:
1322 :    
1323 :     Added README files.
1324 :    
1325 :     ----------------------------------------------------------------------
1326 :     Name: Matthias Blume
1327 : blume 929 Date: 2001/09/18 11:45:00 EDT
1328 :     Tag: Release_110_36 (retag)
1329 :     Description:
1330 :    
1331 :     Fixed mistake in config/preloads. Retagged as 110.36.
1332 :    
1333 :     ----------------------------------------------------------------------
1334 :     Name: Matthias Blume
1335 : blume 928 Date: 2001/09/18 09:40:00 EDT
1336 : blume 929 Tag: Release_110_36_orig (tag changed)
1337 : blume 928 Description:
1338 :    
1339 :     New version (110.36). New bootfiles.
1340 :    
1341 :     ----------------------------------------------------------------------
1342 :     Name: Matthias Blume
1343 : blume 927 Date: 2001/09/14 16:15:00 EDT
1344 :     Tag: blume-20010914-x86fastfp
1345 :     Description:
1346 :    
1347 :     John committed some changes that Allen made, in particular a (hopefully)
1348 :     correctly working version of the x86-fp module.
1349 :    
1350 :     I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
1351 :     flag to "true". Everything seems to compile to a fixpoint ok, and
1352 :     "mandelbrot" speeds up by about 15%.
1353 :    
1354 :     ----------------------------------------------------------------------
1355 :     Name: Matthias Blume
1356 : blume 923 Date: 2001/09/13 11:20:00 EDT
1357 :     Tag: blume-20010913-minimal
1358 :     Description:
1359 :    
1360 :     1. Stefan Monnier's patch to fix a miscompilation problem that
1361 :     was brought to light by John Reppy's work on Moby.
1362 :    
1363 :     2. Implemented a minimal "structure Compiler" that contains just
1364 :     "version" and "architecture". The minimal version will be
1365 :     available when the full version is not. This is for backward-
1366 :     compatibility with code that wants to test Compiler.version.
1367 :    
1368 :     ----------------------------------------------------------------------
1369 :     Name: Matthias Blume
1370 : blume 920 Date: 2001/08/28 14:03:00 EDT
1371 :     Tag: blume-20010828-ml-lex
1372 :     Description:
1373 :    
1374 :     Fix for bug 1581, received from Neophytos Michael.
1375 :    
1376 :     ----------------------------------------------------------------------
1377 :     Name: Matthias Blume
1378 : blume 918 Date: 2001/08/27 11:20:00 EDT
1379 :     Tag: blume-20010827-readme11035
1380 :     Description:
1381 :    
1382 :     Fleshed out the README file for 110.35.
1383 :    
1384 :     ----------------------------------------------------------------------
1385 :     Name: Matthias Blume
1386 : blume 916 Date: 2001/08/24 17:10:00 EDT
1387 :     Tag: Release_110_35
1388 :     Description:
1389 :    
1390 :     New version number (110.35). New bootfiles.
1391 :    
1392 :     ----------------------------------------------------------------------
1393 : george 915 Name: Lal George
1394 : george 909 Date: 2001/08/24 13:47:18 EDT 2001
1395 :     Tag: george-20010824-MLRISC-graphs
1396 :     Description:
1397 :    
1398 :     removed clusters from MLRISC completely and replaced with graphs.
1399 :    
1400 :     ----------------------------------------------------------------------
1401 :     Name: Matthias Blume
1402 : blume 905 Date: 2001/08/23 17:50:00 EDT
1403 :     Tag: blume-20010823-toplevel
1404 :     Description:
1405 :    
1406 :     - some reorganization of the code that implements various kinds of
1407 :     environments in the compiler (static, dynamic, symbolic, combined)
1408 :     - re-implemented the EnvRef module so that evalStream works properly
1409 :     (if the stream contains references to "use", "CM.make", etc.)
1410 :     - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
1411 :    
1412 :     ----------------------------------------------------------------------
1413 :     Name: Matthias Blume
1414 : blume 904 Date: 2001/08/20 15:50 EDT
1415 :     Tag: blume20010820-slipup
1416 :     Description:
1417 :    
1418 :     I forgot to commit a few files. Here they are...
1419 :    
1420 :     ----------------------------------------------------------------------
1421 :     Name: Matthias Blume
1422 : blume 903 Date: 2001/08/20 15:35:00 EDT
1423 :     Tag: blume-20010820-debugprof
1424 :     Description:
1425 :    
1426 :     !!!! NEW BOOTFILES !!!!
1427 :    
1428 :     This is another round of reorganizing the compiler sources. This
1429 :     time the main goal was to factor out all the "instrumentation"
1430 :     passes (for profiling and backtracing) into their own library.
1431 :     The difficulty was to do it in such a way that it does not depend
1432 :     on elaborate.cm but only on elabdata.cm.
1433 :    
1434 :     Therefore there have been further changes to both elaborate.cm and
1435 :     elabdata.cm -- more "generic" things have been moved from the former
1436 :     to the latter. As a result, I was forced to split the assignment
1437 :     of numbers indicating "primtyc"s into two portions: SML-generic and
1438 :     SML/NJ-specific. Since it would have been awkward to maintain,
1439 :     I bit the bullet and actually _changed_ the mapping between these
1440 :     numbers and primtycs. The bottom line of this is that you need
1441 :     a new set of bin- and bootfiles.
1442 :    
1443 :     I have built new bootfiles for all architectures, so doing a fresh
1444 :     checkout and config/install.sh should be all you need.
1445 :    
1446 :     The newly created library's name is
1447 :    
1448 :     $smlnj/viscomp/debugprof.cm
1449 :    
1450 :     and its sources live under
1451 :    
1452 :     src/compiler/DebugProf
1453 :    
1454 :     ----------------------------------------------------------------------
1455 :     Name: Matthias Blume
1456 : blume 902 Date: 2001/08/15 17:15:00 EDT
1457 :     Tag: blume-20010815-compreorg
1458 :     Description:
1459 :    
1460 :     This is a first cut at reorganizing the CM libraries that make up the
1461 :     core of the compiler. The idea is to separate out pieces that could
1462 :     be used independently by tools, e.g., the parser, the typechecker, etc.
1463 :    
1464 :     The current status is a step in this direction, but it is not quite
1465 :     satisfactory yet. Expect more changes in the future.
1466 :    
1467 :     Here is the current (new) organization...
1468 :    
1469 :     What used to be $smlnj/viscomp/core.cm is now divided into
1470 :     six CM libraries:
1471 :    
1472 :     $smlnj/viscomp/basics.cm
1473 :     /parser.cm
1474 :     /elabdata.cm
1475 :     /elaborate.cm
1476 :     /execute.cm
1477 :     /core.cm
1478 :    
1479 :     The CM files for these libraries live under src/system/smlnj/viscomp.
1480 :     All these libraries are proxy libraries that contain precisely
1481 :     one CM library component. Here are the locations of the components
1482 :     (all within the src/compiler tree):
1483 :    
1484 :     Basics/basics.cm
1485 :     Parse/parser.cm
1486 :     ElabData/elabdata.cm
1487 :     Elaborator/elaborate.cm
1488 :     Execution/execute.cm
1489 :     core.cm
1490 :    
1491 :     [This organization is the same that has been used already
1492 :     for a while for the architecture-specific parts of the visible
1493 :     compiler and for the old version of core.cm.]
1494 :    
1495 :     As you will notice, many source files have been moved from their
1496 :     respective original locations to a new home in one of the above
1497 :     subtrees.
1498 :    
1499 :     The division of labor between the new libraries is the following:
1500 :    
1501 :     basics.cm:
1502 :     - Simple, basic definitions that pertain to many (or all) of
1503 :     the other libraries.
1504 :     parser.cm:
1505 :     - The SML parser, producing output of type Ast.dec.
1506 :     - The type family for Ast is also defined and exported here.
1507 :     elabdata.cm:
1508 :     - The datatypes that describe input and output of the elaborator.
1509 :     This includes types, absyn, and static environments.
1510 :     elaborator.cm:
1511 :     - The SML/NJ type checker and elaborator.
1512 :     This maps an Ast.dec (with a given static environment) to
1513 :     an Absyn.dec (with a new static environment).
1514 :     - This libraries implements certain modules that used to be
1515 :     structures as functors (to remove dependencies on FLINT).
1516 :     execute.cm:
1517 :     - Everything having to do with executing binary code objects.
1518 :     - Dynamic environments.
1519 :     core.cm:
1520 :     - SML/NJ-specific instantiations of the elaborator and MLRISC.
1521 :     - Top-level modules.
1522 :     - FLINT (this should eventually become its own library)
1523 :    
1524 :     Notes:
1525 :    
1526 :     I am not 100% happy with the way I separated the elaborator (and its
1527 :     data structures) from FLINT. Two instances of the same problem:
1528 :    
1529 :     1. Data structures contain certain fields that carry FLINT-specific
1530 :     information. I hacked around this using exn and the property list
1531 :     module from smlnj-lib. But the fact that there are middle-end
1532 :     specific fields around at all is a bit annoying.
1533 :    
1534 :     2. The elaborator calculates certain FLINT-related information. I tried
1535 :     to make this as abstract as I could using functorization, but, again,
1536 :     the fact that the elaborator has to perform calculations on behalf
1537 :     of the middle-end at all is not nice.
1538 :    
1539 :     3. Having to used exn and property lists is unfortunate because it
1540 :     weakens type checking. The other alternative (parameterizing
1541 :     nearly *everything*) is not appealing, though.
1542 :    
1543 :     I removed the "rebinding =" warning hack because due to the new organization
1544 :     it was awkward to maintain it. As a result, the compiler now issues some of
1545 :     these warnings when compiling init.cmi during bootstrap compilation. On
1546 :     the plus side, you also get a warning when you do, for example:
1547 :     val op = = Int32.+
1548 :     which was not the case up to now.
1549 :    
1550 :     I placed "assign" and "deref" into the _Core structure so that the
1551 :     code that deals with the "lazy" keyword can find them there. This
1552 :     removes the need for having access to the primitive environment
1553 :     during elaboration.
1554 :    
1555 :     ----------------------------------------------------------------------
1556 :     Name: Matthias Blume
1557 : blume 898 Date: 2001/08/13
1558 :     Tag: blume-20010813-closures
1559 :     Description:
1560 :    
1561 :     This fix was sent to us by Zhong Shao. It is supposed to improve the
1562 :     performance of certain loops by avoiding needless closure allocation.
1563 :    
1564 :     ----------------------------------------------------------------------
1565 : george 890 Name: Lal George
1566 : george 896 Date: 2001/07/31 10:03:23 EDT 2001
1567 :     Tag: george-20010731-x86-fmalloc
1568 :     Description: Fixed bug in x86 calls
1569 :    
1570 :     There was a bug where call instructions would mysteriously
1571 :     vanish. The call instruction had to be one that returned
1572 :     a floating point value.
1573 :    
1574 :     ----------------------------------------------------------------------
1575 :     Name: Lal George
1576 : george 890 Date: 2001/07/19 16:36:29 EDT 2001
1577 :     Tag: george-20010719-simple-cells
1578 :     Description:
1579 :    
1580 :     I have dramatically simplified the interface for CELLS in MLRISC.
1581 :    
1582 :     In summary, the cells interface is broken up into three parts:
1583 :    
1584 :     1. CellsBasis : CELLS_BASIS
1585 :    
1586 :     CellsBasis is a top level structure and common for all
1587 :     architectures. it contains the definitions of basic datatypes
1588 :     and utility functions over these types.
1589 :    
1590 :     2. functor Cells() : CELLS
1591 :    
1592 :     Cells generates an interface for CELLS that incorporates the
1593 :     specific resources on the target architecture, such as the
1594 :     presence of special register classes, their number and size,
1595 :     and various useful substructures.
1596 :    
1597 :     3. <ARCH>CELLS
1598 :    
1599 :     e.g. SparcCells: SPARCCELLS
1600 :    
1601 :     <ARCH>CELLS usually contains additional bindings for special
1602 :     registers on the architecture, such as:
1603 :    
1604 :     val r0 : cell (* register zero *)
1605 :     val y : cell (* Y register *)
1606 :     val psr : cell (* processor status register *)
1607 :     ...
1608 :    
1609 :     The structure returned by applying the Cells functor is opened
1610 :     in this interface.
1611 :    
1612 :     The main implication of all this is that the datatypes for cells is
1613 :     split between CellsBasis and CELLS -- a fairly simple change for user
1614 :     code.
1615 :    
1616 :     In the old scheme the CELLS interface had a definitional binding of
1617 :     the form:
1618 :    
1619 :     signature CELLS = sig
1620 :    
1621 :     structure CellsBasis = CellsBasis
1622 :    
1623 :     ...
1624 :    
1625 :     end
1626 :    
1627 :     With all the sharing constraints that goes on in MLRISC, this old
1628 :     design quickly leads to errors such as:
1629 :    
1630 :     "structure definition spec inside of sharing ... "
1631 :    
1632 :    
1633 :     and appears to require an unacceptable amount of sharing and where
1634 :     constraint hackery.
1635 :    
1636 :     I think this error message (the interaction of definitional specs and
1637 :     sharing) requires more explanation on our web page.
1638 :    
1639 :     ----------------------------------------------------------------------
1640 : blume 826 Name: Matthias Blume
1641 : blume 879 Date: 2001/07/19 15:00:00 EDT
1642 :     Tag: blume-20010719-libreorg
1643 :     Description:
1644 :    
1645 :     This update puts together a fairly extensive but straightforward change
1646 :     to the way the libraries that implement the interactive system are
1647 :     organized:
1648 :    
1649 :     The biggest change is the elimination of structure Compiler. As a
1650 :     replacement for this structure, there is now a CM library
1651 :     (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1652 :     that exports all the substructures of the original structure Compiler
1653 :     directly. So instead of saying Compiler.Foo.bar one now simply
1654 :     says Foo.bar. (The CM libraries actually export a collection of
1655 :     structures that is richer than the collection of substructures of
1656 :     structure Compiler.)
1657 :    
1658 :     To make the transition smooth, there is a separate library called
1659 :     $smlnj/compiler/compiler.cm which puts together and exports the
1660 :     original structure Compiler (or at least something very close to it).
1661 :    
1662 :     There are five members of the original structure Compiler
1663 :     that are not exported directly but which instead became members
1664 :     of a new structure Backend (described by signature BACKEND). These are:
1665 :     structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1666 :     Interact (: INTERACT), structure Machine (: MACHINE), and val
1667 :     architecture (: string).
1668 :    
1669 :     Structure Compiler.Version has become structure CompilerVersion.
1670 :    
1671 :     Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1672 :     by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1673 :     or x86, respectively.
1674 :     Each of these exports the same frontend structures that
1675 :     $smlnj/compiler.cm exports. But they do not have a structure Backend
1676 :     and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1677 :     Hppa, PPC, Sparc, or X86, respectively.
1678 :    
1679 :     Library $smlnj/compiler/all.cm exports the union of the exports of
1680 :     $smlnj/compiler/<arch>.cm
1681 :    
1682 :     There are no structures <Arch>Compiler anymore, use
1683 :     $smlnj/compiler/<arch>.cm instead.
1684 :    
1685 :     Library host-compiler-0.cm is gone. Instead, the internal library
1686 :     that instantiates CM is now called cm0.cm. Selection of the host
1687 :     compiler (backend) is no longer done here but. (Responsibility for it
1688 :     now lies with $smlnj/compiler/current.cm. This seems to be more
1689 :     logical.)
1690 :    
1691 :     Many individual files have been moved or renamed. Some files have
1692 :     been split into multiple files, and some "dead" files have been deleted.
1693 :    
1694 :     Aside from these changes to library organization, there are also changes
1695 :     to the way the code itself is organized:
1696 :    
1697 :     Structure Binfile has been re-implemented in such a way that it no
1698 :     longer needs any knowledge of the compiler. It exclusively deals
1699 :     with the details of binfile layout. It no longer invokes the
1700 :     compiler (for the purpose of creating new prospective binfile
1701 :     content), and it no longer has any knowledge of how to interpret
1702 :     pickles.
1703 :    
1704 :     Structure Compile (: COMPILE) has been stripped down to the bare
1705 :     essentials of compilation. It no longer deals with linking/execution.
1706 :     The interface has been cleaned up considerably.
1707 :    
1708 :     Utility routines for dealing with linking and execution have been
1709 :     moved into their own substructures.
1710 :    
1711 :     (The ultimate goal of these changes is to provide a light-weight
1712 :     binfile loader/linker (at least for, e.g., stable libraries) that
1713 :     does not require CM or the compiler to be present.)
1714 :    
1715 :     CM documentation has been updated to reflect the changes to library
1716 :     organization.
1717 :    
1718 :     ----------------------------------------------------------------------
1719 :     Name: Matthias Blume
1720 : blume 875 Date: 2001/07/10 17:30:00 EDT
1721 :     Tag: Release_110_34
1722 :     Description:
1723 :    
1724 : blume 876 Minor tweak to 110.34 (re-tagged):
1725 :    
1726 :     - README.html file added to CVS repository
1727 :     - runtime compiles properly under FreeBSD 3.X and 4.X
1728 :    
1729 :     ----------------------------------------------------------------------
1730 :     Name: Matthias Blume
1731 :     Date: 2001/07/10 17:30:00 EDT
1732 :     Tag: Release_110_34
1733 :     Description:
1734 :    
1735 : blume 875 New version number (110.34). New bootfiles.
1736 :    
1737 :     ----------------------------------------------------------------------
1738 :     Name: Matthias Blume
1739 : blume 874 Date: 2001/07/09 16:00:00 EDT
1740 :     Tag: blume-20010709-more-varargs
1741 :     Description:
1742 :    
1743 :     I changed the handling of varargs in ml-nlffigen again:
1744 :     The ellipsis ... will now simply be ignored (with an accompanying warning).
1745 :    
1746 :     The immediate effect is that you can actually call a varargs function
1747 :     from ML -- but you can't actually supply any arguments beyond the ones
1748 :     specified explicitly. (For example, you can call printf with its format
1749 :     string, but you cannot pass additional arguments.)
1750 :    
1751 :     This behavior is only marginally more useful than the one before, but
1752 :     it has the advantage that a function or, more importantly, a function
1753 :     type never gets dropped on the floor, thus avoiding follow-up problems with
1754 :     other types that refer to the offending one.
1755 :    
1756 :     ----------------------------------------------------------------------
1757 :     Name: Matthias Blume
1758 : blume 873 Date: 2001/07/09 11:25:00 EDT
1759 :     Tag: blume-20010709-varargs
1760 :     Description:
1761 :    
1762 :     1. ckit-lib.cm now exports structure Error
1763 :     2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1764 :     with a warning accompanied by a source location. Moreover, it
1765 :     merely skips the offending function or type and proceeds with the
1766 :     rest of its work.u As a result, one can safely feed C code containing
1767 :     "..." to ml-nlffigen.
1768 :     3. There are some internal improvements to CM, providing slightly
1769 :     more general string substitutions in the tools subsystem.
1770 :    
1771 :     ----------------------------------------------------------------------
1772 :     Name: Matthias Blume
1773 : blume 854 Date: 2001/06/27 15:10:00 EDT
1774 :     Tag: blume-20010627-concur
1775 :     Description:
1776 :    
1777 :     Fixed a small bug in CM's handling of parallel compilation.
1778 :     (You could observe the bug by Control-C-interrupting an ordinary
1779 :     CMB.make or CM.stabilize and then attaching some compile servers.
1780 :     The result was that all of a sudden the previously interrupted
1781 :     compilation would continue on its own. This was because of
1782 :     an over-optimization: CM did not bother to clean out certain queues
1783 :     when no servers were attached "anyway", resulting in the contents
1784 :     of these queues to grab control when new servers did get attached.)
1785 :    
1786 :     There is also another minor update to the CM manual.
1787 :    
1788 :     ----------------------------------------------------------------------
1789 :     Name: Matthias Blume
1790 : blume 853 Date: 2001/06/26 16:15:00 EDT
1791 :     Tag: blume-20010626-cmdoc
1792 :     Description:
1793 :    
1794 :     Minor typo fixed in CM manual (syntax diagram for libraries).
1795 :    
1796 :     ----------------------------------------------------------------------
1797 :     Name: Matthias Blume
1798 : blume 852 Date: 2001/06/25 22:55:00 EDT
1799 :     Tag: blume-20010625-x86pc
1800 :     Description:
1801 :    
1802 :     Fixed a nasty bug in the X86 assembly code that caused signal
1803 :     handlers to fail (crash) randomly.
1804 :    
1805 :     ----------------------------------------------------------------------
1806 :     Name: Matthias Blume
1807 : blume 846 Date: 2001/06/25 12:05:00 EDT
1808 :     Tag: blume-20010625-nlffigen
1809 :     Description:
1810 :    
1811 :     This update fixes a number of minor bugs in ml-nlffigen as reported by
1812 :     Nick Carter <nbc@andrew.cmu.edu>.
1813 :    
1814 :     1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1815 :     2. Default names for generated files are now derived from the name of
1816 :     the C file *without its directory*. In particular, this causes generated
1817 :     files to be placed locally even if the C file is in some system directory.
1818 :     3. Default names for generated signatures and structures are also derived
1819 :     from the C file name without its directory. This avoids silly things
1820 :     like "structure GL/GL".
1821 :     (Other silly names are still possible because ml-nlffigen does not do
1822 : blume 847 a thorough check of whether generated names are legal ML identifiers.
1823 :     When in doubt, use command line arguments to force particular names.)
1824 : blume 846
1825 :     ----------------------------------------------------------------------
1826 :     Name: Matthias Blume
1827 : blume 845 Date: 2001/06/21 12:25:00 EDT
1828 :     Tag: blume-20010621-eXene
1829 :     Description:
1830 :    
1831 :     eXene now compiles and (sort of) works again.
1832 :    
1833 :     The library name (for version > 110.33) is $/eXene.cm.
1834 :    
1835 :     I also added an new example in src/eXene/examples/nbody. See the
1836 :     README file there for details.
1837 :    
1838 :     ----------------------------------------------------------------------
1839 :     Name: Matthias Blume
1840 : blume 844 Date: 2001/06/20 16:40:00 EDT
1841 :     Tag: blume-20010620-cml
1842 :     Description:
1843 :    
1844 :     CML now compiles and works again.
1845 :    
1846 :     Libraries (for version > 110.33):
1847 :    
1848 :     $cml/cml.cm Main CML library.
1849 :     $cml/basis.cm CML's version of $/basis.cm.
1850 :     $cml/cml-internal.cm Internal helper library.
1851 :     $cml/core-cml.cm Internal helper library.
1852 :     $cml-lib/trace-cml.cm Tracing facility.
1853 :     $cml-lib/smlnj-lib.cm CML's version of $/smlnj-lib.cm
1854 :    
1855 :     The installer (config/install.sh) has been taught how to properly
1856 :     install this stuff.
1857 :    
1858 :     ----------------------------------------------------------------------
1859 :     Name: Matthias Blume
1860 : blume 843 Date: 2001/06/19 17:55:00 EDT
1861 :     Tag: blume-20010619-instantiate
1862 :     Description:
1863 :    
1864 :     This un-breaks the fix for bug 1432.
1865 :     (The bug was originally fixed in 110.9 but I broke it again some
1866 :     time after that.)
1867 :    
1868 :     ----------------------------------------------------------------------
1869 :     Name: Matthias Blume
1870 : blume 842 Date: 2001/06/19 17:25:00 EDT
1871 :     Tag: blume-20010619-signals
1872 :     Description:
1873 :    
1874 :     This should (hopefully) fix the long-standing signal handling bug.
1875 :     (The runtime system was constructing a continuation record with an
1876 :     incorrect descriptor which would cause the GC to drop data on the floor...)
1877 :    
1878 :     ----------------------------------------------------------------------
1879 :     Name: Matthias Blume
1880 : blume 840 Date: 2001/06/15 15:05:00 EDT
1881 : blume 841 Tag: blume-20010615-moresparc
1882 :     Description:
1883 :    
1884 :     Here is a short late-hour update related to Sparc c-calls:
1885 :    
1886 :     -- made handling of double-word arguments a bit smarter
1887 :    
1888 :     -- instruction selection phase tries to collapse certain clumsily
1889 :     constructed ML-Trees; typical example:
1890 :    
1891 :     ADD(ty,ADD(_,e,LI d1),LI d2) -> ADD(ty,e,LI(d1+d2))
1892 :    
1893 :     This currently has no further impact on SML/NJ since mlriscGen does
1894 :     not seem to generate such patterns in the first place, and c-calls
1895 :     (which did generate them in the beginning) has meanwhile been fixed
1896 :     so as to avoid them as well.
1897 :    
1898 :     ----------------------------------------------------------------------
1899 :     Name: Matthias Blume
1900 :     Date: 2001/06/15 15:05:00 EDT
1901 : blume 840 Tag: blume-20010615-sparc
1902 :     Description:
1903 :    
1904 :     The purpose of this update is to provide an implementation of NLFFI
1905 :     on Sparc machines.
1906 :    
1907 :     Here are the changes in detail:
1908 :    
1909 :     * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1910 :     the Sparc implementation of the c-calls API.
1911 :     * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1912 :     for accessing the ML frame. Thus, we have a real frame pointer and
1913 :     can freely modify %sp without need for an omit-frame-ptr phase.
1914 :     The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1915 :     this case.
1916 :     * ml-nlffigen has been taught to produce code for different architectures
1917 :     and calling conventions.
1918 :     * In a way similar to what was done in the x86 case, the Sparc
1919 :     backend uses its own specific extension to mltree. (For example,
1920 :     it needs to be able to generate UNIMP instructions which are part
1921 :     of the calling convention.)
1922 :     * ml-nlffi-lib was reorganized to make it more modular (in particular,
1923 :     to make it easier to plug in new machine- and os-dependent parts).
1924 :    
1925 :     There are some other fairly unrelated bug fixes and cleanups as well:
1926 :    
1927 :     * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1928 :     that they properly share their libraries with existing SML/NJ libraries.
1929 :     * I fixed a minor cosmetic bug in CM, supressing certain spurious
1930 :     follow-up error messages.
1931 :     * Updates to CM/CMB documentation.
1932 :    
1933 :     TODO items:
1934 :    
1935 :     * MLRISC should use a different register as its asmTemp on the Sparc.
1936 :     (The current %o2 is a really bad choice because it is part of the
1937 :     calling conventions, so things might interfere in unexpected ways.)
1938 :    
1939 :     ----------------------------------------------------------------------
1940 :     Name: Matthias Blume
1941 : blume 839 Date: 2001/06/07
1942 :     Tag: blume-20010607-calls
1943 :     Description:
1944 :    
1945 :     A number of internal changes related to C calls and calling conventions:
1946 :    
1947 :     1. ML-Tree CALL statements now carry a "pops" field. It indicates the
1948 :     number of bytes popped implicitly (by the callee). In most cases
1949 :     this field is 0 but on x86/win32 it is some non-zero value. This
1950 :     is information provided for the benefit of the "omit-frameptr" pass.
1951 :     2. The CALL instruction on the x86 carries a similar "pops" field.
1952 :     The instruction selection phase copies its value from the ML-Tree
1953 :     CALL statement.
1954 :     3. On all other architectures, the instruction selection phase checks
1955 :     whether "pops=0" and complains if not.
1956 :     4. The c-calls implementation for x86 now accepts two calling conventions:
1957 :     "ccall" and "stdcall". When "ccall" is selected, the caller cleans
1958 :     up after the call and pops is set to 0. For "stdcall", the caller
1959 :     does nothing, leaving the cleanup to the callee; pops is set to
1960 :     the number of bytes that were pushed onto the stack.
1961 :     5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1962 :     distinguish between "ccall" and "stdcall".
1963 :     6. The UNIMP instruction has been added to the supported Sparc instruction
1964 :     set. (This is needed for implementing the official C calling convention
1965 :     on this architecture.)
1966 :     7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1967 :     work with the latest CM.
1968 :    
1969 :     ----------------------------------------------------------------------
1970 :     Name: Matthias Blume
1971 : blume 838 Date: 2001/06/05 15:10:00 EDT
1972 :     Tag: blume-20010605-cm-index
1973 :     Description:
1974 :    
1975 :     0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1976 :    
1977 :     1. CM can now generate "index files". These are human-readable files
1978 :     that list on a per-.cm-file basis each toplevel symbol defined or
1979 :     imported. The location of the index file for
1980 :     <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1981 :     To enable index-file generation, set CM.Control.generate_index to true
1982 :     or export an environment-symbol: export CM_GENERATE_INDEX=true.
1983 :    
1984 :     The CM manual has been updated accordingly.
1985 :    
1986 :     2. I made some slight modifications to the c-calls API in MLRISC.
1987 :    
1988 :     a) There is now a callback to support saving/restoring of
1989 :     dedicated but caller-save registers around the actual call
1990 :     instruction.
1991 :     b) One can optionally specify a comment-annotation for the
1992 :     call instruction.
1993 :    
1994 :     3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1995 :     (For example, the comment annotation shows the C prototype of
1996 :     the function being called.)
1997 :    
1998 :     ----------------------------------------------------------------------
1999 :     Name: Matthias Blume
2000 : blume 837 Date: 2001/06/01 13:30:00 EDT
2001 :     Tag: blume-20010601-nlffi-cleanup
2002 :     Description:
2003 :    
2004 :     This is mostly a cleanup of MLFFI stuff:
2005 :    
2006 :     - some signature files have been put into a more exposed place
2007 :     - the ugly 'f type parameter is gone (simplifies types tremendously!)
2008 :     - ml-nlffigen changed accordingly
2009 :     - tutorial updated
2010 :    
2011 :     Other changes:
2012 :    
2013 :     - author's affiliation in CM manual(s) updated
2014 :     - some more recognized keywords added to Allen's sml.sty
2015 :    
2016 :     ----------------------------------------------------------------------
2017 :     Name: Matthias Blume
2018 : blume 836 Date: 2001/05/25 15:30:00 EDT
2019 :     Tag: blume-20010525-iptr
2020 :     Description:
2021 :    
2022 :     - put the official 110.33-README (as it appears on the ftp server) under
2023 :     CVS
2024 :     - fixed a small bug related to incomplete pointer types in
2025 :     ml-nlffigen
2026 :     - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
2027 :     (it does not need the 'f type parameter)
2028 :    
2029 :     ----------------------------------------------------------------------
2030 :     Name: Matthias Blume
2031 : blume 834 Date: 2001/05/23 14:30:00 EDT
2032 :     Tag: Release_110_33
2033 :     Description:
2034 :    
2035 :     New version number (110.33). New bootfiles.
2036 :    
2037 :     ----------------------------------------------------------------------
2038 :     Name: Matthias Blume
2039 : blume 833 Date: 2001/05/22 18:06:00 EDT
2040 :     Tag: blume-20010522-targets
2041 :     Description:
2042 :    
2043 :     Made install.sh use file config/targets.customized if it exists, falling
2044 :     back to config/targets if it doesn't. This way one can have a customized
2045 :     version of the targets file without touching the "real thing", thus
2046 :     eliminating the constant fear of accidentally checking something bogus
2047 :     back into the CVS repository... (File config/targets.customized must
2048 :     not be added to the repository!)
2049 :    
2050 :     ----------------------------------------------------------------------
2051 :     Name: Matthias Blume
2052 : blume 832 Date: 2001/05/22 16:30:00 EDT
2053 :     Tag: blume-20010522-minitut
2054 :     Description:
2055 :    
2056 :     1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
2057 :     struct returns.
2058 :     2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt. This is some very
2059 :     incomplete, preliminary documentation for NLFFI.
2060 :    
2061 :     ----------------------------------------------------------------------
2062 :     Name: Matthias Blume
2063 : blume 829 Date: 2001/05/14 11:30:00 EDT
2064 :     Tag: blume-20010514-script
2065 :     Description:
2066 :    
2067 :     Some bugs in install script fixed.
2068 :    
2069 :     In addition to that I also made a slight change to the NLFFI API:
2070 :     Functors generated by ml-nlffigen now take the dynamic library as a
2071 :     straight functor argument, not as a suspended one. (The original
2072 :     functor code used to force the suspension right away anyway, so there
2073 :     was nothing gained by this complication of the interface.)
2074 :    
2075 :     ----------------------------------------------------------------------
2076 :     Name: Matthias Blume
2077 : blume 828 Date: 2001/05/11 14:35:00 EDT
2078 :     Tag: blume-20010511-ml-nlffi
2079 :     Description:
2080 :    
2081 :     I finally took the plunge and added my new FFI code to the main
2082 :     repository. For x86-linux it is now ready for prime-time.
2083 :    
2084 :     There are two new subdirectories of "src":
2085 :    
2086 :     - ml-nlffi-lib:
2087 :     The utility library for programs using the FFI interface.
2088 :     Here is the implementation of $/c.cm and its associated low-level
2089 :     partners $/c-int.cm and $/memory.cm.
2090 :     - ml-nlffigen:
2091 :     A stand-alone program for generating ML glue code from C source
2092 :     code.
2093 :    
2094 :     Building ml-nlffigen requires $/ckit-lib.cm.
2095 :    
2096 :     The config/install.sh script has been updates to do the Right Thing
2097 :     (hopefully).
2098 :    
2099 :     Notice that the source tree for the C-Kit will not be put under "src"
2100 :     but directly under the installation root directory. (This is the
2101 :     structure that currently exists on the CVS server when you check out
2102 :     module "sml".) Fortunately, config/install.sh knows about this oddity.
2103 :    
2104 :     Bugs: No documentation yet.
2105 :    
2106 :     ----------------------------------------------------------------------
2107 :     Name: Matthias Blume
2108 : blume 826 Date: 2001/05/09 16:35:00 EDT
2109 :     Tag: blume-20010509-cpscontract
2110 :     Description:
2111 :    
2112 :     Fixed a bug in the accounting code in cpsopt/contract.sml. (The
2113 :     wrapper/unwrapper elimination did not decrement usage counts and some
2114 :     dead variables got overlooked by the dead-up logic.)
2115 :    
2116 :     ----------------------------------------------------------------------
2117 : george 824 Name: Lal George
2118 :     Date: 2001/05/08 17:26:09 EDT
2119 :     Tag: george-20010508-omit-frameptr
2120 :     Description:
2121 :    
2122 :     Changes to implement the omit-frame-pointer optimization to support
2123 :     raw C calls. For now, there is only support on the Intel x86, but
2124 :     other architectures will follow as more experience is gained with this.
2125 :    
2126 :    
2127 :     ----------------------------------------------------------------------
2128 : blume 818 Name: Matthias Blume
2129 : blume 819 Date: 2001/05/07 14:40:00 EDT
2130 :     Tag: blume-20010507-proxies
2131 :     Description:
2132 :    
2133 :     I made into "proxy libraries" all libraries that qualify for such a
2134 :     change. (A qualifying library is a library that has another library or
2135 :     groups as its sole member and repeats that member's export list
2136 :     verbatim. A proxy library avoids this repetition by omitting its export
2137 :     list, effectively inheriting the list that its (only) member exports.
2138 :     See the CM manual for more explanation.)
2139 :     The main effect is that explicit export lists for these libraries
2140 :     do not have to be kepts in sync, making maintenance a bit easier.
2141 :    
2142 :     I also added copyright notices to many .cm-files.
2143 :    
2144 :     Last but not least, I made a new set of bootfiles.
2145 :    
2146 :     ----------------------------------------------------------------------
2147 :     Name: Matthias Blume
2148 : blume 818 Date: 2001/05/04 17:00:00 EDT
2149 :     Tag: blume-20010504-cm-lsplit
2150 :     Description:
2151 :    
2152 :     0. John merged pending changes to $/smlnj-lib.cm
2153 :    
2154 :     1. Allen's previous change accidentally backed out of one of Lal's
2155 :     earlier changes. I undid this mistake (re-introducing Lal's change).
2156 :    
2157 :     2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
2158 :     within the compiler where applicable. There is some code simplification
2159 :     because of that.
2160 :    
2161 :     3. The "split" phase (in FLINT) is now part of the default list of phases.
2162 :     Compiler.Control.LambdaSplitting.* can be used to globally control the
2163 :     lambda-splitting (cross-module-inlining) engine. In addition to that,
2164 :     it can now also be controlled on a per-source basis: CM has been taught
2165 :     a new tool parameter applicable to ML source files.
2166 :    
2167 :     - To turn lambda-splitting off completely:
2168 :     local open Compiler.Control.LambdaSplitting in
2169 :     val _ = set Off
2170 :     end
2171 :     - To make "no lambda-splitting" the global default (but allow per-source
2172 :     overriding); this is the initial setting:
2173 :     local open Compiler.Control.LambdaSplitting in
2174 :     val _ = set (Default NONE)
2175 :     end
2176 :     - To make "lambda-splitting with aggressiveness a" the global default
2177 :     (and allow per-source overriding):
2178 :     local open Compiler.Control.LambdaSplitting in
2179 :     val _ = set (Default (SOME a))
2180 :     end
2181 :    
2182 :     - To turn lambda-splitting off for a given ML souce file (say: a.sml)
2183 :     write (in the respective .cm-file):
2184 :     a.sml (lambdasplitting:off)
2185 :     - To turn lambda-splitting for a.sml on with minimal aggressiveness:
2186 :     a.sml (lambdasplitting:on)
2187 :     - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
2188 :     <a> is a decimal non-negative integer):
2189 :     a.sml (lambdasplitting:<a>)
2190 :     - To turn lambda-splitting for a.sml on with maximal aggressiveness:
2191 :     a.sml (lambdasplitting:infinity)
2192 :     - To use the global default for a.sml:
2193 :     a.sml (lambdasplitting:default)
2194 :     or simply
2195 :     a.sml
2196 :    
2197 :     ----------------------------------------------------------------------
2198 : leunga 815 Name: Allen Leung
2199 :     Date: 2001/05/04 01:57:00 EDT
2200 :     Tag: leunga-20010504-sync
2201 :     Description:
2202 : leunga 788
2203 : leunga 815 MLRISC features.
2204 :    
2205 :     1. Fix to CMPXCHG instructions.
2206 :     2. Changed RA interface to allow annotations in callbacks.
2207 :     3. Added a new method to the stream interface to allow annotations updates.
2208 :    
2209 : george 749 ----------------------------------------------------------------------
2210 : blume 812 Name: Matthias Blume
2211 : blume 814 Date: 2001/05/01 11:45:00 EDT
2212 :     Tag: blume-20010501-pcedittmp
2213 :     Description:
2214 :    
2215 :     Changed install.sh to use the current working directory instead of
2216 :     /usr/tmp for a temporary file (pcedittmp). The previous choice
2217 :     of /usr/tmp caused trouble with MacOS X because of file premission
2218 :     problems.
2219 :    
2220 :     ----------------------------------------------------------------------
2221 :     Name: Matthias Blume
2222 : blume 812 Date: 2001/04/20 11:10:00 EDT
2223 :     Tag: blume-20010420-inMLflag
2224 :     Description:
2225 :    
2226 :     - added vp_limitPtrMask to vproc-state.h
2227 :     (for use by the raw-C-calls mechanism to implement proper interrupt
2228 :     handling)
2229 :     - made the ML compiler aware of various data-structure offsets so it
2230 :     can generate code for accessing the vp_inML flag and vp_limitPtrMask
2231 :     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
2232 :     raw C-calls
2233 :    
2234 :     ----------------------------------------------------------------------
2235 : george 810 Name: Lal George
2236 :     Date: 2001/04/20 09:15:28 EDT
2237 :     Tag: george-20010420-macosX
2238 :     Description:
2239 :    
2240 :     - Changes to port to Mac OS X; Darwin.
2241 :    
2242 :     - In the process I found that sqrt was broken on the PPC, because the
2243 :     fsqrt instruction is not implemented.
2244 :    
2245 :     ----------------------------------------------------------------------
2246 : blume 799 Name: Matthias Blume
2247 : blume 808 Date: 2001/04/18 12:45:00 EDT
2248 :     Tag: blume-20010418-ccalls
2249 :     Description:
2250 :    
2251 :     - fixed two off-by-4 errors in the x86-specific c-calls implementation
2252 :     (this bug prevented structure arguments containing pointers from being
2253 :     passed correctly)
2254 :     - changed the raw-C-call code in mlriscGen.sml in such a way that
2255 :     structure arguments are represented as a pointer to the beginning
2256 :     of the structure (instead of having a series of synthesized arguments,
2257 :     one for each structure member)
2258 :    
2259 :     - made makeml script's verbosity level configurable via environment
2260 :     variable (MAKEML_VERBOSITY)
2261 :    
2262 : blume 813 - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
2263 : blume 808 in rawmem-x86.sml; we are now using the real thing
2264 :    
2265 :     ----------------------------------------------------------------------
2266 :     Name: Matthias Blume
2267 : blume 806 Date: 2001/03/22 16:25:00 EST
2268 :     Tag: blume-20010322-bootfiles
2269 :     Description:
2270 :    
2271 :     Created a new set of bootfiles (for your automatic installation convenience).
2272 :    
2273 :     ----------------------------------------------------------------------
2274 :     Name: Matthias Blume
2275 : blume 805 Date: 2001/03/22 15:10:00 EST
2276 :     Tag: blume-20010322-rawmem-parcm
2277 :     Description:
2278 :    
2279 :     1. All "raw memory access" primitives for the new FFI are implemented now
2280 :     (at least on the x86).
2281 :     2. Some further cleanup of CM's parallel make mechanism.
2282 :    
2283 :     ----------------------------------------------------------------------
2284 :     Name: Matthias Blume
2285 : blume 801 Date: 2001/03/19 17:53:00 EST
2286 :     Tag: blume-20010319-parallel
2287 :     Description:
2288 :    
2289 :     Parallel make (using compile servers) now works again.
2290 :    
2291 :     To this end, CM.stabilize and CMB.make have been modified to work in
2292 :     two passes when compile servers are attached:
2293 :     1. Compile everything, do not perform stabilization; this pass
2294 :     uses compile servers
2295 :     2. Stabilize everything; this pass does not use compile servers
2296 :     If there are no compile servers, the two passes are combined into one
2297 :     (as before). Splitting the passes increases the inherent parallelism
2298 :     in the dependency graph because the entire graph including all
2299 :     libraries is available at the same time. This, in turn, improves
2300 :     server utilization. The downside is that the master process will
2301 :     have to do some extra work after compilation is done (because for
2302 :     technical reasons it must re-read all the binfiles during stabilization).
2303 :    
2304 :     ----------------------------------------------------------------------
2305 :     Name: Matthias Blume
2306 : blume 800 Date: 2001/03/16 12:22:00 EST
2307 :     Tag: blume-20010316-bootfiles
2308 :     Description:
2309 :    
2310 :     Created a new set of bootfiles (for your automatic installation convenience).
2311 :    
2312 :     ----------------------------------------------------------------------
2313 :     Name: Matthias Blume
2314 : blume 799 Date: 2001/03/16 11:00:00 EST
2315 :     Tag: blume-20010316-MLTREE-fixup
2316 :     Description:
2317 :    
2318 :     This is a minor fixup for an (untagged) earlier commit by Allen.
2319 :     (A file was missing).
2320 :    
2321 :     ----------------------------------------------------------------------
2322 : leunga 788 Name: Allen Leung
2323 : leunga 796 Date: Mon Mar 5 18:54:57 EST 2001
2324 :     Tag: leunga-20010305-cut-support
2325 :    
2326 :     1. New support for alternative control-flow in MLTREE.
2327 :     Currently we support
2328 :    
2329 :     FLOW_TO(CALL ...., [k1,...,kn])
2330 :    
2331 :     This is needed for 'cuts to' in C-- and try/handle-like constructs
2332 :     in Moby
2333 :    
2334 :     New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
2335 :    
2336 :     2. Register Allocator
2337 :    
2338 :     Changes in interface [from Fermin, John]
2339 :    
2340 :     3. Alpha 8-bit SLL support [Fermin]
2341 :    
2342 :     4. All architectures
2343 :    
2344 :     A new module (ClusterExpandCopies) for expanding parallel copies.
2345 :    
2346 :     ----------------------------------------------------------------------
2347 :     Name: Allen Leung
2348 : leunga 788 Date: 2001/02/27 23:07:00 EST
2349 :     Tag: leunga-20010227-minor-stuff
2350 :    
2351 :     1. Alpha bug fix for CMOVNE
2352 :     2. Handle mltree COND(..,FCMP ...,...)
2353 :     3. Bug fix in simplifier
2354 :    
2355 :     ----------------------------------------------------------------------
2356 : blume 777 Name: Matthias Blume
2357 : blume 781 Date: 2001/01/30 17:50:00 EST
2358 :     Tag: blume-20010130-sync
2359 :     Description:
2360 :    
2361 :     This is just a minor update to sync my devel branch with the main brach.
2362 :     The only visible change is the addition of some README files.
2363 :    
2364 :     ----------------------------------------------------------------------
2365 :     Name: Matthias Blume
2366 : blume 778 Date: 2001/01/12 23:30:00 JST
2367 :     Tag: blume-20010112-bootfiles
2368 :     Description:
2369 :    
2370 :     Made a new set of bootfiles that goes with the current state of the
2371 :     repository.
2372 :    
2373 :     ----------------------------------------------------------------------
2374 :     Name: Matthias Blume
2375 : blume 777 Date: 2001/01/12 21:20:00 JST
2376 :     Tag: blume-20010112-sync
2377 :     Description:
2378 :    
2379 :     I am just flushing out some minor changes that had accumulated in
2380 :     my private branch in order to sync with the main tree. (This is
2381 :     mainly because I had CVS trouble when trying to merge _into_ my
2382 :     private branch.)
2383 :    
2384 :     Most people should be completely unaffected by this.
2385 :    
2386 :     ----------------------------------------------------------------------
2387 : leunga 775 Name: Allen Leung
2388 :     Date: Thu Jan 11 21:03:00 EST 2001
2389 :     Tag: leunga-20010111-labexp=mltree
2390 :     Description:
2391 :    
2392 :     1. Removed the type LabelExp and replace it by MLTree.
2393 :     2. Rewritten mltree-simplify with the pattern matcher tool.
2394 :     3. There were some bugs in alpha code generator which would break
2395 :     64-bit code generation.
2396 :     4. Redo the tools to generate code with the
2397 :     5. The CM files in MLRISC (and in src/system/smlnj/MLRISC)
2398 :     are now generated by perl scripts.
2399 :    
2400 :     ----------------------------------------------------------------------
2401 : blume 771 Name: Matthias Blume
2402 : blume 774 Date: 2001/01/10 21:55:00 JST
2403 :     Tag: blume-20010110-rcc
2404 :     Description:
2405 :    
2406 :     The RCC stuff now seems to work (but only on the x86).
2407 :     This required hacking of the c-calls interface (and -implementation) in
2408 :     MLRISC.
2409 :    
2410 :     Normal compiler users should be unaffected.
2411 :    
2412 :     ----------------------------------------------------------------------
2413 :     Name: Matthias Blume
2414 : blume 773 Date: 2001/01/09 01:20:00 JST
2415 :     Tag: blume-20010109-rcc
2416 :     Description:
2417 :    
2418 :     This is a fairly big patch, flushing out a large number of pending
2419 :     changes that I made to my development copy over the last couple of days.
2420 :    
2421 :     Of practical relevance at this moment is a workaround for a pickling
2422 :     bug that Allen ran into the other day. The cause of the bug itself is
2423 :     still unknown and it might be hard to fix it properly, but the
2424 :     workaround has some merits of its own (namely somewhat reducing pickling
2425 :     overhead for certain libraries). Therefore, I think this solution should
2426 :     be satisfactory at this time.
2427 :    
2428 :     The rest of the changes (i.e., the vast majority) has to do with my
2429 :     ongoing efforts of providing direct support for C function calls from
2430 :     ML. At the moment there is a new primop "RAW_CCALL", typing magic
2431 :     in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
2432 :     case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
2433 :     translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
2434 :     handle RCC.
2435 :    
2436 :     The last part (the changes to mlriscGen.sml) are still known to be
2437 :     wrong on the x86 and not implemented on all other architectures. But
2438 :     the infrastructure is in place. I had to change a few functor
2439 :     signatures in the backend to be able to route the CCalls interface
2440 :     from MLRISC there, and I had to specialize the mltree type (on the
2441 :     x86) to include the necessary extensions. (The extensions themselves
2442 :     were already there and redy to go in MLRISC/x86).
2443 :    
2444 :     Everything should be very happy as soon as someone helps me with
2445 :     mlriscGen.sml...
2446 :    
2447 :     In any case, nothing of this should matter to anyone as long as the
2448 :     new primop is not being used (which is going to be the case unless you
2449 :     find it where I hid it :). The rest of the compiler is completely
2450 :     unaffected.
2451 :    
2452 :     ----------------------------------------------------------------------
2453 :     Name: Matthias Blume
2454 : blume 772 Date: 2001/01/05 00:30:00 JST
2455 :     Tag: blume-20010105-primops
2456 :     Description:
2457 :    
2458 :     Added some experimental support for work that I am doing right now.
2459 :     These changes mostly concern added primops, but there is also a new
2460 :     experimental C library in the runtime system (but currently not enabled
2461 :     anywhere except on Linux/X86).
2462 :    
2463 :     In the course of adding primops (and playing with them), I discovered that
2464 :     Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
2465 :     broken. (Zhong was very right he labeled this stuff as "major gross hack".)
2466 :     To recover, I made type information in INL_PRIM mandatory and changed
2467 :     prim.sml as well as built-in.sml accordingly. The InLine structure now
2468 :     has complete, correct type information (i.e., no bottom types).
2469 :    
2470 :     Since all these changes mean that we need new binfiles, I also bumped the
2471 :     version number to 110.32.1.
2472 :    
2473 :     ----------------------------------------------------------------------
2474 :     Name: Matthias Blume
2475 : blume 771 Date: 2000/12/30 22:10:00 JST
2476 :     Tag: blume-20001230-various
2477 :     Description:
2478 :    
2479 :     Added proxy libraries for MLRISC and let MLRISC libraries refer
2480 :     to each other using path anchors. (See CM manual for explanation.)
2481 :    
2482 :     Updated CM documentation.
2483 :    
2484 :     Fixed some bugs in CM.
2485 :    
2486 :     Implemented "proxy" libraries (= syntactic sugar for CM).
2487 :    
2488 :     Added "-quiet" option to makeml and changed runtime system accordingly.
2489 :    
2490 :     Added cleanup handler for exportML to reset timers and compiler stats.
2491 :    
2492 :     ----------------------------------------------------------------------
2493 : george 761 Name: Lal George
2494 :     Date: 2000/12/22 22:22:58 EST 2000
2495 :     Tag: Release_110_32
2496 :     Description:
2497 :    
2498 :     Infinite precision used throughout MLRISC.
2499 :     see MLRISC/mltree/machine-int.sig
2500 :    
2501 :     ----------------------------------------------------------------------
2502 : blume 756 Name: Matthias Blume
2503 : blume 760 Date: 2000/12/22 23:16:00 JST
2504 :     Tag: blume-20001222-warn
2505 :     Description:
2506 : blume 759
2507 : blume 760 Corrected wording and formatting of some CM warning message which I
2508 :     broke in my previous patch.
2509 : blume 759
2510 :     ----------------------------------------------------------------------
2511 :     Name: Matthias Blume
2512 : blume 763 Date: 2000/12/22 21:20:00 JST
2513 :     Tag: blume-20001222-anchorenv
2514 :     Description:
2515 :    
2516 :     Fixed CM's handling of anchor environments in connection with CMB.make.
2517 :    
2518 :     ----------------------------------------------------------------------
2519 :     Name: Matthias Blume
2520 : blume 758 Date: 2000/12/22 13:15:00 JST
2521 :     Tag: blume-20001222-cleanup
2522 :     Description:
2523 :    
2524 :     Removed src/cm/ffi which does not (and did not) belong here.
2525 :    
2526 :     ----------------------------------------------------------------------
2527 :     Name: Matthias Blume
2528 : blume 757 Date: 2000/12/21 23:55:00 JST
2529 :     Tag: blume-20001221-exn
2530 :     Description:
2531 :    
2532 :     Probably most important: CM no longer silently swallows all exceptions
2533 :     in the compiler.
2534 :     Plus: some other minor CM changes. For example, CM now reports some
2535 :     sizes for generated binfiles (code, data, envpickle, lambdapickle).
2536 :    
2537 :     ----------------------------------------------------------------------
2538 :     Name: Matthias Blume
2539 : blume 756 Date: 2000/12/15 00:01:05 JST
2540 :     Tag: blume-20001215-dirtool
2541 :     Description:
2542 :    
2543 :     - "dir" tool added.
2544 :     - improvements and cleanup to Tools structure
2545 :     - documentation updates
2546 :    
2547 :     ----------------------------------------------------------------------
2548 : leunga 752 Name: Allen Leung
2549 : leunga 755 Date: Thu Dec 14 03:45:24 EST 2000
2550 :     Description:
2551 :     Tag: leunga-20001214-int-inf
2552 :     Description:
2553 :    
2554 :     In IntInf, added these standard functions, which are missing from our
2555 :     implementation:
2556 :    
2557 :     andb : int * int -> int
2558 :     xorb : int * int -> int
2559 :     orb : int * int -> int
2560 :     notb : int -> int
2561 :     << : int * word -> int
2562 :     ~>> : int * word -> int
2563 :    
2564 :     Not tested, I hope they are correct.
2565 :    
2566 :     ----------------------------------------------------------------------
2567 :     Name: Allen Leung
2568 : leunga 752 Date: Fri Dec 8 19:23:26 EST 2000
2569 :     Description:
2570 :     Tag: leunga-20001208-nowhere
2571 :     Description:
2572 :    
2573 :     Slight improvements to the 'nowhere' tool to handle OR-patterns,
2574 :     to generate better error messages etc. Plus a brief manual.
2575 :    
2576 :     ----------------------------------------------------------------------
2577 : george 749 Name: Lal George
2578 :     Date: 2000/12/08 09:54:02 EST 2000
2579 :     Tag: Release_110_31
2580 :     Description:
2581 : george 705
2582 : george 749 - Version 110.31
2583 : leunga 591 ----------------------------------------------------------------------
2584 : leunga 744 Name: Allen Leung
2585 :     Date: Thu Dec 7 22:01:04 EST 2000
2586 :     Tag: leunga-20001207-cell-monster-hack
2587 :     Description:
2588 :    
2589 :     Major MLRISC internal changes. Affect all clients.
2590 :     Summary:
2591 :    
2592 :     1. Type CELLS.cell = int is now replaced by a datatype.
2593 :     As a result, the old regmap is now gone. Almost all interfaces
2594 :     in MLRISC change as a consequence.
2595 :    
2596 :     2. A new brand version of machine description tool (v3.0) that generates
2597 :     modules expecting the new interface. The old version is removed.
2598 :    
2599 :     3. The RA interface has been further abstracted into two new functors.
2600 :     RISC_RA and X86RA. These functors have much simpler interfaces.
2601 :     [See also directory MLRISC/demo.]
2602 :    
2603 :     4. Some other new source->source code generation tools are available:
2604 :    
2605 :     a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
2606 :     b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
2607 :     I use this tool to generate the peephole optimizers---with the new
2608 :     cell type changes, peephole rules are becoming difficult to write
2609 :     without conditional pattern matching.
2610 :    
2611 :     5. More Intmap -> IntHashTable change. Previous changes by Matthias didn't
2612 :     cover the entire MLRISC source tree so many things broke.
2613 :    
2614 :     6. CM files have been moved to the subdirectory MLRISC/cm.
2615 :     They are moved because there are a lot of them and they clutter up the
2616 :     root dir.
2617 :    
2618 :     7. More detailed documentation to come...
2619 :    
2620 :     NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
2621 :     a makeml -rebuild first. This is because of other other
2622 :     changes that Matthias has made (see below).
2623 :    
2624 :    
2625 :     ----------------------------------------------------------------------
2626 : blume 733 Name: Matthias Blume
2627 : blume 742 Date: 2000/11/30 23:12:00 JST
2628 :     Tag: blume-20001130-filereorg
2629 :     Description:
2630 :    
2631 :     Some manual updates and some file reorganizations in CM.
2632 :    
2633 :     ----------------------------------------------------------------------
2634 :     Name: Matthias Blume
2635 : blume 737 Date: 2000/11/24 17:45:00 JST
2636 :     Tag: blume-20001124-link
2637 :     Description:
2638 :    
2639 :     Drastically improved link traversal code for the case that the dynamic
2640 :     value was already loaded at bootstrap time. As a result, CM and CMB
2641 :     now both load blazingly fast -- even on a very slow machine. Also,
2642 :     memory consumption has been further reduced by this.
2643 :    
2644 :     Warning: The format of the PIDMAP file has changed. THerefore, to
2645 :     bootstrap you have to do this:
2646 :    
2647 :     1. Run CMB.make
2648 :     2. Make a symbolic link for the boot directory:
2649 :     ln -s sml.boot.ARCH-OS xxx
2650 :     3. "Rebuild" the boot directory:
2651 :     ./makeml -boot xxx -rebuild sml ; rm xxx
2652 :     4. Boot normally:
2653 :     ./makeml
2654 :    
2655 :     ----------------------------------------------------------------------
2656 :     Name: Matthias Blume
2657 : blume 735 Date: 2000/11/21 21:20:00 JST
2658 :     Tag: blume-20001121-tools
2659 :     Description:
2660 :    
2661 :     Continued hacking on autoloading problem -- with success this time.
2662 :     Also changed tool-plugin mechanism. See new CM manual.
2663 :    
2664 :     ----------------------------------------------------------------------
2665 :     Name: Matthias Blume
2666 : blume 734 Date: 2000/11/19 14:30:00 JST
2667 :     Tag: blume-20001119-autoload
2668 :     Description:
2669 :    
2670 :     Some hacking to make autoloading faster. Success for CMB, no success
2671 :     so far for CM. There is a reduced structure CM' that autoloads faster.
2672 :     (This is a temporary, non-documented hack to be eliminated again when
2673 :     the general problem is solved.)
2674 :    
2675 :     ----------------------------------------------------------------------
2676 :     Name: Matthias Blume
2677 : blume 733 Date: 2000/11/17 14:10:00 JST
2678 :     Tag: blume-20001117-pickle-lib
2679 :     Description:
2680 :    
2681 :     1. Eliminated comp-lib.cm
2682 :     2. Made pickle-lib.cm
2683 :     3. Eliminated all uses of intset.sml (from comp-lib.cm)
2684 :     4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
2685 :     equivalent constructs from smlnj-lib.cm (INtHashTable).
2686 :     5. Point 4. also goes for those uses of intmap.* in MLRISC.
2687 :     Duplicated intmap modules thrown out.
2688 :     6. Hunted down all duplicated SCC code and replaced it with
2689 :     equivalent stuff (GraphSCCFn from smlnj-lib.cm).
2690 :     7. Rewrote Feedback module.
2691 :     8. Moved sortedlist.sml into viscomp-lib.cm. Eventually it
2692 :     should be thrown out and equivalent modules from smlnj-lib.cm
2693 :     should be used (IntRedBlackSet, IntListSet, ...).
2694 :    
2695 :     Confirmed that compiler compiles to fixpoint.
2696 :    
2697 :     ----------------------------------------------------------------------
2698 : leunga 731 Name: Allen Leung
2699 :     Date: 2000/11/10 18:00:00
2700 :     Tag: leunga-20001110-new-x86-fp
2701 :    
2702 :     A new x86 floating point code generator has been added.
2703 :     By default this is turned off. To turn this on, do:
2704 :    
2705 :     CM.autoload "$smlnj/compiler.cm";
2706 :     Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
2707 :    
2708 :     Changes:
2709 :    
2710 :     1. Changed FTAN to FPTAN so that the assembly output is correct.
2711 :     2. Changed the extension callback for FTANGENT to generate:
2712 :    
2713 :     fptan
2714 :     fstp %st(0)
2715 :     instead of
2716 :     fptan
2717 :     fstpl ftempmem
2718 :    
2719 :     3. Numerous assembly fixes for x86.
2720 :    
2721 :     5. Cleaned up the machine code output module x86/x86MC.sml and added
2722 :     support for a whole bunch of instructions and addressing modes:
2723 :    
2724 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st, %st(n)
2725 :     faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp %st, %st(n)
2726 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st(n), %st
2727 :     fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
2728 :     fxch %st(n)
2729 :     fld %st(n)
2730 :     fst %st(n)
2731 :     fst mem
2732 :     fstp %st(n)
2733 :     fucom %st(n)
2734 :     fucomp %st(n)
2735 :    
2736 :     All these are now generated when the fast fp mode is turned on.
2737 :    
2738 :     6. Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
2739 :    
2740 :     ----------------------------------------------------------------------
2741 : blume 729 Name: Matthias Blume
2742 :     Date: 2000/11/09 11:20:00 JST
2743 :     Tag: blume-20001109-scc
2744 :     Description:
2745 :    
2746 :     Eliminated some code duplication:
2747 :    
2748 :     1. Added "where" clause to GraphSCCFn in SML/NJ Library.
2749 :     (Otherwise the functor is useless.)
2750 :     2. Used GraphSCCFn where SCCUtilFun was used previously.
2751 :     3. Got rid of SCCUtilFun (in comp-lib.cm).
2752 :    
2753 :     ----------------------------------------------------------------------
2754 : george 721 Name: Lal George
2755 :     Date: 2000/11/06 09:02:21 EST 2000
2756 :     Tag: Release_110_30
2757 :     Description:
2758 :    
2759 :     - Version 110.30
2760 :     ----------------------------------------------------------------------
2761 : blume 715 Name: Matthias Blume
2762 : blume 716 Date: 2000/11/04 14:45:00
2763 :     Tag: blume-20001104-mlbuild
2764 :     Description:
2765 :    
2766 :     - Made ml-build faster on startup.
2767 :     - Documentation fixes.
2768 :    
2769 :     ----------------------------------------------------------------------
2770 :     Name: Matthias Blume
2771 : blume 715 Date: 2000/11/02 17:00:00 JST
2772 :     Tag: blume-20001102-condcomp
2773 :     Description:
2774 :    
2775 :     - Small tweaks to pickler -- new BOOTFILES!
2776 :     - Version bumped to 110.29.2.
2777 :     - Added conditional compilation facility to init.cmi (see comment there).
2778 :     ----------------------------------------------------------------------
2779 : leunga 713 Name: Allen Leung
2780 :     Date: 2000/10/23 19:31:00
2781 :     Tag: leunga-20001023-demo-ra
2782 :    
2783 :     1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
2784 :     2. Test programs for the graph library updated
2785 :     3. Some new MLRISC demo programs added
2786 :    
2787 :     ----------------------------------------------------------------------
2788 : blume 710 Name: Matthias Blume
2789 : blume 711 Date: 2000/08/31 22:15:00 JST
2790 :     Tag: blume-20001017-errmsg
2791 :     Description:
2792 :    
2793 :     More error message grief: Where there used to be no messages, there
2794 :     now were some that had bogus error regions. Fixed.
2795 :    
2796 :     ----------------------------------------------------------------------
2797 :     Name: Matthias Blume
2798 : blume 710 Date: 2000/08/31 17:30:00 JST
2799 :     Tag: blume-20001017-v110p29p1
2800 :     Description:
2801 :    
2802 :     I made a version 110.29.1 with new bootfiles.
2803 :    
2804 :     Changes: Modified pickler/unpickler for faster and leaner unpickling.
2805 :     CM documentation changes and a small bugfix in CM's error reporting.
2806 :    
2807 :     ----------------------------------------------------------------------
2808 : george 705 Name: Lal George
2809 :     Date: 2000/09/27 14:42:35 EDT
2810 :     Tag: george-20000927-nodestatus
2811 :     Description:
2812 :    
2813 :     Changed the type of the nodestatus, so that:
2814 :    
2815 :     SPILLED(~1) is now SPILLED
2816 :     SPILLED(m) where m>=0 is now MEMREG(m)
2817 :     SPILLED(s) where s<~1 is now SPILL_LOC(~s)
2818 :    
2819 :     ----------------------------------------------------------------------
2820 : blume 697 Name: Matthias Blume
2821 : blume 703 Date: 2000/09/07 14:45:00 JST
2822 :     Tag: blume-20000907-cmerrmsg
2823 :     Description:
2824 :    
2825 :     Small tweak to CM to avoid getting ML syntax error messages twice.
2826 :    
2827 :     ----------------------------------------------------------------------
2828 :     Name: Matthias Blume
2829 : blume 701 Date: 2000/08/31 18:00:00 JST
2830 :     Tag: blume-20000831-cvsbootfiles
2831 :     Description:
2832 :    
2833 :     New URL for boot files (because the 110.29 files on the BL server do
2834 :     now work correctly with my updated install scripts for yacc and lex).
2835 :    
2836 :     ----------------------------------------------------------------------
2837 :     Name: Matthias Blume
2838 : blume 697 Date: 2000/08/08 12:33:00 JST
2839 :     Tag: blume-20000808-manual
2840 :     Description:
2841 :    
2842 :     Tiny update to CM manual.
2843 :    
2844 :     ----------------------------------------------------------------------
2845 : leunga 695 Name: Allen Leung
2846 :     Date: 2000/08/7 19:31:00
2847 :     Tag: leunga-20000807-a-whole-bunch-of-stuff
2848 :    
2849 :     Moby, C--, SSA, x86, machine descriptions etc. Should only affect C--
2850 :     and Mobdy.
2851 :    
2852 :     1. x86
2853 :    
2854 :     a. Fixes to peephole module by John and Dan.
2855 :     b. Assembly fix to SETcc by Allen.
2856 :     c. Fix to c-call by John.
2857 :     d. Fix to spilling by John. (This one deals with the missing FSTPT case)
2858 :     e. Instruction selection optimization to SETcc as suggested by John.
2859 :    
2860 :     For example,
2861 :    
2862 :     MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
2863 :    
2864 :     should generate:
2865 :    
2866 :     MOVL a, x
2867 :     SUBL b, x
2868 :     SHRL 31, x
2869 :    
2870 :     2. IR stuff
2871 :    
2872 :     A bunch of new DJ-graph related algorithms added. These
2873 :     speed up SSA construction.
2874 :    
2875 :     3. SSA + Scheduling
2876 :    
2877 :     Added code for SSA and scheduling to the repository
2878 :    
2879 :     ----------------------------------------------------------------------
2880 : george 694 Name: Lal George
2881 :     Date: 2000/07/27 11:53:14 EDT
2882 :    
2883 :     Tag: lal-20000727-linux-ppc
2884 :     Description:
2885 :    
2886 :     Made changes to support Linux PPC.
2887 :     p.s. I have confirmation that the 110.29 boot files work fine.
2888 :    
2889 :     ----------------------------------------------------------------------
2890 : blume 684 Name: Matthias Blume
2891 : blume 692 Date: 2000/07/27 17:40:00 JST
2892 :     Tag: blume-20000727-scripts
2893 :     Description:
2894 :    
2895 :     !!!! WARNING !!!!
2896 :     You must recompile the runtime system!
2897 :     !!!! WARNING !!!!
2898 :    
2899 :     This is basically another round of script-enhancements:
2900 :    
2901 :     1. sml, ml-build, and ml-makedepend accept options -D and -U to define
2902 :     and undefine CM preprocessor symbols.
2903 :    
2904 :     2. ml-build avoids generating a new heap image if it finds that the
2905 :     existing one is still ok. (The condition is that no ML file had to
2906 :     be recompiled and all ML files are found to be older that the heap
2907 :     file.)
2908 :    
2909 :     To make this work smoothly, I also hacked the runtime system as
2910 :     well as SMLofNJ.SysInfo to get access to the heap image suffix
2911 :     (.sparc-solaris, ...) that is currently being used.
2912 :    
2913 :     Moreover, the signature of CM.mk_standalone has changed. See the
2914 :     CM manual.
2915 :    
2916 :     3. ml-makedepend accepts additional options -n, -a, and -o. (See the
2917 :     CM manual for details.)
2918 :    
2919 :     4. More CM manual updates:
2920 :     - all of the above has been documented.
2921 :     - there is now a section describing the (CM-related) command line
2922 :     arguments that are accepted by the "sml" command
2923 :    
2924 :     ----------------------------------------------------------------------
2925 :     Name: Matthias Blume
2926 : blume 691 Date: 2000/07/25 16:20:00 JST
2927 :     Tag: blume-20000725-makedepend
2928 :     Description:
2929 :    
2930 :     Added a script called ml-makedepend. This can be used in makefiles
2931 :     for Unix' make in a way very similar to the "makedepend" command for
2932 :     C.
2933 :    
2934 :     The script internally uses function CM.sources.
2935 :    
2936 :     Synopsis:
2937 :    
2938 :     ml-makedepend [-f makefile] cmfile targetname
2939 :    
2940 :     The default for the makefile is "makefile" (or "Makefile" should
2941 :     "makefile" not exist).
2942 :    
2943 :     ml-makedepend adds a cmfile/targetname-specific section to this
2944 :     makefile (after removing the previous version of this section). The
2945 :     section contains a single dependency specification with targetname on
2946 :     the LHS (targetname is an arbitrary name), and a list of files derived
2947 :     from the cmfile on the RHS. Some of the files on the RHS are
2948 :     ARCH/OPSYS-specific. Therefore, ml-makedepend inserts references to
2949 :     "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
2950 :     path names. The makefile writer is responsible for making sure that
2951 :     these variables have correct at the time "make" is invoked.
2952 :    
2953 :     ----------------------------------------------------------------------
2954 :     Name: Matthias Blume
2955 : blume 690 Date: 2000/07/22 23:30:00 JST
2956 :     Tag: blume-20000722-urlupdate
2957 :     Description:
2958 :    
2959 :     Changed BOOT and config/srcarchiveurl to point to BL server:
2960 :    
2961 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
2962 :    
2963 :     ----------------------------------------------------------------------
2964 :     Name: Matthias Blume
2965 : blume 686 Date: 2000/07/18 18:00:00 JST
2966 :     Tag: blume-20000718-Version_110_29
2967 :     Description:
2968 :    
2969 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
2970 :    
2971 :     2. Updated config/version to 110.29
2972 :    
2973 :     3. Updated config/srcarchiveurl
2974 :    
2975 :     3. New boot files!
2976 :     ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
2977 :    
2978 :     ----------------------------------------------------------------------
2979 :     Name: Matthias Blume
2980 : blume 684 Date: 2000/07/11 13:58:00 JST
2981 :     Tag: blume-20000711-doctypo
2982 :     Description:
2983 :    
2984 :     Fixed a few typos in CM manual.
2985 :    
2986 :     ----------------------------------------------------------------------
2987 : leunga 683 Name: Allen Leung
2988 :     Date: 2000/06/15 00:38:00
2989 :     Tag: leunga-20000704-sparc-x86
2990 :    
2991 :     1. x86 peephole improvement sp += k; sp -= k => nop [from John]
2992 :     2. fix to x86 RET bug [found by Dan Grossman]
2993 :     3. sparc assembly bug fix for ticc instructions [found by Fermin]
2994 :    
2995 :     Affects c-- and moby only
2996 :    
2997 :     ----------------------------------------------------------------------
2998 : blume 680 Name: Matthias Blume
2999 : blume 682 Date: 2000/07/04 15:26:00
3000 :     Tag: blume-20000704-trigger
3001 :     Description:
3002 :    
3003 :     1. Improvements to CM manual.
3004 :     2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
3005 :     of getting a back-trace. The function, when called, raises an
3006 :     internal exception which explicitly carries the full back-trace history,
3007 :     so it is unaffected by any intervening handle-raise pairs ("trivial"
3008 :     or not). The interactive loop will print that history once it arrives
3009 :     at top level.
3010 :     Short of having all exceptions implicitly carry the full history, the
3011 :     recommended way of using this facility is:
3012 :     - compile your program with instrumentation "on"
3013 :     - run it, when it raises an exception, look at the history
3014 :     - if the history is "cut off" because of some handler, go and modify
3015 :     your program so that it explicitly calls BTrace.trigger
3016 :     - recompile (still instrumented), and rerun; look at the full history
3017 :    
3018 :     ----------------------------------------------------------------------
3019 :     Name: Matthias Blume
3020 :     Date: 2000/07/03 15:36:00 JST
3021 : blume 680 Tag: blume-20000702-manual
3022 :     Description:
3023 :    
3024 :     Small corrections and updates to CM manual.
3025 :    
3026 :     ----------------------------------------------------------------------
3027 : blume 682 Name: Matthias Blume
3028 : blume 679 Date: 2000/06/29 16:04:00 JST
3029 :     Tag: blume-20000629-yacctool
3030 :     Description:
3031 :    
3032 :     Changes:
3033 :    
3034 :     1. Class "mlyacc" now takes separate arguments to pass options to
3035 :     generated .sml- and .sig-files independently.
3036 :     2. Corresponding CM manual updates.
3037 :     3. BTrace module now also reports call sites. (However, for loop clusters
3038 :     it only shows from where the cluster was entered.) There are associated
3039 :     modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
3040 :    
3041 :     ----------------------------------------------------------------------
3042 : blume 675 Name: Matthias Blume
3043 : blume 678 Date: 2000/06/27 16:51:00 JST
3044 :     Tag: blume-20000627-noweb
3045 :     Description:
3046 :    
3047 :     Changes:
3048 :    
3049 :     1. Implemented "subdir" and "witness" options for noweb tool.
3050 :     This caused some slight internal changes in CM's tool implementation.
3051 :     2. Fixed bug in "tool plugin" mechanism. This is essentially cleaning
3052 :     some remaining issues from earlier path anchor changes.
3053 :     3. Updated CM manual accordingly.
3054 :    
3055 :     4. Changed implementation of back-tracing so that I now consider it
3056 :     ready for prime-time.
3057 :    
3058 :     In particular, you don't have to explicitly trigger the back-trace
3059 :     anymore. Instead, if you are running BTrace-instrumented code and
3060 :     there is an uncaught exception (regardless of whether or not it was
3061 :     raised in instrumented code), the top-level evalloop will print
3062 :     the back-trace.
3063 :    
3064 :     Features:
3065 :    
3066 :     - Instrumented and uninstrumented code work together seemlessly.
3067 :     (Of course, uninstrumented code is never mentioned in actual
3068 :     back-traces.)
3069 :    
3070 :     - Asymptotic time- and space-complexity of instrumented code is
3071 :     equal to that of uninstrumented code. (This means that
3072 :     tail-recursion is preserved by the instrumentation phase.)
3073 :    
3074 :     - Modules whose code has been instrumented in different sessions
3075 :     work together without problem.
3076 :    
3077 :     - There is no penalty whatsoever on uninstrumented code.
3078 :    
3079 :     - There is no penalty on "raise" expressions, even in
3080 :     instrumented code.
3081 :    
3082 :     A potential bug (or perhaps it is a feature, too):
3083 :    
3084 :     A back-trace reaches no further than the outermost instrumented
3085 :     non-trivial "raise". Here, a "trivial" raise is one that is the
3086 :     sole RHS of a "handle" rule. Thus, back-traces reach trough
3087 :    
3088 :     <exp> handle e => raise e
3089 :    
3090 :     and even
3091 :    
3092 :     <exp> handle Foo => raise Bar
3093 :    
3094 :     and, of course, through
3095 :    
3096 :     <exp> handle Foo => ...
3097 :    
3098 :     if the exception was not Foo.
3099 :    
3100 :     Back-traces always reach right through any un-instrumented code
3101 :     including any of its "handle" expressions, trivial or not.
3102 :    
3103 :     To try this out, do the following:
3104 :    
3105 :     - Erase all existing binfiles for your program.
3106 :     (You may keep binfiles for those modules where you think you
3107 :     definitely don't need back-tracing.)
3108 :     - Turn on back-trace instrumentation:
3109 :     SMLofNJ.Internals.BTrace.mode (SOME true);
3110 :     - Recompile your program. (I.e., run "CM.make" or "use".)
3111 :     - You may now turn instrumentation off again (if you want):
3112 :     SMLofNJ.Internals.BTrace.mode (SOME false);
3113 :     - Run your program as usual. If it raises an exception that
3114 :     reaches the interactive toplevel, then a back-trace will
3115 :     automatically be printed. After that, the toplevel loop
3116 :     will print the exception history as usual.
3117 :    
3118 :     ----------------------------------------------------------------------
3119 :     Name: Matthias Blume
3120 : blume 677 Date: 2000/06/26 09:56:46 JST
3121 :     Tag: blume-20000626-setup
3122 :     Description:
3123 :    
3124 :     CM: - setup-parameter to "sml" added; this can be used to run arbitrary
3125 :     ML code before and after compiling a file (e.g., to set compiler
3126 :     flags)
3127 :    
3128 :     Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
3129 :     - associated changes to btrace.sml (BTrace instrumentation pass)
3130 :     - cleaner implementation of btimp.sml (BTrace tracing and report
3131 :     module)
3132 :    
3133 :     CM manual: * new path encoding documented
3134 :     * description of setup-parameter to "sml" added
3135 :    
3136 :     The biggest user-visible change to back-tracing is that it is no
3137 :     longer necessary to compile all traced modules within the same
3138 :     session. (This was a real limitation.)
3139 :    
3140 :     ----------------------------------------------------------------------
3141 :     Name: Matthias Blume
3142 : blume 676 Date: 2000/06/24 12:40:00 JST
3143 :     Tag: blume-20000624-startup
3144 :     Description:
3145 :    
3146 :     Fixes startup slowdown problem. (I was calling SrcPath.sync a _tad_
3147 :     bit too often -- to put it mildly. :)
3148 :    
3149 :     ----------------------------------------------------------------------
3150 :     Name: Matthias Blume
3151 : blume 675 Date: 2000/06/23 18:20:00 JST
3152 :     Tag: blume-20000623-btrace
3153 :     Description:
3154 :    
3155 :     This updates adds a backtrace facility to aid programmers in debugging
3156 :     their programs. This involves the following changes:
3157 :    
3158 :     1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
3159 :     keeping track of the current call stack. When programs are compiled
3160 :     in a special mode, the compiler will insert calls to these hooks
3161 :     into the user program.
3162 :     "Hook" means that it is possible for different implementations of
3163 :     back-tracing to register themselves (at different times).
3164 :    
3165 :     2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
3166 :     as an Absyn.dec->Absyn.dec rewrite. Normally this phase is turned off.
3167 :     It can be turned on using this call:
3168 :     SMLofNJ.Internals.BTrace.mode (SOME true);
3169 :     Turning it off again:
3170 :     SMLofNJ.Internals.BTrace.mode (SOME false);
3171 :     Querying the current status:
3172 :     SMLofNJ.Internals.BTrace.mode NONE;
3173 :     Annotated programs are about twice as big as normal ones, and they
3174 :     run a factor of 2 to 4 slower with a dummy back-trace plugin (one
3175 :     where all hooks do nothing). The slowdown with a plugin that is
3176 :     actually useful (such as the one supplied by default) is even greater,
3177 :     but in the case of the default plugin it is still only an constant
3178 :     factor (amortized).
3179 :    
3180 :     3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
3181 :     with a sub-structure BTrace for controlling back-tracing. In particular,
3182 :     the above-mentioned function "mode" controls whether the annotation
3183 :     phase is invoked by the compiler. Another important function is
3184 :     "trigger": when called it aborts the current execution and causes
3185 :     the top-level loop to print a full back-trace.
3186 :    
3187 :     4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
3188 :     for back-tracing. It keeps track of the dynamic call stack and in
3189 :     addition to that it keeps a partial history at each "level" of that
3190 :     stack. For example, if a tail-calls b, b tail-calls c, and c tail-calls
3191 :     d and b (at separate times, dynamically), then the report will show:
3192 :    
3193 :     GOTO d
3194 :     /c
3195 :     GOTO \b
3196 :     CALL a
3197 :    
3198 :     This shows that there was an initial non-tail call of a, then a
3199 :     tail-call to b or c, looping behavior in a cluster of functions that
3200 :     consist of b and c, and then a goto from that cluster (i.e., either from
3201 :     b or from c) to d.
3202 :    
3203 :     Note that (depending on the user program) the amount of information
3204 :     that the back-trace module has to keep track of at each level is bounded
3205 :     by a constant. Thus, the whole implementation has the same asymptotical
3206 :     complexity as the original program (both in space and in time).
3207 :    
3208 :     5. compiler/TopLevel/interact/evalloop.sml has been modified to
3209 :     handle the special exception SMLofNJ.Internals.BTrace.BTrace
3210 :     which is raised by the "trigger" function mentioned above.
3211 :    
3212 :     Notes on usage:
3213 :    
3214 :     - Annotated code works well together with unannotated code:
3215 :     Unannotated calls simply do not show up at all in the backtrace.
3216 :    
3217 :     - It is not a good idea to let modules that were annotated during
3218 :     different sessions run at the same time. This is because the compiler
3219 :     chooses small integers to identify individual functions, and there
3220 :     will be clashes if different modules were compiled in separate sessions.
3221 :     (Nothing will crash, and you will even be told about the clashes, but
3222 :     back-trace information will in general not be useful.)
3223 :    
3224 :     - Back-tracing can be confused by callcc and capture.
3225 :    
3226 :     - The only way of getting a back-trace right now is to explicitly
3227 :     invoke the "trigger" function from your user program. Eventually, we
3228 :     should make every exception carry back-trace information (if
3229 :     available). But since this creates more overhead at "raise"-time
3230 :     (similar to the current exnHistory overhead), I have not yet
3231 :     implemented this. (The implementation will be rather easy.) With
3232 :     exceptions carrying back-trace information, this facility will be even
3233 :     more useful because users don't need to modify their programs...
3234 :    
3235 :     - While it is possible to compile the compiler with back-trace
3236 :     annotations turned on (I did it to get some confidence in
3237 :     correctness), you must make absolutely sure that core.sml and
3238 :     btimp.sml are compiled WITHOUT annotation! (core.sml cannot actually
3239 :     be compiled with annotation because there is no core access yet, but
3240 :     if you compile btimp.sml with annotation, then the system will go into
3241 :     an infinite recursion and crash.)
3242 :     Since CM currently does not know about BTrace, the only way to turn
3243 :     annotations on and off for different modules of the compiler is to
3244 :     interrupt CMB.make, change the settings, and re-invoke it. Of course,
3245 :     this is awkward and clumsy.
3246 :    
3247 :     Sample sessions:
3248 :    
3249 :     Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
3250 :     - SMLofNJ.Internals.BTrace.mode (SOME true);
3251 :     [autoloading]
3252 :     [autoloading done]
3253 :     val it = false : bool
3254 :     - structure X = struct
3255 :     - fun main n = let
3256 :     - fun a (x, 0) = d x
3257 :     - | a (x, n) = b (x, n - 1)
3258 :     - and b (x, n) = c (x, n)
3259 :     - and c (x, n) = a (x, n)
3260 :     - and d x = e (x, 3)
3261 :     - and e (x, 0) = f x
3262 :     - | e (x, n) = e (x, n - 1)
3263 :     - and f 0 = SMLofNJ.Internals.BTrace.trigger ()
3264 :     - | f n = n * g (n - 1)
3265 :     - and g n = a (n, 3)
3266 :     - in
3267 :     - f n
3268 :     - end
3269 :     - end;
3270 :     structure X : sig val main : int -> int end
3271 :     - X.main 3;
3272 :     *** BACK-TRACE ***
3273 :     GOTO stdIn:4.2-13.20: X.main[2].f
3274 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
3275 :     GOTO stdIn:4.2-13.20: X.main[2].d
3276 :     / stdIn:4.2-13.20: X.main[2].a
3277 :     | stdIn:4.2-13.20: X.main[2].b
3278 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
3279 :     CALL stdIn:4.2-13.20: X.main[2].g
3280 :     GOTO stdIn:4.2-13.20: X.main[2].f
3281 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
3282 :     GOTO stdIn:4.2-13.20: X.main[2].d
3283 :     / stdIn:4.2-13.20: X.main[2].a
3284 :     | stdIn:4.2-13.20: X.main[2].b
3285 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
3286 :     CALL stdIn:4.2-13.20: X.main[2].g
3287 :     GOTO stdIn:4.2-13.20: X.main[2].f
3288 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
3289 :     GOTO stdIn:4.2-13.20: X.main[2].d
3290 :     / stdIn:4.2-13.20: X.main[2].a
3291 :     | stdIn:4.2-13.20: X.main[2].b
3292 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
3293 :     CALL stdIn:4.2-13.20: X.main[2].g
3294 :     GOTO stdIn:4.2-13.20: X.main[2].f
3295 :     CALL stdIn:2.15-17.4: X.main[2]
3296 :     -
3297 :    
3298 :     (Note that because of a FLINt bug the above code currently does not
3299 :     compile without BTrace turned on.)
3300 :    
3301 :     Here is another example, using my modified Tiger compiler:
3302 :    
3303 :     Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
3304 :     - SMLofNJ.Internals.BTrace.mode (SOME true);
3305 :     [autoloading]
3306 :     [autoloading done]
3307 :     val it = false : bool
3308 :     - CM.make "sources.cm";
3309 :     [autoloading]
3310 :     ...
3311 :     [autoloading done]
3312 :     [scanning sources.cm]
3313 :     [parsing (sources.cm):parse.sml]
3314 :     [creating directory CM/SKEL ...]
3315 :     [parsing (sources.cm):tiger.lex.sml]
3316 :     ...
3317 :     [wrote CM/sparc-unix/semant.sml]
3318 :     [compiling (sources.cm):main.sml]
3319 :     [wrote CM/sparc-unix/main.sml]
3320 :     [New bindings added.]
3321 :     val it = true : bool
3322 :     - Main.compile ("../testcases/merge.tig", "foo.out");
3323 :     *** BACK-TRACE ***
3324 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
3325 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3326 :     CALL lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
3327 :     GOTO lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
3328 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3329 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3330 :     CALL lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
3331 :     / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
3332 :     CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
3333 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3334 :     CALL lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
3335 :     CALL main.sml:1.18-118.4: Main.compile[2]
3336 :     -
3337 :    
3338 :     ----------------------------------------------------------------------
3339 : blume 673 Name: Matthias Blumen
3340 :     Date: 2000/06/21 18:00:00 JST
3341 :     Tag: blume-20000621-manual
3342 :     Description:
3343 :    
3344 :     CM manual update: Path environments documented.
3345 :    
3346 :     ----------------------------------------------------------------------
3347 : blume 666 Name: Matthias Blume
3348 : blume 672 Date: 2000/06/19 13:40:00
3349 :     Tag: blume-20000619-manual
3350 :     Description:
3351 :    
3352 :     CM manual and system/README update. This only covers the fact that
3353 :     there are no more implicit anchors. (Path environments and the "bind"
3354 :     option to "cm" have yet to be documented.)
3355 :    
3356 :     ----------------------------------------------------------------------
3357 :     Name: Matthias Blume
3358 : blume 671 Date: 2000/06/19 11:05:00 JST
3359 :     Tag: blume-20000619-chdir-bugfix
3360 :     Description:
3361 :    
3362 :     Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
3363 :    
3364 :     ----------------------------------------------------------------------
3365 :     Name: Matthias Blume
3366 : blume 670 Date: 2000/06/18 22:00:10 JST
3367 :     Tag: blume-20000618-implicit-anchors-really-gone
3368 :     Description:
3369 :    
3370 :     I updates the previous HISTORY entry where I forgot to mention that
3371 :     implicit anchors are no longer with us.
3372 :    
3373 :     The current update also gets rid of the (now useless) controller
3374 :     CM.Control.implicit_anchors.
3375 :    
3376 :     ----------------------------------------------------------------------
3377 :     Name: Matthias Blume
3378 : blume 666 Date: 2000/06/16 17:30:00 JST
3379 :     Tag: blume-20000616-anchorenv
3380 :     Description:
3381 :    
3382 :     This patch implements the long anticipated (just kidding :) "anchor
3383 :     environment" mechanism. In the course of doing this, I also
3384 :     re-implemented CM's internal "SrcPath" module from scratch. The new
3385 :     one should be more robust in certain boundary cases. In any case, it
3386 :     is a lot cleaner than its predecessor (IMHO).
3387 :    
3388 :     This time, although there is yet another boot file format change, I
3389 :     kept the unpickler backward-compatible. As a result, no new bootfiles
3390 :     are necessary and bootstrapping is straightforward. (You cannot read
3391 :     new bootfiles into an old system, but the other way around is no
3392 :     problem.)
3393 :    
3394 :     Visible changes:
3395 :    
3396 : blume 670 ** 0. Implicit path anchors (without the leading $-symbol) are no
3397 :     longer recognized at all. This means that such path names are not
3398 :     illegal either. For example, the name basis.cm simply refers to a
3399 :     local file called "basis.cm" (i.e, the name is an ordinary path
3400 :     relative to .cm-files directory). Or, to put it differently, only
3401 :     names that start with $ are anchored paths.
3402 : blume 666
3403 : blume 670 ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
3404 :     vanished.
3405 :    
3406 :     John (Reppy) had critizised this as soon as I originally proposed and
3407 : blume 666 implemented it, but at that time I did not really deeply believe
3408 :     him. :) Now I came full-circle because I need the $<singlearc> syntax
3409 :     in another place where it cannot be seen as an abbreviation for
3410 :     $/<singlearc>. To avoid the confusion, $<singlearc> now means what it
3411 :     seems to mean (i.e., it "expands" into the corresponding anchor
3412 :     value).
3413 :    
3414 :     However, when paths are used as members in CM description files, it
3415 :     continues to be true that there must be at least another arc after the
3416 :     anchor. This is now enforced separately during semantic analysis
3417 : blume 670 (i.e., from a lexical/syntactical point of view, the notation is ok.)
3418 : blume 666
3419 :     ** 2. The "cm" class now accepts an option "bind". The option's value
3420 :     is a sub-option list of precisely two items -- one labeled "anchor"
3421 :     and the other one labeled "value". As you might expect, "anchor" is
3422 :     used to specify an anchor name to be bound, and "value" specifies what
3423 :     the anchor is being bound to.
3424 :    
3425 :     The value must be a directory name and can be given in either standard
3426 :     syntax (including the possibility that it is itself an anchored path)
3427 :     or native syntax.
3428 :    
3429 :     Examples:
3430 :    
3431 :     foo.cm (bind:(anchor:bar value:$mystuff/bar))
3432 :     lib.cm (bind:(anchor:a value:"H:\\x\\y\\z")) (* only works under windows *)
3433 :    
3434 :     and so on.
3435 :    
3436 :     The meaning of this is that the .cm-file will be processed with an
3437 :     augmented anchor environment where the given anchor(s) is/are bound to
3438 :     the given values(s).
3439 :    
3440 :     The rationale for having this feature is this: Suppose you are trying
3441 :     to use two different (already stable) libraries a.cm and b.cm (that
3442 :     you perhaps didn't write yourself). Further, suppose each of these
3443 :     two libraries internally uses its own auxiliary library $aux/lib.cm.
3444 :     Normally you would now have a problem because the anchor "lib" can not
3445 :     be bound to more than one value globally. Therefore, the project that
3446 :     uses both a.cm and b.cm must locally redirect the anchor to some other
3447 :     place:
3448 :    
3449 :     a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
3450 :     b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
3451 :    
3452 :     This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
3453 :     /usr/lib/smlnj/b-stuff/aux.cm, respectively.
3454 :    
3455 :     Hard-wiring path names is a bit inflexible (and CM will verbosely warn
3456 :     you when you do so at the time of CM.stabilize). Therefore, you can
3457 :     also use an anchored path as the value:
3458 :    
3459 :     a.cm (bind:(anchor:lib value:$a-lib))
3460 :     b.cm (bind:(anchor:lib value:$b-lib))
3461 :    
3462 :     Now you can globally configure (using the usual CM.Anchor.anchor or
3463 :     pathconfig machinery) bindings for "a-lib" and "b-lib". Since "lib"
3464 :     itself is always locally bound, setting it globally is no longer
3465 :     meaningful or necessary (but it does not hurt either). In fact, "lib"
3466 :     can still be used as a global anchor for separate purposes. As a
3467 :     matter of fact, one can locally define "lib" in terms of a global
3468 :     "lib":
3469 :    
3470 :     a.cm (bind:(anchor:lib value:$lib/a))
3471 :     b.cm (bind:(anchor:lib value:$lib/b))
3472 :    
3473 :     ** 3: The encoding of path names has changed. This affects the way
3474 :     path names are shown in CM's progress report and also the internal
3475 :     protocol encoding used for parallel make.
3476 :    
3477 :     The encoding now uses one or more ':'-separated segments. Each
3478 :     segments corresponds to a file that has been specified relative to the
3479 :     file given by its preceding segment. The first segment is either
3480 :     relative to the CWD, absolute, or anchored. Each segment itself is
3481 :     basically a Unix pathname; all segments but the first are relative.
3482 :    
3483 :     Example:
3484 :    
3485 :     $foo/bar/baz.cm:a/b/c.sml
3486 :    
3487 :     This path denotes the file bar/a/b/c.sml relative to the directory
3488 :     denoted by anchor "foo". Notice that the encoding also includes
3489 :     baz.cm which is the .cm-file that listed a/b/c.sml. As usual, such
3490 :     paths are resolved relative to the .cm-files directory, so baz.cm must
3491 :     be ignored to get the "real" pathname.
3492 :    
3493 :     To make this fact more obvious, CM puts the names of such "virtual
3494 :     arcs" into parentheses when they appear in progress reports. (No
3495 :     parentheses will appear in the internal protocol encoding.) Thus,
3496 :     what you really see is:
3497 :    
3498 :     $foo/bar/(baz.cm):a/b/c.sml
3499 :    
3500 :     I find this notation to be much more informative than before.
3501 :    
3502 :     Another new feature of the encoding is that special characters
3503 :     including parentheses, colons, (back)slashes, and white space are
3504 :     written as \ddd (where ddd is the decimal encoding of the character).
3505 :    
3506 : blume 670 *** The CM manual still needs to be updated.
3507 :    
3508 : blume 666 ----------------------------------------------------------------------
3509 : leunga 660 Name: Allen Leung
3510 :     Date: 2000/06/15 00:38:00
3511 :     Tag: leunga-20000615-x86-peephole
3512 :    
3513 :     x86 Peephole fix by Fermin. Affects c-- and moby only.
3514 :    
3515 :     ----------------------------------------------------------------------
3516 : blume 658 Name: Matthias Blume
3517 :     Date: 2000/06/12 11:40:00
3518 :     Tag: blume-20000612-parmakefix
3519 :     Description:
3520 :    
3521 :     More cleanup after changing the file naming scheme: This time I
3522 :     repaired the parallel make mechanism for CMB.make which I broke earlier.
3523 :    
3524 :     ----------------------------------------------------------------------
3525 : leunga 657 Name: Allen Leung
3526 :     Date: 2000/06/09 01:25:00
3527 :     Tag: leunga-20000609-various
3528 :    
3529 :     None of these things should affect normal SML/NJ operations
3530 :    
3531 :     1. Peephole improvements provided by Fermin (c--)
3532 :     2. New annotation DEFUSE for adding extra dependence (moby)
3533 :     3. New X86 LOCK instructions (moby)
3534 :     4. New machine description language for reservation tables (scheduling)
3535 :     5. Fixes to various optimization/analysis modules (branch chaining, dominator
3536 :     trees etc.)
3537 :     6. I've changed the CM files so that they can work with versions
3538 :     110.0.6, 110.25 and 110.28
3539 :    
3540 :     ----------------------------------------------------------------------
3541 : blume 652 Name: Matthias Blume
3542 : blume 656 Date: 2000/06/09 12:40:00
3543 :     Tag: blume-20000609-log
3544 :     Description:
3545 :    
3546 :     - Removed all(?) remaining RCS Log entries from sources.
3547 :    
3548 :     - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
3549 :     anchored paths).
3550 :    
3551 :     ----------------------------------------------------------------------
3552 :     Name: Matthias Blume
3553 : blume 653 Date: 2000/06/07 17:00:00 JST
3554 :     Tag: blume-20000607-no-implicit-anchors
3555 :     Description:
3556 :    
3557 :     1. This update changes the default setting for
3558 :     CM.Control.implicit_anchors from true to false. This means that
3559 :     implicit anchors are no longer permitted by default. I also tried to
3560 :     make sure that nothing else still relies on implicit anchors.
3561 :     (This is the next step on the schedule towards a CM that does not even
3562 :     have the notion of implicit anchors anymore.)
3563 :    
3564 :     2. More CM manual updates.
3565 :    
3566 :     3. I managed to track down and fix the pickling bug I mentioned last
3567 :     time. Because of the previously existing workaround, this entails no
3568 :     immediate practical changes.
3569 :    
3570 :     ----------------------------------------------------------------------
3571 :     Name: Matthias Blume
3572 : blume 652 Date: 2000/06/06 11:15:00 JST
3573 :     Tag: blume-20000606-lazierpickle
3574 :     Description:
3575 :    
3576 :     !!!! NEW BOOT FILES !!!!
3577 :    
3578 :     * The main purpose of this update is to make library pickles lazier in
3579 :     order to reduce the initial space penalty for autoloading a library.
3580 :     As a result, it is now possible to have $smlnj/compiler.cm
3581 :     pre-registered. This should take care of the many complaints or
3582 :     inquiries about missing structure Compiler. This required changes to
3583 :     CM's internal data structures and small tweaks to some algorithms.
3584 :    
3585 :     As a neat additional effect, it is no longer necessary (for the sake
3586 :     of lean heap image files) to distinguish between a "minimal" CM and a
3587 :     "full" CM. Now, there is only one CM (i.e., the "full" version:
3588 :     $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
3589 :     interactive top level. ($smlnj/cm/minimal.cm is gone.)
3590 :    
3591 :     To make the life of compiler-hackers easier, "makeml" now also
3592 :     pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm). In other
3593 :     words, after you bootstrap a new sml for the first time, you will not
3594 :     have to autoload $smlnj/cmb.cm again afterwards. (The first time
3595 :     around you will still have to do it, though.)
3596 :    
3597 :     * A second change consists of major updates to the CM manual. There
3598 :     are now several appendices with summary information and also a full
3599 :     specification of the CM description file syntax.
3600 :    
3601 :     * In directory src/system I added the script "allcross". This script
3602 :     invokes sml and cross-compiles the compiler for all supported
3603 :     architectures. (Useful when providing a new set of boot files.)
3604 :    
3605 :     * There seems to be a latent bug in my "lazy pickles" mechanism. I
3606 :     added a small tweak to pickle-util.sml to work around this problem,
3607 :     but it is not a proper fix yet. I will investigate further. (The
3608 :     effect of the bug was an inflation of library pickle size.)
3609 :    
3610 :     * Version number increased to 110.28.1 (to avoid compatibility problems).
3611 :    
3612 :     ----------------------------------------------------------------------
3613 : leunga 646 Name: Allen Leung
3614 : leunga 648 Date: 2000/05/25 17:28 EDT
3615 :     Tag: leunga-20000525-ra
3616 :     Description:
3617 :    
3618 :     Fixed a bug in freezing phase of the register allocator.
3619 :    
3620 :     ----------------------------------------------------------------------
3621 :     Name: Allen Leung
3622 : leunga 647 Date: 2000/05/15 22:53 EDT
3623 : leunga 646 Tag: leunga-20000515-alpha-x86-ra
3624 :     Description:
3625 :    
3626 :     1. Alpha
3627 :    
3628 :     Slight cleanup. Removed the instruction SGNXL
3629 :    
3630 :     2. X86
3631 :    
3632 :     Added the following instructions to the instruction set:
3633 :    
3634 :     ROLx, RORx,
3635 :     BTx, BTSx, BTLx, BTRx,
3636 :     XCHGx, and variants with the LOCK prefix
3637 :    
3638 :     3. Register Allocation
3639 :    
3640 :     The module ra-rewrite-with-renaming has been improved.
3641 :    
3642 : leunga 647 These have no effect on SML/NJ.
3643 :    
3644 : leunga 646 ----------------------------------------------------------------------
3645 : blume 642 Name: Matthias Blume
3646 : blume 645 Date: 2000/05/15 16:20:00 JST
3647 :     Tag: blume-20000515-lightrebuild
3648 :     Description:
3649 :    
3650 :     1. I added an alternative to "-rebuild" to "makeml". The difference is
3651 :     that prior to calling CMB.make' the CM-variable "LIGHT" will be
3652 :     defined. In effect, the command will not build any cross-compiler
3653 :     backends and therefore finish more quickly.
3654 :    
3655 :     The "fixpt" script also takes a "-light" switch to be able to use
3656 :     this new facility while compiling for a fixpoint.
3657 :    
3658 :     2. I replaced all mentions of anchored paths in group owner specifications
3659 :     with simple relative paths (usually starting with "..").
3660 :     The rationale is that a library's internal workings should not be
3661 :     compromised by the lack of some anchor. (An anchor is necessary
3662 :     for someone who wants to refer to the library by an anchored path,
3663 :     but it should not be necessary to build the same library in the first
3664 :     place.)
3665 :    
3666 :     3. I changed the way CM's tool mechanism determines the shell command
3667 :     string used for things like ml-yacc etc. so that it does not break
3668 :     when CM.Control.implicit_anchors is turned off.
3669 :    
3670 :     ----------------------------------------------------------------------
3671 :     Name: Matthias Blume
3672 : blume 644 Date: 2000/05/12 18:20:00 JST
3673 :     Tag: blume-20000512-ml-build
3674 :     Description:
3675 :    
3676 :     Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
3677 :     getting installed properly (by config/install.sh).
3678 :    
3679 :     ----------------------------------------------------------------------
3680 :     Name: Matthias Blume
3681 : blume 643 Date: 2000/05/12 17:30:00 JST
3682 :     Tag: blume-20000512-anchors
3683 :     Description:
3684 :    
3685 :     !!! NEW BOOT FILES !!!
3686 :    
3687 :     This change is in preparation of fading out support for "implicitly
3688 :     anchored path names". I went through all sources and used the
3689 :     explicit (and relatively new) $-notation. See system/README and the
3690 :     CM manual for more info on this.
3691 :    
3692 :     I also modified the anchoring scheme for some things such as "smlnj",
3693 :     "MLRISC", "cm", etc. to take advantage of the fact that explicit
3694 :     anchors are more expressive: anchor name and first arc do not have to
3695 :     coincide. This entails the following user-visible change:
3696 :    
3697 :     You have to write $smlnj/foo/bar instead of smlnj/foo/bar. In
3698 :     particular, when you fire up sml with a command-line argument, say,
3699 :     e.g.:
3700 :    
3701 :     sml '$smlnj/cmb.cm'
3702 :    
3703 :     At the ML toplevel prompt:
3704 :    
3705 :     CM.autoload "$smlnj/cmb.cm";
3706 :    
3707 :     There is also a new controller in CM.Control that can be used to turn
3708 :     off all remaining support for implicit anchors by saying:
3709 :    
3710 :     CM.autoload "$smlnj/
3711 :     #set CM.Control.implicit_anchors false;
3712 :    
3713 :     This causes CM to reject implicitly anchored paths. This is (for the
3714 :     time being) less permissive than the "final" version where there will
3715 :     be no more such implicit anchors and relative paths will be just that:
3716 :     relative.
3717 :    
3718 :     The next step (version after next version?) will be to make the
3719 :     default for CM.Control.implicit_anchors false. After the dust has
3720 :     settled, I can then produce the "final" version of this...
3721 :    
3722 :     Note: Since bootstrapping is a bit tricky, I provided new boot files.
3723 :    
3724 :     ----------------------------------------------------------------------
3725 :     Name: Matthias Blume
3726 : blume 642 Date: 2000/05/11 16:30:00 JST
3727 :     Tag: blume-20000511-sources
3728 :     Description:
3729 :    
3730 :     The main change is that I added function CM.sources as a generalized
3731 :     version of the earlier CM.makedepend. This entails the following
3732 :     additional changes:
3733 :    
3734 :     - CM.makedepend has been dropped.
3735 :    
3736 :     - CM manual has been updated.
3737 :    
3738 :     - TOOLS signature and API have been changed.
3739 :    
3740 :     ----------------------------------------------------------------------
3741 : leunga 640 Name: Allen Leung
3742 : leunga 641 Date: 2000/05/10 21:17 EDT
3743 :     Tag: leunga-20000510-moby-c--ssa
3744 :     Description:
3745 :    
3746 :     Various bug fixes and new features for C--, Moby and MLRISC optimizations.
3747 :     None of these affect SML/NJ.
3748 :    
3749 :     1. Register Allocation
3750 :    
3751 :     a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
3752 :     This module tries to remove local (i.e. basic block level) redundancies
3753 :     during spilling.
3754 :    
3755 :     b. A new framework for performing region based register allocation.
3756 :     Not yet entirely functional.
3757 :    
3758 :     2. X86
3759 :    
3760 :     a. DefUse for POP was missing the stack pointer [found by Lal]
3761 :     b. Reload for CALL was incorrect in X86Spill [found by John]
3762 :     c. Various fixes in X86Spill so that it can be used correctly for
3763 :     the new spilling module.
3764 :    
3765 :     3. SSA/IR
3766 :    
3767 :     a. New module ir/dj-dataflow.sml implements elimination based
3768 :     data flow analysis.
3769 :    
3770 :     4. MLRiscGen
3771 :    
3772 :     a. Fix for gc type annotation
3773 :    
3774 :     5. MDGen
3775 :    
3776 :     Various fixes for machine description -> ml code translation. For ssa
3777 :     only.
3778 :    
3779 :     ----------------------------------------------------------------------
3780 :     Name: Allen Leung
3781 : leunga 640 Date: 2000/05/08 22:17 EDT
3782 :     Tag: leunga-20000508-labexp
3783 :     Description:
3784 :    
3785 :     Fermin has found a few assembly problems with constant expressions
3786 :     generated in LabelExp. Mostly, the problems involve extra parentheses,
3787 :     which choke on dumb assemblers. This is his fix.
3788 :    
3789 :     ----------------------------------------------------------------------
3790 : dbm 635 Name: Dave MacQueen
3791 :     Date: 2000/04/09 14:00 EDT
3792 :     Tag: dbm-20000502-Version_110_28
3793 :     Description:
3794 :    
3795 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
3796 :    
3797 :     2. Updated config/version to 110.28
3798 :    
3799 :     3. Updated config/srcarchiveurl
3800 :    
3801 :     3. New boot files!
3802 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
3803 :    
3804 :     ----------------------------------------------------------------------
3805 : blume 634 Name: Matthias Blume
3806 :     Date: 2000/05/01 19:05:00 JST
3807 :     Tag: blume-20000501-noweb
3808 :     Description:
3809 :    
3810 :     A new noweb tool has been added. The existing system is entirely
3811 :     unaffected by this, but some CM users have asked for renewed noweb
3812 :     support. Everything is documented in the CM manual.
3813 :    
3814 :     New (plugin) libraries:
3815 :    
3816 :     noweb-tool.cm
3817 :     nw-ext.cm
3818 :    
3819 :     ----------------------------------------------------------------------
3820 : dbm 633 Name: Dave MacQueen
3821 :     Date: 2000/04/30 12:40PM EDT
3822 :     Tag: dbm-20000430-bug_fixes
3823 :     Description:
3824 :    
3825 :     1. Fix for bug 1498
3826 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sig
3827 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sml
3828 :     added toRealArray function
3829 :     smlnj/src/compiler/MiscUtil/print/ppobj.sml
3830 :     added check for tag Obj.RealArray to array printing case in ppObj
3831 :    
3832 :     2. Fix for bug 1510
3833 :     smlnj/src/compiler/Semant/types/typesutil.sml
3834 :     fixed definition of dummyargs (used by equalTycon) so that
3835 :     dummy args are distinct types
3836 :    
3837 :     ----------------------------------------------------------------------
3838 : blume 629 Name: Matthias Blume
3839 : blume 632 Date: 2000/04/30 01:00:00 JST
3840 :     Tag: blume-20000430-versions
3841 :     Description:
3842 :    
3843 :     1. CM version numbering added. This is an implementation of Lal's
3844 :     proposal for adding version numbers and version checking to .cm
3845 :     files. Lal said that his proposal was just that -- a proposal.
3846 :     For the time being I went ahead and implemented it so that people
3847 :     can comment on it. Everything is completely backward-compatible
3848 :     (except for the stable library format, i.e., new bootfiles!).
3849 :    
3850 :     As usual, see the CM manual for details.
3851 :    
3852 :     2. An alternative syntax for anchored paths has been implemented.
3853 :     Dave has recently voiced the same concerns that I had when I did
3854 :     this, so there should be some support. My take is that eventually
3855 :     I will let support for the current syntax (where anchors are
3856 :     "implicit") fade out in favor of the new, explicit syntax.
3857 :     In order to be backward-compatible, both old and new syntax are
3858 :     currently supported.
3859 :    
3860 :     Again, see the CM manual for details.
3861 :    
3862 :     3. Parallel make is trying to be slightly smarter: When the master
3863 :     process finds a "bottleneck", i.e., when there is only one
3864 :     compilation unit that can be compiled and everybody else is
3865 :     waiting on it, then it will simply compile it directly instead
3866 :     of clumsily telling one of the slaves to do it.
3867 :    
3868 :     4. Support for "unsharing" added. This is necessary in order to be
3869 :     able to have two different versions of the same library running
3870 :     at the same time (e.g., for trying out a new MLRISC while still
3871 :     having the old MLRISC linked into the current compiler, etc.)
3872 :     See the CM manual.
3873 :    
3874 :     5. Simple "makedepend" functionality added for generating Makefile
3875 :     dependency information. (This is rather crude at the moment.
3876 :     Expect some changes here in the future.)
3877 :    
3878 :     6. ".fun" added as a recognized suffix for ML files. Also documented
3879 :     explicitly in the manual that the fallback behavior (unknown suffix
3880 :     -> ML file) is not an official feature!
3881 :    
3882 :     7. Small changes to the pickler for stable libraries.
3883 :    
3884 :     8. Several internal changes to CM (for cleanup/improvement).
3885 :    
3886 :    
3887 :     !!!! NEW BINFILES !!!!
3888 :    
3889 :     ----------------------------------------------------------------------
3890 :     Name: Matthias Blume
3891 : blume 631 Date: 2000/04/28 17:30:00 JST
3892 :     Tag: blume-20000428-pathconfig
3893 :     Description:
3894 :    
3895 :     1. I changed config/install.sh to remove duplicate entries from the
3896 :     lib/pathconfig file at the end. Moreover, the final version of
3897 :     lib/pathconfig is sorted alphabetically. The same (sorting) is done
3898 :     in src/system/installml.
3899 :    
3900 :     2. The config/install.sh script now consistently uses relative
3901 :     pathnames in lib/pathconfig whenever the anchor is in the lib
3902 :     directory. (So far this was true for the libraries that come
3903 :     pre-compiled and bundled as part of the bootfiles but not for
3904 :     libraries that are compiled by the script itself.)
3905 :    
3906 :     ----------------------------------------------------------------------
3907 :     Name: Matthias Blume
3908 : blume 629 Date: 2000/04/26 13:10:00 JST
3909 :     Tag: blume-20000426-fun_suffix
3910 :     Description:
3911 :    
3912 :     Added ".fun" as a recognized file name suffix (for ML code).
3913 :    
3914 :     ----------------------------------------------------------------------
3915 : leunga 628 Name: Allen Leung
3916 :     Date: 2000/04/25 17:00:00 EST
3917 :     Tag: leunga-20000425-alpha-ra
3918 :     Description:
3919 :    
3920 :     1. Alpha
3921 :    
3922 :     PSEUDOARITH was missing in AlphaRewrite. This causes an endless loop
3923 :     in C--.
3924 :    
3925 :     2. RA
3926 :    
3927 :     Added a flag "ra-dump-size" to print out the size of the flowgraph
3928 :     and the interference graph.
3929 :    
3930 :     ----------------------------------------------------------------------
3931 : dbm 627 Name: Dave MacQueen
3932 :     Date: 2000/04/25/
3933 :     Tag: dbm-20000425-mlyacc_doc_examples
3934 :     Description:
3935 :     Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
3936 :     Updated all three examples in src/ml-yacc/examples to run
3937 :     under 110.* using CM.make.
3938 :    
3939 :     ----------------------------------------------------------------------
3940 : leunga 624 Name: Allen Leung
3941 :     Date: 2000/04/20 23:04:00 EST
3942 :     Tag: leunga-20000420-ssa-c---stuff
3943 :     Description:
3944 :    
3945 :     This update synchronizes my repository with Yale's. Most of these
3946 :     changes, however, do not affect SML/NJ at all (the RA is an exception).
3947 :    
3948 :     1. Register Allocator
3949 :    
3950 :     a. An improvement in the interference graph construction:
3951 :     Given a copy
3952 :    
3953 :     s <- t
3954 :    
3955 : leunga 628 no interference edge between s and t is added for this definition of s.
3956 : leunga 624
3957 :     b. I've added two new spill heuristic modules that Fermin and I developed
3958 :     (in the new library RA.cm). These are unused in SML/NJ but maybe
3959 :     useful for others (Moby?)
3960 :    
3961 :     2. X86
3962 :    
3963 :     a. Various fixes in the backend provided by Fermin [C--] and Lal.
3964 :    
3965 :     3. Alpha
3966 :    
3967 :     a. Added the BSR instruction and code generation that goes with it [C--]
3968 :     b. Other fixes too numerous to recount provided by Fermin [C--]
3969 :    
3970 :     4. Regmaps
3971 :    
3972 :     a. The regmaps are not initialized with the identity physical bindings
3973 :     at creation time. This is unneeded.
3974 :    
3975 :     5. MLRISC Optimizations
3976 :    
3977 :     a. The DJ-Graph module can now compute the iterated dominance frontiers
3978 :     intersects with liveness incrementally in linear time! Woohoo!
3979 :     This is now used in my new SSA construction algorithm.
3980 :    
3981 :     b. THe branch reorganization module is now smarter about linear chains of
3982 :     basic blocks.
3983 :    
3984 :    
3985 :     ----------------------------------------------------------------------
3986 : blume 615 Name: Matthias Blume
3987 :     Date: 2000/04/12 13:52:00 JST
3988 :     Tag: blume_main_v110p27_1
3989 :     Description:
3990 :    
3991 :     Changed install.sh script to handle archive files without version number
3992 :     and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
3993 :     name of the boot file archive.
3994 :    
3995 :     ----------------------------------------------------------------------
3996 : dbm 607 Name: Dave MacQueen
3997 :     Date: 2000/04/09 14:00 EDT
3998 :     Tag: dbm-20000410-Version_110_27
3999 :     Description:
4000 :    
4001 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
4002 :    
4003 :     2. Updated src/config/version to 110.27
4004 :    
4005 :     3. New boot files!
4006 :    
4007 :     ----------------------------------------------------------------------
4008 : leunga 606 Name: Allen Leung
4009 :     Date: 2000/04/09 19:09:00 EST
4010 :     Tag: leunga-20000409-misc
4011 :     Description:
4012 :    
4013 :     1. Yet another fix for x86 assembly for idivl, imull, mull and friends.
4014 :    
4015 :     2. Miscellaneous improvements to MLRISC (unused in sml/nj)
4016 :    
4017 :     ----------------------------------------------------------------------
4018 : monnier 605 Name: Stefan
4019 :     Date: 2000/04/07 10:00:00 EDT
4020 :     Tag: monnier-20000406-branch-handling
4021 :     Description:
4022 :    
4023 :     Improved handling of branches (mostly those generated from
4024 :     polymorphic equality), removed switchoff and changed the
4025 :     default optimization settings (more cpsopt and less flintopt).
4026 :    
4027 :     ----------------------------------------------------------------------
4028 : leunga 601 Name: Allen Leung
4029 : leunga 602 Date: 2000/04/06 01:30:00 EST
4030 :     Tag: leunga-20000406-peephole-x86-SSA-2
4031 :     Description:
4032 :    
4033 :     Forgot a few files.
4034 :    
4035 :     ----------------------------------------------------------------------
4036 :     Name: Allen Leung
4037 : leunga 601 Date: 2000/04/06 00:36:00 EST
4038 :     Tag: leunga-20000406-peephole-x86-SSA
4039 :     Description:
4040 :    
4041 :     1. New Peephole code
4042 :    
4043 :     2. Minor improvement to X86 instruction selection
4044 :    
4045 :     3. Various fixes to SSA and machine description -> code translator
4046 :    
4047 :     ----------------------------------------------------------------------
4048 : blume 595 Name: Matthias Blume
4049 :     Date: 2000/04/05 12:30:00 JST
4050 :     Tag: blume_main_v110p26p2_3
4051 :     Description:
4052 :    
4053 :     This update just merges three minor cosmetic updates to CM's sources
4054 :     to get ready for the 110.27 code freeze on Friday. No functionality
4055 :     has changed.
4056 :    
4057 :     ----------------------------------------------------------------------
4058 : leunga 593 Name: Allen Leung
4059 : leunga 594 Date: 2000/04/04 19:39:00 EST
4060 :     Tag: leunga-20000404-x86-asm
4061 :     Description:
4062 :    
4063 :     1. Fixed a problem in X86 assembly.
4064 :    
4065 :     Things like
4066 :    
4067 :     jmp %eax
4068 :     jmp (%eax)
4069 :    
4070 :     should be output as
4071 :    
4072 :     jmp *%eax
4073 :     jmp *(%eax)
4074 :    
4075 :     2. Assembly output
4076 :    
4077 :     Added a new flag
4078 :    
4079 :     "asm-indent-copies" (default to false)
4080 :    
4081 :     When this flag is on, parallel copies will be indented an extra level.
4082 :    
4083 :     ----------------------------------------------------------------------
4084 :     Name: Allen Leung
4085 : leunga 593 Date: 2000/04/04 03:18:00 EST
4086 :     Tag: leunga-20000404-C--Moby
4087 :     Description:
4088 :    
4089 :     All of these fixes are related to C--, Moby, and my own optimization
4090 :     stuff; so they shouldn't affect SML/NJ.
4091 :    
4092 :     1. X86
4093 :    
4094 :     Various fixes related floating point, and extensions.
4095 :    
4096 :     2. Alpha
4097 :    
4098 :     Some extra patterns related to loads with signed/zero extension
4099 :     provided by Fermin.
4100 :    
4101 :     3. Assembly
4102 :    
4103 : monnier 659 When generating assembly, resolve the value of client defined constants,
4104 : leunga 593 instead of generating symbolic values. This is controlled by the
4105 :     new flag "asm-resolve-constants", which is default to true.
4106 :    
4107 :     4. Machine Descriptions
4108 :    
4109 :     a. The precedence parser was slightly broken when parsing infixr symbols.
4110 :     b. The type generalizing code had the bound variables reversed, resulting
4111 :     in a problem during arity raising.
4112 :     c. Various fixes in machine descriptions.
4113 :    
4114 :     ----------------------------------------------------------------------
4115 : blume 592 Name: Matthias Blume
4116 :     Date: 2000/04/03 16:05:00 JST
4117 :     Tag: blume_main_v110p26p2_2
4118 :     Description:
4119 :    
4120 :     I eliminated coreEnv from compInfo. Access to the "Core" structure is
4121 :     now done via the ordinary static environment that is context to each
4122 :     compilation unit.
4123 :    
4124 :     To this end, I arranged that instead of "structure Core" as "structure
4125 :     _Core" is bound in the pervasive environment. Core access is done via
4126 : monnier 659 _Core (which can never be accidentally rebound because _Core is not a
4127 : blume 592 legal surface-syntax symbol).
4128 :    
4129 :     The current solution is much cleaner because the core environment is
4130 :     now simply part of the pervasive environment which is part of every
4131 :     compilation unit's context anyway. In particular, this eliminates all
4132 :     special-case handling that was necessary until now in order to deal
4133 :     with dynamic and symbolic parts of the core environment.
4134 :    
4135 :     Remaining hackery (to bind the "magic" symbol _Core) is localized in the
4136 : monnier 659 compilation manager's bootstrap compiler (actually: in the "init group"
4137 : blume 592 handling). See the comments in src/system/smlnj/init/init.cmi for
4138 :     more details.
4139 :    
4140 :     I also tried to track down all mentions of "Core" (as string argument
4141 :     to Symbol.strSymbol) in the compiler and replaced them with a
4142 :     reference to the new CoreSym.coreSym. Seems cleaner since the actual
4143 :     name appears in one place only.
4144 :    
4145 :     Binfile and bootfile format have not changed, but the switchover from
4146 :     the old "init.cmi" to the new one is a bit tricky, so I supplied new
4147 :     bootfiles anyway.
4148 :    
4149 :     ----------------------------------------------------------------------
4150 : leunga 591 Name: Allen Leung
4151 :     Date: 2000/04/02 21:17:00 EST
4152 :     Tag: leunga-20000402-mltree
4153 :     Description:
4154 : leunga 585
4155 : leunga 591 1. Renamed the constructor CALL in MLTREE by popular demand.
4156 :     2. Added a bunch of files from my repository. These are currently
4157 :     used by other non-SMLNJ backends.
4158 :    
4159 : leunga 576 ----------------------------------------------------------------------
4160 : leunga 590 Name: Allen Leung
4161 :     Date: 2000/03/31 21:15:00 EST
4162 :     Tag: leunga-20000331-aliasing
4163 :     Description:
4164 :    
4165 :     This update contains a rewritten (and hopefully more correct) module
4166 :     for extracting aliasing information from CPS.
4167 :    
4168 :     To turn on this feature:
4169 :    
4170 :     Compiler.Control.CG.memDisambiguate := true
4171 :    
4172 :     To pretty print the region information with assembly
4173 :    
4174 :     Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
4175 :    
4176 :     To control how many levels of aliasing information are printed, use:
4177 :    
4178 :     Compiler.Control.MLRISC.getInt "points-to-show-level" := n
4179 :    
4180 :     The default of n is 3.
4181 :    
4182 :     ----------------------------------------------------------------------
4183 : dbm 589 Name: David MacQueen
4184 :     Date: 2000/03/31 11:15:00 EST
4185 :     Tag: dbm-20000331-runtime_fix
4186 :     Description:
4187 :    
4188 :     This update contains:
4189 :    
4190 :     1. runtime/c-lib/c-libraries.c
4191 :     includes added in revision 1.2 caused compilation errors on hppa-hpux
4192 :    
4193 :     2. fix for bug 1556
4194 :     system/Basis/Implementation/NJ/internal-signals.sml
4195 :    
4196 :     ----------------------------------------------------------------------
4197 : blume 587 Name: Matthias Blume
4198 : blume 588 Date: 2000/03/31 18:00:00 JST
4199 :     Tag: blume_main_v110p26p2_1
4200 :     Description:
4201 :    
4202 :     This update contains:
4203 :    
4204 :     1. A small change to CM's handling of stable libraries:
4205 :     CM now maintains one "global" modmap that is used for all stable
4206 :     libraries. The use of such a global modmap maximizes sharing and
4207 :     minimizes the need for re-traversing parts of environments during
4208 :     modmap construction. (However, this has minor impact since modmap
4209 :     construction seems to account for just one percent or less of total
4210 :     compile time.)
4211 :    
4212 :     2. I added a "genmap" phase to the statistics. This is where I got the
4213 :     "one percent" number (see above).
4214 :    
4215 :     3. CM's new tool parameter mechanism just became _even_ better. :)
4216 :     - The parser understands named parameters and recursive options.
4217 :     - The "make" and "shell" tools use these new features.
4218 :     (This makes it a lot easier to cascade these tools.)
4219 :     - There is a small syntax change: named parameters use a
4220 :    
4221 :     <name> : ( <option> ... ) or
4222 :     <name> : <string>
4223 :    
4224 :     syntax. Previously, named parameters were implemented in an
4225 :     ad-hoc fashion by each tool individually (by parsing strings)
4226 :     and had the form
4227 :    
4228 :     <name>=<string>
4229 :    
4230 :     See the CM manual for a full description of these issues.
4231 :    
4232 :     ----------------------------------------------------------------------
4233 :     Name: Matthias Blume
4234 : blume 587 Date: 2000/03/30 18:00:00 JST
4235 :     Tag: blume_main_v110p26p2_0
4236 :     Description:
4237 :    
4238 :     !!!!! WARNING !!!!!!
4239 :     !! New binfiles !!
4240 :     !!!!!!!!!!!!!!!!!!!!
4241 :    
4242 :     This update contains:
4243 :    
4244 :     1. Moderate changes to CM:
4245 :    
4246 :     - Changes to CM's tools mechanism. In particular, it is now possible
4247 :     to have tools that accept additional "command line" parameters
4248 :     (specified in the .cm file at each instance where the tool's class is
4249 :     used).
4250 :    
4251 : monnier 659 This was done to accommodate the new "make" and "shell" tools which
4252 :     facilitate fairly seamless hookup to portions of code managed using
4253 : blume 587 Makefiles or Shell scripts.
4254 :    
4255 :     There are no classes "shared" or "private" anymore. Instead, the
4256 :     sharing annotation is now a parameter to the "sml" class.
4257 :    
4258 :     There is a bit of generic machinery for implementing one's own
4259 :     tools that accept command-line parameters. However, I am not yet fully
4260 :     satisfied with that part, so expect changes here in the future.
4261 :    
4262 :     All existing tools are described in the CM manual.
4263 :    
4264 : monnier 659 - Slightly better error handling. (CM now suppresses many followup
4265 : blume 587 error messages that tended to be more annoying than helpful.)
4266 :    
4267 :     2. Major changes to the compiler's static environment data structures.
4268 :    
4269 :     - no CMStaticEnv anymore.
4270 :     - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
4271 :     but it is called Environment), no conversions between different
4272 :     kinds of static environments
4273 :    
4274 :     - There is still a notion of a "modmap", but such modmaps are generated
4275 :     on demand at the time when they are needed. This sounds slow, but I
4276 :     sped up the code that generates modmaps enough for this not to lead to
4277 :     a slowdown of the compiler (at least I didn't detect any).
4278 :    
4279 :     - To facilitate rapid modmap generation, static environments now
4280 :     contain an (optional) "modtree" structure. Modtree annotations are
4281 :     constructed by the unpickler during unpickling. (This means that
4282 :     the elaborator does not have to worry about modtrees at all.)
4283 :     Modtrees have the advantage that they are compositional in the same
4284 :     way as the environment data structure itself is compositional.
4285 :     As a result, modtrees never hang on to parts of an environment that
4286 :     has already been rendered "stale" by filtering or rebinding.
4287 :    
4288 :     - I went through many, many trials and errors before arriving at the
4289 :     current solution. (The initial idea of "linkpaths" did not work.)
4290 :     But the result of all this is that I have touched a lot of files that
4291 :     depend on the "modules" and "types" data structures (most of the
4292 :     elaborator). There were a lot of changes during my "linkpath" trials
4293 :     that could have been reverted to their original state but weren't.
4294 :     Please, don't be too harsh on me for messing with this code a bit more
4295 : monnier 921 than what was strictly necessary... (I _did_ resist the temptation
4296 : blume 587 of doing any "global reformatting" to avoid an untimely death at
4297 :     Dave's hands. :)
4298 :    
4299 :     - One positive aspect of the previous point: At least I made sure that
4300 :     all files that I touched now compile without warnings (other than
4301 :     "polyEqual").
4302 :    
4303 :     - compiler now tends to run "leaner" (i.e., ties up less memory in
4304 :     redundant modmaps)
4305 :    
4306 :     ----------------------------------------------------------------------
4307 : leunga 580 Name: Allen Leung
4308 : leunga 585 Date: 2000/03/29 18:00:00
4309 :     Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
4310 : leunga 586 Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
4311 : leunga 585 Description:
4312 :    
4313 :     This update contains *MAJOR* changes to the way code is generated from CPS
4314 :     in the module mlriscGen, and in various backend modules.
4315 :    
4316 :     CHANGES
4317 :     =======
4318 :    
4319 :     1. MLRiscGen: forward propagation fix.
4320 :    
4321 :     There was a bug in forward propagat