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

SCM Repository

[smlnj] Annotation of /sml/branches/primop-branch-3/NOTES/HISTORY
ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1145 - (view) (download)
Original Path: sml/trunk/HISTORY

1 : dbm 570 This is the HISTORY file for the Yale SML/NJ CVS repository.
2 :    
3 :     An entry should be made for _every_ commit to the repository.
4 :     The entries in this file will be used when creating the README
5 :     for new versions, so keep that in mind when writing the
6 :     description.
7 :    
8 :     The form of an entry should be:
9 :    
10 :     Name:
11 : george 1003 Date: yyyy/mm/dd
12 : dbm 570 Tag: <post-commit CVS tag>
13 :     Description:
14 : leunga 1142
15 : george 1136 ----------------------------------------------------------------------
16 : blume 1145 Name: Matthias Blume
17 :     Date: 2002/03/14 21:30:00 EST
18 :     Tag: blume-20020314-controls
19 :     Description:
20 :    
21 :     Controls:
22 :     ---------
23 :    
24 :     1. Factored out the recently-added Controls : CONTROLS stuff and put
25 :     it into its own library $/controls-lib.cm. The source tree for
26 :     this is under src/smlnj-lib/Controls.
27 :    
28 :     2. Changed the names of types and functions in this interface, so they
29 :     make a bit more "sense":
30 :    
31 :     module -> registry
32 :     'a registry -> 'a group
33 :    
34 :     3. The interface now deals in ref cells only. The getter/setter interface
35 :     is (mostly) gone.
36 :    
37 :     4. Added a function that lets one register an already-existing ref cell.
38 :    
39 :     5. Made the corresponding modifications to the rest of the code so that
40 :     everything compiles again.
41 :    
42 :     6. Changed the implementation of Controls.MLRISC back to something closer
43 :     to the original. In particular, this module (and therefore MLRISC)
44 :     does not depend on Controls. There now is some link-time code in
45 :     int-sys.sml that registers the MLRISC controls with the Controls
46 :     module.
47 :    
48 :     CM:
49 :     ---
50 :    
51 :     * One can now specify the lambda-split aggressiveness in init.cmi.
52 :    
53 :     ----------------------------------------------------------------------
54 : leunga 1142 Name: Allen Leung
55 :     Date: 2002/03/13 17:30:00 EST
56 :     Tag: leunga-20020313-x86-fp-unary
57 :     Description:
58 :    
59 :     Bug fix for:
60 :    
61 :     > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
62 :     > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
63 :     > - fun f(x,(y,z)) = Real.~ y;
64 :     > [autoloading]
65 :     > [autoloading done]
66 :     > fchsl (%eax), 184(%esp)
67 :     > Error: MLRisc bug: X86MCEmitter.emitInstr
68 :     >
69 :     > uncaught exception Error
70 :     > raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
71 :    
72 :     The problem was that the code generator did not generate any fp registers
73 :     in this case, and the ra didn't know that it needed to run the X86FP phase to
74 :     translate the pseudo fp instruction. This only happened with unary fp
75 :     operators in certain situations.
76 :    
77 :     ----------------------------------------------------------------------
78 : blume 1137 Name: Matthias Blume
79 : blume 1140 Date: 2002/03/13 14:00:00 EST
80 :     Tag: blume-20020313-overload-etc
81 :     Description:
82 :    
83 :     1. Added _overload as a synonym for overload for backward compatibility.
84 :     (Control.overloadKW must be true for either version to be accepted.)
85 :    
86 :     2. Fixed bug in install script that caused more things to be installed
87 :     than what was requested in config/targets.
88 :    
89 :     3. Made CM aware of the (_)overload construct so that autoloading
90 :     works.
91 :    
92 :     ----------------------------------------------------------------------
93 :     Name: Matthias Blume
94 : blume 1138 Date: 2002/03/12 22:03:00 EST
95 :     Tag: blume-20020312-url
96 :     Description:
97 :    
98 :     Forgot to update BOOT and srcarchiveurl.
99 :    
100 :     ----------------------------------------------------------------------
101 :     Name: Matthias Blume
102 : blume 1137 Date: 2002/03/12 17:30:00 EST
103 :     Tag: blume-20020312-version110392
104 :     Description:
105 :    
106 :     Yet another version number bump (because of small changes to the
107 :     binfile format). Version number is now 110.39.2. NEW BOOTFILES!
108 :    
109 :     Changes:
110 :    
111 :     The new pid generation scheme described a few weeks ago was overly
112 :     complicated. I implemented a new mechanism that is simpler and
113 :     provides a bit more "stability": Once CM has seen a compilation
114 :     unit, it keeps its identity constant (as long as you do not delete
115 :     those crucial CM/GUID/* files). This means that when you change
116 : blume 1145 an interface, compile, then go back to the old interface, and
117 : blume 1137 compile again, you arrive at the original pid.
118 :    
119 :     There now also is a mechanism that instructs CM to use the plain
120 :     environment hash as a module's pid (effectively making its GUID
121 :     the empty string). For this, "noguid" must be specified as an
122 :     option to the .sml file in question within its .cm file.
123 :     This is most useful for code that is being generated by tools such
124 :     as ml-nlffigen (because during development programmers tend to
125 :     erase the tool's entire output directory tree including CM's cached
126 :     GUIDs). "noguid" is somewhat dangerous (since it can be used to locally
127 :     revert to the old, broken behavior of SML/NJ, but in specific cases
128 :     where there is no danger of interface confusion, its use is ok
129 :     (I think).
130 :    
131 :     ml-nlffigen by default generates "noguid" annotations. They can be
132 :     turned off by specifying -guid in its command line.
133 :    
134 :     ----------------------------------------------------------------------
135 : george 1136 Name: Lal George
136 :     Date: 2002/03/12 12 14:42:36 EST
137 :     Tag: george-20020312-frequency-computation
138 :     Description:
139 :    
140 :     Integrated jump chaining and static block frequency into the
141 :     compiler. More details and numbers later.
142 : leunga 1025
143 : george 1003 ----------------------------------------------------------------------
144 : george 1133 Name: Lal George
145 :     Date: 2002/03/11 11 22:38:53 EST
146 :     Tag: george-20020311-jump-chain-elim
147 :     Description:
148 :    
149 :     Tested the jump chain elimination on all architectures (except the
150 :     hppa). This is on by default right now and is profitable for the
151 :     alpha and x86, however, it may not be profitable for the sparc and ppc
152 :     when compiling the compiler.
153 :    
154 :     The gc test will typically jump to a label at the end of the cluster,
155 :     where there is another jump to an external cluster containing the actual
156 :     code to invoke gc. This is to allow factoring of common gc invocation
157 :     sequences. That is to say, we generate:
158 :    
159 :     f:
160 :     testgc
161 :     ja L1 % jump if above to L1
162 :    
163 :     L1:
164 :     jmp L2
165 :    
166 :    
167 :     After jump chain elimination the 'ja L1' instructions is converted to
168 :     'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
169 :     up being implemented in their long form (if L2 is far away) using:
170 :    
171 :     jbe L3 % jump if below or equal to L3
172 :     jmp L2
173 :     L3:
174 :     ...
175 :    
176 :    
177 :     For large compilation units L2 may be far away.
178 :    
179 :    
180 :     ----------------------------------------------------------------------
181 : blume 1132 Name: Matthias Blume
182 :     Date: 2002/03/11 13:30:00 EST
183 :     Tag: blume-20020311-mltreeeval
184 :     Description:
185 :    
186 :     A functor parameter was missing.
187 :    
188 :     ----------------------------------------------------------------------
189 : leunga 1130 Name: Allen Leung
190 : leunga 1131 Date: 2002/03/11 10:30:00 EST
191 : leunga 1142 Tag: leunga-20020311-runtime-string0
192 : leunga 1131 Description:
193 :    
194 :     The representation of the empty string now points to a
195 :     legal null terminated C string instead of unit. It is now possible
196 :     to convert an ML string into C string with InlineT.CharVector.getData.
197 :     This compiles into one single machine instruction.
198 :    
199 :     ----------------------------------------------------------------------
200 :     Name: Allen Leung
201 : leunga 1130 Date: 2002/03/10 23:55:00 EST
202 :     Tag: leunga-20020310-x86-call
203 :     Description:
204 :    
205 :     Added machine generation for CALL instruction (relative displacement mode)
206 :    
207 :     ----------------------------------------------------------------------
208 : blume 1128 Name: Matthias Blume
209 :     Date: 2002/03/08 16:05:00
210 :     Tag: blume-20020308-entrypoints
211 :     Description:
212 :    
213 :     Version number bumped to 110.39.1. NEW BOOTFILES!
214 :    
215 :     Entrypoints: non-zero offset into a code object where execution should begin.
216 :    
217 :     - Added the notion of an entrypoint to CodeObj.
218 :     - Added reading/writing of entrypoint info to Binfile.
219 :     - Made runtime system bootloader aware of entrypoints.
220 :     - Use the address of the label of the first function given to mlriscGen
221 :     as the entrypoint. This address is currently always 0, but it will
222 :     not be 0 once we turn on block placement.
223 :     - Removed the linkage cluster code (which was The Other Way(tm) of dealing
224 :     with entry points) from mlriscGen.
225 :    
226 :     ----------------------------------------------------------------------
227 : leunga 1127 Name: Allen Leung
228 :     Date: 2002/03/07 20:45:00 EST
229 :     Tag: leunga-20020307-x86-cmov
230 :     Description:
231 :    
232 :     Bug fixes for CMOVcc on x86.
233 :    
234 :     1. Added machine code generation for CMOVcc
235 :     2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
236 :     3. CMOVcc cannot have an immediate operand as argument.
237 :    
238 :     ----------------------------------------------------------------------
239 : blume 1126 Name: Matthias Blume
240 :     Date: 2002/03/07 16:15:00 EST
241 :     Tag: blume-20020307-controls
242 :     Description:
243 :    
244 :     This is a very large but mostly boring patch which makes (almost)
245 :     every tuneable compiler knob (i.e., pretty much everything under
246 :     Control.* plus a few other things) configurable via both the command
247 :     line and environment variables in the style CM did its configuration
248 :     until now.
249 :    
250 :     Try starting sml with '-h' (or, if you are brave, '-H')
251 :    
252 :     To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
253 :     implements the underlying generic mechanism.
254 :    
255 :     The interface to some of the existing such facilities has changed somewhat.
256 :     For example, the MLRiscControl module now provides mkFoo instead of getFoo.
257 :     (The getFoo interface is still there for backward-compatibility, but its
258 :     use is deprecated.)
259 :    
260 :     The ml-build script passes -Cxxx=yyy command-line arguments through so
261 :     that one can now twiddle the compiler settings when using this "batch"
262 :     compiler.
263 :    
264 :     TODO items:
265 :    
266 :     We should go through and throw out all controls that are no longer
267 :     connected to anything. Moreover, we should go through and provide
268 :     meaningful (and correct!) documentation strings for those controls
269 :     that still are connected.
270 :    
271 :     Currently, multiple calls to Controls.new are accepted (only the first
272 :     has any effect). Eventually we should make sure that every control
273 :     is being made (via Controls.new) exactly once. Future access can then
274 :     be done using Controls.acc.
275 :    
276 :     Finally, it would probably be a good idea to use the getter-setter
277 :     interface to controls rather than ref cells. For the time being, both
278 :     styles are provided by the Controls module, but getter-setter pairs are
279 :     better if thread-safety is of any concern because they can be wrapped.
280 :    
281 :     *****************************************
282 :    
283 :     One bug fix: The function blockPlacement in three of the MLRISC
284 :     backpatch files used to be hard-wired to one of two possibilities at
285 :     link time (according to the value of the placementFlag). But (I
286 :     think) it should rather sense the flag every time.
287 :    
288 :     *****************************************
289 :    
290 :     Other assorted changes (by other people who did not supply a HISTORY entry):
291 :    
292 :     1. the cross-module inliner now works much better (Monnier)
293 :     2. representation of weights, frequencies, and probabilities in MLRISC
294 :     changed in preparation of using those for weighted block placement
295 :     (Reppy, George)
296 :    
297 :     ----------------------------------------------------------------------
298 : george 1116 Name: Lal George
299 : george 1124 Date: 2002/03/07 14:44:24 EST 2002
300 :     Tag: george-20020307-weighted-block-placement
301 :    
302 :     Tested the weighted block placement optimization on all architectures
303 :     (except the hppa) using AMPL to generate the block and edge frequencies.
304 :     Changes were required in the machine properties to correctly
305 :     categorize trap instructions. There is an MLRISC flag
306 :     "weighted-block-placement" that can be used to enable weighted block
307 :     placement, but this will be ineffective without block/edge
308 :     frequencies (coming soon).
309 :    
310 :    
311 :     ----------------------------------------------------------------------
312 :     Name: Lal George
313 : george 1116 Date: 2002/03/05 17:24:48 EST
314 :     Tag: george-20020305-linkage-cluster
315 :    
316 : george 1124 In order to support the block placement optimization, a new cluster
317 :     is generated as the very first cluster (called the linkage cluster).
318 :     It contains a single jump to the 'real' entry point for the compilation
319 :     unit. Block placement has no effect on the linkage cluster itself, but
320 :     all the other clusters have full freedom in the manner in which they
321 :     reorder blocks or functions.
322 : george 1116
323 :     On the x86 the typical linkage code that is generated is:
324 :     ----------------------
325 :     .align 2
326 :     L0:
327 :     addl $L1-L0, 72(%esp)
328 : george 1124 jmp L1
329 : george 1116
330 :    
331 :     .align 2
332 :     L1:
333 :     ----------------------
334 :    
335 :     72(%esp) is the memory location for the stdlink register. This
336 :     must contain the address of the CPS function being called. In the
337 : george 1124 above example, it contains the address of L0; before
338 : george 1116 calling L1 (the real entry point for the compilation unit), it
339 :     must contain the address for L1, and hence
340 :    
341 :     addl $L1-L0, 72(%esp)
342 :    
343 : george 1124 I have tested this on all architectures except the hppa.The increase
344 :     in code size is of course negligible
345 : george 1116
346 :     ----------------------------------------------------------------------
347 : leunga 1115 Name: Allen Leung
348 :     Date: 2002/03/03 13:20:00 EST
349 :     Tag: leunga-20020303-mlrisc-tools
350 :    
351 :     Added #[ ... ] expressions to mlrisc tools
352 :    
353 :     ----------------------------------------------------------------------
354 : blume 1096 Name: Matthias Blume
355 : blume 1098 Date: 2002/02/27 12:29:00 EST
356 :     Tag: blume-20020227-cdebug
357 :     Description:
358 :    
359 :     - made types in structure C and C_Debug to be equal
360 :     - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
361 :     - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
362 :    
363 :     ----------------------------------------------------------------------
364 :     Name: Matthias Blume
365 : blume 1096 Date: 2002/02/26 12:00:00 EST
366 :     Tag: blume-20020226-ffi
367 :     Description:
368 :    
369 :     1. Fixed a minor bug in CM's "noweb" tool:
370 :     If numbering is turned off, then truly don't number (i.e., do not
371 :     supply the -L option to noweb). The previous behavior was to supply
372 :     -L'' -- which caused noweb to use the "default" line numbering scheme.
373 :     Thanks to Chris Richards for pointing this out (and supplying the fix).
374 :    
375 :     2. Once again, I reworked some aspects of the FFI:
376 :    
377 :     A. The incomplete/complete type business:
378 :    
379 :     - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
380 :     gone!
381 :     - ML types representing an incomplete type are now *equal* to
382 :     ML types representing their corresponding complete types (just like
383 :     in C). This is still safe because ml-nlffigen will not generate
384 :     RTTI for incomplete types, nor will it generate functions that
385 :     require access to such RTTI. But when ML code generated from both
386 :     incomplete and complete versions of the C type meet, the ML types
387 :     are trivially interoperable.
388 :    
389 :     NOTE: These changes restore the full generality of the translation
390 :     (which was previously lost when I eliminated functorization)!
391 :    
392 :     B. Enum types:
393 :    
394 :     - Structure C now has a type constructor "enum" that is similar to
395 :     how the "su" constructor works. However, "enum" is not a phantom
396 :     type because each "T enum" has values (and is isomorphic to
397 :     MLRep.Signed.int).
398 :     - There are generic access operations for enum objects (using
399 :     MLRep.Signed.int).
400 :     - ml-nlffigen will generate a structure E_foo for each "enum foo".
401 :     * The structure contains the definition of type "mlrep" (the ML-side
402 :     representation type of the enum). Normally, mlrep is the same
403 :     as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
404 :     then mlrep will be defined as a datatype -- thus facilitating
405 :     pattern matching on mlrep values.
406 :     ("-ec" will be suppressed if there are duplicate values in an
407 :     enumeration.)
408 :     * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
409 :     will be generated for each C enum constant xxx.
410 :     * Conversion functions m2i and i2m convert between mlrep and
411 :     MLRep.Signed.int. (Without "-ec", these functions are identities.)
412 :     * Coversion functions c and ml convert between mlrep and "tag enum".
413 :     * Access functions (get/set) fetch and store mlrep values.
414 :     - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
415 :     enumerations are merged into one single enumeration represented by
416 :     structure E_'.
417 :    
418 :     ----------------------------------------------------------------------
419 : leunga 1080 Name: Allen Leung
420 : leunga 1094 Date: 2002/02/25 04:45:00 EST
421 :     Tag: leunga-20020225-cps-spill
422 :    
423 :     This is a new implementation of the CPS spill phase.
424 :     The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
425 :     In case of problems, replace it with the old file spill.sml
426 :    
427 :     The current compiler runs into some serious performance problems when
428 :     constructing a large record. This can happen when we try to compile a
429 :     structure with many items. Even a very simple structure like the following
430 :     makes the compiler slow down.
431 :    
432 :     structure Foo = struct
433 :     val x_1 = 0w1 : Word32.int
434 :     val x_2 = 0w2 : Word32.int
435 :     val x_3 = 0w3 : Word32.int
436 :     ...
437 :     val x_N = 0wN : Word32.int
438 :     end
439 :    
440 :     The following table shows the compile time, from N=1000 to N=4000,
441 :     with the old compiler:
442 :    
443 :     N
444 :     1000 CPS 100 spill 0.04u 0.00s 0.00g
445 :     MLRISC ra 0.06u 0.00s 0.05g
446 :     (spills = 0 reloads = 0)
447 :     TOTAL 0.63u 0.07s 0.21g
448 :    
449 :     1100 CPS 100 spill 8.25u 0.32s 0.64g
450 :     MLRISC ra 5.68u 0.59s 3.93g
451 :     (spills = 0 reloads = 0)
452 :     TOTAL 14.71u 0.99s 4.81g
453 :    
454 :     1500 CPS 100 spill 58.55u 2.34s 1.74g
455 :     MLRISC ra 5.54u 0.65s 3.91g
456 :     (spills = 543 reloads = 1082)
457 :     TOTAL 65.40u 3.13s 6.00g
458 :    
459 :     2000 CPS 100 spill 126.69u 4.84s 3.08g
460 :     MLRISC ra 0.80u 0.10s 0.55g
461 :     (spills = 42 reloads = 84)
462 :     TOTAL 129.42u 5.10s 4.13g
463 :    
464 :     3000 CPS 100 spill 675.59u 19.03s 11.64g
465 :     MLRISC ra 2.69u 0.27s 1.38g
466 :     (spills = 62 reloads = 124)
467 :     TOTAL 682.48u 19.61s 13.99g
468 :    
469 :     4000 CPS 100 spill 2362.82u 56.28s 43.60g
470 :     MLRISC ra 4.96u 0.27s 2.72g
471 :     (spills = 85 reloads = 170)
472 :     TOTAL 2375.26u 57.21s 48.00g
473 :    
474 :     As you can see the old cps spill module suffers from some serious
475 : leunga 1115 performance problem. But since I cannot decipher the old code fully,
476 :     instead of patching the problems up, I'm reimplementing it
477 : leunga 1094 with a different algorithm. The new code is more modular,
478 :     smaller when compiled, and substantially faster
479 :     (O(n log n) time and O(n) space). Timing of the new spill module:
480 :    
481 :     4000 CPS 100 spill 0.02u 0.00s 0.00g
482 :     MLRISC ra 0.25u 0.02s 0.15g
483 :     (spills=1 reloads=3)
484 :     TOTAL 7.74u 0.34s 1.62g
485 :    
486 :     Implementation details:
487 :    
488 :     As far as I can tell, the purpose of the CPS spill module is to make sure the
489 :     number of live variables at any program point (the bandwidth)
490 :     does not exceed a certain limit, which is determined by the
491 :     size of the spill area.
492 :    
493 :     When the bandwidth is too large, we decrease the register pressure by
494 :     packing live variables into spill records. How we achieve this is
495 :     completely different than what we did in the old code.
496 :    
497 :     First, there is something about the MLRiscGen code generator
498 :     that we should be aware of:
499 :    
500 :     o MLRiscGen performs code motion!
501 :    
502 :     In particular, it will move floating point computations and
503 :     address computations involving only the heap pointer to
504 :     their use sites (if there is only a single use).
505 :     What this means is that if we have a CPS record construction
506 :     statement
507 :    
508 :     RECORD(k,vl,w,e)
509 :    
510 :     we should never count the new record address w as live if w
511 :     has only one use (which is often the case).
512 :    
513 :     We should do something similar to floating point, but the transformation
514 :     there is much more complex, so I won't deal with that.
515 :    
516 :     Secondly, there are now two new cps primops at our disposal:
517 :    
518 :     1. rawrecord of record_kind option
519 :     This pure operator allocates some uninitialized storage from the heap.
520 :     There are two forms:
521 :    
522 :     rawrecord NONE [INT n] allocates a tagless record of length n
523 :     rawrecord (SOME rk) [INT n] allocates a tagged record of length n
524 :     and initializes the tag.
525 :    
526 :     2. rawupdate of cty
527 :     rawupdate cty (v,i,x)
528 :     Assigns to x to the ith component of record v.
529 :     The storelist is not updated.
530 :    
531 :     We use these new primops for both spilling and increment record construction.
532 :    
533 :     1. Spilling.
534 :    
535 :     This is implemented with a linear scan algorithm (but generalized
536 :     to trees). The algorithm will create a single spill record at the
537 :     beginning of the cps function and use rawupdate to spill to it,
538 :     and SELECT or SELp to reload from it. So both spills and reloads
539 :     are fine-grain operations. In contrast, in the old algorithm
540 :     "spills" have to be bundled together in records.
541 :    
542 :     Ideally, we should sink the spill record construction to where
543 :     it is needed. We can even split the spill record into multiple ones
544 :     at the places where they are needed. But CPS is not a good
545 :     representation for global code motion, so I'll keep it simple and
546 :     am not attempting this.
547 :    
548 :     2. Incremental record construction (aka record splitting).
549 :    
550 :     Long records with many component values which are simulatenously live
551 :     (recall that single use record addresses are not considered to
552 :     be live) are constructed with rawrecord and rawupdate.
553 :     We allocate space on the heap with rawrecord first, then gradually
554 :     fill it in with rawupdate. This is the technique suggested to me
555 :     by Matthias.
556 :    
557 :     Some restrictions on when this is applicable:
558 :     1. It is not a VECTOR record. The code generator currently does not handle
559 :     this case. VECTOR record uses double indirection like arrays.
560 :     2. All the record component values are defined in the same "basic block"
561 :     as the record constructor. This is to prevent speculative
562 :     record construction.
563 :    
564 :     ----------------------------------------------------------------------
565 :     Name: Allen Leung
566 : leunga 1086 Date: 2002/02/22 01:02:00 EST
567 :     Tag: leunga-20020222-mlrisc-tools
568 :    
569 :     Minor bug fixes in the parser and rewriter
570 :    
571 :     ----------------------------------------------------------------------
572 :     Name: Allen Leung
573 : leunga 1085 Date: 2002/02/21 20:20:00 EST
574 :     Tag: leunga-20020221-peephole
575 :    
576 :     Regenerated the peephole files. Some contained typos in the specification
577 :     and some didn't compile because of pretty printing bugs in the old version
578 :     of 'nowhere'.
579 :    
580 :     ----------------------------------------------------------------------
581 :     Name: Allen Leung
582 : leunga 1079 Date: 2002/02/19 20:20:00 EST
583 :     Tag: leunga-20020219-mlrisc-tools
584 :     Description:
585 :    
586 :     Minor bug fixes to the mlrisc-tools library:
587 :    
588 :     1. Fixed up parsing colon suffixed keywords
589 :     2. Added the ability to shut the error messages up
590 :     3. Reimplemented the pretty printer and fixed up/improved
591 :     the pretty printing of handle and -> types.
592 :     4. Fixed up generation of literal symbols in the nowhere tool.
593 :     5. Added some SML keywords to to sml.sty
594 :    
595 :     ----------------------------------------------------------------------
596 :     Name: Matthias Blume
597 : blume 1078 Date: 2002/02/19 16:20:00 EST
598 :     Tag: blume-20020219-cmffi
599 :     Description:
600 :    
601 :     A wild mix of changes, some minor, some major:
602 :    
603 :     * All C FFI-related libraries are now anchored under $c:
604 :     $/c.cm --> $c/c.cm
605 :     $/c-int.cm --> $c/internals/c-int.cm
606 :     $/memory.cm --> $c/memory/memory.cm
607 :    
608 :     * "make" tool (in CM) now treats its argument pathname slightly
609 :     differently:
610 :     1. If the native expansion is an absolute name, then before invoking
611 :     the "make" command on it, CM will apply OS.Path.mkRelative
612 :     (with relativeTo = OS.FileSys.getDir()) to it.
613 :     2. The argument will be passed through to subsequent phases of CM
614 :     processing without "going native". In particular, if the argument
615 :     was an anchored path, then "make" will not lose track of that anchor.
616 :    
617 :     * Compiler backends now "know" their respective C calling conventions
618 :     instead of having to be told about it by ml-nlffigen. This relieves
619 :     ml-nlffigen from one of its burdens.
620 :    
621 :     * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
622 :    
623 :     * Export C_DEBUG and C_Debug from $c/c.cm.
624 :    
625 :     * C type encoding in ml-nlffi-lib has been improved to model the conceptual
626 :     subtyping relationship between incomplete pointers and their complete
627 :     counterparts. For this, ('t, 'c) ptr has been changed to 'o ptr --
628 :     with the convention of instantiating 'o with ('t, 'c) obj whenever
629 :     the pointer target type is complete. In the incomplete case, 'o
630 :     will be instantiated with some "'c iobj" -- a type obtained by
631 :     using one of the functors PointerToIncompleteType or PointerToCompleteType.
632 :    
633 :     Operations that work on both incomplete and complete pointer types are
634 :     typed as taking an 'o ptr while operations that require the target to
635 :     be known are typed as taking some ('t, 'c) obj ptr.
636 :    
637 :     voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
638 :     where void is an eqtype without any values. This makes it possible
639 :     to work on voidptr values using functions meant to operate on light
640 :     incomplete pointers.
641 :    
642 :     * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
643 :     been vastly simplified.
644 :    
645 :     ----------------------------------------------------------------------
646 :     Name: Matthias Blume
647 : blume 1076 Date: 2002/02/19 10:48:00 EST
648 :     Tag: blume-20020219-pqfix
649 :     Description:
650 :    
651 :     Applied Chris Okasaki's bug fix for priority queues.
652 :    
653 :     ----------------------------------------------------------------------
654 :     Name: Matthias Blume
655 : blume 1073 Date: 2002/02/15 17:05:00
656 : blume 1069 Tag: Release_110_39
657 :     Description:
658 :    
659 : blume 1073 Last-minute retagging is becoming a tradition... :-(
660 :    
661 :     This is the working release 110.39.
662 :    
663 :     ----------------------------------------------------------------------
664 :     Name: Matthias Blume
665 :     Date: 2002/02/15 16:00:00 EST
666 :     Tag: Release_110_39-orig
667 :     Description:
668 :    
669 : blume 1069 Working release 110.39. New bootfiles.
670 :    
671 : blume 1073 (Update: There was a small bug in the installer so it wouldn't work
672 :     with all shells. So I retagged. -Matthias)
673 :    
674 : blume 1069 ----------------------------------------------------------------------
675 :     Name: Matthias Blume
676 : blume 1068 Date: 2002/02/15 14:17:00 EST
677 :     Tag: blume-20020215-showbindings
678 :     Description:
679 :    
680 :     Added EnvRef.listBoundSymbols and CM.State.showBindings. Especially
681 :     the latter can be useful for exploring what bindings are available at
682 :     the interactive prompt. (The first function returns only the list
683 :     of symbols that are really bound, the second prints those but also the
684 :     ones that CM's autoloading mechanism knows about.)
685 :    
686 :     ----------------------------------------------------------------------
687 :     Name: Matthias Blume
688 : blume 1067 Date: 2002/02/15 12:08:00 EST
689 :     Tag: blume-20020215-iptrs
690 :     Description:
691 :    
692 :     Two improvements to ml-nlffigen:
693 :    
694 :     1. Write files only if they do not exist or if their current contents
695 :     do not coincide with what's being written. (That is, avoid messing
696 :     with the time stamps unless absolutely necessary.)
697 :    
698 :     2. Implement a "repository" mechanism for generated files related
699 :     to "incomplete pointer types". See the README file for details.
700 :    
701 :     ----------------------------------------------------------------------
702 :     Name: Matthias Blume
703 : blume 1066 Date: 2002/02/14 11:50:00 EST
704 :     Tag: blume-20020214-quote
705 :     Description:
706 :    
707 :     Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm). This is required
708 :     because of the new and improved tag generation scheme. (Thanks to Allen
709 :     Leung for pointing it out.)
710 :    
711 :     ----------------------------------------------------------------------
712 : george 1065 Name: Lal George
713 :     Date: 2002/02/14 09:55:27 EST 2002
714 :     Tag: george-20020214-isabelle-bug
715 :     Description:
716 :    
717 :     Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
718 :     of Isabelle on the x86.
719 :    
720 :     From Allen:
721 :     -----------
722 :     I've found the problem:
723 :    
724 :     in ra-core.sml, I use the counter "blocked" to keep track of the
725 :     true number of elements in the freeze queue. When the counter goes
726 :     to zero, I skip examining the queue. But I've messed up the
727 :     bookkeeping in combine():
728 :    
729 :     else ();
730 :     case !ucol of
731 :     PSEUDO => (if !cntv > 0 then
732 :     (if !cntu > 0 then blocked := !blocked - 1 else ();
733 :     ^^^^^^^^^^^^^^^^^^^^^^^
734 :     moveu := mergeMoveList(!movev, !moveu)
735 :     )
736 :     else ();
737 :    
738 :     combine() is called to coalesce two nodes u and v.
739 :     I think I was thinking that if the move counts of u and v are both
740 :     greater than zero then after they are coalesced then one node is
741 :     removed from the freeze queue. Apparently I was thinking that
742 :     both u and v are of low degree, but that's clearly not necessarily true.
743 :    
744 :    
745 :     02/12/2002:
746 :     Here's the patch. HOL now compiles.
747 :    
748 :     I don't know how this impact on performance (compile
749 :     time or runtime). This bug caused the RA (especially on the x86)
750 :     to go thru the potential spill phase when there are still nodes on the
751 :     freeze queue.
752 :    
753 :    
754 :    
755 :    
756 :     ----------------------------------------------------------------------
757 : blume 1058 Name: Matthias Blume
758 : blume 1064 Date: 2002/02/13 22:40:00 EST
759 :     Tag: blume-20020213-fptr-rtti
760 :     Description:
761 :    
762 :     Fixed a bug in ml-nlffigen that was introduced with one of the previous
763 :     updates.
764 :    
765 :     ----------------------------------------------------------------------
766 :     Name: Matthias Blume
767 : blume 1063 Date: 2002/02/13 16:41:00 EST
768 :     Tag: blume-20020213-cmlpq
769 :     Description:
770 :    
771 :     Added new priority queue export symbols (which have just been added to
772 :     smlnj-lib.cm) to CML's version of smlnj-lib.cm. (Otherwise CML would
773 :     not compile and the installer would choke.)
774 :    
775 :     ----------------------------------------------------------------------
776 :     Name: Matthias Blume
777 : blume 1062 Date: 2002/02/13 16:15:00 EST
778 :     Tag: blume-20020213-various
779 :     Description:
780 :    
781 :     1. More tweaks to ml-nlffigen:
782 :    
783 :     - better internal datastructures (resulting in slight speedup)
784 :     - "-match" option requires exact match
785 :     - "localized" gensym counters (untagged structs/unions nested within
786 :     other structs/unions or within typedefs get a fresh counter; their
787 :     tag will be prefixed by a concatenation of their parents' tags)
788 :     - bug fixes (related to calculation of transitive closure of types
789 :     to be included in the output)
790 :    
791 :     2. Minor Basis updates:
792 :    
793 :     - added implementations for List.collate and Option.app
794 :    
795 :     ----------------------------------------------------------------------
796 :     Name: Matthias Blume
797 : blume 1060 Date: 2002/02/11 15:55:00 EST
798 :     Tag: blume-20020211-gensym
799 :     Description:
800 :    
801 :     Added a "-gensym" option to command line of ml-nlffigen. This can be
802 :     used to specify a "stem" -- a string that is inserted in all "gensym'd"
803 :     names (ML structure names that correspond to unnamed C structs, unions,
804 :     and enums), so that separate runs of ml-nlffigen do not clash.
805 :    
806 :     ----------------------------------------------------------------------
807 :     Name: Matthias Blume
808 : blume 1059 Date: 2002/02/11 12:05:00 EST
809 :     Tag: blume-20020211-gensml
810 :     Description:
811 :    
812 :     A quick fix for a problem with GenSML (in the pgraph-util library):
813 :     Make generation of toplevel "local" optional. (Strictly speaking,
814 :     signature definitions within "local" are not legal SML.)
815 :    
816 :     Other than that: updates to INSTALL and cm/TODO.
817 :    
818 :     ----------------------------------------------------------------------
819 :     Name: Matthias Blume
820 : blume 1058 Date: 2002/02/08 15:00:00 EST
821 :     Tag: blume-20020208-uniquepid
822 :     Description:
823 :    
824 :     0. Version number has been bumped to 110.38.1. NEW BOOTFILES!!!
825 :    
826 :     1. The installer (config/install.sh) has gotten smarter:
827 :    
828 :     - Configuration options are a bit easier to specify now
829 :     (in config/targets).
830 :     - Bug in recognizing .tar.bz2 files fixed.
831 :     - Installer automatically resolves dependencies between
832 :     configuration options (e.g., if you ask for eXene, you will
833 :     also get cml -- regardless whether you asked for it or not).
834 :     - Installer can run in "quieter mode" by setting the environment
835 :     variable INSTALL_QUIETLY to "true". "Quieter" does not mean
836 :     "completely silent", though.
837 :     - Build HashCons library as part of smlnj-lib.
838 :    
839 :     2. A new scheme for assigning persistent identifiers to compilation
840 :     units (and, by extension, to types etc.) has been put into place.
841 :     This fixes a long-standing bug where types and even dynamic values
842 :     can get internally confused, thereby compromising type safety
843 :     (abstraction) and dynamic correctness. See
844 :    
845 :     http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
846 :    
847 :     for an example of how things could go wrong until now.
848 :    
849 :     The downside of the new scheme is that pids are not quite as
850 :     persistent as they used to be: CM will generate a fresh pid
851 :     for every compilation unit that it thinks it sees for the first
852 :     time. That means that if you compile starting from a clean, fresh
853 :     source tree at two different times, you end up with different
854 :     binaries.
855 :    
856 :     Cutoff recompilation, however, has not been compromised because
857 :     CM keeps pid information in special caches between runs.
858 :    
859 :     ----------------------------------------------------------------------
860 : george 1053 Name: Lal George
861 : george 1055 Date: 2002/02/07 15:34:13 EST 2002
862 :     Tag: <none>
863 :     Description:
864 :    
865 :     Compilers that generate assembly code may produce global labels
866 :     whose value is resolved at link time. The various peephole optimization
867 :     modules did not take this in account.
868 :    
869 :     TODO. The Labels.addrOf function should really return an option
870 :     type so that clients are forced to deal with this issue, rather
871 :     than an exception being raised.
872 :    
873 :     ----------------------------------------------------------------------
874 :     Name: Lal George
875 : george 1053 Date: 2002/02/06 13:55:02 EST
876 :     Tag: george-20020206-ra-breakup
877 :     Description:
878 :    
879 :     1. A bug fix from Allen.
880 :    
881 :     A typo causes extra fstp %st(0)'s to be generated at compensation
882 :     edges, which might cause stack underflow traps at runtime. This
883 :     occurs in fft where there are extraneous fstps right before the 'into'
884 :     trap instruction (in this case they are harmless since none of the
885 :     integers overflow.)
886 :    
887 : george 1054 2. Pulled out various utility modules that were embedded in the modules
888 : george 1053 of the register allocator. I need these modules for other purposes, but
889 : george 1054 they are not complete enough to put into a library (just yet).
890 : george 1053 ----------------------------------------------------------------------
891 : blume 1049 Name: Matthias Blume
892 :     Date: 2002/01/31 16:05:00 EST
893 :     Tag: blume-20020131-sparc-ccalls
894 :     Description:
895 :    
896 :     1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
897 :     of extra stack space by mistake. Fixed.
898 :    
899 :     2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
900 :    
901 :     ----------------------------------------------------------------------
902 : leunga 1048 Name: Allen Leung
903 :     Date: 2002/01/30
904 :     Tag: leunga-20020130-nowhere-bug-fix
905 :     Description:
906 :    
907 :     MLRISC bug fixes:
908 :     1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
909 :     2. MachineInt.fromString was negating its value.
910 :    
911 :     ----------------------------------------------------------------------
912 : blume 1039 Name: Matthias Blume
913 : blume 1047 Date: 2002/01/29
914 :     Tag: blume-20020129-INSTALL
915 :     Description:
916 :    
917 :     - Added somewhat detailed installation instructions (file INSTALL).
918 :     - Fixed curl-detection bug in config/install.sh.
919 :     - It is now possible to select the URL getter using the URLGETTER
920 :     environment variable:
921 :    
922 :     not set / "unknown" --> automatic detection (script tries wget,
923 :     curl, and lynx)
924 :     "wget" / "curl" / "lynx" --> use the specified program (script "knows"
925 :     how to properly invoke them)
926 :     other --> use $URLGETTER directly, it must take
927 :     precisely two command-line arguments
928 :     (source URL and destination file name)
929 :    
930 :     ----------------------------------------------------------------------
931 :     Name: Matthias Blume
932 : blume 1044 Date: 2002/01/28
933 : blume 1045 Tag: blume-20020128-sparc-ccalls
934 :     Description:
935 :    
936 :     - Fixed problem with calculation of "used" registers in sparc-c-calls.
937 :     - Make use of the allocParam argument in sparc-c-calls.
938 :    
939 :     ----------------------------------------------------------------------
940 :     Name: Matthias Blume
941 :     Date: 2002/01/28
942 : blume 1044 Tag: blume-20020128-allocParam
943 :     Description:
944 :    
945 :     John Reppy: Changes c-calls API to accept client-callback for
946 :     allocating extra stack space.
947 :     me: Corresponding changes to mlriscGen (using a dummy argument that
948 :     does not change the current behavior).
949 :    
950 :     ----------------------------------------------------------------------
951 :     Name: Matthias Blume
952 : blume 1040 Date: 2002/01/28 12:00:00
953 :     Tag: Release_110_38
954 :     Description:
955 :    
956 :     This time for real!!!
957 :    
958 :     ----------------------------------------------------------------------
959 :     Name: Matthias Blume
960 : blume 1039 Date: 2002/01/28 10:56:00 EST
961 :     Tag: blume-20020128-retraction
962 :     Description:
963 :    
964 :     0. Retracted earlier 110.38. (The Release_110_38 tag has been replaced
965 :     with blume-Release_110_38-retracted.)
966 :    
967 :     1. Fixed a problem with incorrect rounding modes in real64.sml.
968 :     (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
969 :    
970 :     2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
971 :     and enums fixed. The naming of corresponding ML identifiers should
972 :     now be consistent again.
973 :    
974 :     ----------------------------------------------------------------------
975 : leunga 1038 Name: Allen Leung
976 :     Date: 2002/01/27
977 :     Tag: leunga-20020127-nowhere
978 :     Description:
979 :    
980 :     Added a target called nowhere in the configuration scripts.
981 :     Enabling this will build the MLRISC 'nowhere' tool (for translating
982 :     programs with where-clauses into legal SML code) during installation.
983 :    
984 :     ----------------------------------------------------------------------
985 : blume 1036 Name: Matthias Blume
986 : leunga 1038 Date: 2002/01/25 21:27:00 EST
987 : blume 1039 Tag: blume-Release_110_38-retracted
988 : blume 1037 Description:
989 :    
990 :     Call it a (working) release! Version is 110.38. Bootfiles are ready.
991 :    
992 :     README will be added later.
993 :    
994 : blume 1039 !!! NOTE: Re-tagged as blume-Release_110_38-retracted. Original tag
995 :     (Release_110_38) removed. Reason: Last-minute bug fixes.
996 :    
997 : blume 1037 ----------------------------------------------------------------------
998 :     Name: Matthias Blume
999 : blume 1036 Date: 2002/01/25
1000 :     Tag: blume-20020125-ffi
1001 :     Description:
1002 :    
1003 :     A large number of tweaks and improvements to ml-nlffi-lib and
1004 :     ml-nlffigen:
1005 :    
1006 :     - ML represenation types have been streamlined
1007 :     - getter and setter functions work with concrete values, not abstract
1008 :     ones where possible
1009 :     - ml-nlffigen command line more flexible (see README file there)
1010 :     - some bugs have been fixed (hopefully)
1011 :    
1012 :     ----------------------------------------------------------------------
1013 : george 1034 Name: Lal George
1014 :     Date: 2002/01/24
1015 :     Tag: george-20020124-risc-ra-interface
1016 :     Description:
1017 :    
1018 :     There is a dramatic simplification in the interface to the
1019 :     register allocator for RISC architectures as a result of making
1020 :     parallel copy instructions explicit.
1021 :    
1022 :     ----------------------------------------------------------------------
1023 : blume 1028 Name: Matthias Blume
1024 : blume 1032 Date: 2002/01/22
1025 :     Tag: blume-20020122-x86-ccalls
1026 :     Description:
1027 :    
1028 :     Bug fix for c-calls on x86 (having to do with how char- and
1029 :     short-arguments are being handled).
1030 :    
1031 :     ----------------------------------------------------------------------
1032 :     Name: Matthias Blume
1033 : blume 1031 Date: 2002/01/21
1034 :     Tag: blume-20020121-ff
1035 :     Description:
1036 :    
1037 :     Another day of fiddling with the FFI...
1038 :    
1039 :     1. Bug fix/workaround: CKIT does not complain about negative array
1040 :     dimensions, so ml-nlffigen has to guard itself against this possibility.
1041 :     (Otherwise a negative dimension would send it into an infinite loop.)
1042 :    
1043 :     2. Some of the abstract types (light objects, light pointers, most "base"
1044 :     types) in structure C are now eqtypes.
1045 :    
1046 :     3. Added constructors and test functions for NULL function pointers.
1047 :    
1048 :     ----------------------------------------------------------------------
1049 :     Name: Matthias Blume
1050 : blume 1028 Date: 2002/01/18
1051 : blume 1030 Tag: blume-20020118-ready-for-new-release
1052 :     Description:
1053 :    
1054 :     Made config/srcarchiveurl point to a new place. (Will provide boot
1055 :     files shortly.)
1056 :    
1057 :     Maybe we christen this to be 110.38?
1058 :    
1059 :     ----------------------------------------------------------------------
1060 :     Name: Matthias Blume
1061 :     Date: 2002/01/18
1062 : blume 1029 Tag: blume-20020118-more-ffifiddle
1063 :     Description:
1064 :    
1065 :     Today's FFI fiddling:
1066 :    
1067 :     - Provided a structure CGetSet with "convenient" versions of C.Get.* and
1068 :     C.Set.* that use concrete (MLRep.*) arguments and results instead
1069 :     of abstract ones.
1070 :    
1071 :     - Provided word-style bit operations etc. for "int" representation
1072 :     types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
1073 :     and Long.
1074 :    
1075 :     ----------------------------------------------------------------------
1076 :     Name: Matthias Blume
1077 :     Date: 2002/01/18
1078 : blume 1028 Tag: blume-20020118-use-x86-fp
1079 :     Description:
1080 :    
1081 :     Now that x86-fast-fp seems to be working, I turned it back on again
1082 :     by default. (Seems to work fine now, even with the FFI.)
1083 :    
1084 :     Other than that, I added some documentation about the FFI to
1085 :     src/ml-nlffigen/README and updated the FFI test examples in
1086 :     src/ml-nlffi-lib/Tests/*.
1087 :    
1088 :     ----------------------------------------------------------------------
1089 : leunga 1025 Name: Allen Leung
1090 :     Date: 2002/01/17
1091 :     Tag: leunga-20020117-x86-fast-fp-call
1092 :     Description:
1093 :    
1094 :     1. Fixed a problem with handling return fp values when x86's fast fp
1095 :     mode is turned on.
1096 :    
1097 :     2. Minor pretty printing fix for cellset. Print %st(0) as %st(0) instead
1098 :     of %f32.
1099 :    
1100 :     3. Added a constructor INT32lit to the ast of MLRISC tools.
1101 :    
1102 :     ----------------------------------------------------------------------
1103 : blume 1021 Name: Matthias Blume
1104 :     Date: 2002/01/16
1105 :     Tag: blume-20020116-ffifiddle
1106 :     Description:
1107 :    
1108 :     More fiddling with the FFI interface:
1109 :    
1110 :     - Make constness 'c instead of rw wherever possible. This eliminates
1111 :     the need for certain explicit coercions. (However, due to ML's
1112 :     value polymorphism, there will still be many cases where explicit
1113 :     coercions are necessary. Phantom types are not the whole answer
1114 :     to modeling a subtyping relationship in ML.)
1115 :    
1116 :     - ro/rw coersions for pointers added. (Avoids the detour through */&.)
1117 :    
1118 :     - "printf" test example added to src/ml-nlffi-lib/Tests. (Demonstrates
1119 :     clumsy workaround for varargs problem.)
1120 :    
1121 :     ----------------------------------------------------------------------
1122 : george 1016 Name: Lal George
1123 :     Date: 2002/01/15
1124 :     Tag: <none>
1125 :     Description:
1126 :    
1127 :     1. Since COPY instructions are no longer native to the architecture,
1128 :     a generic functor can be used to implement the expandCopies function.
1129 :    
1130 :     2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
1131 :     TEXT segment.
1132 :    
1133 :     ----------------------------------------------------------------------
1134 : blume 1011 Name: Matthias Blume
1135 : blume 1015 Date: 2002/01/15
1136 :     Tag: blume-20020115-ffiupdates
1137 :     Description:
1138 :    
1139 :     1. Fix for bug resulting in single-precision float values being returned
1140 :     incorrectly from FFI calls.
1141 :    
1142 :     2. Small modifications to C FFI API:
1143 :    
1144 :     - memory-allocation routines return straight objects (no options)
1145 :     and raise an exception in out-of-memory situations
1146 :     - unsafe extensions to cast between function pointers and pointers
1147 :     from/to ints
1148 :     - added structure C_Debug as an alternative to structure C where
1149 :     pointer-dereferencing (|*| and |*!) always check for null-pointers
1150 :     - added open_lib' to DynLinkage; open_lib' works like open_lib
1151 :     but also takes a (possibly empty) list of existing library handles
1152 :     that the current library depends on
1153 :    
1154 :     ----------------------------------------------------------------------
1155 :     Name: Matthias Blume
1156 : blume 1011 Date: 2002/01/10
1157 :     Tag: blume-20020110-newffigen
1158 :     Description:
1159 :    
1160 :     1. Updates to portable graph code.
1161 :    
1162 :     2. Major update to ml-nlffigen and ml-nlffi-lib. Things are much
1163 :     more scalable now so that even huge interfaces such as the one
1164 :     for GTK compile in finite time and space. :-)
1165 :     See src/ml-nlffigen/README for details on what's new.
1166 :    
1167 :     ----------------------------------------------------------------------
1168 : george 1003 Name: Lal George
1169 : george 1009 Date: 2001/01/09 14:31:35 EST 2002
1170 :     Tag: george-20011206-rm-native-copy
1171 :     Description:
1172 :    
1173 :     Removed the native COPY and FCOPY instructions
1174 :     from all the architectures and replaced it with the
1175 :     explicit COPY instruction from the previous commit.
1176 :    
1177 :     It is now possible to simplify many of the optimizations
1178 :     modules that manipulate copies. This has not been
1179 :     done in this change.
1180 :    
1181 :     ----------------------------------------------------------------------
1182 :     Name: Lal George
1183 : george 1003 Date: 2001/12/06 16:50:13 EST 2001
1184 :     Tag: george-20011206-mlrisc-instruction
1185 :     Description:
1186 : george 824
1187 : george 1003 Changed the representation of instructions from being fully abstract
1188 :     to being partially concrete. That is to say:
1189 :    
1190 :     from
1191 :     type instruction
1192 :    
1193 :     to
1194 :     type instr (* machine instruction *)
1195 :    
1196 :     datatype instruction =
1197 :     LIVE of {regs: C.cellset, spilled: C.cellset}
1198 :     | KILL of {regs: C.cellset, spilled: C.cellset}
1199 :     | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
1200 :     | ANNOTATION of {i: instruction, a: Annotations.annotation}
1201 :     | INSTR of instr
1202 :    
1203 :     This makes the handling of certain special instructions that appear on
1204 :     all architectures easier and uniform.
1205 :    
1206 :     LIVE and KILL say that a list of registers are live or killed at the
1207 :     program point where they appear. No spill code is generated when an
1208 :     element of the 'regs' field is spilled, but the register is moved to
1209 :     the 'spilled' (which is present, more for debugging than anything else).
1210 :    
1211 :     LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
1212 :     We used to generate:
1213 :    
1214 :     DEFFREG f1
1215 :     f1 := f2 + f3
1216 :     trapb
1217 :    
1218 :     but now generate:
1219 :    
1220 :     f1 := f2 + f3
1221 :     trapb
1222 :     LIVE {regs=[f1,f2,f3], spilled=[]}
1223 :    
1224 :     Furthermore, the DEFFREG (hack) required that all floating point instruction
1225 :     use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
1226 :     defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
1227 :     in a cleaner alpha implementation. (Hopefully, intel will not get rid of
1228 :     this architecture).
1229 :    
1230 :     COPYXXX is intended to replace the parallel COPY and FCOPY available on
1231 :     all the architectures. This will result in further simplification of the
1232 :     register allocator that must be aware of them for coalescing purposes, and
1233 :     will also simplify certain aspects of the machine description that provides
1234 :     callbacks related to parallel copies.
1235 :    
1236 :     ANNOTATION should be obvious, and now INSTR represents the honest to God
1237 :     machine instruction set!
1238 :    
1239 :     The <arch>/instructions/<arch>Instr.sml files define certain utility
1240 :     functions for making porting easier -- essentially converting upper case
1241 :     to lower case. All machine instructions (of type instr) are in upper case,
1242 :     and the lower case form generates an MLRISC instruction. For example on
1243 :     the alpha we have:
1244 :    
1245 :     datatype instr =
1246 :     LDA of {r:cell, b:cell, d:operand}
1247 :     | ...
1248 :    
1249 :     val lda : {r:cell, b:cell, d:operand} -> instruction
1250 :     ...
1251 :    
1252 :     where lda is just (INSTR o LDA), etc.
1253 :    
1254 : leunga 815 ----------------------------------------------------------------------
1255 : blume 986 Name: Matthias Blume
1256 : blume 993 Date: 2001/11/22 21:40:00 EST
1257 :     Tag: Release_110_37
1258 :     Description:
1259 :    
1260 :     Release 110.37. This time for real.
1261 :    
1262 :     ----------------------------------------------------------------------
1263 :     Name: Matthias Blume
1264 : blume 988 Date: 2001/11/21 16:35:00 EST
1265 :     Tag: blume-20011121-foot-in-mouth
1266 :     Description:
1267 :    
1268 :     Removed the "Release_110_37" tag because of a serious bug.
1269 :     This will be re-tagged once the bug is fixed.
1270 :    
1271 :     ----------------------------------------------------------------------
1272 :     Name: Matthias Blume
1273 : blume 987 Date: 2001/11/21 16:14:00 EST
1274 :     Tag: blume-20011121-forgottenfile
1275 :     Description:
1276 :    
1277 :     Forgot to add a file. (Just a .tex-file -- part of
1278 :     the CM manual source.)
1279 :    
1280 :     ----------------------------------------------------------------------
1281 :     Name: Matthias Blume
1282 : blume 986 Date: 2001/11/21 16:10:00 EST
1283 : blume 988 Tag: blume-20011121-invalid_110_37
1284 : blume 986 Description:
1285 :    
1286 : blume 988 Note: I removed the original tag "Release_110_37" from this commit
1287 :     because we found a serious bug in all non-x86 backends.
1288 :     - Matthias
1289 :    
1290 : blume 986 1. Modifications to the SML/NJ code generator and to the runtime system
1291 :     so that code object name strings are directly inserted into code
1292 :     objects at code generation time. The only business the runtime system
1293 :     has with this is now to read the name strings on occasions.
1294 :     (The encoding of the name string has also changed somewhat.)
1295 :    
1296 :     2. CM now implements a simple "set calculus" for specifying export lists.
1297 :     In particular, it is now possible to refer to the export lists of
1298 :     other libraries/groups/sources and form unions as well as differences.
1299 :     See the latest CM manual for details.
1300 :    
1301 :     3. An separate notion of "proxy" libraries has again be eliminated from
1302 :     CM's model. (Proxy libraries are now simply a special case of using
1303 :     the export list calculus.)
1304 :    
1305 :     4. Some of the existing libraries now take advantage of the new set
1306 :     calculus.
1307 :     (Notice that not all libraries have been converted because some
1308 :     of the existing .cm-files are supposed to be backward compatible
1309 :     with 110.0.x.)
1310 :    
1311 :     5. Some cleanup in stand-alone programs. (Don't use "exnMessage" -- use
1312 :     "General.exnMessage"! The former relies on a certain hook to be
1313 :     initialized, and that often does not happen in the stand-alone case.)
1314 :    
1315 :     ----------------------------------------------------------------------
1316 : george 984 Name: Lal George
1317 :     Date: 2001/11/21 13:56:18 EST
1318 :     Tag: george-2001121-pseudo-ops
1319 :     Description:
1320 :    
1321 :     Implemented a complete redesign of MLRISC pseudo-ops. Now there
1322 :     ought to never be any question of incompatabilities with
1323 :     pseudo-op syntax expected by host assemblers.
1324 :    
1325 :     For now, only modules supporting GAS syntax are implemented
1326 :     but more should follow, such as MASM, and vendor assembler
1327 :     syntax, e.g. IBM as, Sun as, etc.
1328 :    
1329 :     ----------------------------------------------------------------------
1330 : blume 977 Name: Matthias Blume
1331 :     Date: 2001/11/14 11:52:00 EST
1332 :     Tag: blume-20011114-srcname
1333 :     Description:
1334 :    
1335 :     1. Routed the name of the current source file to mlriscgen where it
1336 :     should be directly emitted into the code object. (This last part
1337 :     is yet to be done.)
1338 :    
1339 :     2. Some cleanup of the pgraph code to make it match the proposal that
1340 :     I put out the other day. (The proposal notwithstanding, things are
1341 :     still in flux here.)
1342 :    
1343 :     ----------------------------------------------------------------------
1344 : george 976 Name: Lal George
1345 :     Date: 2001/11/14 09:44:04 EST
1346 :     Tag:
1347 :     Description:
1348 :    
1349 :     Fix for a backpatching bug reported by Allen.
1350 :    
1351 :     Because the boundary between short and long span-dependent
1352 :     instructions is +/- 128, there are an astounding number of
1353 :     span-dependent instructions whose size is over estimated.
1354 :    
1355 :     Allen came up with the idea of letting the size of span
1356 :     dependent instructions be non-monotonic, for a maxIter
1357 :     number of times, after which the size must be monotonically
1358 :     increasing.
1359 :    
1360 :     This table shows the number of span-dependent instructions
1361 :     whose size was over-estimated as a function of maxIter, for the
1362 :     file Parse/parse/ml.grm.sml:
1363 :    
1364 :     maxIter # of instructions:
1365 :     10 687
1366 :     20 438
1367 :     30 198
1368 :     40 0
1369 :    
1370 :     In compiling the compiler, there is no significant difference in
1371 :     compilation speed between maxIter=10 and maxIter=40. Actually,
1372 :     my measurements showed that maxIter=40 was a tad faster than
1373 :     maxIter=10! Also 96% of the files in the compiler reach a fix
1374 :     point within 13 iterations, so fixing maxIter at 40, while high,
1375 :     is okay.
1376 :    
1377 :     ----------------------------------------------------------------------
1378 : blume 975 Name: Matthias Blume
1379 :     Date: 2001/10/31 15:25:00 EST
1380 :     Tag: blume-20011031-pgraph
1381 :     Description:
1382 :    
1383 :     CKIT:
1384 :     * Changed the "Function" constructor of type Ast.ctype to carry optional
1385 :     argument identifiers.
1386 :     * Changed the return type of TypeUtil.getFunction accordingly.
1387 :     * Type equality ignores the argument names.
1388 :     * TypeUtil.composite tries to preserve argument names but gives up quickly
1389 :     if there is a mismatch.
1390 :    
1391 :     installation script:
1392 :     * attempts to use "curl" if available (unless "wget" is available as well)
1393 :    
1394 :     CM:
1395 :     * has an experimental implementation of "portable graphs" which I will
1396 :     soon propose as an implementation-independent library format
1397 :     * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
1398 :    
1399 :     NLFFI-LIB:
1400 :     * some cleanup (all cosmetic)
1401 :    
1402 :     NLFFIGEN:
1403 :     * temporarily disabled the mechanism that suppresses ML output for
1404 :     C definitions whose identifiers start with an underscore character
1405 :     * generate val bindings for enum constants
1406 :     * user can request that only one style (light or heavy) is being used;
1407 :     default is to use both (command-line arguments: -heavy and -light)
1408 :     * fixed bug in handling of function types involving incomplete pointers
1409 :     * generate ML entry points that take record arguments (i.e., using
1410 :     named arguments) for C functions that have a prototype with named
1411 :     arguments
1412 :     (see changes to CKIT)
1413 :    
1414 :     ----------------------------------------------------------------------
1415 : leunga 972 Name: Allen Leung
1416 :     Date: 2001/10/27 20:34:00 EDT
1417 :     Tag: leunga-20011027-x86-fast-fp-call
1418 :     Description:
1419 :    
1420 :     Fixed the bug described in blume-20010920-slowfp.
1421 :    
1422 :     The fix involves
1423 :     1. generating FCOPYs in FSTP in ia32-svid
1424 :     2. marking a CALL with the appropriate annotation
1425 :    
1426 :     ----------------------------------------------------------------------
1427 : blume 961 Name: Matthias Blume
1428 :     Date: 2001/10/16 11:32:00 EDT
1429 :     Tag: blume-20011016-netbsd
1430 :     Description:
1431 :    
1432 :     Underscore patch from Chris Richards (fixing problem with compiling
1433 :     runtime system under recent NetBSD).
1434 :    
1435 :     ----------------------------------------------------------------------
1436 : george 959 Name: Allen Leung
1437 :     Date: 2001/10/12 17:18:32 EDT 2001
1438 :     Tag: leung-20011012-x86-printflowgraph
1439 :     Description:
1440 :    
1441 :     X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
1442 :    
1443 :     ----------------------------------------------------------------------
1444 : george 952 Name: Lal George
1445 : george 958 Date: 2001/10/11 23:51:34 EDT
1446 :     Tag: george-20011011-too-many-instrs
1447 :     Description:
1448 :    
1449 :     The representation of a program point never expected to see more
1450 :     than 65536 instructions in a basic block!
1451 :    
1452 :     ----------------------------------------------------------------------
1453 :     Name: Lal George
1454 : george 952 Date: 2001/10/09 09:41:37 EDT
1455 :     Tag: george-20011008-mlrisc-labels
1456 :     Description:
1457 :    
1458 :     Changed the machine description files to support printing of
1459 :     local and global labels in assembly code, based on host assembler
1460 :     conventions.
1461 :    
1462 :     ----------------------------------------------------------------------
1463 : blume 916 Name: Matthias Blume
1464 : blume 939 Date: 2001/09/25 15:25:00 EDT
1465 :     Tag: blume-20010925-exninfo
1466 :     Description:
1467 :    
1468 :     I provided a non-hook implementation of exnName (at the toplevel) and
1469 :     made the "dummy" implementation of exnMessage (at the toplevel) more
1470 :     useful: if nothing gets "hooked in", then at least you are going to
1471 :     see the exception name and a message indicating why you don't see more.
1472 :    
1473 :     [For the time being, programs that need exnMessage and want to use
1474 :     ml-build should either use General.exnMessage (strongly recommended) or
1475 :     refer to structure General at some other point so that CM sees a
1476 :     static dependency.]
1477 :    
1478 :     [Similar remarks go for "print" and "use": If you want to use their
1479 :     functionality in stand-alone programs generated by ml-build, then use
1480 :     TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
1481 :    
1482 :     ----------------------------------------------------------------------
1483 :     Name: Matthias Blume
1484 : blume 938 Date: 2001/09/20 17:28:00 EDT
1485 :     Tag: blume-20010920-slowfp
1486 :     Description:
1487 :    
1488 :     Allen says that x86-fast-fp is not safe yet, so I turned it off again...
1489 :    
1490 :     ----------------------------------------------------------------------
1491 :     Name: Matthias Blume
1492 : blume 937 Date: 2001/09/20 17:20:00 EDT
1493 :     Tag: blume-20010920-canonicalpaths
1494 :     Description:
1495 :    
1496 :     0. Updated the BOOT file (something that I forgot to do earlier).
1497 :    
1498 :     1. Small internal change to CM so that it avoids "/../" in filenames
1499 :     as much as possible (but only where it is safe).
1500 :    
1501 :     2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
1502 :     that arguments that contain delimiters are passed through correctly.
1503 :     This change also means that all "special" arguments of the form
1504 :     @SMLxxx... must come first.
1505 :    
1506 :     3. Changed install script to put relative anchor names for tool commands
1507 :     into pathconfig.
1508 :    
1509 :     ----------------------------------------------------------------------
1510 :     Name: Matthias Blume
1511 : blume 932 Date: 2001/09/18 15:35:00 EDT
1512 :     Tag: blume-20010918-readme11036
1513 :     Description:
1514 :    
1515 :     Added README files.
1516 :    
1517 :     ----------------------------------------------------------------------
1518 :     Name: Matthias Blume
1519 : blume 929 Date: 2001/09/18 11:45:00 EDT
1520 :     Tag: Release_110_36 (retag)
1521 :     Description:
1522 :    
1523 :     Fixed mistake in config/preloads. Retagged as 110.36.
1524 :    
1525 :     ----------------------------------------------------------------------
1526 :     Name: Matthias Blume
1527 : blume 928 Date: 2001/09/18 09:40:00 EDT
1528 : blume 929 Tag: Release_110_36_orig (tag changed)
1529 : blume 928 Description:
1530 :    
1531 :     New version (110.36). New bootfiles.
1532 :    
1533 :     ----------------------------------------------------------------------
1534 :     Name: Matthias Blume
1535 : blume 927 Date: 2001/09/14 16:15:00 EDT
1536 :     Tag: blume-20010914-x86fastfp
1537 :     Description:
1538 :    
1539 :     John committed some changes that Allen made, in particular a (hopefully)
1540 :     correctly working version of the x86-fp module.
1541 :    
1542 :     I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
1543 :     flag to "true". Everything seems to compile to a fixpoint ok, and
1544 :     "mandelbrot" speeds up by about 15%.
1545 :    
1546 :     ----------------------------------------------------------------------
1547 :     Name: Matthias Blume
1548 : blume 923 Date: 2001/09/13 11:20:00 EDT
1549 :     Tag: blume-20010913-minimal
1550 :     Description:
1551 :    
1552 :     1. Stefan Monnier's patch to fix a miscompilation problem that
1553 :     was brought to light by John Reppy's work on Moby.
1554 :    
1555 :     2. Implemented a minimal "structure Compiler" that contains just
1556 :     "version" and "architecture". The minimal version will be
1557 :     available when the full version is not. This is for backward-
1558 :     compatibility with code that wants to test Compiler.version.
1559 :    
1560 :     ----------------------------------------------------------------------
1561 :     Name: Matthias Blume
1562 : blume 920 Date: 2001/08/28 14:03:00 EDT
1563 :     Tag: blume-20010828-ml-lex
1564 :     Description:
1565 :    
1566 :     Fix for bug 1581, received from Neophytos Michael.
1567 :    
1568 :     ----------------------------------------------------------------------
1569 :     Name: Matthias Blume
1570 : blume 918 Date: 2001/08/27 11:20:00 EDT
1571 :     Tag: blume-20010827-readme11035
1572 :     Description:
1573 :    
1574 :     Fleshed out the README file for 110.35.
1575 :    
1576 :     ----------------------------------------------------------------------
1577 :     Name: Matthias Blume
1578 : blume 916 Date: 2001/08/24 17:10:00 EDT
1579 :     Tag: Release_110_35
1580 :     Description:
1581 :    
1582 :     New version number (110.35). New bootfiles.
1583 :    
1584 :     ----------------------------------------------------------------------
1585 : george 915 Name: Lal George
1586 : george 909 Date: 2001/08/24 13:47:18 EDT 2001
1587 :     Tag: george-20010824-MLRISC-graphs
1588 :     Description:
1589 :    
1590 :     removed clusters from MLRISC completely and replaced with graphs.
1591 :    
1592 :     ----------------------------------------------------------------------
1593 :     Name: Matthias Blume
1594 : blume 905 Date: 2001/08/23 17:50:00 EDT
1595 :     Tag: blume-20010823-toplevel
1596 :     Description:
1597 :    
1598 :     - some reorganization of the code that implements various kinds of
1599 :     environments in the compiler (static, dynamic, symbolic, combined)
1600 :     - re-implemented the EnvRef module so that evalStream works properly
1601 :     (if the stream contains references to "use", "CM.make", etc.)
1602 :     - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
1603 :    
1604 :     ----------------------------------------------------------------------
1605 :     Name: Matthias Blume
1606 : blume 904 Date: 2001/08/20 15:50 EDT
1607 :     Tag: blume20010820-slipup
1608 :     Description:
1609 :    
1610 :     I forgot to commit a few files. Here they are...
1611 :    
1612 :     ----------------------------------------------------------------------
1613 :     Name: Matthias Blume
1614 : blume 903 Date: 2001/08/20 15:35:00 EDT
1615 :     Tag: blume-20010820-debugprof
1616 :     Description:
1617 :    
1618 :     !!!! NEW BOOTFILES !!!!
1619 :    
1620 :     This is another round of reorganizing the compiler sources. This
1621 :     time the main goal was to factor out all the "instrumentation"
1622 :     passes (for profiling and backtracing) into their own library.
1623 :     The difficulty was to do it in such a way that it does not depend
1624 :     on elaborate.cm but only on elabdata.cm.
1625 :    
1626 :     Therefore there have been further changes to both elaborate.cm and
1627 :     elabdata.cm -- more "generic" things have been moved from the former
1628 :     to the latter. As a result, I was forced to split the assignment
1629 :     of numbers indicating "primtyc"s into two portions: SML-generic and
1630 :     SML/NJ-specific. Since it would have been awkward to maintain,
1631 :     I bit the bullet and actually _changed_ the mapping between these
1632 :     numbers and primtycs. The bottom line of this is that you need
1633 :     a new set of bin- and bootfiles.
1634 :    
1635 :     I have built new bootfiles for all architectures, so doing a fresh
1636 :     checkout and config/install.sh should be all you need.
1637 :    
1638 :     The newly created library's name is
1639 :    
1640 :     $smlnj/viscomp/debugprof.cm
1641 :    
1642 :     and its sources live under
1643 :    
1644 :     src/compiler/DebugProf
1645 :    
1646 :     ----------------------------------------------------------------------
1647 :     Name: Matthias Blume
1648 : blume 902 Date: 2001/08/15 17:15:00 EDT
1649 :     Tag: blume-20010815-compreorg
1650 :     Description:
1651 :    
1652 :     This is a first cut at reorganizing the CM libraries that make up the
1653 :     core of the compiler. The idea is to separate out pieces that could
1654 :     be used independently by tools, e.g., the parser, the typechecker, etc.
1655 :    
1656 :     The current status is a step in this direction, but it is not quite
1657 :     satisfactory yet. Expect more changes in the future.
1658 :    
1659 :     Here is the current (new) organization...
1660 :    
1661 :     What used to be $smlnj/viscomp/core.cm is now divided into
1662 :     six CM libraries:
1663 :    
1664 :     $smlnj/viscomp/basics.cm
1665 :     /parser.cm
1666 :     /elabdata.cm
1667 :     /elaborate.cm
1668 :     /execute.cm
1669 :     /core.cm
1670 :    
1671 :     The CM files for these libraries live under src/system/smlnj/viscomp.
1672 :     All these libraries are proxy libraries that contain precisely
1673 :     one CM library component. Here are the locations of the components
1674 :     (all within the src/compiler tree):
1675 :    
1676 :     Basics/basics.cm
1677 :     Parse/parser.cm
1678 :     ElabData/elabdata.cm
1679 :     Elaborator/elaborate.cm
1680 :     Execution/execute.cm
1681 :     core.cm
1682 :    
1683 :     [This organization is the same that has been used already
1684 :     for a while for the architecture-specific parts of the visible
1685 :     compiler and for the old version of core.cm.]
1686 :    
1687 :     As you will notice, many source files have been moved from their
1688 :     respective original locations to a new home in one of the above
1689 :     subtrees.
1690 :    
1691 :     The division of labor between the new libraries is the following:
1692 :    
1693 :     basics.cm:
1694 :     - Simple, basic definitions that pertain to many (or all) of
1695 :     the other libraries.
1696 :     parser.cm:
1697 :     - The SML parser, producing output of type Ast.dec.
1698 :     - The type family for Ast is also defined and exported here.
1699 :     elabdata.cm:
1700 :     - The datatypes that describe input and output of the elaborator.
1701 :     This includes types, absyn, and static environments.
1702 :     elaborator.cm:
1703 :     - The SML/NJ type checker and elaborator.
1704 :     This maps an Ast.dec (with a given static environment) to
1705 :     an Absyn.dec (with a new static environment).
1706 :     - This libraries implements certain modules that used to be
1707 :     structures as functors (to remove dependencies on FLINT).
1708 :     execute.cm:
1709 :     - Everything having to do with executing binary code objects.
1710 :     - Dynamic environments.
1711 :     core.cm:
1712 :     - SML/NJ-specific instantiations of the elaborator and MLRISC.
1713 :     - Top-level modules.
1714 :     - FLINT (this should eventually become its own library)
1715 :    
1716 :     Notes:
1717 :    
1718 :     I am not 100% happy with the way I separated the elaborator (and its
1719 :     data structures) from FLINT. Two instances of the same problem:
1720 :    
1721 :     1. Data structures contain certain fields that carry FLINT-specific
1722 :     information. I hacked around this using exn and the property list
1723 :     module from smlnj-lib. But the fact that there are middle-end
1724 :     specific fields around at all is a bit annoying.
1725 :    
1726 :     2. The elaborator calculates certain FLINT-related information. I tried
1727 :     to make this as abstract as I could using functorization, but, again,
1728 :     the fact that the elaborator has to perform calculations on behalf
1729 :     of the middle-end at all is not nice.
1730 :    
1731 :     3. Having to used exn and property lists is unfortunate because it
1732 :     weakens type checking. The other alternative (parameterizing
1733 :     nearly *everything*) is not appealing, though.
1734 :    
1735 :     I removed the "rebinding =" warning hack because due to the new organization
1736 :     it was awkward to maintain it. As a result, the compiler now issues some of
1737 :     these warnings when compiling init.cmi during bootstrap compilation. On
1738 :     the plus side, you also get a warning when you do, for example:
1739 :     val op = = Int32.+
1740 :     which was not the case up to now.
1741 :    
1742 :     I placed "assign" and "deref" into the _Core structure so that the
1743 :     code that deals with the "lazy" keyword can find them there. This
1744 :     removes the need for having access to the primitive environment
1745 :     during elaboration.
1746 :    
1747 :     ----------------------------------------------------------------------
1748 :     Name: Matthias Blume
1749 : blume 898 Date: 2001/08/13
1750 :     Tag: blume-20010813-closures
1751 :     Description:
1752 :    
1753 :     This fix was sent to us by Zhong Shao. It is supposed to improve the
1754 :     performance of certain loops by avoiding needless closure allocation.
1755 :    
1756 :     ----------------------------------------------------------------------
1757 : george 890 Name: Lal George
1758 : george 896 Date: 2001/07/31 10:03:23 EDT 2001
1759 :     Tag: george-20010731-x86-fmalloc
1760 :     Description: Fixed bug in x86 calls
1761 :    
1762 :     There was a bug where call instructions would mysteriously
1763 :     vanish. The call instruction had to be one that returned
1764 :     a floating point value.
1765 :    
1766 :     ----------------------------------------------------------------------
1767 :     Name: Lal George
1768 : george 890 Date: 2001/07/19 16:36:29 EDT 2001
1769 :     Tag: george-20010719-simple-cells
1770 :     Description:
1771 :    
1772 :     I have dramatically simplified the interface for CELLS in MLRISC.
1773 :    
1774 :     In summary, the cells interface is broken up into three parts:
1775 :    
1776 :     1. CellsBasis : CELLS_BASIS
1777 :    
1778 :     CellsBasis is a top level structure and common for all
1779 :     architectures. it contains the definitions of basic datatypes
1780 :     and utility functions over these types.
1781 :    
1782 :     2. functor Cells() : CELLS
1783 :    
1784 :     Cells generates an interface for CELLS that incorporates the
1785 :     specific resources on the target architecture, such as the
1786 :     presence of special register classes, their number and size,
1787 :     and various useful substructures.
1788 :    
1789 :     3. <ARCH>CELLS
1790 :    
1791 :     e.g. SparcCells: SPARCCELLS
1792 :    
1793 :     <ARCH>CELLS usually contains additional bindings for special
1794 :     registers on the architecture, such as:
1795 :    
1796 :     val r0 : cell (* register zero *)
1797 :     val y : cell (* Y register *)
1798 :     val psr : cell (* processor status register *)
1799 :     ...
1800 :    
1801 :     The structure returned by applying the Cells functor is opened
1802 :     in this interface.
1803 :    
1804 :     The main implication of all this is that the datatypes for cells is
1805 :     split between CellsBasis and CELLS -- a fairly simple change for user
1806 :     code.
1807 :    
1808 :     In the old scheme the CELLS interface had a definitional binding of
1809 :     the form:
1810 :    
1811 :     signature CELLS = sig
1812 :    
1813 :     structure CellsBasis = CellsBasis
1814 :    
1815 :     ...
1816 :    
1817 :     end
1818 :    
1819 :     With all the sharing constraints that goes on in MLRISC, this old
1820 :     design quickly leads to errors such as:
1821 :    
1822 :     "structure definition spec inside of sharing ... "
1823 :    
1824 :    
1825 :     and appears to require an unacceptable amount of sharing and where
1826 :     constraint hackery.
1827 :    
1828 :     I think this error message (the interaction of definitional specs and
1829 :     sharing) requires more explanation on our web page.
1830 :    
1831 :     ----------------------------------------------------------------------
1832 : blume 826 Name: Matthias Blume
1833 : blume 879 Date: 2001/07/19 15:00:00 EDT
1834 :     Tag: blume-20010719-libreorg
1835 :     Description:
1836 :    
1837 :     This update puts together a fairly extensive but straightforward change
1838 :     to the way the libraries that implement the interactive system are
1839 :     organized:
1840 :    
1841 :     The biggest change is the elimination of structure Compiler. As a
1842 :     replacement for this structure, there is now a CM library
1843 :     (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1844 :     that exports all the substructures of the original structure Compiler
1845 :     directly. So instead of saying Compiler.Foo.bar one now simply
1846 :     says Foo.bar. (The CM libraries actually export a collection of
1847 :     structures that is richer than the collection of substructures of
1848 :     structure Compiler.)
1849 :    
1850 :     To make the transition smooth, there is a separate library called
1851 :     $smlnj/compiler/compiler.cm which puts together and exports the
1852 :     original structure Compiler (or at least something very close to it).
1853 :    
1854 :     There are five members of the original structure Compiler
1855 :     that are not exported directly but which instead became members
1856 :     of a new structure Backend (described by signature BACKEND). These are:
1857 :     structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1858 :     Interact (: INTERACT), structure Machine (: MACHINE), and val
1859 :     architecture (: string).
1860 :    
1861 :     Structure Compiler.Version has become structure CompilerVersion.
1862 :    
1863 :     Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1864 :     by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1865 :     or x86, respectively.
1866 :     Each of these exports the same frontend structures that
1867 :     $smlnj/compiler.cm exports. But they do not have a structure Backend
1868 :     and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1869 :     Hppa, PPC, Sparc, or X86, respectively.
1870 :    
1871 :     Library $smlnj/compiler/all.cm exports the union of the exports of
1872 :     $smlnj/compiler/<arch>.cm
1873 :    
1874 :     There are no structures <Arch>Compiler anymore, use
1875 :     $smlnj/compiler/<arch>.cm instead.
1876 :    
1877 :     Library host-compiler-0.cm is gone. Instead, the internal library
1878 :     that instantiates CM is now called cm0.cm. Selection of the host
1879 :     compiler (backend) is no longer done here but. (Responsibility for it
1880 :     now lies with $smlnj/compiler/current.cm. This seems to be more
1881 :     logical.)
1882 :    
1883 :     Many individual files have been moved or renamed. Some files have
1884 :     been split into multiple files, and some "dead" files have been deleted.
1885 :    
1886 :     Aside from these changes to library organization, there are also changes
1887 :     to the way the code itself is organized:
1888 :    
1889 :     Structure Binfile has been re-implemented in such a way that it no
1890 :     longer needs any knowledge of the compiler. It exclusively deals
1891 :     with the details of binfile layout. It no longer invokes the
1892 :     compiler (for the purpose of creating new prospective binfile
1893 :     content), and it no longer has any knowledge of how to interpret
1894 :     pickles.
1895 :    
1896 :     Structure Compile (: COMPILE) has been stripped down to the bare
1897 :     essentials of compilation. It no longer deals with linking/execution.
1898 :     The interface has been cleaned up considerably.
1899 :    
1900 :     Utility routines for dealing with linking and execution have been
1901 :     moved into their own substructures.
1902 :    
1903 :     (The ultimate goal of these changes is to provide a light-weight
1904 :     binfile loader/linker (at least for, e.g., stable libraries) that
1905 :     does not require CM or the compiler to be present.)
1906 :    
1907 :     CM documentation has been updated to reflect the changes to library
1908 :     organization.
1909 :    
1910 :     ----------------------------------------------------------------------
1911 :     Name: Matthias Blume
1912 : blume 875 Date: 2001/07/10 17:30:00 EDT
1913 :     Tag: Release_110_34
1914 :     Description:
1915 :    
1916 : blume 876 Minor tweak to 110.34 (re-tagged):
1917 :    
1918 :     - README.html file added to CVS repository
1919 :     - runtime compiles properly under FreeBSD 3.X and 4.X
1920 :    
1921 :     ----------------------------------------------------------------------
1922 :     Name: Matthias Blume
1923 :     Date: 2001/07/10 17:30:00 EDT
1924 :     Tag: Release_110_34
1925 :     Description:
1926 :    
1927 : blume 875 New version number (110.34). New bootfiles.
1928 :    
1929 :     ----------------------------------------------------------------------
1930 :     Name: Matthias Blume
1931 : blume 874 Date: 2001/07/09 16:00:00 EDT
1932 :     Tag: blume-20010709-more-varargs
1933 :     Description:
1934 :    
1935 :     I changed the handling of varargs in ml-nlffigen again:
1936 :     The ellipsis ... will now simply be ignored (with an accompanying warning).
1937 :    
1938 :     The immediate effect is that you can actually call a varargs function
1939 :     from ML -- but you can't actually supply any arguments beyond the ones
1940 :     specified explicitly. (For example, you can call printf with its format
1941 :     string, but you cannot pass additional arguments.)
1942 :    
1943 :     This behavior is only marginally more useful than the one before, but
1944 :     it has the advantage that a function or, more importantly, a function
1945 :     type never gets dropped on the floor, thus avoiding follow-up problems with
1946 :     other types that refer to the offending one.
1947 :    
1948 :     ----------------------------------------------------------------------
1949 :     Name: Matthias Blume
1950 : blume 873 Date: 2001/07/09 11:25:00 EDT
1951 :     Tag: blume-20010709-varargs
1952 :     Description:
1953 :    
1954 :     1. ckit-lib.cm now exports structure Error
1955 :     2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1956 :     with a warning accompanied by a source location. Moreover, it
1957 :     merely skips the offending function or type and proceeds with the
1958 :     rest of its work.u As a result, one can safely feed C code containing
1959 :     "..." to ml-nlffigen.
1960 :     3. There are some internal improvements to CM, providing slightly
1961 :     more general string substitutions in the tools subsystem.
1962 :    
1963 :     ----------------------------------------------------------------------
1964 :     Name: Matthias Blume
1965 : blume 854 Date: 2001/06/27 15:10:00 EDT
1966 :     Tag: blume-20010627-concur
1967 :     Description:
1968 :    
1969 :     Fixed a small bug in CM's handling of parallel compilation.
1970 :     (You could observe the bug by Control-C-interrupting an ordinary
1971 :     CMB.make or CM.stabilize and then attaching some compile servers.
1972 :     The result was that all of a sudden the previously interrupted
1973 :     compilation would continue on its own. This was because of
1974 :     an over-optimization: CM did not bother to clean out certain queues
1975 :     when no servers were attached "anyway", resulting in the contents
1976 :     of these queues to grab control when new servers did get attached.)
1977 :    
1978 :     There is also another minor update to the CM manual.
1979 :    
1980 :     ----------------------------------------------------------------------
1981 :     Name: Matthias Blume
1982 : blume 853 Date: 2001/06/26 16:15:00 EDT
1983 :     Tag: blume-20010626-cmdoc
1984 :     Description:
1985 :    
1986 :     Minor typo fixed in CM manual (syntax diagram for libraries).
1987 :    
1988 :     ----------------------------------------------------------------------
1989 :     Name: Matthias Blume
1990 : blume 852 Date: 2001/06/25 22:55:00 EDT
1991 :     Tag: blume-20010625-x86pc
1992 :     Description:
1993 :    
1994 :     Fixed a nasty bug in the X86 assembly code that caused signal
1995 :     handlers to fail (crash) randomly.
1996 :    
1997 :     ----------------------------------------------------------------------
1998 :     Name: Matthias Blume
1999 : blume 846 Date: 2001/06/25 12:05:00 EDT
2000 :     Tag: blume-20010625-nlffigen
2001 :     Description:
2002 :    
2003 :     This update fixes a number of minor bugs in ml-nlffigen as reported by
2004 :     Nick Carter <nbc@andrew.cmu.edu>.
2005 :    
2006 :     1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
2007 :     2. Default names for generated files are now derived from the name of
2008 :     the C file *without its directory*. In particular, this causes generated
2009 :     files to be placed locally even if the C file is in some system directory.
2010 :     3. Default names for generated signatures and structures are also derived
2011 :     from the C file name without its directory. This avoids silly things
2012 :     like "structure GL/GL".
2013 :     (Other silly names are still possible because ml-nlffigen does not do
2014 : blume 847 a thorough check of whether generated names are legal ML identifiers.
2015 :     When in doubt, use command line arguments to force particular names.)
2016 : blume 846
2017 :     ----------------------------------------------------------------------
2018 :     Name: Matthias Blume
2019 : blume 845 Date: 2001/06/21 12:25:00 EDT
2020 :     Tag: blume-20010621-eXene
2021 :     Description:
2022 :    
2023 :     eXene now compiles and (sort of) works again.
2024 :    
2025 :     The library name (for version > 110.33) is $/eXene.cm.
2026 :    
2027 :     I also added an new example in src/eXene/examples/nbody. See the
2028 :     README file there for details.
2029 :    
2030 :     ----------------------------------------------------------------------
2031 :     Name: Matthias Blume
2032 : blume 844 Date: 2001/06/20 16:40:00 EDT
2033 :     Tag: blume-20010620-cml
2034 :     Description:
2035 :    
2036 :     CML now compiles and works again.
2037 :    
2038 :     Libraries (for version > 110.33):
2039 :    
2040 :     $cml/cml.cm Main CML library.
2041 :     $cml/basis.cm CML's version of $/basis.cm.
2042 :     $cml/cml-internal.cm Internal helper library.
2043 :     $cml/core-cml.cm Internal helper library.
2044 :     $cml-lib/trace-cml.cm Tracing facility.
2045 :     $cml-lib/smlnj-lib.cm CML's version of $/smlnj-lib.cm
2046 :    
2047 :     The installer (config/install.sh) has been taught how to properly
2048 :     install this stuff.
2049 :    
2050 :     ----------------------------------------------------------------------
2051 :     Name: Matthias Blume
2052 : blume 843 Date: 2001/06/19 17:55:00 EDT
2053 :     Tag: blume-20010619-instantiate
2054 :     Description:
2055 :    
2056 :     This un-breaks the fix for bug 1432.
2057 :     (The bug was originally fixed in 110.9 but I broke it again some
2058 :     time after that.)
2059 :    
2060 :     ----------------------------------------------------------------------
2061 :     Name: Matthias Blume
2062 : blume 842 Date: 2001/06/19 17:25:00 EDT
2063 :     Tag: blume-20010619-signals
2064 :     Description:
2065 :    
2066 :     This should (hopefully) fix the long-standing signal handling bug.
2067 :     (The runtime system was constructing a continuation record with an
2068 :     incorrect descriptor which would cause the GC to drop data on the floor...)
2069 :    
2070 :     ----------------------------------------------------------------------
2071 :     Name: Matthias Blume
2072 : blume 840 Date: 2001/06/15 15:05:00 EDT
2073 : blume 841 Tag: blume-20010615-moresparc
2074 :     Description:
2075 :    
2076 :     Here is a short late-hour update related to Sparc c-calls:
2077 :    
2078 :     -- made handling of double-word arguments a bit smarter
2079 :    
2080 :     -- instruction selection phase tries to collapse certain clumsily
2081 :     constructed ML-Trees; typical example:
2082 :    
2083 :     ADD(ty,ADD(_,e,LI d1),LI d2) -> ADD(ty,e,LI(d1+d2))
2084 :    
2085 :     This currently has no further impact on SML/NJ since mlriscGen does
2086 :     not seem to generate such patterns in the first place, and c-calls
2087 :     (which did generate them in the beginning) has meanwhile been fixed
2088 :     so as to avoid them as well.
2089 :    
2090 :     ----------------------------------------------------------------------
2091 :     Name: Matthias Blume
2092 :     Date: 2001/06/15 15:05:00 EDT
2093 : blume 840 Tag: blume-20010615-sparc
2094 :     Description:
2095 :    
2096 :     The purpose of this update is to provide an implementation of NLFFI
2097 :     on Sparc machines.
2098 :    
2099 :     Here are the changes in detail:
2100 :    
2101 :     * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
2102 :     the Sparc implementation of the c-calls API.
2103 :     * The Sparc backend of SML/NJ has been modified to uniformely use %fp
2104 :     for accessing the ML frame. Thus, we have a real frame pointer and
2105 :     can freely modify %sp without need for an omit-frame-ptr phase.
2106 :     The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
2107 :     this case.
2108 :     * ml-nlffigen has been taught to produce code for different architectures
2109 :     and calling conventions.
2110 :     * In a way similar to what was done in the x86 case, the Sparc
2111 :     backend uses its own specific extension to mltree. (For example,
2112 :     it needs to be able to generate UNIMP instructions which are part
2113 :     of the calling convention.)
2114 :     * ml-nlffi-lib was reorganized to make it more modular (in particular,
2115 :     to make it easier to plug in new machine- and os-dependent parts).
2116 :    
2117 :     There are some other fairly unrelated bug fixes and cleanups as well:
2118 :    
2119 :     * I further hacked the .cm files for MLRISC tools (like MDLGen) so
2120 :     that they properly share their libraries with existing SML/NJ libraries.
2121 :     * I fixed a minor cosmetic bug in CM, supressing certain spurious
2122 :     follow-up error messages.
2123 :     * Updates to CM/CMB documentation.
2124 :    
2125 :     TODO items:
2126 :    
2127 :     * MLRISC should use a different register as its asmTemp on the Sparc.
2128 :     (The current %o2 is a really bad choice because it is part of the
2129 :     calling conventions, so things might interfere in unexpected ways.)
2130 :    
2131 :     ----------------------------------------------------------------------
2132 :     Name: Matthias Blume
2133 : blume 839 Date: 2001/06/07
2134 :     Tag: blume-20010607-calls
2135 :     Description:
2136 :    
2137 :     A number of internal changes related to C calls and calling conventions:
2138 :    
2139 :     1. ML-Tree CALL statements now carry a "pops" field. It indicates the
2140 :     number of bytes popped implicitly (by the callee). In most cases
2141 :     this field is 0 but on x86/win32 it is some non-zero value. This
2142 :     is information provided for the benefit of the "omit-frameptr" pass.
2143 :     2. The CALL instruction on the x86 carries a similar "pops" field.
2144 :     The instruction selection phase copies its value from the ML-Tree
2145 :     CALL statement.
2146 :     3. On all other architectures, the instruction selection phase checks
2147 :     whether "pops=0" and complains if not.
2148 :     4. The c-calls implementation for x86 now accepts two calling conventions:
2149 :     "ccall" and "stdcall". When "ccall" is selected, the caller cleans
2150 :     up after the call and pops is set to 0. For "stdcall", the caller
2151 :     does nothing, leaving the cleanup to the callee; pops is set to
2152 :     the number of bytes that were pushed onto the stack.
2153 :     5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
2154 :     distinguish between "ccall" and "stdcall".
2155 :     6. The UNIMP instruction has been added to the supported Sparc instruction
2156 :     set. (This is needed for implementing the official C calling convention
2157 :     on this architecture.)
2158 :     7. I fixed some of the .cm files under src/MLRISC/Tools to make them
2159 :     work with the latest CM.
2160 :    
2161 :     ----------------------------------------------------------------------
2162 :     Name: Matthias Blume
2163 : blume 838 Date: 2001/06/05 15:10:00 EDT
2164 :     Tag: blume-20010605-cm-index
2165 :     Description:
2166 :    
2167 :     0. The "lambdasplit" parameter for class "sml" in CM has been documented.
2168 :    
2169 :     1. CM can now generate "index files". These are human-readable files
2170 :     that list on a per-.cm-file basis each toplevel symbol defined or
2171 :     imported. The location of the index file for
2172 :     <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
2173 :     To enable index-file generation, set CM.Control.generate_index to true
2174 :     or export an environment-symbol: export CM_GENERATE_INDEX=true.
2175 :    
2176 :     The CM manual has been updated accordingly.
2177 :    
2178 :     2. I made some slight modifications to the c-calls API in MLRISC.
2179 :    
2180 :     a) There is now a callback to support saving/restoring of
2181 :     dedicated but caller-save registers around the actual call
2182 :     instruction.
2183 :     b) One can optionally specify a comment-annotation for the
2184 :     call instruction.
2185 :    
2186 :     3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
2187 :     (For example, the comment annotation shows the C prototype of
2188 :     the function being called.)
2189 :    
2190 :     ----------------------------------------------------------------------
2191 :     Name: Matthias Blume
2192 : blume 837 Date: 2001/06/01 13:30:00 EDT
2193 :     Tag: blume-20010601-nlffi-cleanup
2194 :     Description:
2195 :    
2196 :     This is mostly a cleanup of MLFFI stuff:
2197 :    
2198 :     - some signature files have been put into a more exposed place
2199 :     - the ugly 'f type parameter is gone (simplifies types tremendously!)
2200 :     - ml-nlffigen changed accordingly
2201 :     - tutorial updated
2202 :    
2203 :     Other changes:
2204 :    
2205 :     - author's affiliation in CM manual(s) updated
2206 :     - some more recognized keywords added to Allen's sml.sty
2207 :    
2208 :     ----------------------------------------------------------------------
2209 :     Name: Matthias Blume
2210 : blume 836 Date: 2001/05/25 15:30:00 EDT
2211 :     Tag: blume-20010525-iptr
2212 :     Description:
2213 :    
2214 :     - put the official 110.33-README (as it appears on the ftp server) under
2215 :     CVS
2216 :     - fixed a small bug related to incomplete pointer types in
2217 :     ml-nlffigen
2218 :     - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
2219 :     (it does not need the 'f type parameter)
2220 :    
2221 :     ----------------------------------------------------------------------
2222 :     Name: Matthias Blume
2223 : blume 834 Date: 2001/05/23 14:30:00 EDT
2224 :     Tag: Release_110_33
2225 :     Description:
2226 :    
2227 :     New version number (110.33). New bootfiles.
2228 :    
2229 :     ----------------------------------------------------------------------
2230 :     Name: Matthias Blume
2231 : blume 833 Date: 2001/05/22 18:06:00 EDT
2232 :     Tag: blume-20010522-targets
2233 :     Description:
2234 :    
2235 :     Made install.sh use file config/targets.customized if it exists, falling
2236 :     back to config/targets if it doesn't. This way one can have a customized
2237 :     version of the targets file without touching the "real thing", thus
2238 :     eliminating the constant fear of accidentally checking something bogus
2239 :     back into the CVS repository... (File config/targets.customized must
2240 :     not be added to the repository!)
2241 :    
2242 :     ----------------------------------------------------------------------
2243 :     Name: Matthias Blume
2244 : blume 832 Date: 2001/05/22 16:30:00 EDT
2245 :     Tag: blume-20010522-minitut
2246 :     Description:
2247 :    
2248 :     1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
2249 :     struct returns.
2250 :     2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt. This is some very
2251 :     incomplete, preliminary documentation for NLFFI.
2252 :    
2253 :     ----------------------------------------------------------------------
2254 :     Name: Matthias Blume
2255 : blume 829 Date: 2001/05/14 11:30:00 EDT
2256 :     Tag: blume-20010514-script
2257 :     Description:
2258 :    
2259 :     Some bugs in install script fixed.
2260 :    
2261 :     In addition to that I also made a slight change to the NLFFI API:
2262 :     Functors generated by ml-nlffigen now take the dynamic library as a
2263 :     straight functor argument, not as a suspended one. (The original
2264 :     functor code used to force the suspension right away anyway, so there
2265 :     was nothing gained by this complication of the interface.)
2266 :    
2267 :     ----------------------------------------------------------------------
2268 :     Name: Matthias Blume
2269 : blume 828 Date: 2001/05/11 14:35:00 EDT
2270 :     Tag: blume-20010511-ml-nlffi
2271 :     Description:
2272 :    
2273 :     I finally took the plunge and added my new FFI code to the main
2274 :     repository. For x86-linux it is now ready for prime-time.
2275 :    
2276 :     There are two new subdirectories of "src":
2277 :    
2278 :     - ml-nlffi-lib:
2279 :     The utility library for programs using the FFI interface.
2280 :     Here is the implementation of $/c.cm and its associated low-level
2281 :     partners $/c-int.cm and $/memory.cm.
2282 :     - ml-nlffigen:
2283 :     A stand-alone program for generating ML glue code from C source
2284 :     code.
2285 :    
2286 :     Building ml-nlffigen requires $/ckit-lib.cm.
2287 :    
2288 :     The config/install.sh script has been updates to do the Right Thing
2289 :     (hopefully).
2290 :    
2291 :     Notice that the source tree for the C-Kit will not be put under "src"
2292 :     but directly under the installation root directory. (This is the
2293 :     structure that currently exists on the CVS server when you check out
2294 :     module "sml".) Fortunately, config/install.sh knows about this oddity.
2295 :    
2296 :     Bugs: No documentation yet.
2297 :    
2298 :     ----------------------------------------------------------------------
2299 :     Name: Matthias Blume
2300 : blume 826 Date: 2001/05/09 16:35:00 EDT
2301 :     Tag: blume-20010509-cpscontract
2302 :     Description:
2303 :    
2304 :     Fixed a bug in the accounting code in cpsopt/contract.sml. (The
2305 :     wrapper/unwrapper elimination did not decrement usage counts and some
2306 :     dead variables got overlooked by the dead-up logic.)
2307 :    
2308 :     ----------------------------------------------------------------------
2309 : george 824 Name: Lal George
2310 :     Date: 2001/05/08 17:26:09 EDT
2311 :     Tag: george-20010508-omit-frameptr
2312 :     Description:
2313 :    
2314 :     Changes to implement the omit-frame-pointer optimization to support
2315 :     raw C calls. For now, there is only support on the Intel x86, but
2316 :     other architectures will follow as more experience is gained with this.
2317 :    
2318 :    
2319 :     ----------------------------------------------------------------------
2320 : blume 818 Name: Matthias Blume
2321 : blume 819 Date: 2001/05/07 14:40:00 EDT
2322 :     Tag: blume-20010507-proxies
2323 :     Description:
2324 :    
2325 :     I made into "proxy libraries" all libraries that qualify for such a
2326 :     change. (A qualifying library is a library that has another library or
2327 :     groups as its sole member and repeats that member's export list
2328 :     verbatim. A proxy library avoids this repetition by omitting its export
2329 :     list, effectively inheriting the list that its (only) member exports.
2330 :     See the CM manual for more explanation.)
2331 :     The main effect is that explicit export lists for these libraries
2332 :     do not have to be kepts in sync, making maintenance a bit easier.
2333 :    
2334 :     I also added copyright notices to many .cm-files.
2335 :    
2336 :     Last but not least, I made a new set of bootfiles.
2337 :    
2338 :     ----------------------------------------------------------------------
2339 :     Name: Matthias Blume
2340 : blume 818 Date: 2001/05/04 17:00:00 EDT
2341 :     Tag: blume-20010504-cm-lsplit
2342 :     Description:
2343 :    
2344 :     0. John merged pending changes to $/smlnj-lib.cm
2345 :    
2346 :     1. Allen's previous change accidentally backed out of one of Lal's
2347 :     earlier changes. I undid this mistake (re-introducing Lal's change).
2348 :    
2349 :     2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
2350 :     within the compiler where applicable. There is some code simplification
2351 :     because of that.
2352 :    
2353 :     3. The "split" phase (in FLINT) is now part of the default list of phases.
2354 :     Compiler.Control.LambdaSplitting.* can be used to globally control the
2355 :     lambda-splitting (cross-module-inlining) engine. In addition to that,
2356 :     it can now also be controlled on a per-source basis: CM has been taught
2357 :     a new tool parameter applicable to ML source files.
2358 :    
2359 :     - To turn lambda-splitting off completely:
2360 :     local open Compiler.Control.LambdaSplitting in
2361 :     val _ = set Off
2362 :     end
2363 :     - To make "no lambda-splitting" the global default (but allow per-source
2364 :     overriding); this is the initial setting:
2365 :     local open Compiler.Control.LambdaSplitting in
2366 :     val _ = set (Default NONE)
2367 :     end
2368 :     - To make "lambda-splitting with aggressiveness a" the global default
2369 :     (and allow per-source overriding):
2370 :     local open Compiler.Control.LambdaSplitting in
2371 :     val _ = set (Default (SOME a))
2372 :     end
2373 :    
2374 :     - To turn lambda-splitting off for a given ML souce file (say: a.sml)
2375 :     write (in the respective .cm-file):
2376 :     a.sml (lambdasplitting:off)
2377 :     - To turn lambda-splitting for a.sml on with minimal aggressiveness:
2378 :     a.sml (lambdasplitting:on)
2379 :     - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
2380 :     <a> is a decimal non-negative integer):
2381 :     a.sml (lambdasplitting:<a>)
2382 :     - To turn lambda-splitting for a.sml on with maximal aggressiveness:
2383 :     a.sml (lambdasplitting:infinity)
2384 :     - To use the global default for a.sml:
2385 :     a.sml (lambdasplitting:default)
2386 :     or simply
2387 :     a.sml
2388 :    
2389 :     ----------------------------------------------------------------------
2390 : leunga 815 Name: Allen Leung
2391 :     Date: 2001/05/04 01:57:00 EDT
2392 :     Tag: leunga-20010504-sync
2393 :     Description:
2394 : leunga 788
2395 : leunga 815 MLRISC features.
2396 :    
2397 :     1. Fix to CMPXCHG instructions.
2398 :     2. Changed RA interface to allow annotations in callbacks.
2399 :     3. Added a new method to the stream interface to allow annotations updates.
2400 :    
2401 : george 749 ----------------------------------------------------------------------
2402 : blume 812 Name: Matthias Blume
2403 : blume 814 Date: 2001/05/01 11:45:00 EDT
2404 :     Tag: blume-20010501-pcedittmp
2405 :     Description:
2406 :    
2407 :     Changed install.sh to use the current working directory instead of
2408 :     /usr/tmp for a temporary file (pcedittmp). The previous choice
2409 :     of /usr/tmp caused trouble with MacOS X because of file premission
2410 :     problems.
2411 :    
2412 :     ----------------------------------------------------------------------
2413 :     Name: Matthias Blume
2414 : blume 812 Date: 2001/04/20 11:10:00 EDT
2415 :     Tag: blume-20010420-inMLflag
2416 :     Description:
2417 :    
2418 :     - added vp_limitPtrMask to vproc-state.h
2419 :     (for use by the raw-C-calls mechanism to implement proper interrupt
2420 :     handling)
2421 :     - made the ML compiler aware of various data-structure offsets so it
2422 :     can generate code for accessing the vp_inML flag and vp_limitPtrMask
2423 :     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
2424 :     raw C-calls
2425 :    
2426 :     ----------------------------------------------------------------------
2427 : george 810 Name: Lal George
2428 :     Date: 2001/04/20 09:15:28 EDT
2429 :     Tag: george-20010420-macosX
2430 :     Description:
2431 :    
2432 :     - Changes to port to Mac OS X; Darwin.
2433 :    
2434 :     - In the process I found that sqrt was broken on the PPC, because the
2435 :     fsqrt instruction is not implemented.
2436 :    
2437 :     ----------------------------------------------------------------------
2438 : blume 799 Name: Matthias Blume
2439 : blume 808 Date: 2001/04/18 12:45:00 EDT
2440 :     Tag: blume-20010418-ccalls
2441 :     Description:
2442 :    
2443 :     - fixed two off-by-4 errors in the x86-specific c-calls implementation
2444 :     (this bug prevented structure arguments containing pointers from being
2445 :     passed correctly)
2446 :     - changed the raw-C-call code in mlriscGen.sml in such a way that
2447 :     structure arguments are represented as a pointer to the beginning
2448 :     of the structure (instead of having a series of synthesized arguments,
2449 :     one for each structure member)
2450 :    
2451 :     - made makeml script's verbosity level configurable via environment
2452 :     variable (MAKEML_VERBOSITY)
2453 :    
2454 : blume 813 - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
2455 : blume 808 in rawmem-x86.sml; we are now using the real thing
2456 :    
2457 :     ----------------------------------------------------------------------
2458 :     Name: Matthias Blume
2459 : blume 806 Date: 2001/03/22 16:25:00 EST
2460 :     Tag: blume-20010322-bootfiles
2461 :     Description:
2462 :    
2463 :     Created a new set of bootfiles (for your automatic installation convenience).
2464 :    
2465 :     ----------------------------------------------------------------------
2466 :     Name: Matthias Blume
2467 : blume 805 Date: 2001/03/22 15:10:00 EST
2468 :     Tag: blume-20010322-rawmem-parcm
2469 :     Description:
2470 :    
2471 :     1. All "raw memory access" primitives for the new FFI are implemented now
2472 :     (at least on the x86).
2473 :     2. Some further cleanup of CM's parallel make mechanism.
2474 :    
2475 :     ----------------------------------------------------------------------
2476 :     Name: Matthias Blume
2477 : blume 801 Date: 2001/03/19 17:53:00 EST
2478 :     Tag: blume-20010319-parallel
2479 :     Description:
2480 :    
2481 :     Parallel make (using compile servers) now works again.
2482 :    
2483 :     To this end, CM.stabilize and CMB.make have been modified to work in
2484 :     two passes when compile servers are attached:
2485 :     1. Compile everything, do not perform stabilization; this pass
2486 :     uses compile servers
2487 :     2. Stabilize everything; this pass does not use compile servers
2488 :     If there are no compile servers, the two passes are combined into one
2489 :     (as before). Splitting the passes increases the inherent parallelism
2490 :     in the dependency graph because the entire graph including all
2491 :     libraries is available at the same time. This, in turn, improves
2492 :     server utilization. The downside is that the master process will
2493 :     have to do some extra work after compilation is done (because for
2494 :     technical reasons it must re-read all the binfiles during stabilization).
2495 :    
2496 :     ----------------------------------------------------------------------
2497 :     Name: Matthias Blume
2498 : blume 800 Date: 2001/03/16 12:22:00 EST
2499 :     Tag: blume-20010316-bootfiles
2500 :     Description:
2501 :    
2502 :     Created a new set of bootfiles (for your automatic installation convenience).
2503 :    
2504 :     ----------------------------------------------------------------------
2505 :     Name: Matthias Blume
2506 : blume 799 Date: 2001/03/16 11:00:00 EST
2507 :     Tag: blume-20010316-MLTREE-fixup
2508 :     Description:
2509 :    
2510 :     This is a minor fixup for an (untagged) earlier commit by Allen.
2511 :     (A file was missing).
2512 :    
2513 :     ----------------------------------------------------------------------
2514 : leunga 788 Name: Allen Leung
2515 : leunga 796 Date: Mon Mar 5 18:54:57 EST 2001
2516 :     Tag: leunga-20010305-cut-support
2517 :    
2518 :     1. New support for alternative control-flow in MLTREE.
2519 :     Currently we support
2520 :    
2521 :     FLOW_TO(CALL ...., [k1,...,kn])
2522 :    
2523 :     This is needed for 'cuts to' in C-- and try/handle-like constructs
2524 :     in Moby
2525 :    
2526 :     New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
2527 :    
2528 :     2. Register Allocator
2529 :    
2530 :     Changes in interface [from Fermin, John]
2531 :    
2532 :     3. Alpha 8-bit SLL support [Fermin]
2533 :    
2534 :     4. All architectures
2535 :    
2536 :     A new module (ClusterExpandCopies) for expanding parallel copies.
2537 :    
2538 :     ----------------------------------------------------------------------
2539 :     Name: Allen Leung
2540 : leunga 788 Date: 2001/02/27 23:07:00 EST
2541 :     Tag: leunga-20010227-minor-stuff
2542 :    
2543 :     1. Alpha bug fix for CMOVNE
2544 :     2. Handle mltree COND(..,FCMP ...,...)
2545 :     3. Bug fix in simplifier
2546 :    
2547 :     ----------------------------------------------------------------------
2548 : blume 777 Name: Matthias Blume
2549 : blume 781 Date: 2001/01/30 17:50:00 EST
2550 :     Tag: blume-20010130-sync
2551 :     Description:
2552 :    
2553 :     This is just a minor update to sync my devel branch with the main brach.
2554 :     The only visible change is the addition of some README files.
2555 :    
2556 :     ----------------------------------------------------------------------
2557 :     Name: Matthias Blume
2558 : blume 778 Date: 2001/01/12 23:30:00 JST
2559 :     Tag: blume-20010112-bootfiles
2560 :     Description:
2561 :    
2562 :     Made a new set of bootfiles that goes with the current state of the
2563 :     repository.
2564 :    
2565 :     ----------------------------------------------------------------------
2566 :     Name: Matthias Blume
2567 : blume 777 Date: 2001/01/12 21:20:00 JST
2568 :     Tag: blume-20010112-sync
2569 :     Description:
2570 :    
2571 :     I am just flushing out some minor changes that had accumulated in
2572 :     my private branch in order to sync with the main tree. (This is
2573 :     mainly because I had CVS trouble when trying to merge _into_ my
2574 :     private branch.)
2575 :    
2576 :     Most people should be completely unaffected by this.
2577 :    
2578 :     ----------------------------------------------------------------------
2579 : leunga 775 Name: Allen Leung
2580 :     Date: Thu Jan 11 21:03:00 EST 2001
2581 :     Tag: leunga-20010111-labexp=mltree
2582 :     Description:
2583 :    
2584 :     1. Removed the type LabelExp and replace it by MLTree.
2585 :     2. Rewritten mltree-simplify with the pattern matcher tool.
2586 :     3. There were some bugs in alpha code generator which would break
2587 :     64-bit code generation.
2588 :     4. Redo the tools to generate code with the
2589 :     5. The CM files in MLRISC (and in src/system/smlnj/MLRISC)
2590 :     are now generated by perl scripts.
2591 :    
2592 :     ----------------------------------------------------------------------
2593 : blume 771 Name: Matthias Blume
2594 : blume 774 Date: 2001/01/10 21:55:00 JST
2595 :     Tag: blume-20010110-rcc
2596 :     Description:
2597 :    
2598 :     The RCC stuff now seems to work (but only on the x86).
2599 :     This required hacking of the c-calls interface (and -implementation) in
2600 :     MLRISC.
2601 :    
2602 :     Normal compiler users should be unaffected.
2603 :    
2604 :     ----------------------------------------------------------------------
2605 :     Name: Matthias Blume
2606 : blume 773 Date: 2001/01/09 01:20:00 JST
2607 :     Tag: blume-20010109-rcc
2608 :     Description:
2609 :    
2610 :     This is a fairly big patch, flushing out a large number of pending
2611 :     changes that I made to my development copy over the last couple of days.
2612 :    
2613 :     Of practical relevance at this moment is a workaround for a pickling
2614 :     bug that Allen ran into the other day. The cause of the bug itself is
2615 :     still unknown and it might be hard to fix it properly, but the
2616 :     workaround has some merits of its own (namely somewhat reducing pickling
2617 :     overhead for certain libraries). Therefore, I think this solution should
2618 :     be satisfactory at this time.
2619 :    
2620 :     The rest of the changes (i.e., the vast majority) has to do with my
2621 :     ongoing efforts of providing direct support for C function calls from
2622 :     ML. At the moment there is a new primop "RAW_CCALL", typing magic
2623 :     in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
2624 :     case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
2625 :     translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
2626 :     handle RCC.
2627 :    
2628 :     The last part (the changes to mlriscGen.sml) are still known to be
2629 :     wrong on the x86 and not implemented on all other architectures. But
2630 :     the infrastructure is in place. I had to change a few functor
2631 :     signatures in the backend to be able to route the CCalls interface
2632 :     from MLRISC there, and I had to specialize the mltree type (on the
2633 :     x86) to include the necessary extensions. (The extensions themselves
2634 :     were already there and redy to go in MLRISC/x86).
2635 :    
2636 :     Everything should be very happy as soon as someone helps me with
2637 :     mlriscGen.sml...
2638 :    
2639 :     In any case, nothing of this should matter to anyone as long as the
2640 :     new primop is not being used (which is going to be the case unless you
2641 :     find it where I hid it :). The rest of the compiler is completely
2642 :     unaffected.
2643 :    
2644 :     ----------------------------------------------------------------------
2645 :     Name: Matthias Blume
2646 : blume 772 Date: 2001/01/05 00:30:00 JST
2647 :     Tag: blume-20010105-primops
2648 :     Description:
2649 :    
2650 :     Added some experimental support for work that I am doing right now.
2651 :     These changes mostly concern added primops, but there is also a new
2652 :     experimental C library in the runtime system (but currently not enabled
2653 :     anywhere except on Linux/X86).
2654 :    
2655 :     In the course of adding primops (and playing with them), I discovered that
2656 :     Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
2657 :     broken. (Zhong was very right he labeled this stuff as "major gross hack".)
2658 :     To recover, I made type information in INL_PRIM mandatory and changed
2659 :     prim.sml as well as built-in.sml accordingly. The InLine structure now
2660 :     has complete, correct type information (i.e., no bottom types).
2661 :    
2662 :     Since all these changes mean that we need new binfiles, I also bumped the
2663 :     version number to 110.32.1.
2664 :    
2665 :     ----------------------------------------------------------------------
2666 :     Name: Matthias Blume
2667 : blume 771 Date: 2000/12/30 22:10:00 JST
2668 :     Tag: blume-20001230-various
2669 :     Description:
2670 :    
2671 :     Added proxy libraries for MLRISC and let MLRISC libraries refer
2672 :     to each other using path anchors. (See CM manual for explanation.)
2673 :    
2674 :     Updated CM documentation.
2675 :    
2676 :     Fixed some bugs in CM.
2677 :    
2678 :     Implemented "proxy" libraries (= syntactic sugar for CM).
2679 :    
2680 :     Added "-quiet" option to makeml and changed runtime system accordingly.
2681 :    
2682 :     Added cleanup handler for exportML to reset timers and compiler stats.
2683 :    
2684 :     ----------------------------------------------------------------------
2685 : george 761 Name: Lal George
2686 :     Date: 2000/12/22 22:22:58 EST 2000
2687 :     Tag: Release_110_32
2688 :     Description:
2689 :    
2690 :     Infinite precision used throughout MLRISC.
2691 :     see MLRISC/mltree/machine-int.sig
2692 :    
2693 :     ----------------------------------------------------------------------
2694 : blume 756 Name: Matthias Blume
2695 : blume 760 Date: 2000/12/22 23:16:00 JST
2696 :     Tag: blume-20001222-warn
2697 :     Description:
2698 : blume 759
2699 : blume 760 Corrected wording and formatting of some CM warning message which I
2700 :     broke in my previous patch.
2701 : blume 759
2702 :     ----------------------------------------------------------------------
2703 :     Name: Matthias Blume
2704 : blume 763 Date: 2000/12/22 21:20:00 JST
2705 :     Tag: blume-20001222-anchorenv
2706 :     Description:
2707 :    
2708 :     Fixed CM's handling of anchor environments in connection with CMB.make.
2709 :    
2710 :     ----------------------------------------------------------------------
2711 :     Name: Matthias Blume
2712 : blume 758 Date: 2000/12/22 13:15:00 JST
2713 :     Tag: blume-20001222-cleanup
2714 :     Description:
2715 :    
2716 :     Removed src/cm/ffi which does not (and did not) belong here.
2717 :    
2718 :     ----------------------------------------------------------------------
2719 :     Name: Matthias Blume
2720 : blume 757 Date: 2000/12/21 23:55:00 JST
2721 :     Tag: blume-20001221-exn
2722 :     Description:
2723 :    
2724 :     Probably most important: CM no longer silently swallows all exceptions
2725 :     in the compiler.
2726 :     Plus: some other minor CM changes. For example, CM now reports some
2727 :     sizes for generated binfiles (code, data, envpickle, lambdapickle).
2728 :    
2729 :     ----------------------------------------------------------------------
2730 :     Name: Matthias Blume
2731 : blume 756 Date: 2000/12/15 00:01:05 JST
2732 :     Tag: blume-20001215-dirtool
2733 :     Description:
2734 :    
2735 :     - "dir" tool added.
2736 :     - improvements and cleanup to Tools structure
2737 :     - documentation updates
2738 :    
2739 :     ----------------------------------------------------------------------
2740 : leunga 752 Name: Allen Leung
2741 : leunga 755 Date: Thu Dec 14 03:45:24 EST 2000
2742 :     Description:
2743 :     Tag: leunga-20001214-int-inf
2744 :     Description:
2745 :    
2746 :     In IntInf, added these standard functions, which are missing from our
2747 :     implementation:
2748 :    
2749 :     andb : int * int -> int
2750 :     xorb : int * int -> int
2751 :     orb : int * int -> int
2752 :     notb : int -> int
2753 :     << : int * word -> int
2754 :     ~>> : int * word -> int
2755 :    
2756 :     Not tested, I hope they are correct.
2757 :    
2758 :     ----------------------------------------------------------------------
2759 :     Name: Allen Leung
2760 : leunga 752 Date: Fri Dec 8 19:23:26 EST 2000
2761 :     Description:
2762 :     Tag: leunga-20001208-nowhere
2763 :     Description:
2764 :    
2765 :     Slight improvements to the 'nowhere' tool to handle OR-patterns,
2766 :     to generate better error messages etc. Plus a brief manual.
2767 :    
2768 :     ----------------------------------------------------------------------
2769 : george 749 Name: Lal George
2770 :     Date: 2000/12/08 09:54:02 EST 2000
2771 :     Tag: Release_110_31
2772 :     Description:
2773 : george 705
2774 : george 749 - Version 110.31
2775 : leunga 591 ----------------------------------------------------------------------
2776 : leunga 744 Name: Allen Leung
2777 :     Date: Thu Dec 7 22:01:04 EST 2000
2778 :     Tag: leunga-20001207-cell-monster-hack
2779 :     Description:
2780 :    
2781 :     Major MLRISC internal changes. Affect all clients.
2782 :     Summary:
2783 :    
2784 :     1. Type CELLS.cell = int is now replaced by a datatype.
2785 :     As a result, the old regmap is now gone. Almost all interfaces
2786 :     in MLRISC change as a consequence.
2787 :    
2788 :     2. A new brand version of machine description tool (v3.0) that generates
2789 :     modules expecting the new interface. The old version is removed.
2790 :    
2791 :     3. The RA interface has been further abstracted into two new functors.
2792 :     RISC_RA and X86RA. These functors have much simpler interfaces.
2793 :     [See also directory MLRISC/demo.]
2794 :    
2795 :     4. Some other new source->source code generation tools are available:
2796 :    
2797 :     a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
2798 :     b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
2799 :     I use this tool to generate the peephole optimizers---with the new
2800 :     cell type changes, peephole rules are becoming difficult to write
2801 :     without conditional pattern matching.
2802 :    
2803 :     5. More Intmap -> IntHashTable change. Previous changes by Matthias didn't
2804 :     cover the entire MLRISC source tree so many things broke.
2805 :    
2806 :     6. CM files have been moved to the subdirectory MLRISC/cm.
2807 :     They are moved because there are a lot of them and they clutter up the
2808 :     root dir.
2809 :    
2810 :     7. More detailed documentation to come...
2811 :    
2812 :     NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
2813 :     a makeml -rebuild first. This is because of other other
2814 :     changes that Matthias has made (see below).
2815 :    
2816 :    
2817 :     ----------------------------------------------------------------------
2818 : blume 733 Name: Matthias Blume
2819 : blume 742 Date: 2000/11/30 23:12:00 JST
2820 :     Tag: blume-20001130-filereorg
2821 :     Description:
2822 :    
2823 :     Some manual updates and some file reorganizations in CM.
2824 :    
2825 :     ----------------------------------------------------------------------
2826 :     Name: Matthias Blume
2827 : blume 737 Date: 2000/11/24 17:45:00 JST
2828 :     Tag: blume-20001124-link
2829 :     Description:
2830 :    
2831 :     Drastically improved link traversal code for the case that the dynamic
2832 :     value was already loaded at bootstrap time. As a result, CM and CMB
2833 :     now both load blazingly fast -- even on a very slow machine. Also,
2834 :     memory consumption has been further reduced by this.
2835 :    
2836 :     Warning: The format of the PIDMAP file has changed. THerefore, to
2837 :     bootstrap you have to do this:
2838 :    
2839 :     1. Run CMB.make
2840 :     2. Make a symbolic link for the boot directory:
2841 :     ln -s sml.boot.ARCH-OS xxx
2842 :     3. "Rebuild" the boot directory:
2843 :     ./makeml -boot xxx -rebuild sml ; rm xxx
2844 :     4. Boot normally:
2845 :     ./makeml
2846 :    
2847 :     ----------------------------------------------------------------------
2848 :     Name: Matthias Blume
2849 : blume 735 Date: 2000/11/21 21:20:00 JST
2850 :     Tag: blume-20001121-tools
2851 :     Description:
2852 :    
2853 :     Continued hacking on autoloading problem -- with success this time.
2854 :     Also changed tool-plugin mechanism. See new CM manual.
2855 :    
2856 :     ----------------------------------------------------------------------
2857 :     Name: Matthias Blume
2858 : blume 734 Date: 2000/11/19 14:30:00 JST
2859 :     Tag: blume-20001119-autoload
2860 :     Description:
2861 :    
2862 :     Some hacking to make autoloading faster. Success for CMB, no success
2863 :     so far for CM. There is a reduced structure CM' that autoloads faster.
2864 :     (This is a temporary, non-documented hack to be eliminated again when
2865 :     the general problem is solved.)
2866 :    
2867 :     ----------------------------------------------------------------------
2868 :     Name: Matthias Blume
2869 : blume 733 Date: 2000/11/17 14:10:00 JST
2870 :     Tag: blume-20001117-pickle-lib
2871 :     Description:
2872 :    
2873 :     1. Eliminated comp-lib.cm
2874 :     2. Made pickle-lib.cm
2875 :     3. Eliminated all uses of intset.sml (from comp-lib.cm)
2876 :     4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
2877 :     equivalent constructs from smlnj-lib.cm (INtHashTable).
2878 :     5. Point 4. also goes for those uses of intmap.* in MLRISC.
2879 :     Duplicated intmap modules thrown out.
2880 :     6. Hunted down all duplicated SCC code and replaced it with
2881 :     equivalent stuff (GraphSCCFn from smlnj-lib.cm).
2882 :     7. Rewrote Feedback module.
2883 :     8. Moved sortedlist.sml into viscomp-lib.cm. Eventually it
2884 :     should be thrown out and equivalent modules from smlnj-lib.cm
2885 :     should be used (IntRedBlackSet, IntListSet, ...).
2886 :    
2887 :     Confirmed that compiler compiles to fixpoint.
2888 :    
2889 :     ----------------------------------------------------------------------
2890 : leunga 731 Name: Allen Leung
2891 :     Date: 2000/11/10 18:00:00
2892 :     Tag: leunga-20001110-new-x86-fp
2893 :    
2894 :     A new x86 floating point code generator has been added.
2895 :     By default this is turned off. To turn this on, do:
2896 :    
2897 :     CM.autoload "$smlnj/compiler.cm";
2898 :     Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
2899 :    
2900 :     Changes:
2901 :    
2902 :     1. Changed FTAN to FPTAN so that the assembly output is correct.
2903 :     2. Changed the extension callback for FTANGENT to generate:
2904 :    
2905 :     fptan
2906 :     fstp %st(0)
2907 :     instead of
2908 :     fptan
2909 :     fstpl ftempmem
2910 :    
2911 :     3. Numerous assembly fixes for x86.
2912 :    
2913 :     5. Cleaned up the machine code output module x86/x86MC.sml and added
2914 :     support for a whole bunch of instructions and addressing modes:
2915 :    
2916 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st, %st(n)
2917 :     faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp %st, %st(n)
2918 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st(n), %st
2919 :     fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
2920 :     fxch %st(n)
2921 :     fld %st(n)
2922 :     fst %st(n)
2923 :     fst mem
2924 :     fstp %st(n)
2925 :     fucom %st(n)
2926 :     fucomp %st(n)
2927 :    
2928 :     All these are now generated when the fast fp mode is turned on.
2929 :    
2930 :     6. Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
2931 :    
2932 :     ----------------------------------------------------------------------
2933 : blume 729 Name: Matthias Blume
2934 :     Date: 2000/11/09 11:20:00 JST
2935 :     Tag: blume-20001109-scc
2936 :     Description:
2937 :    
2938 :     Eliminated some code duplication:
2939 :    
2940 :     1. Added "where" clause to GraphSCCFn in SML/NJ Library.
2941 :     (Otherwise the functor is useless.)
2942 :     2. Used GraphSCCFn where SCCUtilFun was used previously.
2943 :     3. Got rid of SCCUtilFun (in comp-lib.cm).
2944 :    
2945 :     ----------------------------------------------------------------------
2946 : george 721 Name: Lal George
2947 :     Date: 2000/11/06 09:02:21 EST 2000
2948 :     Tag: Release_110_30
2949 :     Description:
2950 :    
2951 :     - Version 110.30
2952 :     ----------------------------------------------------------------------
2953 : blume 715 Name: Matthias Blume
2954 : blume 716 Date: 2000/11/04 14:45:00
2955 :     Tag: blume-20001104-mlbuild
2956 :     Description:
2957 :    
2958 :     - Made ml-build faster on startup.
2959 :     - Documentation fixes.
2960 :    
2961 :     ----------------------------------------------------------------------
2962 :     Name: Matthias Blume
2963 : blume 715 Date: 2000/11/02 17:00:00 JST
2964 :     Tag: blume-20001102-condcomp
2965 :     Description:
2966 :    
2967 :     - Small tweaks to pickler -- new BOOTFILES!
2968 :     - Version bumped to 110.29.2.
2969 :     - Added conditional compilation facility to init.cmi (see comment there).
2970 :     ----------------------------------------------------------------------
2971 : leunga 713 Name: Allen Leung
2972 :     Date: 2000/10/23 19:31:00
2973 :     Tag: leunga-20001023-demo-ra
2974 :    
2975 :     1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
2976 :     2. Test programs for the graph library updated
2977 :     3. Some new MLRISC demo programs added
2978 :    
2979 :     ----------------------------------------------------------------------
2980 : blume 710 Name: Matthias Blume
2981 : blume 711 Date: 2000/08/31 22:15:00 JST
2982 :     Tag: blume-20001017-errmsg
2983 :     Description:
2984 :    
2985 :     More error message grief: Where there used to be no messages, there
2986 :     now were some that had bogus error regions. Fixed.
2987 :    
2988 :     ----------------------------------------------------------------------
2989 :     Name: Matthias Blume
2990 : blume 710 Date: 2000/08/31 17:30:00 JST
2991 :     Tag: blume-20001017-v110p29p1
2992 :     Description:
2993 :    
2994 :     I made a version 110.29.1 with new bootfiles.
2995 :    
2996 :     Changes: Modified pickler/unpickler for faster and leaner unpickling.
2997 :     CM documentation changes and a small bugfix in CM's error reporting.
2998 :    
2999 :     ----------------------------------------------------------------------
3000 : george 705 Name: Lal George
3001 :     Date: 2000/09/27 14:42:35 EDT
3002 :     Tag: george-20000927-nodestatus
3003 :     Description:
3004 :    
3005 :     Changed the type of the nodestatus, so that:
3006 :    
3007 :     SPILLED(~1) is now SPILLED
3008 :     SPILLED(m) where m>=0 is now MEMREG(m)
3009 :     SPILLED(s) where s<~1 is now SPILL_LOC(~s)
3010 :    
3011 :     ----------------------------------------------------------------------
3012 : blume 697 Name: Matthias Blume
3013 : blume 703 Date: 2000/09/07 14:45:00 JST
3014 :     Tag: blume-20000907-cmerrmsg
3015 :     Description:
3016 :    
3017 :     Small tweak to CM to avoid getting ML syntax error messages twice.
3018 :    
3019 :     ----------------------------------------------------------------------
3020 :     Name: Matthias Blume
3021 : blume 701 Date: 2000/08/31 18:00:00 JST
3022 :     Tag: blume-20000831-cvsbootfiles
3023 :     Description:
3024 :    
3025 :     New URL for boot files (because the 110.29 files on the BL server do
3026 :     now work correctly with my updated install scripts for yacc and lex).
3027 :    
3028 :     ----------------------------------------------------------------------
3029 :     Name: Matthias Blume
3030 : blume 697 Date: 2000/08/08 12:33:00 JST
3031 :     Tag: blume-20000808-manual
3032 :     Description:
3033 :    
3034 :     Tiny update to CM manual.
3035 :    
3036 :     ----------------------------------------------------------------------
3037 : leunga 695 Name: Allen Leung
3038 :     Date: 2000/08/7 19:31:00
3039 :     Tag: leunga-20000807-a-whole-bunch-of-stuff
3040 :    
3041 :     Moby, C--, SSA, x86, machine descriptions etc. Should only affect C--
3042 :     and Mobdy.
3043 :    
3044 :     1. x86
3045 :    
3046 :     a. Fixes to peephole module by John and Dan.
3047 :     b. Assembly fix to SETcc by Allen.
3048 :     c. Fix to c-call by John.
3049 :     d. Fix to spilling by John. (This one deals with the missing FSTPT case)
3050 :     e. Instruction selection optimization to SETcc as suggested by John.
3051 :    
3052 :     For example,
3053 :    
3054 :     MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
3055 :    
3056 :     should generate:
3057 :    
3058 :     MOVL a, x
3059 :     SUBL b, x
3060 :     SHRL 31, x
3061 :    
3062 :     2. IR stuff
3063 :    
3064 :     A bunch of new DJ-graph related algorithms added. These
3065 :     speed up SSA construction.
3066 :    
3067 :     3. SSA + Scheduling
3068 :    
3069 :     Added code for SSA and scheduling to the repository
3070 :    
3071 :     ----------------------------------------------------------------------
3072 : george 694 Name: Lal George
3073 :     Date: 2000/07/27 11:53:14 EDT
3074 :    
3075 :     Tag: lal-20000727-linux-ppc
3076 :     Description:
3077 :    
3078 :     Made changes to support Linux PPC.
3079 :     p.s. I have confirmation that the 110.29 boot files work fine.
3080 :    
3081 :     ----------------------------------------------------------------------
3082 : blume 684 Name: Matthias Blume
3083 : blume 692 Date: 2000/07/27 17:40:00 JST
3084 :     Tag: blume-20000727-scripts
3085 :     Description:
3086 :    
3087 :     !!!! WARNING !!!!
3088 :     You must recompile the runtime system!
3089 :     !!!! WARNING !!!!
3090 :    
3091 :     This is basically another round of script-enhancements:
3092 :    
3093 :     1. sml, ml-build, and ml-makedepend accept options -D and -U to define
3094 :     and undefine CM preprocessor symbols.
3095 :    
3096 :     2. ml-build avoids generating a new heap image if it finds that the
3097 :     existing one is still ok. (The condition is that no ML file had to
3098 :     be recompiled and all ML files are found to be older that the heap
3099 :     file.)
3100 :    
3101 :     To make this work smoothly, I also hacked the runtime system as
3102 :     well as SMLofNJ.SysInfo to get access to the heap image suffix
3103 :     (.sparc-solaris, ...) that is currently being used.
3104 :    
3105 :     Moreover, the signature of CM.mk_standalone has changed. See the
3106 :     CM manual.
3107 :    
3108 :     3. ml-makedepend accepts additional options -n, -a, and -o. (See the
3109 :     CM manual for details.)
3110 :    
3111 :     4. More CM manual updates:
3112 :     - all of the above has been documented.
3113 :     - there is now a section describing the (CM-related) command line
3114 :     arguments that are accepted by the "sml" command
3115 :    
3116 :     ----------------------------------------------------------------------
3117 :     Name: Matthias Blume
3118 : blume 691 Date: 2000/07/25 16:20:00 JST
3119 :     Tag: blume-20000725-makedepend
3120 :     Description:
3121 :    
3122 :     Added a script called ml-makedepend. This can be used in makefiles
3123 :     for Unix' make in a way very similar to the "makedepend" command for
3124 :     C.
3125 :    
3126 :     The script internally uses function CM.sources.
3127 :    
3128 :     Synopsis:
3129 :    
3130 :     ml-makedepend [-f makefile] cmfile targetname
3131 :    
3132 :     The default for the makefile is "makefile" (or "Makefile" should
3133 :     "makefile" not exist).
3134 :    
3135 :     ml-makedepend adds a cmfile/targetname-specific section to this
3136 :     makefile (after removing the previous version of this section). The
3137 :     section contains a single dependency specification with targetname on
3138 :     the LHS (targetname is an arbitrary name), and a list of files derived
3139 :     from the cmfile on the RHS. Some of the files on the RHS are
3140 :     ARCH/OPSYS-specific. Therefore, ml-makedepend inserts references to
3141 :     "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
3142 :     path names. The makefile writer is responsible for making sure that
3143 :     these variables have correct at the time "make" is invoked.
3144 :    
3145 :     ----------------------------------------------------------------------
3146 :     Name: Matthias Blume
3147 : blume 690 Date: 2000/07/22 23:30:00 JST
3148 :     Tag: blume-20000722-urlupdate
3149 :     Description:
3150 :    
3151 :     Changed BOOT and config/srcarchiveurl to point to BL server:
3152 :    
3153 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
3154 :    
3155 :     ----------------------------------------------------------------------
3156 :     Name: Matthias Blume
3157 : blume 686 Date: 2000/07/18 18:00:00 JST
3158 :     Tag: blume-20000718-Version_110_29
3159 :     Description:
3160 :    
3161 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
3162 :    
3163 :     2. Updated config/version to 110.29
3164 :    
3165 :     3. Updated config/srcarchiveurl
3166 :    
3167 :     3. New boot files!
3168 :     ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
3169 :    
3170 :     ----------------------------------------------------------------------
3171 :     Name: Matthias Blume
3172 : blume 684 Date: 2000/07/11 13:58:00 JST
3173 :     Tag: blume-20000711-doctypo
3174 :     Description:
3175 :    
3176 :     Fixed a few typos in CM manual.
3177 :    
3178 :     ----------------------------------------------------------------------
3179 : leunga 683 Name: Allen Leung
3180 :     Date: 2000/06/15 00:38:00
3181 :     Tag: leunga-20000704-sparc-x86
3182 :    
3183 :     1. x86 peephole improvement sp += k; sp -= k => nop [from John]
3184 :     2. fix to x86 RET bug [found by Dan Grossman]
3185 :     3. sparc assembly bug fix for ticc instructions [found by Fermin]
3186 :    
3187 :     Affects c-- and moby only
3188 :    
3189 :     ----------------------------------------------------------------------
3190 : blume 680 Name: Matthias Blume
3191 : blume 682 Date: 2000/07/04 15:26:00
3192 :     Tag: blume-20000704-trigger
3193 :     Description:
3194 :    
3195 :     1. Improvements to CM manual.
3196 :     2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
3197 :     of getting a back-trace. The function, when called, raises an
3198 :     internal exception which explicitly carries the full back-trace history,
3199 :     so it is unaffected by any intervening handle-raise pairs ("trivial"
3200 :     or not). The interactive loop will print that history once it arrives
3201 :     at top level.
3202 :     Short of having all exceptions implicitly carry the full history, the
3203 :     recommended way of using this facility is:
3204 :     - compile your program with instrumentation "on"
3205 :     - run it, when it raises an exception, look at the history
3206 :     - if the history is "cut off" because of some handler, go and modify
3207 :     your program so that it explicitly calls BTrace.trigger
3208 :     - recompile (still instrumented), and rerun; look at the full history
3209 :    
3210 :     ----------------------------------------------------------------------
3211 :     Name: Matthias Blume
3212 :     Date: 2000/07/03 15:36:00 JST
3213 : blume 680 Tag: blume-20000702-manual
3214 :     Description:
3215 :    
3216 :     Small corrections and updates to CM manual.
3217 :    
3218 :     ----------------------------------------------------------------------
3219 : blume 682 Name: Matthias Blume
3220 : blume 679 Date: 2000/06/29 16:04:00 JST
3221 :     Tag: blume-20000629-yacctool
3222 :     Description:
3223 :    
3224 :     Changes:
3225 :    
3226 :     1. Class "mlyacc" now takes separate arguments to pass options to
3227 :     generated .sml- and .sig-files independently.
3228 :     2. Corresponding CM manual updates.
3229 :     3. BTrace module now also reports call sites. (However, for loop clusters
3230 :     it only shows from where the cluster was entered.) There are associated
3231 :     modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
3232 :    
3233 :     ----------------------------------------------------------------------
3234 : blume 675 Name: Matthias Blume
3235 : blume 678 Date: 2000/06/27 16:51:00 JST
3236 :     Tag: blume-20000627-noweb
3237 :     Description:
3238 :    
3239 :     Changes:
3240 :    
3241 :     1. Implemented "subdir" and "witness" options for noweb tool.
3242 :     This caused some slight internal changes in CM's tool implementation.
3243 :     2. Fixed bug in "tool plugin" mechanism. This is essentially cleaning
3244 :     some remaining issues from earlier path anchor changes.
3245 :     3. Updated CM manual accordingly.
3246 :    
3247 :     4. Changed implementation of back-tracing so that I now consider it
3248 :     ready for prime-time.
3249 :    
3250 :     In particular, you don't have to explicitly trigger the back-trace
3251 :     anymore. Instead, if you are running BTrace-instrumented code and
3252 :     there is an uncaught exception (regardless of whether or not it was
3253 :     raised in instrumented code), the top-level evalloop will print
3254 :     the back-trace.
3255 :    
3256 :     Features:
3257 :    
3258 :     - Instrumented and uninstrumented code work together seemlessly.
3259 :     (Of course, uninstrumented code is never mentioned in actual
3260 :     back-traces.)
3261 :    
3262 :     - Asymptotic time- and space-complexity of instrumented code is
3263 :     equal to that of uninstrumented code. (This means that
3264 :     tail-recursion is preserved by the instrumentation phase.)
3265 :    
3266 :     - Modules whose code has been instrumented in different sessions
3267 :     work together without problem.
3268 :    
3269 :     - There is no penalty whatsoever on uninstrumented code.
3270 :    
3271 :     - There is no penalty on "raise" expressions, even in
3272 :     instrumented code.
3273 :    
3274 :     A potential bug (or perhaps it is a feature, too):
3275 :    
3276 :     A back-trace reaches no further than the outermost instrumented
3277 :     non-trivial "raise". Here, a "trivial" raise is one that is the
3278 :     sole RHS of a "handle" rule. Thus, back-traces reach trough
3279 :    
3280 :     <exp> handle e => raise e
3281 :    
3282 :     and even
3283 :    
3284 :     <exp> handle Foo => raise Bar
3285 :    
3286 :     and, of course, through
3287 :    
3288 :     <exp> handle Foo => ...
3289 :    
3290 :     if the exception was not Foo.
3291 :    
3292 :     Back-traces always reach right through any un-instrumented code
3293 :     including any of its "handle" expressions, trivial or not.
3294 :    
3295 :     To try this out, do the following:
3296 :    
3297 :     - Erase all existing binfiles for your program.
3298 :     (You may keep binfiles for those modules where you think you
3299 :     definitely don't need back-tracing.)
3300 :     - Turn on back-trace instrumentation:
3301 :     SMLofNJ.Internals.BTrace.mode (SOME true);
3302 :     - Recompile your program. (I.e., run "CM.make" or "use".)
3303 :     - You may now turn instrumentation off again (if you want):
3304 :     SMLofNJ.Internals.BTrace.mode (SOME false);
3305 :     - Run your program as usual. If it raises an exception that
3306 :     reaches the interactive toplevel, then a back-trace will
3307 :     automatically be printed. After that, the toplevel loop
3308 :     will print the exception history as usual.
3309 :    
3310 :     ----------------------------------------------------------------------
3311 :     Name: Matthias Blume
3312 : blume 677 Date: 2000/06/26 09:56:46 JST
3313 :     Tag: blume-20000626-setup
3314 :     Description:
3315 :    
3316 :     CM: - setup-parameter to "sml" added; this can be used to run arbitrary
3317 :     ML code before and after compiling a file (e.g., to set compiler
3318 :     flags)
3319 :    
3320 :     Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
3321 :     - associated changes to btrace.sml (BTrace instrumentation pass)
3322 :     - cleaner implementation of btimp.sml (BTrace tracing and report
3323 :     module)
3324 :    
3325 :     CM manual: * new path encoding documented
3326 :     * description of setup-parameter to "sml" added
3327 :    
3328 :     The biggest user-visible change to back-tracing is that it is no
3329 :     longer necessary to compile all traced modules within the same
3330 :     session. (This was a real limitation.)
3331 :    
3332 :     ----------------------------------------------------------------------
3333 :     Name: Matthias Blume
3334 : blume 676 Date: 2000/06/24 12:40:00 JST
3335 :     Tag: blume-20000624-startup
3336 :     Description:
3337 :    
3338 :     Fixes startup slowdown problem. (I was calling SrcPath.sync a _tad_
3339 :     bit too often -- to put it mildly. :)
3340 :    
3341 :     ----------------------------------------------------------------------
3342 :     Name: Matthias Blume
3343 : blume 675 Date: 2000/06/23 18:20:00 JST
3344 :     Tag: blume-20000623-btrace
3345 :     Description:
3346 :    
3347 :     This updates adds a backtrace facility to aid programmers in debugging
3348 :     their programs. This involves the following changes:
3349 :    
3350 :     1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
3351 :     keeping track of the current call stack. When programs are compiled
3352 :     in a special mode, the compiler will insert calls to these hooks
3353 :     into the user program.
3354 :     "Hook" means that it is possible for different implementations of
3355 :     back-tracing to register themselves (at different times).
3356 :    
3357 :     2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
3358 :     as an Absyn.dec->Absyn.dec rewrite. Normally this phase is turned off.
3359 :     It can be turned on using this call:
3360 :     SMLofNJ.Internals.BTrace.mode (SOME true);
3361 :     Turning it off again:
3362 :     SMLofNJ.Internals.BTrace.mode (SOME false);
3363 :     Querying the current status:
3364 :     SMLofNJ.Internals.BTrace.mode NONE;
3365 :     Annotated programs are about twice as big as normal ones, and they
3366 :     run a factor of 2 to 4 slower with a dummy back-trace plugin (one
3367 :     where all hooks do nothing). The slowdown with a plugin that is
3368 :     actually useful (such as the one supplied by default) is even greater,
3369 :     but in the case of the default plugin it is still only an constant
3370 :     factor (amortized).
3371 :    
3372 :     3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
3373 :     with a sub-structure BTrace for controlling back-tracing. In particular,
3374 :     the above-mentioned function "mode" controls whether the annotation
3375 :     phase is invoked by the compiler. Another important function is
3376 :     "trigger": when called it aborts the current execution and causes
3377 :     the top-level loop to print a full back-trace.
3378 :    
3379 :     4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
3380 :     for back-tracing. It keeps track of the dynamic call stack and in
3381 :     addition to that it keeps a partial history at each "level" of that
3382 :     stack. For example, if a tail-calls b, b tail-calls c, and c tail-calls
3383 :     d and b (at separate times, dynamically), then the report will show:
3384 :    
3385 :     GOTO d
3386 :     /c
3387 :     GOTO \b
3388 :     CALL a
3389 :    
3390 :     This shows that there was an initial non-tail call of a, then a
3391 :     tail-call to b or c, looping behavior in a cluster of functions that
3392 :     consist of b and c, and then a goto from that cluster (i.e., either from
3393 :     b or from c) to d.
3394 :    
3395 :     Note that (depending on the user program) the amount of information
3396 :     that the back-trace module has to keep track of at each level is bounded
3397 :     by a constant. Thus, the whole implementation has the same asymptotical
3398 :     complexity as the original program (both in space and in time).
3399 :    
3400 :     5. compiler/TopLevel/interact/evalloop.sml has been modified to
3401 :     handle the special exception SMLofNJ.Internals.BTrace.BTrace
3402 :     which is raised by the "trigger" function mentioned above.
3403 :    
3404 :     Notes on usage:
3405 :    
3406 :     - Annotated code works well together with unannotated code:
3407 :     Unannotated calls simply do not show up at all in the backtrace.
3408 :    
3409 :     - It is not a good idea to let modules that were annotated during
3410 :     different sessions run at the same time. This is because the compiler
3411 :     chooses small integers to identify individual functions, and there
3412 :     will be clashes if different modules were compiled in separate sessions.
3413 :     (Nothing will crash, and you will even be told about the clashes, but
3414 :     back-trace information will in general not be useful.)
3415 :    
3416 :     - Back-tracing can be confused by callcc and capture.
3417 :    
3418 :     - The only way of getting a back-trace right now is to explicitly
3419 :     invoke the "trigger" function from your user program. Eventually, we
3420 :     should make every exception carry back-trace information (if
3421 :     available). But since this creates more overhead at "raise"-time
3422 :     (similar to the current exnHistory overhead), I have not yet
3423 :     implemented this. (The implementation will be rather easy.) With
3424 :     exceptions carrying back-trace information, this facility will be even
3425 :     more useful because users don't need to modify their programs...
3426 :    
3427 :     - While it is possible to compile the compiler with back-trace
3428 :     annotations turned on (I did it to get some confidence in
3429 :     correctness), you must make absolutely sure that core.sml and
3430 :     btimp.sml are compiled WITHOUT annotation! (core.sml cannot actually
3431 :     be compiled with annotation because there is no core access yet, but
3432 :     if you compile btimp.sml with annotation, then the system will go into
3433 :     an infinite recursion and crash.)
3434 :     Since CM currently does not know about BTrace, the only way to turn
3435 :     annotations on and off for different modules of the compiler is to
3436 :     interrupt CMB.make, change the settings, and re-invoke it. Of course,
3437 :     this is awkward and clumsy.
3438 :    
3439 :     Sample sessions:
3440 :    
3441 :     Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
3442 :     - SMLofNJ.Internals.BTrace.mode (SOME true);
3443 :     [autoloading]
3444 :     [autoloading done]
3445 :     val it = false : bool
3446 :     - structure X = struct
3447 :     - fun main n = let
3448 :     - fun a (x, 0) = d x
3449 :     - | a (x, n) = b (x, n - 1)
3450 :     - and b (x, n) = c (x, n)
3451 :     - and c (x, n) = a (x, n)
3452 :     - and d x = e (x, 3)
3453 :     - and e (x, 0) = f x
3454 :     - | e (x, n) = e (x, n - 1)
3455 :     - and f 0 = SMLofNJ.Internals.BTrace.trigger ()
3456 :     - | f n = n * g (n - 1)
3457 :     - and g n = a (n, 3)
3458 :     - in
3459 :     - f n
3460 :     - end
3461 :     - end;
3462 :     structure X : sig val main : int -> int end
3463 :     - X.main 3;
3464 :     *** BACK-TRACE ***
3465 :     GOTO stdIn:4.2-13.20: X.main[2].f
3466 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
3467 :     GOTO stdIn:4.2-13.20: X.main[2].d
3468 :     / stdIn:4.2-13.20: X.main[2].a
3469 :     | stdIn:4.2-13.20: X.main[2].b
3470 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
3471 :     CALL stdIn:4.2-13.20: X.main[2].g
3472 :     GOTO stdIn:4.2-13.20: X.main[2].f
3473 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
3474 :     GOTO stdIn:4.2-13.20: X.main[2].d
3475 :     / stdIn:4.2-13.20: X.main[2].a
3476 :     | stdIn:4.2-13.20: X.main[2].b
3477 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
3478 :     CALL stdIn:4.2-13.20: X.main[2].g
3479 :     GOTO stdIn:4.2-13.20: X.main[2].f
3480 :     GOTO-( stdIn:4.2-13.20: X.main[2].e
3481 :     GOTO stdIn:4.2-13.20: X.main[2].d
3482 :     / stdIn:4.2-13.20: X.main[2].a
3483 :     | stdIn:4.2-13.20: X.main[2].b
3484 :     GOTO-\ stdIn:4.2-13.20: X.main[2].c
3485 :     CALL stdIn:4.2-13.20: X.main[2].g
3486 :     GOTO stdIn:4.2-13.20: X.main[2].f
3487 :     CALL stdIn:2.15-17.4: X.main[2]
3488 :     -
3489 :    
3490 :     (Note that because of a FLINt bug the above code currently does not
3491 :     compile without BTrace turned on.)
3492 :    
3493 :     Here is another example, using my modified Tiger compiler:
3494 :    
3495 :     Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
3496 :     - SMLofNJ.Internals.BTrace.mode (SOME true);
3497 :     [autoloading]
3498 :     [autoloading done]
3499 :     val it = false : bool
3500 :     - CM.make "sources.cm";
3501 :     [autoloading]
3502 :     ...
3503 :     [autoloading done]
3504 :     [scanning sources.cm]
3505 :     [parsing (sources.cm):parse.sml]
3506 :     [creating directory CM/SKEL ...]
3507 :     [parsing (sources.cm):tiger.lex.sml]
3508 :     ...
3509 :     [wrote CM/sparc-unix/semant.sml]
3510 :     [compiling (sources.cm):main.sml]
3511 :     [wrote CM/sparc-unix/main.sml]
3512 :     [New bindings added.]
3513 :     val it = true : bool
3514 :     - Main.compile ("../testcases/merge.tig", "foo.out");
3515 :     *** BACK-TRACE ***
3516 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
3517 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3518 :     CALL lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
3519 :     GOTO lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
3520 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3521 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3522 :     CALL lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
3523 :     / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
3524 :     CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
3525 :     CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
3526 :     CALL lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
3527 :     CALL main.sml:1.18-118.4: Main.compile[2]
3528 :     -
3529 :    
3530 :     ----------------------------------------------------------------------
3531 : blume 673 Name: Matthias Blumen
3532 :     Date: 2000/06/21 18:00:00 JST
3533 :     Tag: blume-20000621-manual
3534 :     Description:
3535 :    
3536 :     CM manual update: Path environments documented.
3537 :    
3538 :     ----------------------------------------------------------------------
3539 : blume 666 Name: Matthias Blume
3540 : blume 672 Date: 2000/06/19 13:40:00
3541 :     Tag: blume-20000619-manual
3542 :     Description:
3543 :    
3544 :     CM manual and system/README update. This only covers the fact that
3545 :     there are no more implicit anchors. (Path environments and the "bind"
3546 :     option to "cm" have yet to be documented.)
3547 :    
3548 :     ----------------------------------------------------------------------
3549 :     Name: Matthias Blume
3550 : blume 671 Date: 2000/06/19 11:05:00 JST
3551 :     Tag: blume-20000619-chdir-bugfix
3552 :     Description:
3553 :    
3554 :     Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
3555 :    
3556 :     ----------------------------------------------------------------------
3557 :     Name: Matthias Blume
3558 : blume 670 Date: 2000/06/18 22:00:10 JST
3559 :     Tag: blume-20000618-implicit-anchors-really-gone
3560 :     Description:
3561 :    
3562 :     I updates the previous HISTORY entry where I forgot to mention that
3563 :     implicit anchors are no longer with us.
3564 :    
3565 :     The current update also gets rid of the (now useless) controller
3566 :     CM.Control.implicit_anchors.
3567 :    
3568 :     ----------------------------------------------------------------------
3569 :     Name: Matthias Blume
3570 : blume 666 Date: 2000/06/16 17:30:00 JST
3571 :     Tag: blume-20000616-anchorenv
3572 :     Description:
3573 :    
3574 :     This patch implements the long anticipated (just kidding :) "anchor
3575 :     environment" mechanism. In the course of doing this, I also
3576 :     re-implemented CM's internal "SrcPath" module from scratch. The new
3577 :     one should be more robust in certain boundary cases. In any case, it
3578 :     is a lot cleaner than its predecessor (IMHO).
3579 :    
3580 :     This time, although there is yet another boot file format change, I
3581 :     kept the unpickler backward-compatible. As a result, no new bootfiles
3582 :     are necessary and bootstrapping is straightforward. (You cannot read
3583 :     new bootfiles into an old system, but the other way around is no
3584 :     problem.)
3585 :    
3586 :     Visible changes:
3587 :    
3588 : blume 670 ** 0. Implicit path anchors (without the leading $-symbol) are no
3589 :     longer recognized at all. This means that such path names are not
3590 :     illegal either. For example, the name basis.cm simply refers to a
3591 :     local file called "basis.cm" (i.e, the name is an ordinary path
3592 :     relative to .cm-files directory). Or, to put it differently, only
3593 :     names that start with $ are anchored paths.
3594 : blume 666
3595 : blume 670 ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
3596 :     vanished.
3597 :    
3598 :     John (Reppy) had critizised this as soon as I originally proposed and
3599 : blume 666 implemented it, but at that time I did not really deeply believe
3600 :     him. :) Now I came full-circle because I need the $<singlearc> syntax
3601 :     in another place where it cannot be seen as an abbreviation for
3602 :     $/<singlearc>. To avoid the confusion, $<singlearc> now means what it
3603 :     seems to mean (i.e., it "expands" into the corresponding anchor
3604 :     value).
3605 :    
3606 :     However, when paths are used as members in CM description files, it
3607 :     continues to be true that there must be at least another arc after the
3608 :     anchor. This is now enforced separately during semantic analysis
3609 : blume 670 (i.e., from a lexical/syntactical point of view, the notation is ok.)
3610 : blume 666
3611 :     ** 2. The "cm" class now accepts an option "bind". The option's value
3612 :     is a sub-option list of precisely two items -- one labeled "anchor"
3613 :     and the other one labeled "value". As you might expect, "anchor" is
3614 :     used to specify an anchor name to be bound, and "value" specifies what
3615 :     the anchor is being bound to.
3616 :    
3617 :     The value must be a directory name and can be given in either standard
3618 :     syntax (including the possibility that it is itself an anchored path)
3619 :     or native syntax.
3620 :    
3621 :     Examples:
3622 :    
3623 :     foo.cm (bind:(anchor:bar value:$mystuff/bar))
3624 :     lib.cm (bind:(anchor:a value:"H:\\x\\y\\z")) (* only works under windows *)
3625 :    
3626 :     and so on.
3627 :    
3628 :     The meaning of this is that the .cm-file will be processed with an
3629 :     augmented anchor environment where the given anchor(s) is/are bound to
3630 :     the given values(s).
3631 :    
3632 :     The rationale for having this feature is this: Suppose you are trying
3633 :     to use two different (already stable) libraries a.cm and b.cm (that
3634 :     you perhaps didn't write yourself). Further, suppose each of these
3635 :     two libraries internally uses its own auxiliary library $aux/lib.cm.
3636 :     Normally you would now have a problem because the anchor "lib" can not
3637 :     be bound to more than one value globally. Therefore, the project that
3638 :     uses both a.cm and b.cm must locally redirect the anchor to some other
3639 :     place:
3640 :    
3641 :     a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
3642 :     b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
3643 :    
3644 :     This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
3645 :     /usr/lib/smlnj/b-stuff/aux.cm, respectively.
3646 :    
3647 :     Hard-wiring path names is a bit inflexible (and CM will verbosely warn
3648 :     you when you do so at the time of CM.stabilize). Therefore, you can
3649 :     also use an anchored path as the value:
3650 :    
3651 :     a.cm (bind:(anchor:lib value:$a-lib))
3652 :     b.cm (bind:(anchor:lib value:$b-lib))
3653 :    
3654 :     Now you can globally configure (using the usual CM.Anchor.anchor or
3655 :     pathconfig machinery) bindings for "a-lib" and "b-lib". Since "lib"
3656 :     itself is always locally bound, setting it globally is no longer
3657 :     meaningful or necessary (but it does not hurt either). In fact, "lib"
3658 :     can still be used as a global anchor for separate purposes. As a
3659 :     matter of fact, one can locally define "lib" in terms of a global
3660 :     "lib":
3661 :    
3662 :     a.cm (bind:(anchor:lib value:$lib/a))
3663 :     b.cm (bind:(anchor:lib value:$lib/b))
3664 :    
3665 :     ** 3: The encoding of path names has changed. This affects the way
3666 :     path names are shown in CM's progress report and also the internal
3667 :     protocol encoding used for parallel make.
3668 :    
3669 :     The encoding now uses one or more ':'-separated segments. Each
3670 :     segments corresponds to a file that has been specified relative to the
3671 :     file given by its preceding segment. The first segment is either
3672 :     relative to the CWD, absolute, or anchored. Each segment itself is
3673 :     basically a Unix pathname; all segments but the first are relative.
3674 :    
3675 :     Example:
3676 :    
3677 :     $foo/bar/baz.cm:a/b/c.sml
3678 :    
3679 :     This path denotes the file bar/a/b/c.sml relative to the directory
3680 :     denoted by anchor "foo". Notice that the encoding also includes
3681 :     baz.cm which is the .cm-file that listed a/b/c.sml. As usual, such
3682 :     paths are resolved relative to the .cm-files directory, so baz.cm must
3683 :     be ignored to get the "real" pathname.
3684 :    
3685 :     To make this fact more obvious, CM puts the names of such "virtual
3686 :     arcs" into parentheses when they appear in progress reports. (No
3687 :     parentheses will appear in the internal protocol encoding.) Thus,
3688 :     what you really see is:
3689 :    
3690 :     $foo/bar/(baz.cm):a/b/c.sml
3691 :    
3692 :     I find this notation to be much more informative than before.
3693 :    
3694 :     Another new feature of the encoding is that special characters
3695 :     including parentheses, colons, (back)slashes, and white space are
3696 :     written as \ddd (where ddd is the decimal encoding of the character).
3697 :    
3698 : blume 670 *** The CM manual still needs to be updated.
3699 :    
3700 : blume 666 ----------------------------------------------------------------------
3701 : leunga 660 Name: Allen Leung
3702 :     Date: 2000/06/15 00:38:00
3703 :     Tag: leunga-20000615-x86-peephole
3704 :    
3705 :     x86 Peephole fix by Fermin. Affects c-- and moby only.
3706 :    
3707 :     ----------------------------------------------------------------------
3708 : blume 658 Name: Matthias Blume
3709 :     Date: 2000/06/12 11:40:00
3710 :     Tag: blume-20000612-parmakefix
3711 :     Description:
3712 :    
3713 :     More cleanup after changing the file naming scheme: This time I
3714 :     repaired the parallel make mechanism for CMB.make which I broke earlier.
3715 :    
3716 :     ----------------------------------------------------------------------
3717 : leunga 657 Name: Allen Leung
3718 :     Date: 2000/06/09 01:25:00
3719 :     Tag: leunga-20000609-various
3720 :    
3721 :     None of these things should affect normal SML/NJ operations
3722 :    
3723 :     1. Peephole improvements provided by Fermin (c--)
3724 :     2. New annotation DEFUSE for adding extra dependence (moby)
3725 :     3. New X86 LOCK instructions (moby)
3726 :     4. New machine description language for reservation tables (scheduling)
3727 :     5. Fixes to various optimization/analysis modules (branch chaining, dominator
3728 :     trees etc.)
3729 :     6. I've changed the CM files so that they can work with versions
3730 :     110.0.6, 110.25 and 110.28
3731 :    
3732 :     ----------------------------------------------------------------------
3733 : blume 652 Name: Matthias Blume
3734 : blume 656 Date: 2000/06/09 12:40:00
3735 :     Tag: blume-20000609-log
3736 :     Description:
3737 :    
3738 :     - Removed all(?) remaining RCS Log entries from sources.
3739 :    
3740 :     - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
3741 :     anchored paths).
3742 :    
3743 :     ----------------------------------------------------------------------
3744 :     Name: Matthias Blume
3745 : blume 653 Date: 2000/06/07 17:00:00 JST
3746 :     Tag: blume-20000607-no-implicit-anchors
3747 :     Description:
3748 :    
3749 :     1. This update changes the default setting for
3750 :     CM.Control.implicit_anchors from true to false. This means that
3751 :     implicit anchors are no longer permitted by default. I also tried to
3752 :     make sure that nothing else still relies on implicit anchors.
3753 :     (This is the next step on the schedule towards a CM that does not even
3754 :     have the notion of implicit anchors anymore.)
3755 :    
3756 :     2. More CM manual updates.
3757 :    
3758 :     3. I managed to track down and fix the pickling bug I mentioned last
3759 :     time. Because of the previously existing workaround, this entails no
3760 :     immediate practical changes.
3761 :    
3762 :     ----------------------------------------------------------------------
3763 :     Name: Matthias Blume
3764 : blume 652 Date: 2000/06/06 11:15:00 JST
3765 :     Tag: blume-20000606-lazierpickle
3766 :     Description:
3767 :    
3768 :     !!!! NEW BOOT FILES !!!!
3769 :    
3770 :     * The main purpose of this update is to make library pickles lazier in
3771 :     order to reduce the initial space penalty for autoloading a library.
3772 :     As a result, it is now possible to have $smlnj/compiler.cm
3773 :     pre-registered. This should take care of the many complaints or
3774 :     inquiries about missing structure Compiler. This required changes to
3775 :     CM's internal data structures and small tweaks to some algorithms.
3776 :    
3777 :     As a neat additional effect, it is no longer necessary (for the sake
3778 :     of lean heap image files) to distinguish between a "minimal" CM and a
3779 :     "full" CM. Now, there is only one CM (i.e., the "full" version:
3780 :     $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
3781 :     interactive top level. ($smlnj/cm/minimal.cm is gone.)
3782 :    
3783 :     To make the life of compiler-hackers easier, "makeml" now also
3784 :     pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm). In other
3785 :     words, after you bootstrap a new sml for the first time, you will not
3786 :     have to autoload $smlnj/cmb.cm again afterwards. (The first time
3787 :     around you will still have to do it, though.)
3788 :    
3789 :     * A second change consists of major updates to the CM manual. There
3790 :     are now several appendices with summary information and also a full
3791 :     specification of the CM description file syntax.
3792 :    
3793 :     * In directory src/system I added the script "allcross". This script
3794 :     invokes sml and cross-compiles the compiler for all supported
3795 :     architectures. (Useful when providing a new set of boot files.)
3796 :    
3797 :     * There seems to be a latent bug in my "lazy pickles" mechanism. I
3798 :     added a small tweak to pickle-util.sml to work around this problem,
3799 :     but it is not a proper fix yet. I will investigate further. (The
3800 :     effect of the bug was an inflation of library pickle size.)
3801 :    
3802 :     * Version number increased to 110.28.1 (to avoid compatibility problems).
3803 :    
3804 :     ----------------------------------------------------------------------
3805 : leunga 646 Name: Allen Leung
3806 : leunga 648 Date: 2000/05/25 17:28 EDT
3807 :     Tag: leunga-20000525-ra
3808 :     Description:
3809 :    
3810 :     Fixed a bug in freezing phase of the register allocator.
3811 :    
3812 :     ----------------------------------------------------------------------
3813 :     Name: Allen Leung
3814 : leunga 647 Date: 2000/05/15 22:53 EDT
3815 : leunga 646 Tag: leunga-20000515-alpha-x86-ra
3816 :     Description:
3817 :    
3818 :     1. Alpha
3819 :    
3820 :     Slight cleanup. Removed the instruction SGNXL
3821 :    
3822 :     2. X86
3823 :    
3824 :     Added the following instructions to the instruction set:
3825 :    
3826 :     ROLx, RORx,
3827 :     BTx, BTSx, BTLx, BTRx,
3828 :     XCHGx, and variants with the LOCK prefix
3829 :    
3830 :     3. Register Allocation
3831 :    
3832 :     The module ra-rewrite-with-renaming has been improved.
3833 :    
3834 : leunga 647 These have no effect on SML/NJ.
3835 :    
3836 : leunga 646 ----------------------------------------------------------------------
3837 : blume 642 Name: Matthias Blume
3838 : blume 645 Date: 2000/05/15 16:20:00 JST
3839 :     Tag: blume-20000515-lightrebuild
3840 :     Description:
3841 :    
3842 :     1. I added an alternative to "-rebuild" to "makeml". The difference is
3843 :     that prior to calling CMB.make' the CM-variable "LIGHT" will be
3844 :     defined. In effect, the command will not build any cross-compiler
3845 :     backends and therefore finish more quickly.
3846 :    
3847 :     The "fixpt" script also takes a "-light" switch to be able to use
3848 :     this new facility while compiling for a fixpoint.
3849 :    
3850 :     2. I replaced all mentions of anchored paths in group owner specifications
3851 :     with simple relative paths (usually starting with "..").
3852 :     The rationale is that a library's internal workings should not be
3853 :     compromised by the lack of some anchor. (An anchor is necessary
3854 :     for someone who wants to refer to the library by an anchored path,
3855 :     but it should not be necessary to build the same library in the first
3856 :     place.)
3857 :    
3858 :     3. I changed the way CM's tool mechanism determines the shell command
3859 :     string used for things like ml-yacc etc. so that it does not break
3860 :     when CM.Control.implicit_anchors is turned off.
3861 :    
3862 :     ----------------------------------------------------------------------
3863 :     Name: Matthias Blume
3864 : blume 644 Date: 2000/05/12 18:20:00 JST
3865 :     Tag: blume-20000512-ml-build
3866 :     Description:
3867 :    
3868 :     Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
3869 :     getting installed properly (by config/install.sh).
3870 :    
3871 :     ----------------------------------------------------------------------
3872 :     Name: Matthias Blume
3873 : blume 643 Date: 2000/05/12 17:30:00 JST
3874 :     Tag: blume-20000512-anchors
3875 :     Description:
3876 :    
3877 :     !!! NEW BOOT FILES !!!
3878 :    
3879 :     This change is in preparation of fading out support for "implicitly
3880 :     anchored path names". I went through all sources and used the
3881 :     explicit (and relatively new) $-notation. See system/README and the
3882 :     CM manual for more info on this.
3883 :    
3884 :     I also modified the anchoring scheme for some things such as "smlnj",
3885 :     "MLRISC", "cm", etc. to take advantage of the fact that explicit
3886 :     anchors are more expressive: anchor name and first arc do not have to
3887 :     coincide. This entails the following user-visible change:
3888 :    
3889 :     You have to write $smlnj/foo/bar instead of smlnj/foo/bar. In
3890 :     particular, when you fire up sml with a command-line argument, say,
3891 :     e.g.:
3892 :    
3893 :     sml '$smlnj/cmb.cm'
3894 :    
3895 :     At the ML toplevel prompt:
3896 :    
3897 :     CM.autoload "$smlnj/cmb.cm";
3898 :    
3899 :     There is also a new controller in CM.Control that can be used to turn
3900 :     off all remaining support for implicit anchors by saying:
3901 :    
3902 :     CM.autoload "$smlnj/
3903 :     #set CM.Control.implicit_anchors false;
3904 :    
3905 :     This causes CM to reject implicitly anchored paths. This is (for the
3906 :     time being) less permissive than the "final" version where there will
3907 :     be no more such implicit anchors and relative paths will be just that:
3908 :     relative.
3909 :    
3910 :     The next step (version after next version?) will be to make the
3911 :     default for CM.Control.implicit_anchors false. After the dust has
3912 :     settled, I can then produce the "final" version of this...
3913 :    
3914 :     Note: Since bootstrapping is a bit tricky, I provided new boot files.
3915 :    
3916 :     ----------------------------------------------------------------------
3917 :     Name: Matthias Blume
3918 : blume 642 Date: 2000/05/11 16:30:00 JST
3919 :     Tag: blume-20000511-sources
3920 :     Description:
3921 :    
3922 :     The main change is that I added function CM.sources as a generalized
3923 :     version of the earlier CM.makedepend. This entails the following
3924 :     additional changes:
3925 :    
3926 :     - CM.makedepend has been dropped.
3927 :    
3928 :     - CM manual has been updated.
3929 :    
3930 :     - TOOLS signature and API have been changed.
3931 :    
3932 :     ----------------------------------------------------------------------
3933 : leunga 640 Name: Allen Leung
3934 : leunga 641 Date: 2000/05/10 21:17 EDT
3935 :     Tag: leunga-20000510-moby-c--ssa
3936 :     Description:
3937 :    
3938 :     Various bug fixes and new features for C--, Moby and MLRISC optimizations.
3939 :     None of these affect SML/NJ.
3940 :    
3941 :     1. Register Allocation
3942 :    
3943 :     a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
3944 :     This module tries to remove local (i.e. basic block level) redundancies
3945 :     during spilling.
3946 :    
3947 :     b. A new framework for performing region based register allocation.
3948 :     Not yet entirely functional.
3949 :    
3950 :     2. X86
3951 :    
3952 :     a. DefUse for POP was missing the stack pointer [found by Lal]
3953 :     b. Reload for CALL was incorrect in X86Spill [found by John]
3954 :     c. Various fixes in X86Spill so that it can be used correctly for
3955 :     the new spilling module.
3956 :    
3957 :     3. SSA/IR
3958 :    
3959 :     a. New module ir/dj-dataflow.sml implements elimination based
3960 :     data flow analysis.
3961 :    
3962 :     4. MLRiscGen
3963 :    
3964 :     a. Fix for gc type annotation
3965 :    
3966 :     5. MDGen
3967 :    
3968 :     Various fixes for machine description -> ml code translation. For ssa
3969 :     only.
3970 :    
3971 :     ----------------------------------------------------------------------
3972 :     Name: Allen Leung
3973 : leunga 640 Date: 2000/05/08 22:17 EDT
3974 :     Tag: leunga-20000508-labexp
3975 :     Description:
3976 :    
3977 :     Fermin has found a few assembly problems with constant expressions
3978 :     generated in LabelExp. Mostly, the problems involve extra parentheses,
3979 :     which choke on dumb assemblers. This is his fix.
3980 :    
3981 :     ----------------------------------------------------------------------
3982 : dbm 635 Name: Dave MacQueen
3983 :     Date: 2000/04/09 14:00 EDT
3984 :     Tag: dbm-20000502-Version_110_28
3985 :     Description:
3986 :    
3987 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
3988 :    
3989 :     2. Updated config/version to 110.28
3990 :    
3991 :     3. Updated config/srcarchiveurl
3992 :    
3993 :     3. New boot files!
3994 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
3995 :    
3996 :     ----------------------------------------------------------------------
3997 : blume 634 Name: Matthias Blume
3998 :     Date: 2000/05/01 19:05:00 JST
3999 :     Tag: blume-20000501-noweb
4000 :     Description:
4001 :    
4002 :     A new noweb tool has been added. The existing system is entirely
4003 :     unaffected by this, but some CM users have asked for renewed noweb
4004 :     support. Everything is documented in the CM manual.
4005 :    
4006 :     New (plugin) libraries:
4007 :    
4008 :     noweb-tool.cm
4009 :     nw-ext.cm
4010 :    
4011 :     ----------------------------------------------------------------------
4012 : dbm 633 Name: Dave MacQueen
4013 :     Date: 2000/04/30 12:40PM EDT
4014 :     Tag: dbm-20000430-bug_fixes
4015 :     Description:
4016 :    
4017 :     1. Fix for bug 1498
4018 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sig
4019 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sml
4020 :     added toRealArray function
4021 :     smlnj/src/compiler/MiscUtil/print/ppobj.sml
4022 :     added check for tag Obj.RealArray to array printing case in ppObj
4023 :    
4024 :     2. Fix for bug 1510
4025 :     smlnj/src/compiler/Semant/types/typesutil.sml
4026 :     fixed definition of dummyargs (used by equalTycon) so that
4027 :     dummy args are distinct types
4028 :    
4029 :     ----------------------------------------------------------------------
4030 : blume 629 Name: Matthias Blume
4031 : blume 632 Date: 2000/04/30 01:00:00 JST
4032 :     Tag: blume-20000430-versions
4033 :     Description:
4034 :    
4035 :     1. CM version numbering added. This is an implementation of Lal's
4036 :     proposal for adding version numbers and version checking to .cm
4037 :     files. Lal said that his proposal was just that -- a proposal.
4038 :     For the time being I went ahead and implemented it so that people
4039 :     can comment on it. Everything is completely backward-compatible
4040 :     (except for the stable library format, i.e., new bootfiles!).
4041 :    
4042 :     As usual, see the CM manual for details.
4043 :    
4044 :     2. An alternative syntax for anchored paths has been implemented.
4045 :     Dave has recently voiced the same concerns that I had when I did
4046 :     this, so there should be some support. My take is that eventually
4047 :     I will let support for the current syntax (where anchors are
4048 :     "implicit") fade out in favor of the new, explicit syntax.
4049 :     In order to be backward-compatible, both old and new syntax are
4050 :     currently supported.
4051 :    
4052 :     Again, see the CM manual for details.
4053 :    
4054 :     3. Parallel make is trying to be slightly smarter: When the master
4055 :     process finds a "bottleneck", i.e., when there is only one
4056 :     compilation unit that can be compiled and everybody else is
4057 :     waiting on it, then it will simply compile it directly instead
4058 :     of clumsily telling one of the slaves to do it.
4059 :    
4060 :     4. Support for "unsharing" added. This is necessary in order to be
4061 :     able to have two different versions of the same library running
4062 :     at the same time (e.g., for trying out a new MLRISC while still
4063 :     having the old MLRISC linked into the current compiler, etc.)
4064 :     See the CM manual.
4065 :    
4066 :     5. Simple "makedepend" functionality added for generating Makefile
4067 :     dependency information. (This is rather crude at the moment.
4068 :     Expect some changes here in the future.)
4069 :    
4070 :     6. ".fun" added as a recognized suffix for ML files. Also documented
4071 :     explicitly in the manual that the fallback behavior (unknown suffix
4072 :     -> ML file) is not an official feature!
4073 :    
4074 :     7. Small changes to the pickler for stable libraries.
4075 :    
4076 :     8. Several internal changes to CM (for cleanup/improvement).
4077 :    
4078 :    
4079 :     !!!! NEW BINFILES !!!!
4080 :    
4081 :     ----------------------------------------------------------------------
4082 :     Name: Matthias Blume
4083 : blume 631 Date: 2000/04/28 17:30:00 JST
4084 :     Tag: blume-20000428-pathconfig
4085 :     Description:
4086 :    
4087 :     1. I changed config/install.sh to remove duplicate entries from the
4088 :     lib/pathconfig file at the end. Moreover, the final version of
4089 :     lib/pathconfig is sorted alphabetically. The same (sorting) is done
4090 :     in src/system/installml.
4091 :    
4092 :     2. The config/install.sh script now consistently uses relative
4093 :     pathnames in lib/pathconfig whenever the anchor is in the lib
4094 :     directory. (So far this was true for the libraries that come
4095 :     pre-compiled and bundled as part of the bootfiles but not for
4096 :     libraries that are compiled by the script itself.)
4097 :    
4098 :     ----------------------------------------------------------------------
4099 :     Name: Matthias Blume
4100 : blume 629 Date: 2000/04/26 13:10:00 JST
4101 :     Tag: blume-20000426-fun_suffix
4102 :     Description:
4103 :    
4104 :     Added ".fun" as a recognized file name suffix (for ML code).
4105 :    
4106 :     ----------------------------------------------------------------------
4107 : leunga 628 Name: Allen Leung
4108 :     Date: 2000/04/25 17:00:00 EST
4109 :     Tag: leunga-20000425-alpha-ra
4110 :     Description:
4111 :    
4112 :     1. Alpha
4113 :    
4114 :     PSEUDOARITH was missing in AlphaRewrite. This causes an endless loop
4115 :     in C--.
4116 :    
4117 :     2. RA
4118 :    
4119 :     Added a flag "ra-dump-size" to print out the size of the flowgraph
4120 :     and the interference graph.
4121 :    
4122 :     ----------------------------------------------------------------------
4123 : dbm 627 Name: Dave MacQueen
4124 :     Date: 2000/04/25/
4125 :     Tag: dbm-20000425-mlyacc_doc_examples
4126 :     Description:
4127 :     Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
4128 :     Updated all three examples in src/ml-yacc/examples to run
4129 :     under 110.* using CM.make.
4130 :    
4131 :     ----------------------------------------------------------------------
4132 : leunga 624 Name: Allen Leung
4133 :     Date: 2000/04/20 23:04:00 EST
4134 :     Tag: leunga-20000420-ssa-c---stuff
4135 :     Description:
4136 :    
4137 :     This update synchronizes my repository with Yale's. Most of these
4138 :     changes, however, do not affect SML/NJ at all (the RA is an exception).
4139 :    
4140 :     1. Register Allocator
4141 :    
4142 :     a. An improvement in the interference graph construction:
4143 :     Given a copy
4144 :    
4145 :     s <- t
4146 :    
4147 : leunga 628 no interference edge between s and t is added for this definition of s.
4148 : leunga 624
4149 :     b. I've added two new spill heuristic modules that Fermin and I developed
4150 :     (in the new library RA.cm). These are unused in SML/NJ but maybe
4151 :     useful for others (Moby?)
4152 :    
4153 :     2. X86
4154 :    
4155 :     a. Various fixes in the backend provided by Fermin [C--] and Lal.
4156 :    
4157 :     3. Alpha
4158 :    
4159 :     a. Added the BSR instruction and code generation that goes with it [C--]
4160 :     b. Other fixes too numerous to recount provided by Fermin [C--]
4161 :    
4162 :     4. Regmaps
4163 :    
4164 :     a. The regmaps are not initialized with the identity physical bindings
4165 :     at creation time. This is unneeded.
4166 :    
4167 :     5. MLRISC Optimizations
4168 :    
4169 :     a. The DJ-Graph module can now compute the iterated dominance frontiers
4170 :     intersects with liveness incrementally in linear time! Woohoo!
4171 :     This is now used in my new SSA construction algorithm.
4172 :    
4173 :     b. THe branch reorganization module is now smarter about linear chains of
4174 :     basic blocks.
4175 :    
4176 :    
4177 :     ----------------------------------------------------------------------
4178 : blume 615 Name: Matthias Blume
4179 :     Date: 2000/04/12 13:52:00 JST
4180 :     Tag: blume_main_v110p27_1
4181 :     Description:
4182 :    
4183 :     Changed install.sh script to handle archive files without version number
4184 :     and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
4185 :     name of the boot file archive.
4186 :    
4187 :     ----------------------------------------------------------------------
4188 : dbm 607 Name: Dave MacQueen
4189 :     Date: 2000/04/09 14:00 EDT
4190 :     Tag: dbm-20000410-Version_110_27
4191 :     Description:
4192 :    
4193 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
4194 :    
4195 :     2. Updated src/config/version to 110.27
4196 :    
4197 :     3. New boot files!
4198 :    
4199 :     ----------------------------------------------------------------------
4200 : leunga 606 Name: Allen Leung
4201 :     Date: 2000/04/09 19:09:00 EST
4202 :     Tag: leunga-20000409-misc
4203 :     Description:
4204 :    
4205 :     1. Yet another fix for x86 assembly for idivl, imull, mull and friends.
4206 :    
4207 :     2. Miscellaneous improvements to MLRISC (unused in sml/nj)
4208 :    
4209 :     ----------------------------------------------------------------------
4210 : monnier 605 Name: Stefan
4211 :     Date: 2000/04/07 10:00:00 EDT
4212 :     Tag: monnier-20000406-branch-handling
4213 :     Description:
4214 :    
4215 :     Improved handling of branches (mostly those generated from
4216 :     polymorphic equality), removed switchoff and changed the
4217 :     default optimization settings (more cpsopt and less flintopt).
4218 :    
4219 :     ----------------------------------------------------------------------
4220 : leunga 601 Name: Allen Leung
4221 : leunga 602 Date: 2000/04/06 01:30:00 EST
4222 :     Tag: leunga-20000406-peephole-x86-SSA-2
4223 :     Description:
4224 :    
4225 :     Forgot a few files.
4226 :    
4227 :     ----------------------------------------------------------------------
4228 :     Name: Allen Leung
4229 : leunga 601 Date: 2000/04/06 00:36:00 EST
4230 :     Tag: leunga-20000406-peephole-x86-SSA
4231 :     Description:
4232 :    
4233 :     1. New Peephole code
4234 :    
4235 :     2. Minor improvement to X86 instruction selection
4236 :    
4237 :     3. Various fixes to SSA and machine description -> code translator
4238 :    
4239 :     ----------------------------------------------------------------------
4240 : blume 595 Name: Matthias Blume
4241 :     Date: 2000/04/05 12:30:00 JST
4242 :     Tag: blume_main_v110p26p2_3
4243 :     Description:
4244 :    
4245 :     This update just merges three minor cosmetic updates to CM's sources
4246 :     to get ready for the 110.27 code freeze on Friday. No functionality
4247 :     has changed.
4248 :    
4249 :     ----------------------------------------------------------------------
4250 : leunga 593 Name: Allen Leung
4251 : leunga 594 Date: 2000/04/04 19:39:00 EST
4252 :     Tag: leunga-20000404-x86-asm
4253 :     Description:
4254 :    
4255 :     1. Fixed a problem in X86 assembly.
4256 :    
4257 :     Things like
4258 :    
4259 :     jmp %eax
4260 :     jmp (%eax)
4261 :    
4262 :     should be output as
4263 :    
4264 :     jmp *%eax
4265 :     jmp *(%eax)
4266 :    
4267 :     2. Assembly output
4268 :    
4269 :     Added a new flag
4270 :    
4271 :     "asm-indent-copies" (default to false)
4272 :    
4273 :     When this flag is on, parallel copies will be indented an extra level.
4274 :    
4275 :     ----------------------------------------------------------------------
4276 :     Name: Allen Leung
4277 : leunga 593 Date: 2000/04/04 03:18:00 EST
4278 :     Tag: leunga-20000404-C--Moby
4279 :     Description:
4280 :    
4281 :     All of these fixes are related to C--, Moby, and my own optimization
4282 :     stuff; so they shouldn't affect SML/NJ.
4283 :    
4284 :     1. X86
4285 :    
4286 :     Various fixes related floating point, and extensions.
4287 :    
4288 :     2. Alpha
4289 :    
4290 :     Some extra patterns related to loads with signed/zero extension
4291 :     provided by Fermin.
4292 :    
4293 :     3. Assembly
4294 :    
4295 : monnier 659 When generating assembly, resolve the value of client defined constants,
4296 : leunga 593 instead of generating symbolic values. This is controlled by the
4297 :     new flag "asm-resolve-constants", which is default to true.
4298 :    
4299 :     4. Machine Descriptions
4300 :    
4301 :     a. The precedence parser was slightly broken when parsing infixr symbols.
4302 :     b. The type generalizing code had the bound variables reversed, resulting
4303 :     in a problem during arity raising.
4304 :     c. Various fixes in machine descriptions.
4305 :    
4306 :     ----------------------------------------------------------------------
4307 : blume 592 Name: Matthias Blume
4308 :     Date: 2000/04/03 16:05:00 JST
4309 :     Tag: blume_main_v110p26p2_2
4310 :     Description:
4311 :    
4312 :     I eliminated coreEnv from compInfo. Access to the "Core" structure is
4313 :     now done via the ordinary static environment that is context to each
4314 :     compilation unit.
4315 :    
4316 :     To this end, I arranged that instead of "structure Core" as "structure
4317 :     _Core" is bound in the pervasive environment. Core access is done via
4318 : monnier 659 _Core (which can never be accidentally rebound because _Core is not a
4319 : blume 592 legal surface-syntax symbol).
4320 :    
4321 :     The current solution is much cleaner because the core environment is
4322 :     now simply part of the pervasive environment which is part of every
4323 :     compilation unit's context anyway. In particular, this eliminates all
4324 :     special-case handling that was necessary until now in order to deal
4325 :     with dynamic and symbolic parts of the core environment.
4326 :    
4327 :     Remaining hackery (to bind the "magic" symbol _Core) is localized in the
4328 : monnier 659 compilation manager's bootstrap compiler (actually: in the "init group"
4329 : blume 592 handling). See the comments in src/system/smlnj/init/init.cmi for
4330 :     more details.
4331 :    
4332 :     I also tried to track down all mentions of "Core" (as string argument
4333 :     to Symbol.strSymbol) in the compiler and replaced them with a
4334 :     reference to the new CoreSym.coreSym. Seems cleaner since the actual
4335 :     name appears in one place only.
4336 :