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

SCM Repository

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

Annotation of /sml/trunk/NOTES/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1381 - (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 : mblume 1381 Name: Matthias Blume (blume (at) tti - c (dot) org)
17 :     Date: 2003/09/17 16:30:00 CDT
18 :     Tag: blume-20030917-real32-slices
19 :     Description:
20 :    
21 :     Added an entry to the primitive environment
22 :     (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
23 :     added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
24 :    
25 :     Removed some of the "magic" constants in real64.sml and replaced them
26 :     with code that generates these values from their corresponding
27 :     integer counterparts.
28 :    
29 :     Made all(?) the slice-related changes to the Basis and made everything
30 :     compile again...
31 :    
32 :     ----------------------------------------------------------------------
33 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
34 : mblume 1375 Date: 2003/09/15 17:45:00 CDT
35 :     Tag: blume-20030915-rbase
36 :     Description:
37 :    
38 :     Fixed bug in Real.fromLargeInt.
39 :    
40 :     ----------------------------------------------------------------------
41 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
42 : mblume 1369 Date: 2003/09/13 18:11:00 CDT
43 :     Tag: blume-20030913-libinstall
44 :     Description:
45 :    
46 :     Minor bugfix in config/libinstall (set anchor with path to
47 :     standalone tool after installing it, otherwise libraries that
48 :     need ml-lex or ml-yacc won't compile the first time the installer
49 :     runs).
50 :    
51 :     ----------------------------------------------------------------------
52 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
53 : mblume 1368 Date: 2003/09/12 11:45:00 CDT
54 :     Tag: blume-20030912-various
55 :     Description:
56 :    
57 :     - fixed bug in Real.toLargeInt
58 :     - fixed bug in Posix.ProcEnv.times
59 :     - changed inputLine functions to return an option
60 :     - minor installer improvements / bugfixes
61 :     - changed default @SMLalloc parameter for x86/celeron to 64k
62 :    
63 :     ----------------------------------------------------------------------
64 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
65 : mblume 1356 Date: 2003/09/09 22:00:00 CDT
66 :     Tag: Release_110_43
67 :     Description:
68 :    
69 :     New working release 110.43. New bootfiles.
70 :    
71 :     ----------------------------------------------------------------------
72 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
73 : mblume 1354 Date: 2003/09/09 19:20:00 CDT
74 :     Tag: blume-20030909-installer
75 :     Description:
76 :    
77 :     Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
78 :     Modified config/install.bat to take advantage of it. Also modified
79 :     config/install.sh (and called it config/new-install.sh) to take advantage
80 :     of it on Unix systems. (The SML code is (supposed to be) platform-
81 :     independent.)
82 :    
83 :     The installer can now install everything under Win32
84 :     as well as under *nix as long as it compiles.
85 :    
86 :     Other changes:
87 :    
88 :     - made CML compile again under Win32
89 :     - made eXene compile under Win32 (by providing a fake structure UnixSock
90 :     and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
91 :     - fixed a bug in nowhere: it assumed that type OS.Process.status is the
92 :     same as type int; under Win32 it isn't
93 :     - fixed some slice-related problems in the win32-specific parts of CML
94 :     - added a functor argument "sameVol" to os-path-fn.sml in the Basis
95 :     (under Win32, the volume name is case-insensitive, and the
96 :     OS.Path code compares volume names for equality)
97 :    
98 :     ----------------------------------------------------------------------
99 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
100 : mblume 1353 Date: 2003/09/08 11:55:00 CDT
101 :     Tag: blume-20030908-fullpath
102 :     Description:
103 :    
104 :     Made Win32 version of OS.FileSys.fullPath return current directory
105 :     when given an empty string. This is what the spec says, and incidentally,
106 :     CM depends on it. (CM otherwise goes into an infinite loop in certain
107 :     cases when presented with the name of a non-existing .cm file.)
108 :    
109 :     ----------------------------------------------------------------------
110 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
111 : mblume 1350 Date: 2003/09/04 16:30:00 CDT
112 : mblume 1352 Tag: blume-20030905-slices-etc
113 : mblume 1350 Description:
114 :    
115 :     1. Changed interface to vectors and arrays in Basis to match
116 :     (draft) Basis spec.
117 :     2. Added signatures and implementations of slices according to
118 :     Basis spec.
119 :     3. Edited source code throughout the system to make it compile again
120 :     under 1. and 2. (In some cases code had to be added to have it
121 :     match the new signatures.)
122 :     4. MLRISC should be backward-compatible: the copies of the originals
123 :     of files that needed to change under 3. were retained, the .cm files
124 :     check the compiler version number and use old versions when
125 :     appropriate.
126 :     5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
127 :     dirstream -> string option (in accordance with Basis spec).
128 :     6. When generating code that counts lines, ml-lex used function
129 :     CharVector.foldli, taking advantage of its old interface.
130 :     This has been replaced with the corresponding code from
131 :     CharVectorSlice. (html-lex must be re-lexed!)
132 :     7. BitArray in smlnj-lib/Util has been extended/modified to match the
133 :     new MONO_ARRAY signature. (Do we need BitArraySlice?)
134 :     8. Removed temporary additions (fromInternal, toInternal) from the
135 :     (now obsolete) IntInf in smlnj-lib/Util.
136 :     9. Cleaned up structure Byte.
137 :     10. Added localOffset, scan, and fromString to Date (according to spec).
138 :     Cleaned/corrected implementation of Date.
139 :     (Still need to check for correctness; implement better canonicalizeDate.)
140 :     11. Added "scan" to signature IEEE_REAL.
141 :     12. Some improvements to IntInf [in particular: efficiency-hack for
142 :     mod and rem when second operand is 2 (for parity checks).]
143 :     13. Changed representation of type Time.time, using a single IntInf.int
144 :     value counting microseconds. This considerably simplified the
145 :     implementation of structure Time. We now support negative time
146 :     values; scan and fromString handle signs.
147 :     14. Functor PrimIO now takes two additional arguments (VectorSlice and
148 :     ArraySlice).
149 :    
150 :     ----------------------------------------------------------------------
151 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
152 : mblume 1347 Date: 2003/08/28 17:00:00 CDT
153 :     Tag: blume-20030828-intinf
154 :     Description:
155 :    
156 :     This is a major update which comes with a version number bump
157 :     (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
158 :     and an implementation of IntInf in the Basis.
159 :    
160 :     There are a fairly large number of related changes and updates throughout
161 :     the system:
162 :    
163 :     Basis:
164 :     - Implemented IntInf.
165 :     - Made LargeInt a projection of IntInf (by filtering through INTEGER).
166 : mblume 1348 - Added some missing Real64 operations, most notably Real.toLargeInt.
167 : mblume 1347 - Added FixedInt as a synonym for Int32.
168 :    
169 :     compiler:
170 :     * Added support for a built-in intinf type.
171 :     - literals
172 :     - pattern matching
173 :     - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
174 :     - overloading on literals and operations
175 :    
176 :     This required adding a primitive type intinf, some additional
177 :     primops, and implementations for several non-trivial intinf
178 :     operations in Core. (The intinf type is completely abstract
179 :     to the compiler; all operations get delegated back to the Core.)
180 :    
181 :     * Intinf equality is handled by polyequal. However, the compiler
182 :     does not print its usual warning in this case (since polyequal
183 :     is the right thing to do there).
184 :    
185 :     * Improved the organization of structure InlineT.
186 :    
187 :     * A word about conversion primops:
188 :     If conversions involving intinf do not cancel out during
189 : mblume 1348 CPS contract, then the compiler must insert calls to Core functions.
190 : mblume 1347 Since all core access must be resolved already during the FLINT
191 :     translate phase, it would be too late a the time of CPS contract
192 : mblume 1348 to add new Core calls. For this reason, conversion primops
193 : mblume 1347 for intinf carry two arguments: 1. the numeric argument that
194 :     they are supposed to convert, and 2. the Core function that
195 :     can help with this conversion if necessary. If CPS contract
196 : mblume 1348 eliminates a primop, then the associated Core function becomes
197 :     dead and goes away. Intinf conversion primops that do not get
198 :     eliminated by CPS contract get rewritten into calls of their
199 :     core functions by a separate, new phase.
200 : mblume 1347
201 :     interactive system:
202 :     - Control.Print.intinfDepth controls max length of intinf constants
203 :     being printed. (Analogous to Control.Print.stringDepth.)
204 :     - Cleanup in printutil and pputil: got rid of unused stuff and
205 :     duplicates; replaced some of the code with code that makes better
206 :     use of library functionality.
207 :    
208 :     CM:
209 :     Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
210 :     are no longer silent.
211 :    
212 :     CKIT:
213 :     Fixed mismatched uses of Int32 and LargeInt. I always decided
214 :     in favor of LargeInt -- which is now the same as IntInf.
215 :     CKIT-knowledgable people should check whether this is what's
216 :     intended and otherwise change things back to using Int32 or
217 :     FixedInt.
218 :    
219 :     Throughout the code:
220 : mblume 1348 Started using IntInf.int literals and built-in operations
221 : mblume 1347 (e.g., comparison with 0) where this seems appropriate.
222 :    
223 :    
224 :     ----------------------------------------------------------------------
225 : macqueen 1343 Name: Dave MacQueen (dbm@cs.uchicago.edu)
226 :     Date: 2003/08/13 11:36:00 CDT
227 :     Tag: dbm-20030813-mcz-merge1
228 :     Description:
229 :    
230 :     Merging changes from the mcz-branch development branch into trunk.
231 :     These changes involve replacement of the emulated old prettyprinter
232 :     interface with direct use of the SML/NJ Lib PP library, and fixing
233 :     of a couple of bugs (895, 1186) relating to error messages. A new
234 :     prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
235 :     has been added.
236 :    
237 :     ----------------------------------------------------------------------
238 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
239 : mblume 1342 Date: 2003/08/11 15:45:00 CDT
240 :     Tag: blume-20030811-windows
241 :     Description:
242 :    
243 :     Version number bumped to 110.42.9. NEW BOOTFILES!!!
244 :    
245 :     http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
246 :    
247 :     This patch restores SML/NJ's ability to run under win32. There are a
248 :     number of changes, including fixes for several bugs that had gone
249 :     unnoticed until now:
250 :    
251 :     - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
252 :     trying to run the win32 version from within cygwin.)
253 :    
254 :     - There are a number of simple .bat scripts that substitute for their
255 :     corresponding Unix shell-scripts. (See below.)
256 :    
257 :     - The internals of ml-build have been modified slightly. The main
258 :     difference is that instead of calling ".link-sml" (or link-sml.bat)
259 :     using OS.Process.system, the ML process delegates this task back
260 :     to the script. Otherwise problems arise in mixed environments such
261 :     as Cygwin where scripts look and work like Unix scripts, but
262 :     where OS.Process.system cannot run them.
263 :    
264 :     - In CM, the srcpath pickler used native pathname syntax -- which
265 :     is incorrect in the case of cross-compilation. The new pickle format
266 :     is independent of platform-specific naming conventions.
267 :    
268 :     - Path configuration files (such as lib/pathconfig) can now choose
269 :     between native and standard syntax. Placing a line of the form
270 :    
271 :     standard!
272 :    
273 :     into the file causes all subsequent paths to be interpreted using
274 :     CM standard pathname syntax (= Unix conventions); a line
275 :    
276 :     native!
277 :    
278 :     switches back to native style. This was needed so that
279 :     path config files can be written portably, see src/system/pathconfig.
280 :    
281 :     - Runtime system:
282 :    
283 :     - win32-filesys.c: get_file_time and set_file_time now
284 :     access modification time, not creation time.
285 :    
286 :     - I/O code made aware of new array representation.
287 :    
288 :     - Bug fixes in X86.prim.masm.
289 :    
290 :     - src/system/makeml made aware of win32. (For use under cygwin
291 :     and other Unix-environments for windows.)
292 :    
293 :     - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
294 :     which caused BinIO.inputAll to fail consistently.
295 :    
296 :     .bat scripts:
297 :    
298 :     Windows .bat scripts assume that SMLNJ_HOME is defined.
299 :    
300 :     - sml.bat, ml-yacc.bat, ml-lex.bat: Driver scripts for standalone
301 :     applications (sml, ml-yacc, ml-lex).
302 :     - ml-build.bat: analogous to ml-build.
303 :     - config\install.bat: Analogous to config/install.sh. This requires
304 :     that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
305 :     (nmake etc. must be on the path, and vcvars32 must have been run.)
306 :     Moreover, sources for ml-lex and ml-yacc need to exist under src,
307 :     and the bootfile hierarchy must have been unpacked under
308 :     sml.boot.x86-win32.
309 :     The script is very primitive and does a poor job at error checking.
310 :     It only installs the base system, ml-lex, and ml-yacc. No other
311 :     libraries are being installed (i.e., you get only those that
312 :     are part of the compiler.)
313 :     - link-sml.bat: analogous to .link-sml, but not currently used
314 :    
315 :     Unrelated bug fixes:
316 :    
317 :     - ml-nlffigen now exports structures ST_* corresponding to incomplete
318 :     types.
319 :     - Added getDevice to PP/src/pp-debug-fn.sml. (Would not compile
320 :     otherwise.)
321 :    
322 :     ----------------------------------------------------------------------
323 : macqueen 1337 Name: Dave MacQueen (macqueen@cs.uchicago.edu)
324 :     Date: 2003/06/17
325 :     Tag: macqueen-20030617-bug895
326 :     Description:
327 :    
328 :     Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
329 :     Tag will be used for new development branch (mcz-branch) for
330 :     use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
331 :    
332 :     ----------------------------------------------------------------------
333 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
334 : mblume 1335 Date: 2003/05/27 16:55:00 CDT
335 :     Tag: blume-20030527-polyeq
336 :     Description:
337 :    
338 :     Tried to eliminated most cases of polymorphic equality.
339 :    
340 :     ----------------------------------------------------------------------
341 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
342 : mblume 1334 Date: 2003/05/21 17:45:00 CDT
343 :     Tag: blume-20030517-complete
344 :     Description:
345 :    
346 :     Two changes:
347 :    
348 :     1. Added a flag for controlling whether non-exhaustive bindings will
349 :     be treated as errors (default is false).
350 :     2. Cleaned up the *entire* source tree so that CMB.make goes through
351 :     without a single non-exhaustive match- or bind warning.
352 :    
353 :     ----------------------------------------------------------------------
354 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
355 : mblume 1332 Date: 2003/05/17 10:20:00 CDT
356 :     Tag: blume-20030517-absyn
357 :     Description:
358 :    
359 :     1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
360 :    
361 :     This mainly affects the quality of error messages. However, some
362 :     of the code is now more straightforward than before. (Treatment of
363 :     the above four constructs in translate.sml is much simpler than
364 :     the "macro-expansion" that was going on before. Plus, the mach-
365 :     compiler no longer gets invoked just to be able to compile an
366 :     if-expression.)
367 :    
368 :     2. The ErrorMsg.Error exception is now caught and absorbed by the
369 :     interactive loop.
370 :    
371 :     ----------------------------------------------------------------------
372 : allenleung 1327 Name: Allen Leung
373 :     Date: 2003/05/16 13:05:00 CDT
374 :     Tag: leunga-20030516-cygwin-runtime
375 :     Description:
376 :    
377 :     Ported the runtime system to cygwin, which uses the unix
378 :     x86-unix bin files. Missing/buggy features:
379 :    
380 :     o getnetbyname, getnetbyaddr: these functions seem to be missing in
381 :     the Cygwin library.
382 :     o Ctrl-C handling may be flaky.
383 :     o Windows system calls and Windows I/O are not supported.
384 :    
385 :     A new set of binfiles is located at:
386 :    
387 :     http://www.dorsai.org/~leunga/boot.x86-unix.tgz
388 :    
389 :     This is only needed for bootstrapping the cygwin version of smlnj.
390 :     Other x86 versions can use the existing binfiles.
391 :    
392 :     ----------------------------------------------------------------------
393 : mblume 1317 Name: Matthias Blume
394 :     Date: 2003/04/08 15:42:00 CDT
395 :     Tag: blume-20030408-listpair
396 :     Description:
397 :    
398 :     1. Added a target 'mlrisc' to installer.
399 :    
400 :     2. Added missing elements to structure ListPair.
401 :    
402 :     ----------------------------------------------------------------------
403 : leunga 1296 Name: Allen Leung
404 :     Date: 2003/01/07 10:40:00 EST
405 :     Tag: leunga-20030107-int-rem
406 :     Description:
407 :    
408 :     Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
409 :     The arguments to the SUBL instruction were swapped.
410 :    
411 :     ----------------------------------------------------------------------
412 : blume 1229 Name: Matthias Blume
413 : blume 1294 Date: 2002/12/12 16:25:00 EST
414 :     Tag: blume-20021212-risc-ra
415 :     Description:
416 :    
417 :     Fixed a serious bug in the rewrite code for FP spilling/reloading that
418 :     sent the RA into an infinite loop when floating point registers get
419 :     spilled. (Because of this bug, e.g., nucleic stopped compiling between
420 :     110.37 and 110.38.)
421 :     There was another set of potential problems related to the handling of
422 :     MLRISC annotations (but those did not yet cause real problems, apparently).
423 :    
424 :     ----------------------------------------------------------------------
425 :     Name: Matthias Blume
426 : blume 1291 Date: 2002/12/06 22:40:00 EST
427 :     Tag: blume-20021206-cm-fileid
428 :     Description:
429 :    
430 :     Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
431 :     This fixes the problem of CM getting confused by files that suddenly
432 :     change their identity (e.g., by getting unlinked and recreated by some
433 :     text editor such as vi). There might be a better/cheaper/cleaner way
434 :     of doing this, but for now this will have to do.
435 :    
436 :     ----------------------------------------------------------------------
437 :     Name: Matthias Blume
438 : blume 1288 Date: 2002/10/28 09:50:00 EST
439 :     Tag: blume-20021028-typecheck
440 :     Description:
441 :    
442 :     Exported structure Typecheck from $smlnj/viscomp/core.cm.
443 :    
444 :     ----------------------------------------------------------------------
445 :     Name: Matthias Blume
446 : blume 1282 Date: 2002/10/17 09:10:00 EDT
447 : blume 1279 Tag: Release_110_42
448 :     Description:
449 :    
450 : blume 1282 In good old tradition, there has been a slight hiccup so that we have
451 :     to patch 110.42 after the fact. The old release tag has been replaced
452 :     (see below).
453 :    
454 :     The change solves a problem with two competing approaches the
455 :     configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
456 :     each other's way.
457 :    
458 :     This change only affects the runtime system code and the installer script.
459 :     (No new bootfiles.)
460 :    
461 :     ----------------------------------------------------------------------
462 :     Name: Matthias Blume
463 :     Date: 2002/10/16 12:00:00 EDT
464 :     Tag: Release_110_42_removed
465 :     Description:
466 :    
467 : blume 1279 New working release. New bootfiles.
468 :    
469 :     ----------------------------------------------------------------------
470 :     Name: Matthias Blume
471 : blume 1273 Date: 2002/10/10 13:10:00 EDT
472 :     Tag: blume-20021010-ppc-divs
473 :     Description:
474 :    
475 :     The mltree operator DIVS must be implemented with an overflow check on
476 :     the PPC because the hardware indicates divide-by-zero using "overflow" as
477 :     well.
478 :    
479 :     ----------------------------------------------------------------------
480 :     Name: Matthias Blume
481 : blume 1261 Date: 2002/07/23 11:45:00
482 :     Tag: blume-20020723-smlnj-home
483 :     Description:
484 :    
485 :     Sml now senses the SMLNJ_HOME environment variable. If this is set,
486 :     then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
487 :     CM_PATHCONFIG is also set), the path configuration file is assumed
488 :     to be in $SMLNJ_HOME/lib/pathconfig. This way one can easily move
489 :     the entire tree to some other place and everything will "just work".
490 :    
491 :     (Companion commands such as ml-build and ml-makedepend also sense this
492 :     variable.)
493 :    
494 :     ----------------------------------------------------------------------
495 :     Name: Matthias Blume
496 : blume 1259 Date: 2002/07/12 21:19:00 EDT
497 :     Tag: blume-20020712-liveness
498 :     Description:
499 :    
500 :     Exported two useful "step" functions from liveness module (MLRISC).
501 :    
502 :     ----------------------------------------------------------------------
503 :     Name: Matthias Blume
504 : blume 1253 Date: 2002/07/05 16:00 EDT
505 :     Tag: Release_110_41
506 :     Description:
507 :    
508 :     New working release. New bootfiles.
509 :    
510 :     ----------------------------------------------------------------------
511 :     Name: Matthias Blume
512 : blume 1250 Date: 2002/07/05 10:25:00 EDT
513 :     Tag: blume-20020705-btimp
514 :     Description:
515 :    
516 :     Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
517 :     other clients can set up backtracing support.
518 :    
519 :     ----------------------------------------------------------------------
520 :     Name: Matthias Blume
521 : blume 1249 Date: 2002/06/25 17:23:00 EDT
522 :     Tag: blume-20020625-fpmax
523 :     Description:
524 :    
525 :     Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
526 :     case. (The sense of the isNaN test was reversed -- which made min and
527 :     max always return their first argument.)
528 :    
529 :     ----------------------------------------------------------------------
530 :     Name: Matthias Blume
531 : blume 1240 Date: 2002/06/11
532 :     Tag: blume-20020611-unixpath
533 :     Description:
534 :    
535 :     Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
536 :    
537 :     ----------------------------------------------------------------------
538 :     Name: Matthias Blume
539 : blume 1238 Date: 2002/06/10 16:35:00 EDT
540 :     Tag: blume-20020610-ieeereal
541 :     Description:
542 :    
543 :     I back-ported my implementation of IEEEReal.fromString from the
544 :     idlbasis-devel branch so that we can test it.
545 :    
546 :     Another small change is that ppDec tries to give more information
547 :     than just "<sig>" in the case of functors. However, this code is
548 :     broken in some mysterious way if the functor's body's signature
549 :     has not been declared by ascription but gets inferred from the
550 :     implementation. This needs fixing...
551 :    
552 :     ----------------------------------------------------------------------
553 :     Name: Matthias Blume
554 : blume 1229 Date: 2002/05/31
555 :     Tag: blume-20020531-btrace-mode
556 :     Description:
557 :    
558 :     Resurrected SMLofNJ.Internals.BTrace.mode. (It accidentally fell by
559 :     the wayside when I switched over to using Controls everywhere.)
560 :    
561 :     ----------------------------------------------------------------------
562 : george 1222 Name: Lal George
563 :     Date: 2002/05/23 12:21:40 EDT
564 :     Tag: george-20020523-visual-labels
565 :     Description:
566 :    
567 :     Labels are now displayed in the graphical output to make
568 :     the fall-through and target blocks obvious.
569 :    
570 :     ----------------------------------------------------------------------
571 : blume 1204 Name: Matthias Blume
572 : blume 1220 Date: 2002/05/22 11:03:00 EDT
573 :     Tag: blume-20020522-shrink
574 :     Description:
575 :    
576 :     John tweaked yesterday's fix for 1131 to handle an out-of-memory
577 :     situation that comes up when allocating huge arrays.
578 :    
579 :     ----------------------------------------------------------------------
580 :     Name: Matthias Blume
581 : blume 1215 Date: 2002/05/21 16:00:00 EDT
582 :     Tag: Release_110_40
583 :     Description:
584 :    
585 :     New working release (110.40). New bootfiles.
586 :    
587 :     [Also: John Reppy fixed GC bug 1131.]
588 :    
589 :     ----------------------------------------------------------------------
590 :     Name: Matthias Blume
591 : blume 1212 Date: 2002/05/21 12:35:00 EDT
592 :     Tag: blume-20020521-cmdoc
593 :     Description:
594 :    
595 :     CM documentation update.
596 :    
597 :     ----------------------------------------------------------------------
598 :     Name: Matthias Blume
599 : blume 1211 Date: 2002/05/21 10:55:00 EDT
600 :     Tag: blume-20020521-misc
601 :     Description:
602 :    
603 :     - John tweaked runtime to be silent on heap export (except when
604 :     GC messages are on).
605 :     - I added a few more things (cross-compiling versions of CMB) to
606 :     config/preloads (as suggestions).
607 :    
608 :     ----------------------------------------------------------------------
609 :     Name: Matthias Blume
610 : blume 1208 Date: 2002/05/20 22:25:00 EDT
611 :     Tag: blume-20020520-controls
612 :     Description:
613 :    
614 :     - Added ControlUtil structure to control-lib.cm.
615 :     - Use it throughout.
616 :     - Used Controls facility to define MLRISC controls (as opposed to
617 :     registering MLRISC control ref cells with Controls after the fact)
618 :     - Fixed messed-up controls priorities.
619 :    
620 :     * Removed again all the stuff from config/preloads that one wouldn't
621 :     be able to preload at the time the initial heap image is built.
622 :     (Many libraries, e.g., CML, do not exist yet at this time. The
623 :     only libraries that can be preloaded via config/preloads are those
624 :     that come bundled with the bootfiles.)
625 :    
626 :     ----------------------------------------------------------------------
627 :     Name: Matthias Blume
628 : blume 1204 Date: 2002/05/20 10:59:00 EDT
629 :     Tag: blume-20020520-preloads
630 :     Description:
631 :    
632 :     Added a lot of commented-out suggestions for things to be included
633 :     in config/preloads.
634 :    
635 :     ----------------------------------------------------------------------
636 : leunga 1203 Name: Allen Leung
637 :     Date: 2002/05/18 14:20:00 EDT
638 :     Tag: leunga-20020518-mdl
639 :     Description:
640 :    
641 :     o Made the mdl tool stuff compile and run again.
642 :     o I've disabled all the stuff that depends on RTL specifications; they
643 :     are all badly broken anyway.
644 :    
645 :     ----------------------------------------------------------------------
646 : blume 1201 Name: Matthias Blume
647 :     Date: 2002/05/17 16:49:00 EDT
648 :     Tag: blume-20020517-controls
649 :     Description:
650 :    
651 :     0. John Reppy made several modifications to the SML/NJ library.
652 :     In particular, there is a shiny new controls-lib.cm.
653 :    
654 :     1. Pushed new controls interface through compiler so that everything
655 :     compiles again.
656 :    
657 :     2. Added FormatComb and FORMAT_COMB to the CML version of the
658 :     SML/NJ library (so that CML compiles again).
659 :    
660 :     3. Modified init scripts because XXX_DEFAULT environment variables
661 :     are no longer with us. (Boot-time initialization is now done
662 :     using the same environment variables that are also used for
663 :     startup-time initialization of controls.)
664 :    
665 :     ----------------------------------------------------------------------
666 : george 1192 Name: Lal George
667 :     Date: 2002/05/15 09:20:10 EDT
668 :     Tag: george-20020515-pseudo-op-decls
669 :     Description:
670 :    
671 :     All pseudo-ops emitted before the first segment declaration
672 :     such as TEXT, DATA, and BSS directives are assumed to be global
673 :     declarations and are emitted first in the assembly file. This is
674 :     useful in a number of situations where one has pseudo-ops that are not
675 :     specific to any segment, and also works around the constraint that one
676 :     cannot have client pseudo-ops in the TEXT segment.
677 :    
678 :     Because no segment is associated with these declarations it is
679 :     an error to allocate any space or objects before the first segment
680 :     directive and an exception will be raised. However, we cannot make
681 :     this check for client pseudo-ops.
682 :    
683 :     These top level declarations are a field in the CFG graph_info.
684 :     In theory you can continue to add to this field after the CFG has been
685 :     built -- provided you know what you are doing;-)
686 :    
687 :     ----------------------------------------------------------------------
688 : blume 1175 Name: Matthias Blume
689 : blume 1190 Date: 2002/05/13 16:40:00 EDT
690 :     Tag: blume-20020513-pp-etc
691 :     Description:
692 :    
693 :     A few minor bugfixes:
694 :    
695 :     - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
696 :     (Bogus printouts for redefined bindings still occur. Compiler
697 :     bug should no longer occur now. We need to redo the prettyprinter
698 :     from scratch.)
699 :    
700 :     - CM pathname printer now also adds escape sequences for ( and )
701 :    
702 :     - commend and docu fixes for ml-nlffi
703 :    
704 :     ----------------------------------------------------------------------
705 :     Name: Matthias Blume
706 : blume 1189 Date: 2002/05/10 16:40:00 EDT
707 :     Tag: blume-20020510-erg-textio
708 :     Description:
709 :    
710 :     Applied the following bugfix provided by Emden Gansner:
711 :    
712 :     Output is corrupted when outputSubstr is used rather than output.
713 :    
714 :     The problem occurs when a substring
715 :    
716 :     ss = (s, dataStart, dataLen)
717 :    
718 :     where dataStart > 0, fills a stream buffer with avail bytes left.
719 :     avail bytes of s, starting at index dataStart, are copied into the
720 :     buffer, the buffer is flushed, and then the remaining dataLen-avail
721 :     bytes of ss are copied into the beginning of the buffer. Instead of
722 :     starting this copy at index dataStart+avail in s, the current code
723 :     starts the copy at index avail.
724 :    
725 :     Fix:
726 :     In text-io-fn.sml, change line 695 from
727 :     val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
728 :     to
729 :     val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
730 :    
731 :     ----------------------------------------------------------------------
732 :     Name: Matthias Blume
733 : blume 1186 Date: 2002/04/12 13:55:00 EDT
734 :     Tag: blume-20020412-assyntax
735 :     Description:
736 :    
737 :     1. Grabbed newer assyntax.h from the XFree86 project.
738 :     2. Fiddled with how to compile X86.prim.asm without warnings.
739 :     3. (Very) Minor cleanup in CM.
740 :    
741 :     ----------------------------------------------------------------------
742 :     Name: Matthias Blume
743 : blume 1185 Date: 2002/04/01 (no joke!) 17:07:00 EST
744 :     Tag: blume-20020401-x86div
745 :     Description:
746 :    
747 :     Added full support for div/mod/rem/quot on the x86, using the machine
748 :     instruction's two results (without clumsily recomputing the remainder)
749 :     directly where appropriate.
750 :    
751 :     Some more extensive power-of-two support was added to the x86 instruction
752 :     selector (avoiding expensive divs, mods, and muls where they can be
753 :     replaced with cheaper shifts and masks). However, this sort of thing
754 :     ought to be done earlier, e.g., within the CPS optimizer so that
755 :     all architectures benefit from it.
756 :    
757 :     The compiler compiles to a fixed point, but changes might be somewhat
758 :     fragile nevertheless. Please, report any strange things that you might
759 :     see wrt. div/mod/quot/rem...
760 :    
761 :     ----------------------------------------------------------------------
762 :     Name: Matthias Blume
763 : blume 1184 Date: 2002/03/29 17:22:00
764 :     Tag: blume-20020329-div
765 :     Description:
766 :    
767 :     Fixed my broken div/mod logic. Unfortunately, this means that the
768 :     inline code for div/mod now has one more comparison than before.
769 :     Fast paths (quotient > 0 or remainder = 0) are not affected, though.
770 :     The problem was with quotient = 0, because that alone does not tell
771 :     us which way the rounding went. One then has to look at whether
772 :     remainder and divisor have the same sign... :(
773 :    
774 :     Anyway, I replaced the bootfiles with fresh ones...
775 :    
776 :     ----------------------------------------------------------------------
777 :     Name: Matthias Blume
778 : blume 1183 Date: 2002/03/29 14:10:00 EST
779 :     Tag: blume-20020329-inlprims
780 :     Description:
781 :    
782 :     NEW BOOTFILES!!! Version number bumped to 110.39.3.
783 :    
784 :     Primops have changed. This means that the bin/boot-file formats have
785 :     changed as well.
786 :    
787 :     To make sure that there is no confusion, I made a new version.
788 :    
789 :    
790 :     CHANGES:
791 :    
792 :     * removed REMT from mltree (remainder should never overflow).
793 :    
794 :     * added primops to deal with divisions of all flavors to the frontend
795 :    
796 :     * handled these primops all the way through so they map to their respective
797 :     MLRISC support
798 :    
799 :     * used these primops in the implementation of Int, Int32, Word, Word32
800 :    
801 :     * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
802 :    
803 :     * parameterized INLMIN, INLMAX, and INLABS by a numkind
804 :    
805 :     * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
806 :     floating point
807 :    
808 :     * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
809 :     and Real (but Real.abs maps to a separate floating-point-only primop)
810 :    
811 :    
812 :     TODO items:
813 :    
814 :     * Hacked Alpha32 instruction selection, disabling the selection of REMx
815 :     instructions because the machine instruction encoder cannot handle
816 :     them. (Hppa, PPC, and Sparc instruction selection did not handle
817 :     REM in the first place, and REM is supported by the x86 machine coder.)
818 :    
819 :     * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
820 :     selection phase. (The two can be streamlined because the hardware
821 :     delivers both quotient and remainder at the same time anyway.)
822 :    
823 :     * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
824 :     (Currently the behavior is inconsistent both across architectures and
825 :     wrt. the draft Basis spec.)
826 :    
827 :     * Word8 should eventually be handled natively, too.
828 :    
829 :     * There seems to be one serious bug in mltree-gen.sml. It appears, though,
830 :     as if there currently is no execution path that could trigger it in
831 :     SML/NJ. (The assumptions underlying functions arith and promotable do not
832 :     hold for things like multiplication and division.)
833 :    
834 :     ----------------------------------------------------------------------
835 :     Name: Matthias Blume
836 : blume 1181 Date: 2002/03/27 16:27:00 EST
837 :     Tag: blume-20020327-mlrisc-divisions
838 :     Description:
839 :    
840 :     Added support for all four division operations (ML's div, mod, quot,
841 :     and rem) to MLRISC. In the course of doing so, I also rationalized
842 :     the naming (no more annoying switch-around of DIV and QUOT), by
843 :     parameterizing the operation by div_rounding_mode (which can be either
844 :     DIV_TO_ZERO or DIV_TO_NEGINF).
845 :    
846 :     The generic MLTreeGen functor takes care of compiling all four
847 :     operations down to only round-to-zero div.
848 :    
849 :     Missing pieces:
850 :    
851 :     * Doing something smarter than relying on MLTreeGen on architectures
852 :     like, e.g., the x86 where hardware division delivers both quotient and
853 :     remainder at the same time. With this, the implementation of the
854 :     round-to-neginf operations could be further streamlined.
855 :    
856 :     * Remove inlining support for div/mod/rem from the frontend and replace it
857 :     with primops that get carried through to the backend. Do this for all
858 :     int and word types.
859 :    
860 :     ----------------------------------------------------------------------
861 :     Name: Matthias Blume
862 : blume 1180 Date: 2002/03/25 17:25:00 EST
863 :     Tag: blume-20020325-divmod
864 :     Description:
865 :    
866 :     I improved (hopefully without breaking them) the implementation of Int.div,
867 :     Int.mod, and Int.rem. For this, the code in translate.sml now takes
868 :     advantage of the following observations:
869 :    
870 :     Let q = x quot y r = x rem y
871 :     d = x div y m = x mod y
872 :    
873 :     where "quot" is the round-to-zero version of integer division that
874 :     hardware usually provides. Then we have:
875 :    
876 :     r = x - q * y where neither the * nor the - will overflow
877 :     d = if q >= 0 orelse x = q * y then q else q - 1
878 :     where neither the * nor the - will overflow
879 :     m = if q >= 0 orelse r = 0 then r else r + y
880 :     where the + will not overflow
881 :    
882 :     This results in substantial simplification of the generated code.
883 :     The following table shows the number of CFG nodes and edges generated
884 :     for
885 :     fun f (x, y) = x OPER y
886 :     (* with OPER \in div, mod, quot, rem *)
887 :    
888 :    
889 :     OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
890 :     --------------------------------------------------------
891 :     div | 24 | 39 | 12 | 16
892 :     mod | 41 | 71 | 12 | 16
893 :     quot | 8 | 10 | 8 | 10
894 :     rem | 10 | 14 | 8 | 10
895 :    
896 :    
897 :     ----------------------------------------------------------------------
898 :     Name: Matthias Blume
899 : blume 1179 Date: 2002/03/25 22:06:00 EST
900 :     Tag: blume-20020325-cprotobug
901 :     Description:
902 :    
903 :     Fixed a bug in cproto (c prototype decoder).
904 :    
905 :     ----------------------------------------------------------------------
906 :     Name: Matthias Blume
907 : blume 1178 Date: 2002/03/25 16:00:00 EST
908 :     Tag: blume-20020325-raw-primops
909 :     Description:
910 :    
911 :     I did some cleanup to Allen's new primop code and
912 :     replaced yesterday's bootfiles with new ones.
913 :     (But they are stored in the same place.)
914 :    
915 :     ----------------------------------------------------------------------
916 :     Name: Matthias Blume
917 : blume 1175 Date: 2002/03/24 22:40:00 EST
918 :     Tag: blume-20020324-bootfiles
919 :     Description:
920 :    
921 :     Made the bootfiles that Allen asked for.
922 :    
923 :     ----------------------------------------------------------------------
924 : leunga 1174 Name: Allen Leung
925 :     Date: 2002/03/23 15:50:00 EST
926 :     Tag: leunga-20020323-flint-cps-rcc-primops
927 :     Description:
928 :    
929 :     1. Changes to FLINT primops:
930 :    
931 :     (* make a call to a C-function;
932 :     * The primop carries C function prototype information and specifies
933 :     * which of its (ML-) arguments are floating point. C prototype
934 :     * information is for use by the backend, ML information is for
935 :     * use by the CPS converter. *)
936 :     | RAW_CCALL of { c_proto: CTypes.c_proto,
937 :     ml_args: ccall_type list,
938 :     ml_res_opt: ccall_type option,
939 :     reentrant : bool
940 :     } option
941 :     (* Allocate uninitialized storage on the heap.
942 :     * The record is meant to hold short-lived C objects, i.e., they
943 :     * are not ML pointers. With the tag, the representation is
944 :     * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
945 :     *)
946 :     | RAW_RECORD of {tag:bool,sz:int}
947 :     and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
948 :    
949 :     2. These CPS primops are now overloaded:
950 :    
951 :     rawload of {kind:numkind}
952 :     rawstore of {kind:numkind}
953 :    
954 :     The one argument form is:
955 :    
956 :     rawload {kind} address
957 :    
958 :     The two argument form is:
959 :    
960 :     rawload {kind} [ml object, byte-offset]
961 :    
962 :     3. RAW_CCALL/RCC now takes two extra arguments:
963 :    
964 :     a. The first is whether the C call is reentrant, i.e., whether
965 :     ML state should be saved and restored.
966 :     b. The second argument is a string argument specifying the name of
967 :     library and the C function.
968 :    
969 :     These things are currently not handled in the code generator, yet.
970 :    
971 :     4. In CProto,
972 :    
973 :     An encoding type of "bool" means "ml object" and is mapped into
974 :     C prototype of PTR. Note that "bool" is different than "string",
975 :     even though "string" is also mapped into PTR, because "bool"
976 :     is assigned an CPS type of BOGt, while "string" is assigned INT32t.
977 :    
978 :     5. Pickler/unpicker
979 :    
980 :     Changed to handle RAW_RECORD and newest RAW_CCALL
981 :    
982 :     6. MLRiscGen,
983 :    
984 :     1. Changed to handle the new rawload/rawstore/rawrecord operators.
985 :     2. Code for handling C Calls has been moved to a new module CPSCCalls,
986 :     in the file CodeGen/cpscompile/cps-c-calls.sml
987 :    
988 :     7. Added the conditional move operator
989 :    
990 :     condmove of branch
991 :    
992 :     to cps. Generation of this is still buggy so it is currently
993 :     disabled.
994 :    
995 :     ----------------------------------------------------------------------
996 : george 1168 Name: Lal George
997 :     Date: 2002/03/22 14:18:25 EST
998 : george 1171 Tag: george-20020322-cps-branch-prob
999 : george 1168 Description:
1000 :    
1001 : george 1171 Implemented the Ball-Larus branch prediction-heuristics, and
1002 : george 1168 incorporated graphical viewers for control flow graphs.
1003 :    
1004 : george 1171 Ball-Larus Heuristics:
1005 : george 1168 ---------------------
1006 :     See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
1007 :    
1008 :     By design it uses the Dempster-Shafer theory for combining
1009 :     probabilities. For example, in the function:
1010 :    
1011 :     fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
1012 :    
1013 :     the ball-larus heuristics predicts that the n=0 is unlikely
1014 :     (OH-heuristic), and the 'then' branch is unlikely because of the
1015 :     RH-heuristic -- giving the 'then' branch an even lower combined
1016 : george 1171 probability using the Dempster-Shafer theory.
1017 : george 1168
1018 :     Finally, John Reppy's loop analysis in MLRISC, further lowers the
1019 :     probability of the 'then' branch because of the loop in the else
1020 :     branch.
1021 :    
1022 :    
1023 :     Graphical Viewing:
1024 :     ------------------
1025 :     I merely plugged in Allen's graphical viewers into the compiler. The
1026 :     additional code is not much. At the top level, saying:
1027 :    
1028 :     Control.MLRISC.getFlag "cfg-graphical-view" := true;
1029 :    
1030 :     will display the graphical view of the control flow graph just before
1031 :     back-patching. daVinci must be in your path for this to work. If
1032 :     daVinci is not available, then the default viewer can be changed
1033 :     using:
1034 :    
1035 :     Control.MLRISC.getString "viewer"
1036 :    
1037 :     which can be set to "dot" or "vcg" for the corresponding viewers. Of
1038 :     course, these viewers must be in your path.
1039 :    
1040 :     The above will display the compilation unit at the level of clusters,
1041 :     many of which are small, boring, and un-interesting. Also setting:
1042 :    
1043 :     Control.MLRISC.getInt "cfg-graphical-view_size"
1044 :    
1045 :     will display clusters that are larger than the value set by the above.
1046 :    
1047 :    
1048 :     ----------------------------------------------------------------------
1049 : blume 1159 Name: Matthias Blume
1050 :     Date: 2002/03/21 22:20:00 EST
1051 :     Tag: blume-20020321-kmp-bugfix
1052 :     Description:
1053 :    
1054 :     Changed the interface to the KMP routine in PreString and fixed
1055 :     a minor bug in one place where it was used.
1056 :    
1057 :     ----------------------------------------------------------------------
1058 : leunga 1156 Name: Allen Leung
1059 : leunga 1158 Date: 2002/03/21 20:30:00 EST
1060 :     Tag: leunga-20020321-cfg
1061 :     Description:
1062 :    
1063 :     Fixed a potential problem in cfg edge splitting.
1064 :    
1065 :     ----------------------------------------------------------------------
1066 :     Name: Allen Leung
1067 : leunga 1156 Date: 2002/03/21 17:15:00 EST
1068 :     Tag: leunga-20020321-x86-fp-cfg
1069 :     Description:
1070 :    
1071 :     1. Recoded the buggy parts of x86-fp.
1072 :    
1073 :     a. All the block reordering code has been removed.
1074 :     We now depend on the block placement phases to do this work.
1075 :    
1076 :     b. Critical edge splitting code has been simplified and moved into the
1077 :     CFG modules, as where they belong.
1078 :    
1079 :     Both of these were quite buggy and complex. The code is now much, much
1080 :     simpler.
1081 :    
1082 :     2. X86 backend.
1083 :    
1084 :     a. Added instructions for 64-bit support. Instruction selection for
1085 :     64-bit has not been committed, however, since that
1086 :     requires changes to MLTREE which haven't been approved by
1087 :     Lal and John.
1088 :    
1089 :     b. Added support for FUCOMI and FUCOMIP when generating code for
1090 :     PentiumPro and above. We only generate these instructions in
1091 :     the fast-fp mode.
1092 :    
1093 :     c. Added cases for JP and JNP in X86FreqProps.
1094 :    
1095 :     3. CFG
1096 :    
1097 :     CFG now has a bunch of methods for edge splitting and merging.
1098 :    
1099 :     4. Machine description.
1100 :    
1101 :     John's simplification of MLTREE_BASIS.fcond broke a few machine
1102 :     description things:
1103 :    
1104 :     rtl-build.{sig,sml} and hppa.mdl fixed.
1105 :    
1106 :     NOTE: the machine description stuff in the repository is still broken.
1107 :     Again, I can't put my fixes in because that involves
1108 :     changes to MLTREE.
1109 :    
1110 :     ----------------------------------------------------------------------
1111 : blume 1145 Name: Matthias Blume
1112 : blume 1155 Date: 2002/03/20 15:55:00 EST
1113 :     Tag: blume-20020320-kmp
1114 :     Description:
1115 :    
1116 :     Implemented Knuth-Morris-Pratt string matching in PreString and used
1117 :     it for String.isSubstring, Substring.isSubstring, and
1118 :     Substring.position.
1119 :    
1120 :     (Might need some stress-testing. Simple examples worked fine.)
1121 :    
1122 :     ----------------------------------------------------------------------
1123 :     Name: Matthias Blume
1124 : blume 1152 Date: 2002/03/19 16:37:00 EST
1125 :     Tag: blume-20020319-witnesses
1126 :     Description:
1127 :    
1128 :     Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
1129 :    
1130 :     This implements a generic mechanism for changing constness qualifiers
1131 :     anywhere within big C types without resorting to outright "casts".
1132 :     (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
1133 :     modify the constness at the outermost level.)
1134 :     The implementation of "convert" is based on the idea of "witness"
1135 :     values -- values that are not used by the operation but whose types
1136 :     "testify" to their applicability. On the implementation side, "convert"
1137 :     is simply a projection (returning its second curried argument). With
1138 :     cross-module inlining, it should not result in any machine code being
1139 :     generated.
1140 :    
1141 :     ----------------------------------------------------------------------
1142 :     Name: Matthias Blume
1143 : blume 1148 Date: 2002/03/15 16:40:00 EST
1144 :     Tag: blume-20020315-basis
1145 :     Description:
1146 :    
1147 :     Provided (preliminary?) implementations for
1148 :    
1149 :     {String,Substring}.{concatWith,isSuffix,isSubstring}
1150 :    
1151 :     and
1152 :    
1153 :     Substring.full
1154 :    
1155 :     Those are in the Basis spec but they were missing in SML/NJ.
1156 :    
1157 :     ----------------------------------------------------------------------
1158 :     Name: Matthias Blume
1159 : blume 1145 Date: 2002/03/14 21:30:00 EST
1160 :     Tag: blume-20020314-controls
1161 :     Description:
1162 :    
1163 :     Controls:
1164 :     ---------
1165 :    
1166 :     1. Factored out the recently-added Controls : CONTROLS stuff and put
1167 :     it into its own library $/controls-lib.cm. The source tree for
1168 :     this is under src/smlnj-lib/Controls.
1169 :    
1170 :     2. Changed the names of types and functions in this interface, so they
1171 :     make a bit more "sense":
1172 :    
1173 :     module -> registry
1174 :     'a registry -> 'a group
1175 :    
1176 :     3. The interface now deals in ref cells only. The getter/setter interface
1177 :     is (mostly) gone.
1178 :    
1179 :     4. Added a function that lets one register an already-existing ref cell.
1180 :    
1181 :     5. Made the corresponding modifications to the rest of the code so that
1182 :     everything compiles again.
1183 :    
1184 :     6. Changed the implementation of Controls.MLRISC back to something closer
1185 :     to the original. In particular, this module (and therefore MLRISC)
1186 :     does not depend on Controls. There now is some link-time code in
1187 :     int-sys.sml that registers the MLRISC controls with the Controls
1188 :     module.
1189 :    
1190 :     CM:
1191 :     ---
1192 :    
1193 :     * One can now specify the lambda-split aggressiveness in init.cmi.
1194 :    
1195 :     ----------------------------------------------------------------------
1196 : leunga 1142 Name: Allen Leung
1197 :     Date: 2002/03/13 17:30:00 EST
1198 :     Tag: leunga-20020313-x86-fp-unary
1199 :     Description:
1200 :    
1201 :     Bug fix for:
1202 :    
1203 :     > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
1204 :     > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
1205 :     > - fun f(x,(y,z)) = Real.~ y;
1206 :     > [autoloading]
1207 :     > [autoloading done]
1208 :     > fchsl (%eax), 184(%esp)
1209 :     > Error: MLRisc bug: X86MCEmitter.emitInstr
1210 :     >
1211 :     > uncaught exception Error
1212 :     > raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
1213 :    
1214 :     The problem was that the code generator did not generate any fp registers
1215 :     in this case, and the ra didn't know that it needed to run the X86FP phase to
1216 :     translate the pseudo fp instruction. This only happened with unary fp
1217 :     operators in certain situations.
1218 :    
1219 :     ----------------------------------------------------------------------
1220 : blume 1137 Name: Matthias Blume
1221 : blume 1140 Date: 2002/03/13 14:00:00 EST
1222 :     Tag: blume-20020313-overload-etc
1223 :     Description:
1224 :    
1225 :     1. Added _overload as a synonym for overload for backward compatibility.
1226 :     (Control.overloadKW must be true for either version to be accepted.)
1227 :    
1228 :     2. Fixed bug in install script that caused more things to be installed
1229 :     than what was requested in config/targets.
1230 :    
1231 :     3. Made CM aware of the (_)overload construct so that autoloading
1232 :     works.
1233 :    
1234 :     ----------------------------------------------------------------------
1235 :     Name: Matthias Blume
1236 : blume 1138 Date: 2002/03/12 22:03:00 EST
1237 :     Tag: blume-20020312-url
1238 :     Description:
1239 :    
1240 :     Forgot to update BOOT and srcarchiveurl.
1241 :    
1242 :     ----------------------------------------------------------------------
1243 :     Name: Matthias Blume
1244 : blume 1137 Date: 2002/03/12 17:30:00 EST
1245 :     Tag: blume-20020312-version110392
1246 :     Description:
1247 :    
1248 :     Yet another version number bump (because of small changes to the
1249 :     binfile format). Version number is now 110.39.2. NEW BOOTFILES!
1250 :    
1251 :     Changes:
1252 :    
1253 :     The new pid generation scheme described a few weeks ago was overly
1254 :     complicated. I implemented a new mechanism that is simpler and
1255 :     provides a bit more "stability": Once CM has seen a compilation
1256 :     unit, it keeps its identity constant (as long as you do not delete
1257 :     those crucial CM/GUID/* files). This means that when you change
1258 : blume 1145 an interface, compile, then go back to the old interface, and
1259 : blume 1137 compile again, you arrive at the original pid.
1260 :    
1261 :     There now also is a mechanism that instructs CM to use the plain
1262 :     environment hash as a module's pid (effectively making its GUID
1263 :     the empty string). For this, "noguid" must be specified as an
1264 :     option to the .sml file in question within its .cm file.
1265 :     This is most useful for code that is being generated by tools such
1266 :     as ml-nlffigen (because during development programmers tend to
1267 :     erase the tool's entire output directory tree including CM's cached
1268 :     GUIDs). "noguid" is somewhat dangerous (since it can be used to locally
1269 :     revert to the old, broken behavior of SML/NJ, but in specific cases
1270 :     where there is no danger of interface confusion, its use is ok
1271 :     (I think).
1272 :    
1273 :     ml-nlffigen by default generates "noguid" annotations. They can be
1274 :     turned off by specifying -guid in its command line.
1275 :    
1276 :     ----------------------------------------------------------------------
1277 : george 1136 Name: Lal George
1278 :     Date: 2002/03/12 12 14:42:36 EST
1279 :     Tag: george-20020312-frequency-computation
1280 :     Description:
1281 :    
1282 :     Integrated jump chaining and static block frequency into the
1283 :     compiler. More details and numbers later.
1284 : leunga 1025
1285 : george 1003 ----------------------------------------------------------------------
1286 : george 1133 Name: Lal George
1287 :     Date: 2002/03/11 11 22:38:53 EST
1288 :     Tag: george-20020311-jump-chain-elim
1289 :     Description:
1290 :    
1291 :     Tested the jump chain elimination on all architectures (except the
1292 :     hppa). This is on by default right now and is profitable for the
1293 :     alpha and x86, however, it may not be profitable for the sparc and ppc
1294 :     when compiling the compiler.
1295 :    
1296 :     The gc test will typically jump to a label at the end of the cluster,
1297 :     where there is another jump to an external cluster containing the actual
1298 :     code to invoke gc. This is to allow factoring of common gc invocation
1299 :     sequences. That is to say, we generate:
1300 :    
1301 :     f:
1302 :     testgc
1303 :     ja L1 % jump if above to L1
1304 :    
1305 :     L1:
1306 :     jmp L2
1307 :    
1308 :    
1309 :     After jump chain elimination the 'ja L1' instructions is converted to
1310 :     'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
1311 :     up being implemented in their long form (if L2 is far away) using:
1312 :    
1313 :     jbe L3 % jump if below or equal to L3
1314 :     jmp L2
1315 :     L3:
1316 :     ...
1317 :    
1318 :    
1319 :     For large compilation units L2 may be far away.
1320 :    
1321 :    
1322 :     ----------------------------------------------------------------------
1323 : blume 1132 Name: Matthias Blume
1324 :     Date: 2002/03/11 13:30:00 EST
1325 :     Tag: blume-20020311-mltreeeval
1326 :     Description:
1327 :    
1328 :     A functor parameter was missing.
1329 :    
1330 :     ----------------------------------------------------------------------
1331 : leunga 1130 Name: Allen Leung
1332 : leunga 1131 Date: 2002/03/11 10:30:00 EST
1333 : leunga 1142 Tag: leunga-20020311-runtime-string0
1334 : leunga 1131 Description:
1335 :    
1336 :     The representation of the empty string now points to a
1337 :     legal null terminated C string instead of unit. It is now possible
1338 :     to convert an ML string into C string with InlineT.CharVector.getData.
1339 :     This compiles into one single machine instruction.
1340 :    
1341 :     ----------------------------------------------------------------------
1342 :     Name: Allen Leung
1343 : leunga 1130 Date: 2002/03/10 23:55:00 EST
1344 :     Tag: leunga-20020310-x86-call
1345 :     Description:
1346 :    
1347 :     Added machine generation for CALL instruction (relative displacement mode)
1348 :    
1349 :     ----------------------------------------------------------------------
1350 : blume 1128 Name: Matthias Blume
1351 :     Date: 2002/03/08 16:05:00
1352 :     Tag: blume-20020308-entrypoints
1353 :     Description:
1354 :    
1355 :     Version number bumped to 110.39.1. NEW BOOTFILES!
1356 :    
1357 :     Entrypoints: non-zero offset into a code object where execution should begin.
1358 :    
1359 :     - Added the notion of an entrypoint to CodeObj.
1360 :     - Added reading/writing of entrypoint info to Binfile.
1361 :     - Made runtime system bootloader aware of entrypoints.
1362 :     - Use the address of the label of the first function given to mlriscGen
1363 :     as the entrypoint. This address is currently always 0, but it will
1364 :     not be 0 once we turn on block placement.
1365 :     - Removed the linkage cluster code (which was The Other Way(tm) of dealing
1366 :     with entry points) from mlriscGen.
1367 :    
1368 :     ----------------------------------------------------------------------
1369 : leunga 1127 Name: Allen Leung
1370 :     Date: 2002/03/07 20:45:00 EST
1371 :     Tag: leunga-20020307-x86-cmov
1372 :     Description:
1373 :    
1374 :     Bug fixes for CMOVcc on x86.
1375 :    
1376 :     1. Added machine code generation for CMOVcc
1377 :     2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
1378 :     3. CMOVcc cannot have an immediate operand as argument.
1379 :    
1380 :     ----------------------------------------------------------------------
1381 : blume 1126 Name: Matthias Blume
1382 :     Date: 2002/03/07 16:15:00 EST
1383 :     Tag: blume-20020307-controls
1384 :     Description:
1385 :    
1386 :     This is a very large but mostly boring patch which makes (almost)
1387 :     every tuneable compiler knob (i.e., pretty much everything under
1388 :     Control.* plus a few other things) configurable via both the command
1389 :     line and environment variables in the style CM did its configuration
1390 :     until now.
1391 :    
1392 :     Try starting sml with '-h' (or, if you are brave, '-H')
1393 :    
1394 :     To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
1395 :     implements the underlying generic mechanism.
1396 :    
1397 :     The interface to some of the existing such facilities has changed somewhat.
1398 :     For example, the MLRiscControl module now provides mkFoo instead of getFoo.
1399 :     (The getFoo interface is still there for backward-compatibility, but its
1400 :     use is deprecated.)
1401 :    
1402 :     The ml-build script passes -Cxxx=yyy command-line arguments through so
1403 :     that one can now twiddle the compiler settings when using this "batch"
1404 :     compiler.
1405 :    
1406 :     TODO items:
1407 :    
1408 :     We should go through and throw out all controls that are no longer
1409 :     connected to anything. Moreover, we should go through and provide
1410 :     meaningful (and correct!) documentation strings for those controls
1411 :     that still are connected.
1412 :    
1413 :     Currently, multiple calls to Controls.new are accepted (only the first
1414 :     has any effect). Eventually we should make sure that every control
1415 :     is being made (via Controls.new) exactly once. Future access can then
1416 :     be done using Controls.acc.
1417 :    
1418 :     Finally, it would probably be a good idea to use the getter-setter
1419 :     interface to controls rather than ref cells. For the time being, both
1420 :     styles are provided by the Controls module, but getter-setter pairs are
1421 :     better if thread-safety is of any concern because they can be wrapped.
1422 :    
1423 :     *****************************************
1424 :    
1425 :     One bug fix: The function blockPlacement in three of the MLRISC
1426 :     backpatch files used to be hard-wired to one of two possibilities at
1427 :     link time (according to the value of the placementFlag). But (I
1428 :     think) it should rather sense the flag every time.
1429 :    
1430 :     *****************************************
1431 :    
1432 :     Other assorted changes (by other people who did not supply a HISTORY entry):
1433 :    
1434 :     1. the cross-module inliner now works much better (Monnier)
1435 :     2. representation of weights, frequencies, and probabilities in MLRISC
1436 :     changed in preparation of using those for weighted block placement
1437 :     (Reppy, George)
1438 :    
1439 :     ----------------------------------------------------------------------
1440 : george 1116 Name: Lal George
1441 : george 1124 Date: 2002/03/07 14:44:24 EST 2002
1442 :     Tag: george-20020307-weighted-block-placement
1443 :    
1444 :     Tested the weighted block placement optimization on all architectures
1445 :     (except the hppa) using AMPL to generate the block and edge frequencies.
1446 :     Changes were required in the machine properties to correctly
1447 :     categorize trap instructions. There is an MLRISC flag
1448 :     "weighted-block-placement" that can be used to enable weighted block
1449 :     placement, but this will be ineffective without block/edge
1450 :     frequencies (coming soon).
1451 :    
1452 :    
1453 :     ----------------------------------------------------------------------
1454 :     Name: Lal George
1455 : george 1116 Date: 2002/03/05 17:24:48 EST
1456 :     Tag: george-20020305-linkage-cluster
1457 :    
1458 : george 1124 In order to support the block placement optimization, a new cluster
1459 :     is generated as the very first cluster (called the linkage cluster).
1460 :     It contains a single jump to the 'real' entry point for the compilation
1461 :     unit. Block placement has no effect on the linkage cluster itself, but
1462 :     all the other clusters have full freedom in the manner in which they
1463 :     reorder blocks or functions.
1464 : george 1116
1465 :     On the x86 the typical linkage code that is generated is:
1466 :     ----------------------
1467 :     .align 2
1468 :     L0:
1469 :     addl $L1-L0, 72(%esp)
1470 : george 1124 jmp L1
1471 : george 1116
1472 :    
1473 :     .align 2
1474 :     L1:
1475 :     ----------------------
1476 :    
1477 :     72(%esp) is the memory location for the stdlink register. This
1478 :     must contain the address of the CPS function being called. In the
1479 : george 1124 above example, it contains the address of L0; before
1480 : george 1116 calling L1 (the real entry point for the compilation unit), it
1481 :     must contain the address for L1, and hence
1482 :    
1483 :     addl $L1-L0, 72(%esp)
1484 :    
1485 : george 1124 I have tested this on all architectures except the hppa.The increase
1486 :     in code size is of course negligible
1487 : george 1116
1488 :     ----------------------------------------------------------------------
1489 : leunga 1115 Name: Allen Leung
1490 :     Date: 2002/03/03 13:20:00 EST
1491 :     Tag: leunga-20020303-mlrisc-tools
1492 :    
1493 :     Added #[ ... ] expressions to mlrisc tools
1494 :    
1495 :     ----------------------------------------------------------------------
1496 : blume 1096 Name: Matthias Blume
1497 : blume 1098 Date: 2002/02/27 12:29:00 EST
1498 :     Tag: blume-20020227-cdebug
1499 :     Description:
1500 :    
1501 :     - made types in structure C and C_Debug to be equal
1502 :     - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
1503 :     - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
1504 :    
1505 :     ----------------------------------------------------------------------
1506 :     Name: Matthias Blume
1507 : blume 1096 Date: 2002/02/26 12:00:00 EST
1508 :     Tag: blume-20020226-ffi
1509 :     Description:
1510 :    
1511 :     1. Fixed a minor bug in CM's "noweb" tool:
1512 :     If numbering is turned off, then truly don't number (i.e., do not
1513 :     supply the -L option to noweb). The previous behavior was to supply
1514 :     -L'' -- which caused noweb to use the "default" line numbering scheme.
1515 :     Thanks to Chris Richards for pointing this out (and supplying the fix).
1516 :    
1517 :     2. Once again, I reworked some aspects of the FFI:
1518 :    
1519 :     A. The incomplete/complete type business:
1520 :    
1521 :     - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
1522 :     gone!
1523 :     - ML types representing an incomplete type are now *equal* to
1524 :     ML types representing their corresponding complete types (just like
1525 :     in C). This is still safe because ml-nlffigen will not generate
1526 :     RTTI for incomplete types, nor will it generate functions that
1527 :     require access to such RTTI. But when ML code generated from both
1528 :     incomplete and complete versions of the C type meet, the ML types
1529 :     are trivially interoperable.
1530 :    
1531 :     NOTE: These changes restore the full generality of the translation
1532 :     (which was previously lost when I eliminated functorization)!
1533 :    
1534 :     B. Enum types:
1535 :    
1536 :     - Structure C now has a type constructor "enum" that is similar to
1537 :     how the "su" constructor works. However, "enum" is not a phantom
1538 :     type because each "T enum" has values (and is isomorphic to
1539 :     MLRep.Signed.int).
1540 :     - There are generic access operations for enum objects (using
1541 :     MLRep.Signed.int).
1542 :     - ml-nlffigen will generate a structure E_foo for each "enum foo".
1543 :     * The structure contains the definition of type "mlrep" (the ML-side
1544 :     representation type of the enum). Normally, mlrep is the same
1545 :     as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
1546 :     then mlrep will be defined as a datatype -- thus facilitating
1547 :     pattern matching on mlrep values.
1548 :     ("-ec" will be suppressed if there are duplicate values in an
1549 :     enumeration.)
1550 :     * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
1551 :     will be generated for each C enum constant xxx.
1552 :     * Conversion functions m2i and i2m convert between mlrep and
1553 :     MLRep.Signed.int. (Without "-ec", these functions are identities.)
1554 :     * Coversion functions c and ml convert between mlrep and "tag enum".
1555 :     * Access functions (get/set) fetch and store mlrep values.
1556 :     - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
1557 :     enumerations are merged into one single enumeration represented by
1558 :     structure E_'.
1559 :    
1560 :     ----------------------------------------------------------------------
1561 : leunga 1080 Name: Allen Leung
1562 : leunga 1094 Date: 2002/02/25 04:45:00 EST
1563 :     Tag: leunga-20020225-cps-spill
1564 :    
1565 :     This is a new implementation of the CPS spill phase.
1566 :     The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
1567 :     In case of problems, replace it with the old file spill.sml
1568 :    
1569 :     The current compiler runs into some serious performance problems when
1570 :     constructing a large record. This can happen when we try to compile a
1571 :     structure with many items. Even a very simple structure like the following
1572 :     makes the compiler slow down.
1573 :    
1574 :     structure Foo = struct
1575 :     val x_1 = 0w1 : Word32.int
1576 :     val x_2 = 0w2 : Word32.int
1577 :     val x_3 = 0w3 : Word32.int
1578 :     ...
1579 :     val x_N = 0wN : Word32.int
1580 :     end
1581 :    
1582 :     The following table shows the compile time, from N=1000 to N=4000,
1583 :     with the old compiler:
1584 :    
1585 :     N
1586 :     1000 CPS 100 spill 0.04u 0.00s 0.00g
1587 :     MLRISC ra 0.06u 0.00s 0.05g
1588 :     (spills = 0 reloads = 0)
1589 :     TOTAL 0.63u 0.07s 0.21g
1590 :    
1591 :     1100 CPS 100 spill 8.25u 0.32s 0.64g
1592 :     MLRISC ra 5.68u 0.59s 3.93g
1593 :     (spills = 0 reloads = 0)
1594 :     TOTAL 14.71u 0.99s 4.81g
1595 :    
1596 :     1500 CPS 100 spill 58.55u 2.34s 1.74g
1597 :     MLRISC ra 5.54u 0.65s 3.91g
1598 :     (spills = 543 reloads = 1082)
1599 :     TOTAL 65.40u 3.13s 6.00g
1600 :    
1601 :     2000 CPS 100 spill 126.69u 4.84s 3.08g
1602 :     MLRISC ra 0.80u 0.10s 0.55g
1603 :     (spills = 42 reloads = 84)
1604 :     TOTAL 129.42u 5.10s 4.13g
1605 :    
1606 :     3000 CPS 100 spill 675.59u 19.03s 11.64g
1607 :     MLRISC ra 2.69u 0.27s 1.38g
1608 :     (spills = 62 reloads = 124)
1609 :     TOTAL 682.48u 19.61s 13.99g
1610 :    
1611 :     4000 CPS 100 spill 2362.82u 56.28s 43.60g
1612 :     MLRISC ra 4.96u 0.27s 2.72g
1613 :     (spills = 85 reloads = 170)
1614 :     TOTAL 2375.26u 57.21s 48.00g
1615 :    
1616 :     As you can see the old cps spill module suffers from some serious
1617 : leunga 1115 performance problem. But since I cannot decipher the old code fully,
1618 :     instead of patching the problems up, I'm reimplementing it
1619 : leunga 1094 with a different algorithm. The new code is more modular,
1620 :     smaller when compiled, and substantially faster
1621 :     (O(n log n) time and O(n) space). Timing of the new spill module:
1622 :    
1623 :     4000 CPS 100 spill 0.02u 0.00s 0.00g
1624 :     MLRISC ra 0.25u 0.02s 0.15g
1625 :     (spills=1 reloads=3)
1626 :     TOTAL 7.74u 0.34s 1.62g
1627 :    
1628 :     Implementation details:
1629 :    
1630 :     As far as I can tell, the purpose of the CPS spill module is to make sure the
1631 :     number of live variables at any program point (the bandwidth)
1632 :     does not exceed a certain limit, which is determined by the
1633 :     size of the spill area.
1634 :    
1635 :     When the bandwidth is too large, we decrease the register pressure by
1636 :     packing live variables into spill records. How we achieve this is
1637 :     completely different than what we did in the old code.
1638 :    
1639 :     First, there is something about the MLRiscGen code generator
1640 :     that we should be aware of:
1641 :    
1642 :     o MLRiscGen performs code motion!
1643 :    
1644 :     In particular, it will move floating point computations and
1645 :     address computations involving only the heap pointer to
1646 :     their use sites (if there is only a single use).
1647 :     What this means is that if we have a CPS record construction
1648 :     statement
1649 :    
1650 :     RECORD(k,vl,w,e)
1651 :    
1652 :     we should never count the new record address w as live if w
1653 :     has only one use (which is often the case).
1654 :    
1655 :     We should do something similar to floating point, but the transformation
1656 :     there is much more complex, so I won't deal with that.
1657 :    
1658 :     Secondly, there are now two new cps primops at our disposal:
1659 :    
1660 :     1. rawrecord of record_kind option
1661 :     This pure operator allocates some uninitialized storage from the heap.
1662 :     There are two forms:
1663 :    
1664 :     rawrecord NONE [INT n] allocates a tagless record of length n
1665 :     rawrecord (SOME rk) [INT n] allocates a tagged record of length n
1666 :     and initializes the tag.
1667 :    
1668 :     2. rawupdate of cty
1669 :     rawupdate cty (v,i,x)
1670 :     Assigns to x to the ith component of record v.
1671 :     The storelist is not updated.
1672 :    
1673 :     We use these new primops for both spilling and increment record construction.
1674 :    
1675 :     1. Spilling.
1676 :    
1677 :     This is implemented with a linear scan algorithm (but generalized
1678 :     to trees). The algorithm will create a single spill record at the
1679 :     beginning of the cps function and use rawupdate to spill to it,
1680 :     and SELECT or SELp to reload from it. So both spills and reloads
1681 :     are fine-grain operations. In contrast, in the old algorithm
1682 :     "spills" have to be bundled together in records.
1683 :    
1684 :     Ideally, we should sink the spill record construction to where
1685 :     it is needed. We can even split the spill record into multiple ones
1686 :     at the places where they are needed. But CPS is not a good
1687 :     representation for global code motion, so I'll keep it simple and
1688 :     am not attempting this.
1689 :    
1690 :     2. Incremental record construction (aka record splitting).
1691 :    
1692 :     Long records with many component values which are simulatenously live
1693 :     (recall that single use record addresses are not considered to
1694 :     be live) are constructed with rawrecord and rawupdate.
1695 :     We allocate space on the heap with rawrecord first, then gradually
1696 :     fill it in with rawupdate. This is the technique suggested to me
1697 :     by Matthias.
1698 :    
1699 :     Some restrictions on when this is applicable:
1700 :     1. It is not a VECTOR record. The code generator currently does not handle
1701 :     this case. VECTOR record uses double indirection like arrays.
1702 :     2. All the record component values are defined in the same "basic block"
1703 :     as the record constructor. This is to prevent speculative
1704 :     record construction.
1705 :    
1706 :     ----------------------------------------------------------------------
1707 :     Name: Allen Leung
1708 : leunga 1086 Date: 2002/02/22 01:02:00 EST
1709 :     Tag: leunga-20020222-mlrisc-tools
1710 :    
1711 :     Minor bug fixes in the parser and rewriter
1712 :    
1713 :     ----------------------------------------------------------------------
1714 :     Name: Allen Leung
1715 : leunga 1085 Date: 2002/02/21 20:20:00 EST
1716 :     Tag: leunga-20020221-peephole
1717 :    
1718 :     Regenerated the peephole files. Some contained typos in the specification
1719 :     and some didn't compile because of pretty printing bugs in the old version
1720 :     of 'nowhere'.
1721 :    
1722 :     ----------------------------------------------------------------------
1723 :     Name: Allen Leung
1724 : leunga 1079 Date: 2002/02/19 20:20:00 EST
1725 :     Tag: leunga-20020219-mlrisc-tools
1726 :     Description:
1727 :    
1728 :     Minor bug fixes to the mlrisc-tools library:
1729 :    
1730 :     1. Fixed up parsing colon suffixed keywords
1731 :     2. Added the ability to shut the error messages up
1732 :     3. Reimplemented the pretty printer and fixed up/improved
1733 :     the pretty printing of handle and -> types.
1734 :     4. Fixed up generation of literal symbols in the nowhere tool.
1735 :     5. Added some SML keywords to to sml.sty
1736 :    
1737 :     ----------------------------------------------------------------------
1738 :     Name: Matthias Blume
1739 : blume 1078 Date: 2002/02/19 16:20:00 EST
1740 :     Tag: blume-20020219-cmffi
1741 :     Description:
1742 :    
1743 :     A wild mix of changes, some minor, some major:
1744 :    
1745 :     * All C FFI-related libraries are now anchored under $c:
1746 :     $/c.cm --> $c/c.cm
1747 :     $/c-int.cm --> $c/internals/c-int.cm
1748 :     $/memory.cm --> $c/memory/memory.cm
1749 :    
1750 :     * "make" tool (in CM) now treats its argument pathname slightly
1751 :     differently:
1752 :     1. If the native expansion is an absolute name, then before invoking
1753 :     the "make" command on it, CM will apply OS.Path.mkRelative
1754 :     (with relativeTo = OS.FileSys.getDir()) to it.
1755 :     2. The argument will be passed through to subsequent phases of CM
1756 :     processing without "going native". In particular, if the argument
1757 :     was an anchored path, then "make" will not lose track of that anchor.
1758 :    
1759 :     * Compiler backends now "know" their respective C calling conventions
1760 :     instead of having to be told about it by ml-nlffigen. This relieves
1761 :     ml-nlffigen from one of its burdens.
1762 :    
1763 :     * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
1764 :    
1765 :     * Export C_DEBUG and C_Debug from $c/c.cm.
1766 :    
1767 :     * C type encoding in ml-nlffi-lib has been improved to model the conceptual
1768 :     subtyping relationship between incomplete pointers and their complete
1769 :     counterparts. For this, ('t, 'c) ptr has been changed to 'o ptr --
1770 :     with the convention of instantiating 'o with ('t, 'c) obj whenever
1771 :     the pointer target type is complete. In the incomplete case, 'o
1772 :     will be instantiated with some "'c iobj" -- a type obtained by
1773 :     using one of the functors PointerToIncompleteType or PointerToCompleteType.
1774 :    
1775 :     Operations that work on both incomplete and complete pointer types are
1776 :     typed as taking an 'o ptr while operations that require the target to
1777 :     be known are typed as taking some ('t, 'c) obj ptr.
1778 :    
1779 :     voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
1780 :     where void is an eqtype without any values. This makes it possible
1781 :     to work on voidptr values using functions meant to operate on light
1782 :     incomplete pointers.
1783 :    
1784 :     * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
1785 :     been vastly simplified.
1786 :    
1787 :     ----------------------------------------------------------------------
1788 :     Name: Matthias Blume
1789 : blume 1076 Date: 2002/02/19 10:48:00 EST
1790 :     Tag: blume-20020219-pqfix
1791 :     Description:
1792 :    
1793 :     Applied Chris Okasaki's bug fix for priority queues.
1794 :    
1795 :     ----------------------------------------------------------------------
1796 :     Name: Matthias Blume
1797 : blume 1073 Date: 2002/02/15 17:05:00
1798 : blume 1069 Tag: Release_110_39
1799 :     Description:
1800 :    
1801 : blume 1073 Last-minute retagging is becoming a tradition... :-(
1802 :    
1803 :     This is the working release 110.39.
1804 :    
1805 :     ----------------------------------------------------------------------
1806 :     Name: Matthias Blume
1807 :     Date: 2002/02/15 16:00:00 EST
1808 :     Tag: Release_110_39-orig
1809 :     Description:
1810 :    
1811 : blume 1069 Working release 110.39. New bootfiles.
1812 :    
1813 : blume 1073 (Update: There was a small bug in the installer so it wouldn't work
1814 :     with all shells. So I retagged. -Matthias)
1815 :    
1816 : blume 1069 ----------------------------------------------------------------------
1817 :     Name: Matthias Blume
1818 : blume 1068 Date: 2002/02/15 14:17:00 EST
1819 :     Tag: blume-20020215-showbindings
1820 :     Description:
1821 :    
1822 :     Added EnvRef.listBoundSymbols and CM.State.showBindings. Especially
1823 :     the latter can be useful for exploring what bindings are available at
1824 :     the interactive prompt. (The first function returns only the list
1825 :     of symbols that are really bound, the second prints those but also the
1826 :     ones that CM's autoloading mechanism knows about.)
1827 :    
1828 :     ----------------------------------------------------------------------
1829 :     Name: Matthias Blume
1830 : blume 1067 Date: 2002/02/15 12:08:00 EST
1831 :     Tag: blume-20020215-iptrs
1832 :     Description:
1833 :    
1834 :     Two improvements to ml-nlffigen:
1835 :    
1836 :     1. Write files only if they do not exist or if their current contents
1837 :     do not coincide with what's being written. (That is, avoid messing
1838 :     with the time stamps unless absolutely necessary.)
1839 :    
1840 :     2. Implement a "repository" mechanism for generated files related
1841 :     to "incomplete pointer types". See the README file for details.
1842 :    
1843 :     ----------------------------------------------------------------------
1844 :     Name: Matthias Blume
1845 : blume 1066 Date: 2002/02/14 11:50:00 EST
1846 :     Tag: blume-20020214-quote
1847 :     Description:
1848 :    
1849 :     Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm). This is required
1850 :     because of the new and improved tag generation scheme. (Thanks to Allen
1851 :     Leung for pointing it out.)
1852 :    
1853 :     ----------------------------------------------------------------------
1854 : george 1065 Name: Lal George
1855 :     Date: 2002/02/14 09:55:27 EST 2002
1856 :     Tag: george-20020214-isabelle-bug
1857 :     Description:
1858 :    
1859 :     Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
1860 :     of Isabelle on the x86.
1861 :    
1862 :     From Allen:
1863 :     -----------
1864 :     I've found the problem:
1865 :    
1866 :     in ra-core.sml, I use the counter "blocked" to keep track of the
1867 :     true number of elements in the freeze queue. When the counter goes
1868 :     to zero, I skip examining the queue. But I've messed up the
1869 :     bookkeeping in combine():
1870 :    
1871 :     else ();
1872 :     case !ucol of
1873 :     PSEUDO => (if !cntv > 0 then
1874 :     (if !cntu > 0 then blocked := !blocked - 1 else ();
1875 :     ^^^^^^^^^^^^^^^^^^^^^^^
1876 :     moveu := mergeMoveList(!movev, !moveu)
1877 :     )
1878 :     else ();
1879 :    
1880 :     combine() is called to coalesce two nodes u and v.
1881 :     I think I was thinking that if the move counts of u and v are both
1882 :     greater than zero then after they are coalesced then one node is
1883 :     removed from the freeze queue. Apparently I was thinking that
1884 :     both u and v are of low degree, but that's clearly not necessarily true.
1885 :    
1886 :    
1887 :     02/12/2002:
1888 :     Here's the patch. HOL now compiles.
1889 :    
1890 :     I don't know how this impact on performance (compile
1891 :     time or runtime). This bug caused the RA (especially on the x86)
1892 :     to go thru the potential spill phase when there are still nodes on the
1893 :     freeze queue.
1894 :    
1895 :    
1896 :    
1897 :    
1898 :     ----------------------------------------------------------------------
1899 : blume 1058 Name: Matthias Blume
1900 : blume 1064 Date: 2002/02/13 22:40:00 EST
1901 :     Tag: blume-20020213-fptr-rtti
1902 :     Description:
1903 :    
1904 :     Fixed a bug in ml-nlffigen that was introduced with one of the previous
1905 :     updates.
1906 :    
1907 :     ----------------------------------------------------------------------
1908 :     Name: Matthias Blume
1909 : blume 1063 Date: 2002/02/13 16:41:00 EST
1910 :     Tag: blume-20020213-cmlpq
1911 :     Description:
1912 :    
1913 :     Added new priority queue export symbols (which have just been added to
1914 :     smlnj-lib.cm) to CML's version of smlnj-lib.cm. (Otherwise CML would
1915 :     not compile and the installer would choke.)
1916 :    
1917 :     ----------------------------------------------------------------------
1918 :     Name: Matthias Blume
1919 : blume 1062 Date: 2002/02/13 16:15:00 EST
1920 :     Tag: blume-20020213-various
1921 :     Description:
1922 :    
1923 :     1. More tweaks to ml-nlffigen:
1924 :    
1925 :     - better internal datastructures (resulting in slight speedup)
1926 :     - "-match" option requires exact match
1927 :     - "localized" gensym counters (untagged structs/unions nested within
1928 :     other structs/unions or within typedefs get a fresh counter; their
1929 :     tag will be prefixed by a concatenation of their parents' tags)
1930 :     - bug fixes (related to calculation of transitive closure of types
1931 :     to be included in the output)
1932 :    
1933 :     2. Minor Basis updates:
1934 :    
1935 :     - added implementations for List.collate and Option.app
1936 :    
1937 :     ----------------------------------------------------------------------
1938 :     Name: Matthias Blume
1939 : blume 1060 Date: 2002/02/11 15:55:00 EST
1940 :     Tag: blume-20020211-gensym
1941 :     Description:
1942 :    
1943 :     Added a "-gensym" option to command line of ml-nlffigen. This can be
1944 :     used to specify a "stem" -- a string that is inserted in all "gensym'd"
1945 :     names (ML structure names that correspond to unnamed C structs, unions,
1946 :     and enums), so that separate runs of ml-nlffigen do not clash.
1947 :    
1948 :     ----------------------------------------------------------------------
1949 :     Name: Matthias Blume
1950 : blume 1059 Date: 2002/02/11 12:05:00 EST
1951 :     Tag: blume-20020211-gensml
1952 :     Description:
1953 :    
1954 :     A quick fix for a problem with GenSML (in the pgraph-util library):
1955 :     Make generation of toplevel "local" optional. (Strictly speaking,
1956 :     signature definitions within "local" are not legal SML.)
1957 :    
1958 :     Other than that: updates to INSTALL and cm/TODO.
1959 :    
1960 :     ----------------------------------------------------------------------
1961 :     Name: Matthias Blume
1962 : blume 1058 Date: 2002/02/08 15:00:00 EST
1963 :     Tag: blume-20020208-uniquepid
1964 :     Description:
1965 :    
1966 :     0. Version number has been bumped to 110.38.1. NEW BOOTFILES!!!
1967 :    
1968 :     1. The installer (config/install.sh) has gotten smarter:
1969 :    
1970 :     - Configuration options are a bit easier to specify now
1971 :     (in config/targets).
1972 :     - Bug in recognizing .tar.bz2 files fixed.
1973 :     - Installer automatically resolves dependencies between
1974 :     configuration options (e.g., if you ask for eXene, you will
1975 :     also get cml -- regardless whether you asked for it or not).
1976 :     - Installer can run in "quieter mode" by setting the environment
1977 :     variable INSTALL_QUIETLY to "true". "Quieter" does not mean
1978 :     "completely silent", though.
1979 :     - Build HashCons library as part of smlnj-lib.
1980 :    
1981 :     2. A new scheme for assigning persistent identifiers to compilation
1982 :     units (and, by extension, to types etc.) has been put into place.
1983 :     This fixes a long-standing bug where types and even dynamic values
1984 :     can get internally confused, thereby compromising type safety
1985 :     (abstraction) and dynamic correctness. See
1986 :    
1987 :     http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
1988 :    
1989 :     for an example of how things could go wrong until now.
1990 :    
1991 :     The downside of the new scheme is that pids are not quite as
1992 :     persistent as they used to be: CM will generate a fresh pid
1993 :     for every compilation unit that it thinks it sees for the first
1994 :     time. That means that if you compile starting from a clean, fresh
1995 :     source tree at two different times, you end up with different
1996 :     binaries.
1997 :    
1998 :     Cutoff recompilation, however, has not been compromised because
1999 :     CM keeps pid information in special caches between runs.
2000 :    
2001 :     ----------------------------------------------------------------------
2002 : george 1053 Name: Lal George
2003 : george 1055 Date: 2002/02/07 15:34:13 EST 2002
2004 :     Tag: <none>
2005 :     Description:
2006 :    
2007 :     Compilers that generate assembly code may produce global labels
2008 :     whose value is resolved at link time. The various peephole optimization
2009 :     modules did not take this in account.
2010 :    
2011 :     TODO. The Labels.addrOf function should really return an option
2012 :     type so that clients are forced to deal with this issue, rather
2013 :     than an exception being raised.
2014 :    
2015 :     ----------------------------------------------------------------------
2016 :     Name: Lal George
2017 : george 1053 Date: 2002/02/06 13:55:02 EST
2018 :     Tag: george-20020206-ra-breakup
2019 :     Description:
2020 :    
2021 :     1. A bug fix from Allen.
2022 :    
2023 :     A typo causes extra fstp %st(0)'s to be generated at compensation
2024 :     edges, which might cause stack underflow traps at runtime. This
2025 :     occurs in fft where there are extraneous fstps right before the 'into'
2026 :     trap instruction (in this case they are harmless since none of the
2027 :     integers overflow.)
2028 :    
2029 : george 1054 2. Pulled out various utility modules that were embedded in the modules
2030 : george 1053 of the register allocator. I need these modules for other purposes, but
2031 : george 1054 they are not complete enough to put into a library (just yet).
2032 : george 1053 ----------------------------------------------------------------------
2033 : blume 1049 Name: Matthias Blume
2034 :     Date: 2002/01/31 16:05:00 EST
2035 :     Tag: blume-20020131-sparc-ccalls
2036 :     Description:
2037 :    
2038 :     1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
2039 :     of extra stack space by mistake. Fixed.
2040 :    
2041 :     2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
2042 :    
2043 :     ----------------------------------------------------------------------
2044 : leunga 1048 Name: Allen Leung
2045 :     Date: 2002/01/30
2046 :     Tag: leunga-20020130-nowhere-bug-fix
2047 :     Description:
2048 :    
2049 :     MLRISC bug fixes:
2050 :     1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
2051 :     2. MachineInt.fromString was negating its value.
2052 :    
2053 :     ----------------------------------------------------------------------
2054 : blume 1039 Name: Matthias Blume
2055 : blume 1047 Date: 2002/01/29
2056 :     Tag: blume-20020129-INSTALL
2057 :     Description:
2058 :    
2059 :     - Added somewhat detailed installation instructions (file INSTALL).
2060 :     - Fixed curl-detection bug in config/install.sh.
2061 :     - It is now possible to select the URL getter using the URLGETTER
2062 :     environment variable:
2063 :    
2064 :     not set / "unknown" --> automatic detection (script tries wget,
2065 :     curl, and lynx)
2066 :     "wget" / "curl" / "lynx" --> use the specified program (script "knows"
2067 :     how to properly invoke them)
2068 :     other --> use $URLGETTER directly, it must take
2069 :     precisely two command-line arguments
2070 :     (source URL and destination file name)
2071 :    
2072 :     ----------------------------------------------------------------------
2073 :     Name: Matthias Blume
2074 : blume 1044 Date: 2002/01/28
2075 : blume 1045 Tag: blume-20020128-sparc-ccalls
2076 :     Description:
2077 :    
2078 :     - Fixed problem with calculation of "used" registers in sparc-c-calls.
2079 :     - Make use of the allocParam argument in sparc-c-calls.
2080 :    
2081 :     ----------------------------------------------------------------------
2082 :     Name: Matthias Blume
2083 :     Date: 2002/01/28
2084 : blume 1044 Tag: blume-20020128-allocParam
2085 :     Description:
2086 :    
2087 :     John Reppy: Changes c-calls API to accept client-callback for
2088 :     allocating extra stack space.
2089 :     me: Corresponding changes to mlriscGen (using a dummy argument that
2090 :     does not change the current behavior).
2091 :    
2092 :     ----------------------------------------------------------------------
2093 :     Name: Matthias Blume
2094 : blume 1040 Date: 2002/01/28 12:00:00
2095 :     Tag: Release_110_38
2096 :     Description:
2097 :    
2098 :     This time for real!!!
2099 :    
2100 :     ----------------------------------------------------------------------
2101 :     Name: Matthias Blume
2102 : blume 1039 Date: 2002/01/28 10:56:00 EST
2103 :     Tag: blume-20020128-retraction
2104 :     Description:
2105 :    
2106 :     0. Retracted earlier 110.38. (The Release_110_38 tag has been replaced
2107 :     with blume-Release_110_38-retracted.)
2108 :    
2109 :     1. Fixed a problem with incorrect rounding modes in real64.sml.
2110 :     (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
2111 :    
2112 :     2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
2113 :     and enums fixed. The naming of corresponding ML identifiers should
2114 :     now be consistent again.
2115 :    
2116 :     ----------------------------------------------------------------------
2117 : leunga 1038 Name: Allen Leung
2118 :     Date: 2002/01/27
2119 :     Tag: leunga-20020127-nowhere
2120 :     Description:
2121 :    
2122 :     Added a target called nowhere in the configuration scripts.
2123 :     Enabling this will build the MLRISC 'nowhere' tool (for translating
2124 :     programs with where-clauses into legal SML code) during installation.
2125 :    
2126 :     ----------------------------------------------------------------------
2127 : blume 1036 Name: Matthias Blume
2128 : leunga 1038 Date: 2002/01/25 21:27:00 EST
2129 : blume 1039 Tag: blume-Release_110_38-retracted
2130 : blume 1037 Description:
2131 :    
2132 :     Call it a (working) release! Version is 110.38. Bootfiles are ready.
2133 :    
2134 :     README will be added later.
2135 :    
2136 : blume 1039 !!! NOTE: Re-tagged as blume-Release_110_38-retracted. Original tag
2137 :     (Release_110_38) removed. Reason: Last-minute bug fixes.
2138 :    
2139 : blume 1037 ----------------------------------------------------------------------
2140 :     Name: Matthias Blume
2141 : blume 1036 Date: 2002/01/25
2142 :     Tag: blume-20020125-ffi
2143 :     Description:
2144 :    
2145 :     A large number of tweaks and improvements to ml-nlffi-lib and
2146 :     ml-nlffigen:
2147 :    
2148 :     - ML represenation types have been streamlined
2149 :     - getter and setter functions work with concrete values, not abstract
2150 :     ones where possible
2151 :     - ml-nlffigen command line more flexible (see README file there)
2152 :     - some bugs have been fixed (hopefully)
2153 :    
2154 :     ----------------------------------------------------------------------
2155 : george 1034 Name: Lal George
2156 :     Date: 2002/01/24
2157 :     Tag: george-20020124-risc-ra-interface
2158 :     Description:
2159 :    
2160 :     There is a dramatic simplification in the interface to the
2161 :     register allocator for RISC architectures as a result of making
2162 :     parallel copy instructions explicit.
2163 :    
2164 :     ----------------------------------------------------------------------
2165 : blume 1028 Name: Matthias Blume
2166 : blume 1032 Date: 2002/01/22
2167 :     Tag: blume-20020122-x86-ccalls
2168 :     Description:
2169 :    
2170 :     Bug fix for c-calls on x86 (having to do with how char- and
2171 :     short-arguments are being handled).
2172 :    
2173 :     ----------------------------------------------------------------------
2174 :     Name: Matthias Blume
2175 : blume 1031 Date: 2002/01/21
2176 :     Tag: blume-20020121-ff
2177 :     Description:
2178 :    
2179 :     Another day of fiddling with the FFI...
2180 :    
2181 :     1. Bug fix/workaround: CKIT does not complain about negative array
2182 :     dimensions, so ml-nlffigen has to guard itself against this possibility.
2183 :     (Otherwise a negative dimension would send it into an infinite loop.)
2184 :    
2185 :     2. Some of the abstract types (light objects, light pointers, most "base"
2186 :     types) in structure C are now eqtypes.
2187 :    
2188 :     3. Added constructors and test functions for NULL function pointers.
2189 :    
2190 :     ----------------------------------------------------------------------
2191 :     Name: Matthias Blume
2192 : blume 1028 Date: 2002/01/18
2193 : blume 1030 Tag: blume-20020118-ready-for-new-release
2194 :     Description:
2195 :    
2196 :     Made config/srcarchiveurl point to a new place. (Will provide boot
2197 :     files shortly.)
2198 :    
2199 :     Maybe we christen this to be 110.38?
2200 :    
2201 :     ----------------------------------------------------------------------
2202 :     Name: Matthias Blume
2203 :     Date: 2002/01/18
2204 : blume 1029 Tag: blume-20020118-more-ffifiddle
2205 :     Description:
2206 :    
2207 :     Today's FFI fiddling:
2208 :    
2209 :     - Provided a structure CGetSet with "convenient" versions of C.Get.* and
2210 :     C.Set.* that use concrete (MLRep.*) arguments and results instead
2211 :     of abstract ones.
2212 :    
2213 :     - Provided word-style bit operations etc. for "int" representation
2214 :     types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
2215 :     and Long.
2216 :    
2217 :     ----------------------------------------------------------------------
2218 :     Name: Matthias Blume
2219 :     Date: 2002/01/18
2220 : blume 1028 Tag: blume-20020118-use-x86-fp
2221 :     Description:
2222 :    
2223 :     Now that x86-fast-fp seems to be working, I turned it back on again
2224 :     by default. (Seems to work fine now, even with the FFI.)
2225 :    
2226 :     Other than that, I added some documentation about the FFI to
2227 :     src/ml-nlffigen/README and updated the FFI test examples in
2228 :     src/ml-nlffi-lib/Tests/*.
2229 :    
2230 :     ----------------------------------------------------------------------
2231 : leunga 1025 Name: Allen Leung
2232 :     Date: 2002/01/17
2233 :     Tag: leunga-20020117-x86-fast-fp-call
2234 :     Description:
2235 :    
2236 :     1. Fixed a problem with handling return fp values when x86's fast fp
2237 :     mode is turned on.
2238 :    
2239 :     2. Minor pretty printing fix for cellset. Print %st(0) as %st(0) instead
2240 :     of %f32.
2241 :    
2242 :     3. Added a constructor INT32lit to the ast of MLRISC tools.
2243 :    
2244 :     ----------------------------------------------------------------------
2245 : blume 1021 Name: Matthias Blume
2246 :     Date: 2002/01/16
2247 :     Tag: blume-20020116-ffifiddle
2248 :     Description:
2249 :    
2250 :     More fiddling with the FFI interface:
2251 :    
2252 :     - Make constness 'c instead of rw wherever possible. This eliminates
2253 :     the need for certain explicit coercions. (However, due to ML's
2254 :     value polymorphism, there will still be many cases where explicit
2255 :     coercions are necessary. Phantom types are not the whole answer
2256 :     to modeling a subtyping relationship in ML.)
2257 :    
2258 :     - ro/rw coersions for pointers added. (Avoids the detour through */&.)
2259 :    
2260 :     - "printf" test example added to src/ml-nlffi-lib/Tests. (Demonstrates
2261 :     clumsy workaround for varargs problem.)
2262 :    
2263 :     ----------------------------------------------------------------------
2264 : george 1016 Name: Lal George
2265 :     Date: 2002/01/15
2266 :     Tag: <none>
2267 :     Description:
2268 :    
2269 :     1. Since COPY instructions are no longer native to the architecture,
2270 :     a generic functor can be used to implement the expandCopies function.
2271 :    
2272 :     2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
2273 :     TEXT segment.
2274 :    
2275 :     ----------------------------------------------------------------------
2276 : blume 1011 Name: Matthias Blume
2277 : blume 1015 Date: 2002/01/15
2278 :     Tag: blume-20020115-ffiupdates
2279 :     Description:
2280 :    
2281 :     1. Fix for bug resulting in single-precision float values being returned
2282 :     incorrectly from FFI calls.
2283 :    
2284 :     2. Small modifications to C FFI API:
2285 :    
2286 :     - memory-allocation routines return straight objects (no options)
2287 :     and raise an exception in out-of-memory situations
2288 :     - unsafe extensions to cast between function pointers and pointers
2289 :     from/to ints
2290 :     - added structure C_Debug as an alternative to structure C where
2291 :     pointer-dereferencing (|*| and |*!) always check for null-pointers
2292 :     - added open_lib' to DynLinkage; open_lib' works like open_lib
2293 :     but also takes a (possibly empty) list of existing library handles
2294 :     that the current library depends on
2295 :    
2296 :     ----------------------------------------------------------------------
2297 :     Name: Matthias Blume
2298 : blume 1011 Date: 2002/01/10
2299 :     Tag: blume-20020110-newffigen
2300 :     Description:
2301 :    
2302 :     1. Updates to portable graph code.
2303 :    
2304 :     2. Major update to ml-nlffigen and ml-nlffi-lib. Things are much
2305 :     more scalable now so that even huge interfaces such as the one
2306 :     for GTK compile in finite time and space. :-)
2307 :     See src/ml-nlffigen/README for details on what's new.
2308 :    
2309 :     ----------------------------------------------------------------------
2310 : george 1003 Name: Lal George
2311 : george 1009 Date: 2001/01/09 14:31:35 EST 2002
2312 :     Tag: george-20011206-rm-native-copy
2313 :     Description:
2314 :    
2315 :     Removed the native COPY and FCOPY instructions
2316 :     from all the architectures and replaced it with the
2317 :     explicit COPY instruction from the previous commit.
2318 :    
2319 :     It is now possible to simplify many of the optimizations
2320 :     modules that manipulate copies. This has not been
2321 :     done in this change.
2322 :    
2323 :     ----------------------------------------------------------------------
2324 :     Name: Lal George
2325 : george 1003 Date: 2001/12/06 16:50:13 EST 2001
2326 :     Tag: george-20011206-mlrisc-instruction
2327 :     Description:
2328 : george 824
2329 : george 1003 Changed the representation of instructions from being fully abstract
2330 :     to being partially concrete. That is to say:
2331 :    
2332 :     from
2333 :     type instruction
2334 :    
2335 :     to
2336 :     type instr (* machine instruction *)
2337 :    
2338 :     datatype instruction =
2339 :     LIVE of {regs: C.cellset, spilled: C.cellset}
2340 :     | KILL of {regs: C.cellset, spilled: C.cellset}
2341 :     | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
2342 :     | ANNOTATION of {i: instruction, a: Annotations.annotation}
2343 :     | INSTR of instr
2344 :    
2345 :     This makes the handling of certain special instructions that appear on
2346 :     all architectures easier and uniform.
2347 :    
2348 :     LIVE and KILL say that a list of registers are live or killed at the
2349 :     program point where they appear. No spill code is generated when an
2350 :     element of the 'regs' field is spilled, but the register is moved to
2351 :     the 'spilled' (which is present, more for debugging than anything else).
2352 :    
2353 :     LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
2354 :     We used to generate:
2355 :    
2356 :     DEFFREG f1
2357 :     f1 := f2 + f3
2358 :     trapb
2359 :    
2360 :     but now generate:
2361 :    
2362 :     f1 := f2 + f3
2363 :     trapb
2364 :     LIVE {regs=[f1,f2,f3], spilled=[]}
2365 :    
2366 :     Furthermore, the DEFFREG (hack) required that all floating point instruction
2367 :     use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
2368 :     defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
2369 :     in a cleaner alpha implementation. (Hopefully, intel will not get rid of
2370 :     this architecture).
2371 :    
2372 :     COPYXXX is intended to replace the parallel COPY and FCOPY available on
2373 :     all the architectures. This will result in further simplification of the
2374 :     register allocator that must be aware of them for coalescing purposes, and
2375 :     will also simplify certain aspects of the machine description that provides
2376 :     callbacks related to parallel copies.
2377 :    
2378 :     ANNOTATION should be obvious, and now INSTR represents the honest to God
2379 :     machine instruction set!
2380 :    
2381 :     The <arch>/instructions/<arch>Instr.sml files define certain utility
2382 :     functions for making porting easier -- essentially converting upper case
2383 :     to lower case. All machine instructions (of type instr) are in upper case,
2384 :     and the lower case form generates an MLRISC instruction. For example on
2385 :     the alpha we have:
2386 :    
2387 :     datatype instr =
2388 :     LDA of {r:cell, b:cell, d:operand}
2389 :     | ...
2390 :    
2391 :     val lda : {r:cell, b:cell, d:operand} -> instruction
2392 :     ...
2393 :    
2394 :     where lda is just (INSTR o LDA), etc.
2395 :    
2396 : leunga 815 ----------------------------------------------------------------------
2397 : blume 986 Name: Matthias Blume
2398 : blume 993 Date: 2001/11/22 21:40:00 EST
2399 :     Tag: Release_110_37
2400 :     Description:
2401 :    
2402 :     Release 110.37. This time for real.
2403 :    
2404 :     ----------------------------------------------------------------------
2405 :     Name: Matthias Blume
2406 : blume 988 Date: 2001/11/21 16:35:00 EST
2407 :     Tag: blume-20011121-foot-in-mouth
2408 :     Description:
2409 :    
2410 :     Removed the "Release_110_37" tag because of a serious bug.
2411 :     This will be re-tagged once the bug is fixed.
2412 :    
2413 :     ----------------------------------------------------------------------
2414 :     Name: Matthias Blume
2415 : blume 987 Date: 2001/11/21 16:14:00 EST
2416 :     Tag: blume-20011121-forgottenfile
2417 :     Description:
2418 :    
2419 :     Forgot to add a file. (Just a .tex-file -- part of
2420 :     the CM manual source.)
2421 :    
2422 :     ----------------------------------------------------------------------
2423 :     Name: Matthias Blume
2424 : blume 986 Date: 2001/11/21 16:10:00 EST
2425 : blume 988 Tag: blume-20011121-invalid_110_37
2426 : blume 986 Description:
2427 :    
2428 : blume 988 Note: I removed the original tag "Release_110_37" from this commit
2429 :     because we found a serious bug in all non-x86 backends.
2430 :     - Matthias
2431 :    
2432 : blume 986 1. Modifications to the SML/NJ code generator and to the runtime system
2433 :     so that code object name strings are directly inserted into code
2434 :     objects at code generation time. The only business the runtime system
2435 :     has with this is now to read the name strings on occasions.
2436 :     (The encoding of the name string has also changed somewhat.)
2437 :    
2438 :     2. CM now implements a simple "set calculus" for specifying export lists.
2439 :     In particular, it is now possible to refer to the export lists of
2440 :     other libraries/groups/sources and form unions as well as differences.
2441 :     See the latest CM manual for details.
2442 :    
2443 :     3. An separate notion of "proxy" libraries has again be eliminated from
2444 :     CM's model. (Proxy libraries are now simply a special case of using
2445 :     the export list calculus.)
2446 :    
2447 :     4. Some of the existing libraries now take advantage of the new set
2448 :     calculus.
2449 :     (Notice that not all libraries have been converted because some
2450 :     of the existing .cm-files are supposed to be backward compatible
2451 :     with 110.0.x.)
2452 :    
2453 :     5. Some cleanup in stand-alone programs. (Don't use "exnMessage" -- use
2454 :     "General.exnMessage"! The former relies on a certain hook to be
2455 :     initialized, and that often does not happen in the stand-alone case.)
2456 :    
2457 :     ----------------------------------------------------------------------
2458 : george 984 Name: Lal George
2459 :     Date: 2001/11/21 13:56:18 EST
2460 :     Tag: george-2001121-pseudo-ops
2461 :     Description:
2462 :    
2463 :     Implemented a complete redesign of MLRISC pseudo-ops. Now there
2464 :     ought to never be any question of incompatabilities with
2465 :     pseudo-op syntax expected by host assemblers.
2466 :    
2467 :     For now, only modules supporting GAS syntax are implemented
2468 :     but more should follow, such as MASM, and vendor assembler
2469 :     syntax, e.g. IBM as, Sun as, etc.
2470 :    
2471 :     ----------------------------------------------------------------------
2472 : blume 977 Name: Matthias Blume
2473 :     Date: 2001/11/14 11:52:00 EST
2474 :     Tag: blume-20011114-srcname
2475 :     Description:
2476 :    
2477 :     1. Routed the name of the current source file to mlriscgen where it
2478 :     should be directly emitted into the code object. (This last part
2479 :     is yet to be done.)
2480 :    
2481 :     2. Some cleanup of the pgraph code to make it match the proposal that
2482 :     I put out the other day. (The proposal notwithstanding, things are
2483 :     still in flux here.)
2484 :    
2485 :     ----------------------------------------------------------------------
2486 : george 976 Name: Lal George
2487 :     Date: 2001/11/14 09:44:04 EST
2488 :     Tag:
2489 :     Description:
2490 :    
2491 :     Fix for a backpatching bug reported by Allen.
2492 :    
2493 :     Because the boundary between short and long span-dependent
2494 :     instructions is +/- 128, there are an astounding number of
2495 :     span-dependent instructions whose size is over estimated.
2496 :    
2497 :     Allen came up with the idea of letting the size of span
2498 :     dependent instructions be non-monotonic, for a maxIter
2499 :     number of times, after which the size must be monotonically
2500 :     increasing.
2501 :    
2502 :     This table shows the number of span-dependent instructions
2503 :     whose size was over-estimated as a function of maxIter, for the
2504 :     file Parse/parse/ml.grm.sml:
2505 :    
2506 :     maxIter # of instructions:
2507 :     10 687
2508 :     20 438
2509 :     30 198
2510 :     40 0
2511 :    
2512 :     In compiling the compiler, there is no significant difference in
2513 :     compilation speed between maxIter=10 and maxIter=40. Actually,
2514 :     my measurements showed that maxIter=40 was a tad faster than
2515 :     maxIter=10! Also 96% of the files in the compiler reach a fix
2516 :     point within 13 iterations, so fixing maxIter at 40, while high,
2517 :     is okay.
2518 :    
2519 :     ----------------------------------------------------------------------
2520 : blume 975 Name: Matthias Blume
2521 :     Date: 2001/10/31 15:25:00 EST
2522 :     Tag: blume-20011031-pgraph
2523 :     Description:
2524 :    
2525 :     CKIT:
2526 :     * Changed the "Function" constructor of type Ast.ctype to carry optional
2527 :     argument identifiers.
2528 :     * Changed the return type of TypeUtil.getFunction accordingly.
2529 :     * Type equality ignores the argument names.
2530 :     * TypeUtil.composite tries to preserve argument names but gives up quickly
2531 :     if there is a mismatch.
2532 :    
2533 :     installation script:
2534 :     * attempts to use "curl" if available (unless "wget" is available as well)
2535 :    
2536 :     CM:
2537 :     * has an experimental implementation of "portable graphs" which I will
2538 :     soon propose as an implementation-independent library format
2539 :     * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
2540 :    
2541 :     NLFFI-LIB:
2542 :     * some cleanup (all cosmetic)
2543 :    
2544 :     NLFFIGEN:
2545 :     * temporarily disabled the mechanism that suppresses ML output for
2546 :     C definitions whose identifiers start with an underscore character
2547 :     * generate val bindings for enum constants
2548 :     * user can request that only one style (light or heavy) is being used;
2549 :     default is to use both (command-line arguments: -heavy and -light)
2550 :     * fixed bug in handling of function types involving incomplete pointers
2551 :     * generate ML entry points that take record arguments (i.e., using
2552 :     named arguments) for C functions that have a prototype with named
2553 :     arguments
2554 :     (see changes to CKIT)
2555 :    
2556 :     ----------------------------------------------------------------------
2557 : leunga 972 Name: Allen Leung
2558 :     Date: 2001/10/27 20:34:00 EDT
2559 :     Tag: leunga-20011027-x86-fast-fp-call
2560 :     Description:
2561 :    
2562 :     Fixed the bug described in blume-20010920-slowfp.
2563 :    
2564 :     The fix involves
2565 :     1. generating FCOPYs in FSTP in ia32-svid
2566 :     2. marking a CALL with the appropriate annotation
2567 :    
2568 :     ----------------------------------------------------------------------
2569 : blume 961 Name: Matthias Blume
2570 :     Date: 2001/10/16 11:32:00 EDT
2571 :     Tag: blume-20011016-netbsd
2572 :     Description:
2573 :    
2574 :     Underscore patch from Chris Richards (fixing problem with compiling
2575 :     runtime system under recent NetBSD).
2576 :    
2577 :     ----------------------------------------------------------------------
2578 : george 959 Name: Allen Leung
2579 :     Date: 2001/10/12 17:18:32 EDT 2001
2580 :     Tag: leung-20011012-x86-printflowgraph
2581 :     Description:
2582 :    
2583 :     X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
2584 :    
2585 :     ----------------------------------------------------------------------
2586 : george 952 Name: Lal George
2587 : george 958 Date: 2001/10/11 23:51:34 EDT
2588 :     Tag: george-20011011-too-many-instrs
2589 :     Description:
2590 :    
2591 :     The representation of a program point never expected to see more
2592 :     than 65536 instructions in a basic block!
2593 :    
2594 :     ----------------------------------------------------------------------
2595 :     Name: Lal George
2596 : george 952 Date: 2001/10/09 09:41:37 EDT
2597 :     Tag: george-20011008-mlrisc-labels
2598 :     Description:
2599 :    
2600 :     Changed the machine description files to support printing of
2601 :     local and global labels in assembly code, based on host assembler
2602 :     conventions.
2603 :    
2604 :     ----------------------------------------------------------------------
2605 : blume 916 Name: Matthias Blume
2606 : blume 939 Date: 2001/09/25 15:25:00 EDT
2607 :     Tag: blume-20010925-exninfo
2608 :     Description:
2609 :    
2610 :     I provided a non-hook implementation of exnName (at the toplevel) and
2611 :     made the "dummy" implementation of exnMessage (at the toplevel) more
2612 :     useful: if nothing gets "hooked in", then at least you are going to
2613 :     see the exception name and a message indicating why you don't see more.
2614 :    
2615 :     [For the time being, programs that need exnMessage and want to use
2616 :     ml-build should either use General.exnMessage (strongly recommended) or
2617 :     refer to structure General at some other point so that CM sees a
2618 :     static dependency.]
2619 :    
2620 :     [Similar remarks go for "print" and "use": If you want to use their
2621 :     functionality in stand-alone programs generated by ml-build, then use
2622 :     TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
2623 :    
2624 :     ----------------------------------------------------------------------
2625 :     Name: Matthias Blume
2626 : blume 938 Date: 2001/09/20 17:28:00 EDT
2627 :     Tag: blume-20010920-slowfp
2628 :     Description:
2629 :    
2630 :     Allen says that x86-fast-fp is not safe yet, so I turned it off again...
2631 :    
2632 :     ----------------------------------------------------------------------
2633 :     Name: Matthias Blume
2634 : blume 937 Date: 2001/09/20 17:20:00 EDT
2635 :     Tag: blume-20010920-canonicalpaths
2636 :     Description:
2637 :    
2638 :     0. Updated the BOOT file (something that I forgot to do earlier).
2639 :    
2640 :     1. Small internal change to CM so that it avoids "/../" in filenames
2641 :     as much as possible (but only where it is safe).
2642 :    
2643 :     2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
2644 :     that arguments that contain delimiters are passed through correctly.
2645 :     This change also means that all "special" arguments of the form
2646 :     @SMLxxx... must come first.
2647 :    
2648 :     3. Changed install script to put relative anchor names for tool commands
2649 :     into pathconfig.
2650 :    
2651 :     ----------------------------------------------------------------------
2652 :     Name: Matthias Blume
2653 : blume 932 Date: 2001/09/18 15:35:00 EDT
2654 :     Tag: blume-20010918-readme11036
2655 :     Description:
2656 :    
2657 :     Added README files.
2658 :    
2659 :     ----------------------------------------------------------------------
2660 :     Name: Matthias Blume
2661 : blume 929 Date: 2001/09/18 11:45:00 EDT
2662 :     Tag: Release_110_36 (retag)
2663 :     Description:
2664 :    
2665 :     Fixed mistake in config/preloads. Retagged as 110.36.
2666 :    
2667 :     ----------------------------------------------------------------------
2668 :     Name: Matthias Blume
2669 : blume 928 Date: 2001/09/18 09:40:00 EDT
2670 : blume 929 Tag: Release_110_36_orig (tag changed)
2671 : blume 928 Description:
2672 :    
2673 :     New version (110.36). New bootfiles.
2674 :    
2675 :     ----------------------------------------------------------------------
2676 :     Name: Matthias Blume
2677 : blume 927 Date: 2001/09/14 16:15:00 EDT
2678 :     Tag: blume-20010914-x86fastfp
2679 :     Description:
2680 :    
2681 :     John committed some changes that Allen made, in particular a (hopefully)
2682 :     correctly working version of the x86-fp module.
2683 :    
2684 :     I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
2685 :     flag to "true". Everything seems to compile to a fixpoint ok, and
2686 :     "mandelbrot" speeds up by about 15%.
2687 :    
2688 :     ----------------------------------------------------------------------
2689 :     Name: Matthias Blume
2690 : blume 923 Date: 2001/09/13 11:20:00 EDT
2691 :     Tag: blume-20010913-minimal
2692 :     Description:
2693 :    
2694 :     1. Stefan Monnier's patch to fix a miscompilation problem that
2695 :     was brought to light by John Reppy's work on Moby.
2696 :    
2697 :     2. Implemented a minimal "structure Compiler" that contains just
2698 :     "version" and "architecture". The minimal version will be
2699 :     available when the full version is not. This is for backward-
2700 :     compatibility with code that wants to test Compiler.version.
2701 :    
2702 :     ----------------------------------------------------------------------
2703 :     Name: Matthias Blume
2704 : blume 920 Date: 2001/08/28 14:03:00 EDT
2705 :     Tag: blume-20010828-ml-lex
2706 :     Description:
2707 :    
2708 :     Fix for bug 1581, received from Neophytos Michael.
2709 :    
2710 :     ----------------------------------------------------------------------
2711 :     Name: Matthias Blume
2712 : blume 918 Date: 2001/08/27 11:20:00 EDT
2713 :     Tag: blume-20010827-readme11035
2714 :     Description:
2715 :    
2716 :     Fleshed out the README file for 110.35.
2717 :    
2718 :     ----------------------------------------------------------------------
2719 :     Name: Matthias Blume
2720 : blume 916 Date: 2001/08/24 17:10:00 EDT
2721 :     Tag: Release_110_35
2722 :     Description:
2723 :    
2724 :     New version number (110.35). New bootfiles.
2725 :    
2726 :     ----------------------------------------------------------------------
2727 : george 915 Name: Lal George
2728 : george 909 Date: 2001/08/24 13:47:18 EDT 2001
2729 :     Tag: george-20010824-MLRISC-graphs
2730 :     Description:
2731 :    
2732 :     removed clusters from MLRISC completely and replaced with graphs.
2733 :    
2734 :     ----------------------------------------------------------------------
2735 :     Name: Matthias Blume
2736 : blume 905 Date: 2001/08/23 17:50:00 EDT
2737 :     Tag: blume-20010823-toplevel
2738 :     Description:
2739 :    
2740 :     - some reorganization of the code that implements various kinds of
2741 :     environments in the compiler (static, dynamic, symbolic, combined)
2742 :     - re-implemented the EnvRef module so that evalStream works properly
2743 :     (if the stream contains references to "use", "CM.make", etc.)
2744 :     - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
2745 :    
2746 :     ----------------------------------------------------------------------
2747 :     Name: Matthias Blume
2748 : blume 904 Date: 2001/08/20 15:50 EDT
2749 :     Tag: blume20010820-slipup
2750 :     Description:
2751 :    
2752 :     I forgot to commit a few files. Here they are...
2753 :    
2754 :     ----------------------------------------------------------------------
2755 :     Name: Matthias Blume
2756 : blume 903 Date: 2001/08/20 15:35:00 EDT
2757 :     Tag: blume-20010820-debugprof
2758 :     Description:
2759 :    
2760 :     !!!! NEW BOOTFILES !!!!
2761 :    
2762 :     This is another round of reorganizing the compiler sources. This
2763 :     time the main goal was to factor out all the "instrumentation"
2764 :     passes (for profiling and backtracing) into their own library.
2765 :     The difficulty was to do it in such a way that it does not depend
2766 :     on elaborate.cm but only on elabdata.cm.
2767 :    
2768 :     Therefore there have been further changes to both elaborate.cm and
2769 :     elabdata.cm -- more "generic" things have been moved from the former
2770 :     to the latter. As a result, I was forced to split the assignment
2771 :     of numbers indicating "primtyc"s into two portions: SML-generic and
2772 :     SML/NJ-specific. Since it would have been awkward to maintain,
2773 :     I bit the bullet and actually _changed_ the mapping between these
2774 :     numbers and primtycs. The bottom line of this is that you need
2775 :     a new set of bin- and bootfiles.
2776 :    
2777 :     I have built new bootfiles for all architectures, so doing a fresh
2778 :     checkout and config/install.sh should be all you need.
2779 :    
2780 :     The newly created library's name is
2781 :    
2782 :     $smlnj/viscomp/debugprof.cm
2783 :    
2784 :     and its sources live under
2785 :    
2786 :     src/compiler/DebugProf
2787 :    
2788 :     ----------------------------------------------------------------------
2789 :     Name: Matthias Blume
2790 : blume 902 Date: 2001/08/15 17:15:00 EDT
2791 :     Tag: blume-20010815-compreorg
2792 :     Description:
2793 :    
2794 :     This is a first cut at reorganizing the CM libraries that make up the
2795 :     core of the compiler. The idea is to separate out pieces that could
2796 :     be used independently by tools, e.g., the parser, the typechecker, etc.
2797 :    
2798 :     The current status is a step in this direction, but it is not quite
2799 :     satisfactory yet. Expect more changes in the future.
2800 :    
2801 :     Here is the current (new) organization...
2802 :    
2803 :     What used to be $smlnj/viscomp/core.cm is now divided into
2804 :     six CM libraries:
2805 :    
2806 :     $smlnj/viscomp/basics.cm
2807 :     /parser.cm
2808 :     /elabdata.cm
2809 :     /elaborate.cm
2810 :     /execute.cm
2811 :     /core.cm
2812 :    
2813 :     The CM files for these libraries live under src/system/smlnj/viscomp.
2814 :     All these libraries are proxy libraries that contain precisely
2815 :     one CM library component. Here are the locations of the components
2816 :     (all within the src/compiler tree):
2817 :    
2818 :     Basics/basics.cm
2819 :     Parse/parser.cm
2820 :     ElabData/elabdata.cm
2821 :     Elaborator/elaborate.cm
2822 :     Execution/execute.cm
2823 :     core.cm
2824 :    
2825 :     [This organization is the same that has been used already
2826 :     for a while for the architecture-specific parts of the visible
2827 :     compiler and for the old version of core.cm.]
2828 :    
2829 :     As you will notice, many source files have been moved from their
2830 :     respective original locations to a new home in one of the above
2831 :     subtrees.
2832 :    
2833 :     The division of labor between the new libraries is the following:
2834 :    
2835 :     basics.cm:
2836 :     - Simple, basic definitions that pertain to many (or all) of
2837 :     the other libraries.
2838 :     parser.cm:
2839 :     - The SML parser, producing output of type Ast.dec.
2840 :     - The type family for Ast is also defined and exported here.
2841 :     elabdata.cm:
2842 :     - The datatypes that describe input and output of the elaborator.
2843 :     This includes types, absyn, and static environments.
2844 :     elaborator.cm:
2845 :     - The SML/NJ type checker and elaborator.
2846 :     This maps an Ast.dec (with a given static environment) to
2847 :     an Absyn.dec (with a new static environment).
2848 :     - This libraries implements certain modules that used to be
2849 :     structures as functors (to remove dependencies on FLINT).
2850 :     execute.cm:
2851 :     - Everything having to do with executing binary code objects.
2852 :     - Dynamic environments.
2853 :     core.cm:
2854 :     - SML/NJ-specific instantiations of the elaborator and MLRISC.
2855 :     - Top-level modules.
2856 :     - FLINT (this should eventually become its own library)
2857 :    
2858 :     Notes:
2859 :    
2860 :     I am not 100% happy with the way I separated the elaborator (and its
2861 :     data structures) from FLINT. Two instances of the same problem:
2862 :    
2863 :     1. Data structures contain certain fields that carry FLINT-specific
2864 :     information. I hacked around this using exn and the property list
2865 :     module from smlnj-lib. But the fact that there are middle-end
2866 :     specific fields around at all is a bit annoying.
2867 :    
2868 :     2. The elaborator calculates certain FLINT-related information. I tried
2869 :     to make this as abstract as I could using functorization, but, again,
2870 :     the fact that the elaborator has to perform calculations on behalf
2871 :     of the middle-end at all is not nice.
2872 :    
2873 :     3. Having to used exn and property lists is unfortunate because it
2874 :     weakens type checking. The other alternative (parameterizing
2875 :     nearly *everything*) is not appealing, though.
2876 :    
2877 :     I removed the "rebinding =" warning hack because due to the new organization
2878 :     it was awkward to maintain it. As a result, the compiler now issues some of
2879 :     these warnings when compiling init.cmi during bootstrap compilation. On
2880 :     the plus side, you also get a warning when you do, for example:
2881 :     val op = = Int32.+
2882 :     which was not the case up to now.
2883 :    
2884 :     I placed "assign" and "deref" into the _Core structure so that the
2885 :     code that deals with the "lazy" keyword can find them there. This
2886 :     removes the need for having access to the primitive environment
2887 :     during elaboration.
2888 :    
2889 :     ----------------------------------------------------------------------
2890 :     Name: Matthias Blume
2891 : blume 898 Date: 2001/08/13
2892 :     Tag: blume-20010813-closures
2893 :     Description:
2894 :    
2895 :     This fix was sent to us by Zhong Shao. It is supposed to improve the
2896 :     performance of certain loops by avoiding needless closure allocation.
2897 :    
2898 :     ----------------------------------------------------------------------
2899 : george 890 Name: Lal George
2900 : george 896 Date: 2001/07/31 10:03:23 EDT 2001
2901 :     Tag: george-20010731-x86-fmalloc
2902 :     Description: Fixed bug in x86 calls
2903 :    
2904 :     There was a bug where call instructions would mysteriously
2905 :     vanish. The call instruction had to be one that returned
2906 :     a floating point value.
2907 :    
2908 :     ----------------------------------------------------------------------
2909 :     Name: Lal George
2910 : george 890 Date: 2001/07/19 16:36:29 EDT 2001
2911 :     Tag: george-20010719-simple-cells
2912 :     Description:
2913 :    
2914 :     I have dramatically simplified the interface for CELLS in MLRISC.
2915 :    
2916 :     In summary, the cells interface is broken up into three parts:
2917 :    
2918 :     1. CellsBasis : CELLS_BASIS
2919 :    
2920 :     CellsBasis is a top level structure and common for all
2921 :     architectures. it contains the definitions of basic datatypes
2922 :     and utility functions over these types.
2923 :    
2924 :     2. functor Cells() : CELLS
2925 :    
2926 :     Cells generates an interface for CELLS that incorporates the
2927 :     specific resources on the target architecture, such as the
2928 :     presence of special register classes, their number and size,
2929 :     and various useful substructures.
2930 :    
2931 :     3. <ARCH>CELLS
2932 :    
2933 :     e.g. SparcCells: SPARCCELLS
2934 :    
2935 :     <ARCH>CELLS usually contains additional bindings for special
2936 :     registers on the architecture, such as:
2937 :    
2938 :     val r0 : cell (* register zero *)
2939 :     val y : cell (* Y register *)
2940 :     val psr : cell (* processor status register *)
2941 :     ...
2942 :    
2943 :     The structure returned by applying the Cells functor is opened
2944 :     in this interface.
2945 :    
2946 :     The main implication of all this is that the datatypes for cells is
2947 :     split between CellsBasis and CELLS -- a fairly simple change for user
2948 :     code.
2949 :    
2950 :     In the old scheme the CELLS interface had a definitional binding of
2951 :     the form:
2952 :    
2953 :     signature CELLS = sig
2954 :    
2955 :     structure CellsBasis = CellsBasis
2956 :    
2957 :     ...
2958 :    
2959 :     end
2960 :    
2961 :     With all the sharing constraints that goes on in MLRISC, this old
2962 :     design quickly leads to errors such as:
2963 :    
2964 :     "structure definition spec inside of sharing ... "
2965 :    
2966 :    
2967 :     and appears to require an unacceptable amount of sharing and where
2968 :     constraint hackery.
2969 :    
2970 :     I think this error message (the interaction of definitional specs and
2971 :     sharing) requires more explanation on our web page.
2972 :    
2973 :     ----------------------------------------------------------------------
2974 : blume 826 Name: Matthias Blume
2975 : blume 879 Date: 2001/07/19 15:00:00 EDT
2976 :     Tag: blume-20010719-libreorg
2977 :     Description:
2978 :    
2979 :     This update puts together a fairly extensive but straightforward change
2980 :     to the way the libraries that implement the interactive system are
2981 :     organized:
2982 :    
2983 :     The biggest change is the elimination of structure Compiler. As a
2984 :     replacement for this structure, there is now a CM library
2985 :     (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
2986 :     that exports all the substructures of the original structure Compiler
2987 :     directly. So instead of saying Compiler.Foo.bar one now simply
2988 :     says Foo.bar. (The CM libraries actually export a collection of
2989 :     structures that is richer than the collection of substructures of
2990 :     structure Compiler.)
2991 :    
2992 :     To make the transition smooth, there is a separate library called
2993 :     $smlnj/compiler/compiler.cm which puts together and exports the
2994 :     original structure Compiler (or at least something very close to it).
2995 :    
2996 :     There are five members of the original structure Compiler
2997 :     that are not exported directly but which instead became members
2998 :     of a new structure Backend (described by signature BACKEND). These are:
2999 :     structure Profile (: PROFILE), structure Compile (: COMPILE), structure
3000 :     Interact (: INTERACT), structure Machine (: MACHINE), and val
3001 :     architecture (: string).
3002 :    
3003 :     Structure Compiler.Version has become structure CompilerVersion.
3004 :    
3005 :     Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
3006 :     by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
3007 :     or x86, respectively.
3008 :     Each of these exports the same frontend structures that
3009 :     $smlnj/compiler.cm exports. But they do not have a structure Backend
3010 :     and instead export some structure <Arch>Backend where <Arch> is Alpha32,
3011 :     Hppa, PPC, Sparc, or X86, respectively.
3012 :    
3013 :     Library $smlnj/compiler/all.cm exports the union of the exports of
3014 :     $smlnj/compiler/<arch>.cm
3015 :    
3016 :     There are no structures <Arch>Compiler anymore, use
3017 :     $smlnj/compiler/<arch>.cm instead.
3018 :    
3019 :     Library host-compiler-0.cm is gone. Instead, the internal library
3020 :     that instantiates CM is now called cm0.cm. Selection of the host
3021 :     compiler (backend) is no longer done here but. (Responsibility for it
3022 :     now lies with $smlnj/compiler/current.cm. This seems to be more
3023 :     logical.)
3024 :    
3025 :     Many individual files have been moved or renamed. Some files have
3026 :     been split into multiple files, and some "dead" files have been deleted.
3027 :    
3028 :     Aside from these changes to library organization, there are also changes
3029 :     to the way the code itself is organized:
3030 :    
3031 :     Structure Binfile has been re-implemented in such a way that it no
3032 :     longer needs any knowledge of the compiler. It exclusively deals
3033 :     with the details of binfile layout. It no longer invokes the
3034 :     compiler (for the purpose of creating new prospective binfile
3035 :     content), and it no longer has any knowledge of how to interpret
3036 :     pickles.
3037 :    
3038 :     Structure Compile (: COMPILE) has been stripped down to the bare
3039 :     essentials of compilation. It no longer deals with linking/execution.
3040 :     The interface has been cleaned up considerably.
3041 :    
3042 :     Utility routines for dealing with linking and execution have been
3043 :     moved into their own substructures.
3044 :    
3045 :     (The ultimate goal of these changes is to provide a light-weight
3046 :     binfile loader/linker (at least for, e.g., stable libraries) that
3047 :     does not require CM or the compiler to be present.)
3048 :    
3049 :     CM documentation has been updated to reflect the changes to library
3050 :     organization.
3051 :    
3052 :     ----------------------------------------------------------------------
3053 :     Name: Matthias Blume
3054 : blume 875 Date: 2001/07/10 17:30:00 EDT
3055 :     Tag: Release_110_34
3056 :     Description:
3057 :    
3058 : blume 876 Minor tweak to 110.34 (re-tagged):
3059 :    
3060 :     - README.html file added to CVS repository
3061 :     - runtime compiles properly under FreeBSD 3.X and 4.X
3062 :    
3063 :     ----------------------------------------------------------------------
3064 :     Name: Matthias Blume
3065 :     Date: 2001/07/10 17:30:00 EDT
3066 :     Tag: Release_110_34
3067 :     Description:
3068 :    
3069 : blume 875 New version number (110.34). New bootfiles.
3070 :    
3071 :     ----------------------------------------------------------------------
3072 :     Name: Matthias Blume
3073 : blume 874 Date: 2001/07/09 16:00:00 EDT
3074 :     Tag: blume-20010709-more-varargs
3075 :     Description:
3076 :    
3077 :     I changed the handling of varargs in ml-nlffigen again:
3078 :     The ellipsis ... will now simply be ignored (with an accompanying warning).
3079 :    
3080 :     The immediate effect is that you can actually call a varargs function
3081 :     from ML -- but you can't actually supply any arguments beyond the ones
3082 :     specified explicitly. (For example, you can call printf with its format
3083 :     string, but you cannot pass additional arguments.)
3084 :    
3085 :     This behavior is only marginally more useful than the one before, but
3086 :     it has the advantage that a function or, more importantly, a function
3087 :     type never gets dropped on the floor, thus avoiding follow-up problems with
3088 :     other types that refer to the offending one.
3089 :    
3090 :     ----------------------------------------------------------------------
3091 :     Name: Matthias Blume
3092 : blume 873 Date: 2001/07/09 11:25:00 EDT
3093 :     Tag: blume-20010709-varargs
3094 :     Description:
3095 :    
3096 :     1. ckit-lib.cm now exports structure Error
3097 :     2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
3098 :     with a warning accompanied by a source location. Moreover, it
3099 :     merely skips the offending function or type and proceeds with the
3100 :     rest of its work.u As a result, one can safely feed C code containing
3101 :     "..." to ml-nlffigen.
3102 :     3. There are some internal improvements to CM, providing slightly
3103 :     more general string substitutions in the tools subsystem.
3104 :    
3105 :     ----------------------------------------------------------------------
3106 :     Name: Matthias Blume
3107 : blume 854 Date: 2001/06/27 15:10:00 EDT
3108 :     Tag: blume-20010627-concur
3109 :     Description:
3110 :    
3111 :     Fixed a small bug in CM's handling of parallel compilation.
3112 :     (You could observe the bug by Control-C-interrupting an ordinary
3113 :     CMB.make or CM.stabilize and then attaching some compile servers.
3114 :     The result was that all of a sudden the previously interrupted
3115 :     compilation would continue on its own. This was because of
3116 :     an over-optimization: CM did not bother to clean out certain queues
3117 :     when no servers were attached "anyway", resulting in the contents
3118 :     of these queues to grab control when new servers did get attached.)
3119 :    
3120 :     There is also another minor update to the CM manual.
3121 :    
3122 :     ----------------------------------------------------------------------
3123 :     Name: Matthias Blume
3124 : blume 853 Date: 2001/06/26 16:15:00 EDT
3125 :     Tag: blume-20010626-cmdoc
3126 :     Description:
3127 :    
3128 :     Minor typo fixed in CM manual (syntax diagram for libraries).
3129 :    
3130 :     ----------------------------------------------------------------------
3131 :     Name: Matthias Blume
3132 : blume 852 Date: 2001/06/25 22:55:00 EDT
3133 :     Tag: blume-20010625-x86pc
3134 :     Description:
3135 :    
3136 :     Fixed a nasty bug in the X86 assembly code that caused signal
3137 :     handlers to fail (crash) randomly.
3138 :    
3139 :     ----------------------------------------------------------------------
3140 :     Name: Matthias Blume
3141 : blume 846 Date: 2001/06/25 12:05:00 EDT
3142 :     Tag: blume-20010625-nlffigen
3143 :     Description:
3144 :    
3145 :     This update fixes a number of minor bugs in ml-nlffigen as reported by
3146 :     Nick Carter <nbc@andrew.cmu.edu>.
3147 :    
3148 :     1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
3149 :     2. Default names for generated files are now derived from the name of
3150 :     the C file *without its directory*. In particular, this causes generated
3151 :     files to be placed locally even if the C file is in some system directory.
3152 :     3. Default names for generated signatures and structures are also derived
3153 :     from the C file name without its directory. This avoids silly things
3154 :     like "structure GL/GL".
3155 :     (Other silly names are still possible because ml-nlffigen does not do
3156 : blume 847 a thorough check of whether generated names are legal ML identifiers.
3157 :     When in doubt, use command line arguments to force particular names.)
3158 : blume 846
3159 :     ----------------------------------------------------------------------
3160 :     Name: Matthias Blume
3161 : blume 845 Date: 2001/06/21 12:25:00 EDT
3162 :     Tag: blume-20010621-eXene
3163 :     Description:
3164 :    
3165 :     eXene now compiles and (sort of) works again.
3166 :    
3167 :     The library name (for version > 110.33) is $/eXene.cm.
3168 :    
3169 :     I also added an new example in src/eXene/examples/nbody. See the
3170 :     README file there for details.
3171 :    
3172 :     ----------------------------------------------------------------------
3173 :     Name: Matthias Blume
3174 : blume 844 Date: 2001/06/20 16:40:00 EDT
3175 :     Tag: blume-20010620-cml
3176 :     Description:
3177 :    
3178 :     CML now compiles and works again.
3179 :    
3180 :     Libraries (for version > 110.33):
3181 :    
3182 :     $cml/cml.cm Main CML library.
3183 :     $cml/basis.cm CML's version of $/basis.cm.
3184 :     $cml/cml-internal.cm Internal helper library.
3185 :     $cml/core-cml.cm Internal helper library.
3186 :     $cml-lib/trace-cml.cm Tracing facility.
3187 :     $cml-lib/smlnj-lib.cm CML's version of $/smlnj-lib.cm
3188 :    
3189 :     The installer (config/install.sh) has been taught how to properly
3190 :     install this stuff.
3191 :    
3192 :     ----------------------------------------------------------------------
3193 :     Name: Matthias Blume
3194 : blume 843 Date: 2001/06/19 17:55:00 EDT
3195 :     Tag: blume-20010619-instantiate
3196 :     Description:
3197 :    
3198 :     This un-breaks the fix for bug 1432.
3199 :     (The bug was originally fixed in 110.9 but I broke it again some
3200 :     time after that.)
3201 :    
3202 :     ----------------------------------------------------------------------
3203 :     Name: Matthias Blume
3204 : blume 842 Date: 2001/06/19 17:25:00 EDT
3205 :     Tag: blume-20010619-signals
3206 :     Description:
3207 :    
3208 :     This should (hopefully) fix the long-standing signal handling bug.
3209 :     (The runtime system was constructing a continuation record with an
3210 :     incorrect descriptor which would cause the GC to drop data on the floor...)
3211 :    
3212 :     ----------------------------------------------------------------------
3213 :     Name: Matthias Blume
3214 : blume 840 Date: 2001/06/15 15:05:00 EDT
3215 : blume 841 Tag: blume-20010615-moresparc
3216 :     Description:
3217 :    
3218 :     Here is a short late-hour update related to Sparc c-calls:
3219 :    
3220 :     -- made handling of double-word arguments a bit smarter
3221 :    
3222 :     -- instruction selection phase tries to collapse certain clumsily
3223 :     constructed ML-Trees; typical example:
3224 :    
3225 :     ADD(ty,ADD(_,e,LI d1),LI d2) -> ADD(ty,e,LI(d1+d2))
3226 :    
3227 :     This currently has no further impact on SML/NJ since mlriscGen does
3228 :     not seem to generate such patterns in the first place, and c-calls
3229 :     (which did generate them in the beginning) has meanwhile been fixed
3230 :     so as to avoid them as well.
3231 :    
3232 :     ----------------------------------------------------------------------
3233 :     Name: Matthias Blume
3234 :     Date: 2001/06/15 15:05:00 EDT
3235 : blume 840 Tag: blume-20010615-sparc
3236 :     Description:
3237 :    
3238 :     The purpose of this update is to provide an implementation of NLFFI
3239 :     on Sparc machines.
3240 :    
3241 :     Here are the changes in detail:
3242 :    
3243 :     * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
3244 :     the Sparc implementation of the c-calls API.
3245 :     * The Sparc backend of SML/NJ has been modified to uniformely use %fp
3246 :     for accessing the ML frame. Thus, we have a real frame pointer and
3247 :     can freely modify %sp without need for an omit-frame-ptr phase.
3248 :     The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
3249 :     this case.
3250 :     * ml-nlffigen has been taught to produce code for different architectures
3251 :     and calling conventions.
3252 :     * In a way similar to what was done in the x86 case, the Sparc
3253 :     backend uses its own specific extension to mltree. (For example,
3254 :     it needs to be able to generate UNIMP instructions which are part
3255 :     of the calling convention.)
3256 :     * ml-nlffi-lib was reorganized to make it more modular (in particular,
3257 :     to make it easier to plug in new machine- and os-dependent parts).
3258 :    
3259 :     There are some other fairly unrelated bug fixes and cleanups as well:
3260 :    
3261 :     * I further hacked the .cm files for MLRISC tools (like MDLGen) so
3262 :     that they properly share their libraries with existing SML/NJ libraries.
3263 :     * I fixed a minor cosmetic bug in CM, supressing certain spurious
3264 :     follow-up error messages.
3265 :     * Updates to CM/CMB documentation.
3266 :    
3267 :     TODO items:
3268 :    
3269 :     * MLRISC should use a different register as its asmTemp on the Sparc.
3270 :     (The current %o2 is a really bad choice because it is part of the
3271 :     calling conventions, so things might interfere in unexpected ways.)
3272 :    
3273 :     ----------------------------------------------------------------------
3274 :     Name: Matthias Blume
3275 : blume 839 Date: 2001/06/07
3276 :     Tag: blume-20010607-calls
3277 :     Description:
3278 :    
3279 :     A number of internal changes related to C calls and calling conventions:
3280 :    
3281 :     1. ML-Tree CALL statements now carry a "pops" field. It indicates the
3282 :     number of bytes popped implicitly (by the callee). In most cases
3283 :     this field is 0 but on x86/win32 it is some non-zero value. This
3284 :     is information provided for the benefit of the "omit-frameptr" pass.
3285 :     2. The CALL instruction on the x86 carries a similar "pops" field.
3286 :     The instruction selection phase copies its value from the ML-Tree
3287 :     CALL statement.
3288 :     3. On all other architectures, the instruction selection phase checks
3289 :     whether "pops=0" and complains if not.
3290 :     4. The c-calls implementation for x86 now accepts two calling conventions:
3291 :     "ccall" and "stdcall". When "ccall" is selected, the caller cleans
3292 :     up after the call and pops is set to 0. For "stdcall", the caller
3293 :     does nothing, leaving the cleanup to the callee; pops is set to
3294 :     the number of bytes that were pushed onto the stack.
3295 :     5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
3296 :     distinguish between "ccall" and "stdcall".
3297 :     6. The UNIMP instruction has been added to the supported Sparc instruction
3298 :     set. (This is needed for implementing the official C calling convention
3299 :     on this architecture.)
3300 :     7. I fixed some of the .cm files under src/MLRISC/Tools to make them
3301 :     work with the latest CM.
3302 :    
3303 :     ----------------------------------------------------------------------
3304 :     Name: Matthias Blume
3305 : blume 838 Date: 2001/06/05 15:10:00 EDT
3306 :     Tag: blume-20010605-cm-index
3307 :     Description:
3308 :    
3309 :     0. The "lambdasplit" parameter for class "sml" in CM has been documented.
3310 :    
3311 :     1. CM can now generate "index files". These are human-readable files
3312 :     that list on a per-.cm-file basis each toplevel symbol defined or
3313 :     imported. The location of the index file for
3314 :     <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
3315 :     To enable index-file generation, set CM.Control.generate_index to true
3316 :     or export an environment-symbol: export CM_GENERATE_INDEX=true.
3317 :    
3318 :     The CM manual has been updated accordingly.
3319 :    
3320 :     2. I made some slight modifications to the c-calls API in MLRISC.
3321 :    
3322 :     a) There is now a callback to support saving/restoring of
3323 :     dedicated but caller-save registers around the actual call
3324 :     instruction.
3325 :     b) One can optionally specify a comment-annotation for the
3326 :     call instruction.
3327 :    
3328 :     3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
3329 :     (For example, the comment annotation shows the C prototype of
3330 :     the function being called.)
3331 :    
3332 :     ----------------------------------------------------------------------
3333 :     Name: Matthias Blume
3334 : blume 837 Date: 2001/06/01 13:30:00 EDT
3335 :     Tag: blume-20010601-nlffi-cleanup
3336 :     Description:
3337 :    
3338 :     This is mostly a cleanup of MLFFI stuff:
3339 :    
3340 :     - some signature files have been put into a more exposed place
3341 :     - the ugly 'f type parameter is gone (simplifies types tremendously!)
3342 :     - ml-nlffigen changed accordingly
3343 :     - tutorial updated
3344 :    
3345 :     Other changes:
3346 :    
3347 :     - author's affiliation in CM manual(s) updated
3348 :     - some more recognized keywords added to Allen's sml.sty
3349 :    
3350 :     ----------------------------------------------------------------------
3351 :     Name: Matthias Blume
3352 : blume 836 Date: 2001/05/25 15:30:00 EDT
3353 :     Tag: blume-20010525-iptr
3354 :     Description:
3355 :    
3356 :     - put the official 110.33-README (as it appears on the ftp server) under
3357 :     CVS
3358 :     - fixed a small bug related to incomplete pointer types in
3359 :     ml-nlffigen
3360 :     - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
3361 :     (it does not need the 'f type parameter)
3362 :    
3363 :     ----------------------------------------------------------------------
3364 :     Name: Matthias Blume
3365 : blume 834 Date: 2001/05/23 14:30:00 EDT
3366 :     Tag: Release_110_33
3367 :     Description:
3368 :    
3369 :     New version number (110.33). New bootfiles.
3370 :    
3371 :     ----------------------------------------------------------------------
3372 :     Name: Matthias Blume
3373 : blume 833 Date: 2001/05/22 18:06:00 EDT
3374 :     Tag: blume-20010522-targets
3375 :     Description:
3376 :    
3377 :     Made install.sh use file config/targets.customized if it exists, falling
3378 :     back to config/targets if it doesn't. This way one can have a customized
3379 :     version of the targets file without touching the "real thing", thus
3380 :     eliminating the constant fear of accidentally checking something bogus
3381 :     back into the CVS repository... (File config/targets.customized must
3382 :     not be added to the repository!)
3383 :    
3384 :     ----------------------------------------------------------------------
3385 :     Name: Matthias Blume
3386 : blume 832 Date: 2001/05/22 16:30:00 EDT
3387 :     Tag: blume-20010522-minitut
3388 :     Description:
3389 :    
3390 :     1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
3391 :     struct returns.
3392 :     2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt. This is some very
3393 :     incomplete, preliminary documentation for NLFFI.
3394 :    
3395 :     ----------------------------------------------------------------------
3396 :     Name: Matthias Blume
3397 : blume 829 Date: 2001/05/14 11:30:00 EDT
3398 :     Tag: blume-20010514-script
3399 :     Description:
3400 :    
3401 :     Some bugs in install script fixed.
3402 :    
3403 :     In addition to that I also made a slight change to the NLFFI API:
3404 :     Functors generated by ml-nlffigen now take the dynamic library as a
3405 :     straight functor argument, not as a suspended one. (The original
3406 :     functor code used to force the suspension right away anyway, so there
3407 :     was nothing gained by this complication of the interface.)
3408 :    
3409 :     ----------------------------------------------------------------------
3410 :     Name: Matthias Blume
3411 : blume 828 Date: 2001/05/11 14:35:00 EDT
3412 :     Tag: blume-20010511-ml-nlffi
3413 :     Description:
3414 :    
3415 :     I finally took the plunge and added my new FFI code to the main
3416 :     repository. For x86-linux it is now ready for prime-time.
3417 :    
3418 :     There are two new subdirectories of "src":
3419 :    
3420 :     - ml-nlffi-lib:
3421 :     The utility library for programs using the FFI interface.
3422 :     Here is the implementation of $/c.cm and its associated low-level
3423 :     partners $/c-int.cm and $/memory.cm.
3424 :     - ml-nlffigen:
3425 :     A stand-alone program for generating ML glue code from C source
3426 :     code.
3427 :    
3428 :     Building ml-nlffigen requires $/ckit-lib.cm.
3429 :    
3430 :     The config/install.sh script has been updates to do the Right Thing
3431 :     (hopefully).
3432 :    
3433 :     Notice that the source tree for the C-Kit will not be put under "src"
3434 :     but directly under the installation root directory. (This is the
3435 :     structure that currently exists on the CVS server when you check out
3436 :     module "sml".) Fortunately, config/install.sh knows about this oddity.
3437 :    
3438 :     Bugs: No documentation yet.
3439 :    
3440 :     ----------------------------------------------------------------------
3441 :     Name: Matthias Blume
3442 : blume 826 Date: 2001/05/09 16:35:00 EDT
3443 :     Tag: blume-20010509-cpscontract
3444 :     Description:
3445 :    
3446 :     Fixed a bug in the accounting code in cpsopt/contract.sml. (The
3447 :     wrapper/unwrapper elimination did not decrement usage counts and some
3448 :     dead variables got overlooked by the dead-up logic.)
3449 :    
3450 :     ----------------------------------------------------------------------
3451 : george 824 Name: Lal George
3452 :     Date: 2001/05/08 17:26:09 EDT
3453 :     Tag: george-20010508-omit-frameptr
3454 :     Description:
3455 :    
3456 :     Changes to implement the omit-frame-pointer optimization to support
3457 :     raw C calls. For now, there is only support on the Intel x86, but
3458 :     other architectures will follow as more experience is gained with this.
3459 :    
3460 :    
3461 :     ----------------------------------------------------------------------
3462 : blume 818 Name: Matthias Blume
3463 : blume 819 Date: 2001/05/07 14:40:00 EDT
3464 :     Tag: blume-20010507-proxies
3465 :     Description:
3466 :    
3467 :     I made into "proxy libraries" all libraries that qualify for such a
3468 :     change. (A qualifying library is a library that has another library or
3469 :     groups as its sole member and repeats that member's export list
3470 :     verbatim. A proxy library avoids this repetition by omitting its export
3471 :     list, effectively inheriting the list that its (only) member exports.
3472 :     See the CM manual for more explanation.)
3473 :     The main effect is that explicit export lists for these libraries
3474 :     do not have to be kepts in sync, making maintenance a bit easier.
3475 :    
3476 :     I also added copyright notices to many .cm-files.
3477 :    
3478 :     Last but not least, I made a new set of bootfiles.
3479 :    
3480 :     ----------------------------------------------------------------------
3481 :     Name: Matthias Blume
3482 : blume 818 Date: 2001/05/04 17:00:00 EDT
3483 :     Tag: blume-20010504-cm-lsplit
3484 :     Description:
3485 :    
3486 :     0. John merged pending changes to $/smlnj-lib.cm
3487 :    
3488 :     1. Allen's previous change accidentally backed out of one of Lal's
3489 :     earlier changes. I undid this mistake (re-introducing Lal's change).
3490 :    
3491 :     2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
3492 :     within the compiler where applicable. There is some code simplification
3493 :     because of that.
3494 :    
3495 :     3. The "split" phase (in FLINT) is now part of the default list of phases.
3496 :     Compiler.Control.LambdaSplitting.* can be used to globally control the
3497 :     lambda-splitting (cross-module-inlining) engine. In addition to that,
3498 :     it can now also be controlled on a per-source basis: CM has been taught
3499 :     a new tool parameter applicable to ML source files.
3500 :    
3501 :     - To turn lambda-splitting off completely:
3502 :     local open Compiler.Control.LambdaSplitting in
3503 :     val _ = set Off
3504 :     end
3505 :     - To make "no lambda-splitting" the global default (but allow per-source
3506 :     overriding); this is the initial setting:
3507 :     local open Compiler.Control.LambdaSplitting in
3508 :     val _ = set (Default NONE)
3509 :     end
3510 :     - To make "lambda-splitting with aggressiveness a" the global default
3511 :     (and allow per-source overriding):
3512 :     local open Compiler.Control.LambdaSplitting in
3513 :     val _ = set (Default (SOME a))
3514 :     end
3515 :    
3516 :     - To turn lambda-splitting off for a given ML souce file (say: a.sml)
3517 :     write (in the respective .cm-file):
3518 :     a.sml (lambdasplitting:off)
3519 :     - To turn lambda-splitting for a.sml on with minimal aggressiveness:
3520 :     a.sml (lambdasplitting:on)
3521 :     - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
3522 :     <a> is a decimal non-negative integer):
3523 :     a.sml (lambdasplitting:<a>)
3524 :     - To turn lambda-splitting for a.sml on with maximal aggressiveness:
3525 :     a.sml (lambdasplitting:infinity)
3526 :     - To use the global default for a.sml:
3527 :     a.sml (lambdasplitting:default)
3528 :     or simply
3529 :     a.sml
3530 :    
3531 :     ----------------------------------------------------------------------
3532 : leunga 815 Name: Allen Leung
3533 :     Date: 2001/05/04 01:57:00 EDT
3534 :     Tag: leunga-20010504-sync
3535 :     Description:
3536 : leunga 788
3537 : leunga 815 MLRISC features.
3538 :    
3539 :     1. Fix to CMPXCHG instructions.
3540 :     2. Changed RA interface to allow annotations in callbacks.
3541 :     3. Added a new method to the stream interface to allow annotations updates.
3542 :    
3543 : george 749 ----------------------------------------------------------------------
3544 : blume 812 Name: Matthias Blume
3545 : blume 814 Date: 2001/05/01 11:45:00 EDT
3546 :     Tag: blume-20010501-pcedittmp
3547 :     Description:
3548 :    
3549 :     Changed install.sh to use the current working directory instead of
3550 :     /usr/tmp for a temporary file (pcedittmp). The previous choice
3551 :     of /usr/tmp caused trouble with MacOS X because of file premission
3552 :     problems.
3553 :    
3554 :     ----------------------------------------------------------------------
3555 :     Name: Matthias Blume
3556 : blume 812 Date: 2001/04/20 11:10:00 EDT
3557 :     Tag: blume-20010420-inMLflag
3558 :     Description:
3559 :    
3560 :     - added vp_limitPtrMask to vproc-state.h
3561 :     (for use by the raw-C-calls mechanism to implement proper interrupt
3562 :     handling)
3563 :     - made the ML compiler aware of various data-structure offsets so it
3564 :     can generate code for accessing the vp_inML flag and vp_limitPtrMask
3565 :     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
3566 :     raw C-calls
3567 :    
3568 :     ----------------------------------------------------------------------
3569 : george 810 Name: Lal George
3570 :     Date: 2001/04/20 09:15:28 EDT
3571 :     Tag: george-20010420-macosX
3572 :     Description:
3573 :    
3574 :     - Changes to port to Mac OS X; Darwin.
3575 :    
3576 :     - In the process I found that sqrt was broken on the PPC, because the
3577 :     fsqrt instruction is not implemented.
3578 :    
3579 :     ----------------------------------------------------------------------
3580 : blume 799 Name: Matthias Blume
3581 : blume 808 Date: 2001/04/18 12:45:00 EDT
3582 :     Tag: blume-20010418-ccalls
3583 :     Description:
3584 :    
3585 :     - fixed two off-by-4 errors in the x86-specific c-calls implementation
3586 :     (this bug prevented structure arguments containing pointers from being
3587 :     passed correctly)
3588 :     - changed the raw-C-call code in mlriscGen.sml in such a way that
3589 :     structure arguments are represented as a pointer to the beginning
3590 :     of the structure (instead of having a series of synthesized arguments,
3591 :     one for each structure member)
3592 :    
3593 :     - made makeml script's verbosity level configurable via environment
3594 :     variable (MAKEML_VERBOSITY)
3595 :    
3596 : blume 813 - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
3597 : blume 808 in rawmem-x86.sml; we are now using the real thing
3598 :    
3599 :     ----------------------------------------------------------------------
3600 :     Name: Matthias Blume
3601 : blume 806 Date: 2001/03/22 16:25:00 EST
3602 :     Tag: blume-20010322-bootfiles
3603 :     Description:
3604 :    
3605 :     Created a new set of bootfiles (for your automatic installation convenience).
3606 :    
3607 :     ----------------------------------------------------------------------
3608 :     Name: Matthias Blume
3609 : blume 805 Date: 2001/03/22 15:10:00 EST
3610 :     Tag: blume-20010322-rawmem-parcm
3611 :     Description:
3612 :    
3613 :     1. All "raw memory access" primitives for the new FFI are implemented now
3614 :     (at least on the x86).
3615 :     2. Some further cleanup of CM's parallel make mechanism.
3616 :    
3617 :     ----------------------------------------------------------------------
3618 :     Name: Matthias Blume
3619 : blume 801 Date: 2001/03/19 17:53:00 EST
3620 :     Tag: blume-20010319-parallel
3621 :     Description:
3622 :    
3623 :     Parallel make (using compile servers) now works again.
3624 :    
3625 :     To this end, CM.stabilize and CMB.make have been modified to work in
3626 :     two passes when compile servers are attached:
3627 :     1. Compile everything, do not perform stabilization; this pass
3628 :     uses compile servers
3629 :     2. Stabilize everything; this pass does not use compile servers
3630 :     If there are no compile servers, the two passes are combined into one
3631 :     (as before). Splitting the passes increases the inherent parallelism
3632 :     in the dependency graph because the entire graph including all
3633 :     libraries is available at the same time. This, in turn, improves
3634 :     server utilization. The downside is that the master process will
3635 :     have to do some extra work after compilation is done (because for
3636 :     technical reasons it must re-read all the binfiles during stabilization).
3637 :    
3638 :     ----------------------------------------------------------------------
3639 :     Name: Matthias Blume
3640 : blume 800 Date: 2001/03/16 12:22:00 EST
3641 :     Tag: blume-20010316-bootfiles
3642 :     Description:
3643 :    
3644 :     Created a new set of bootfiles (for your automatic installation convenience).
3645 :    
3646 :     ----------------------------------------------------------------------
3647 :     Name: Matthias Blume
3648 : blume 799 Date: 2001/03/16 11:00:00 EST
3649 :     Tag: blume-20010316-MLTREE-fixup
3650 :     Description:
3651 :    
3652 :     This is a minor fixup for an (untagged) earlier commit by Allen.
3653 :     (A file was missing).
3654 :    
3655 :     ----------------------------------------------------------------------
3656 : leunga 788 Name: Allen Leung
3657 : leunga 796 Date: Mon Mar 5 18:54:57 EST 2001
3658 :     Tag: leunga-20010305-cut-support
3659 :    
3660 :     1. New support for alternative control-flow in MLTREE.
3661 :     Currently we support
3662 :    
3663 :     FLOW_TO(CALL ...., [k1,...,kn])
3664 :    
3665 :     This is needed for 'cuts to' in C-- and try/handle-like constructs
3666 :     in Moby
3667 :    
3668 :     New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
3669 :    
3670 :     2. Register Allocator
3671 :    
3672 :     Changes in interface [from Fermin, John]
3673 :    
3674 :     3. Alpha 8-bit SLL support [Fermin]
3675 :    
3676 :     4. All architectures
3677 :    
3678 :     A new module (ClusterExpandCopies) for expanding parallel copies.
3679 :    
3680 :     ----------------------------------------------------------------------
3681 :     Name: Allen Leung
3682 : leunga 788 Date: 2001/02/27 23:07:00 EST
3683 :     Tag: leunga-20010227-minor-stuff
3684 :    
3685 :     1. Alpha bug fix for CMOVNE
3686 :     2. Handle mltree COND(..,FCMP ...,...)
3687 :     3. Bug fix in simplifier
3688 :    
3689 :     ----------------------------------------------------------------------
3690 : blume 777 Name: Matthias Blume
3691 : blume 781 Date: 2001/01/30 17:50:00 EST
3692 :     Tag: blume-20010130-sync
3693 :     Description:
3694 :    
3695 :     This is just a minor update to sync my devel branch with the main brach.
3696 :     The only visible change is the addition of some README files.
3697 :    
3698 :     ----------------------------------------------------------------------
3699 :     Name: Matthias Blume
3700 : blume 778 Date: 2001/01/12 23:30:00 JST
3701 :     Tag: blume-20010112-bootfiles
3702 :     Description:
3703 :    
3704 :     Made a new set of bootfiles that goes with the current state of the
3705 :     repository.
3706 :    
3707 :     ----------------------------------------------------------------------
3708 :     Name: Matthias Blume
3709 : blume 777 Date: 2001/01/12 21:20:00 JST
3710 :     Tag: blume-20010112-sync
3711 :     Description:
3712 :    
3713 :     I am just flushing out some minor changes that had accumulated in
3714 :     my private branch in order to sync with the main tree. (This is
3715 :     mainly because I had CVS trouble when trying to merge _into_ my
3716 :     private branch.)
3717 :    
3718 :     Most people should be completely unaffected by this.
3719 :    
3720 :     ----------------------------------------------------------------------
3721 : leunga 775 Name: Allen Leung
3722 :     Date: Thu Jan 11 21:03:00 EST 2001
3723 :     Tag: leunga-20010111-labexp=mltree
3724 :     Description:
3725 :    
3726 :     1. Removed the type LabelExp and replace it by MLTree.
3727 :     2. Rewritten mltree-simplify with the pattern matcher tool.
3728 :     3. There were some bugs in alpha code generator which would break
3729 :     64-bit code generation.
3730 :     4. Redo the tools to generate code with the
3731 :     5. The CM files in MLRISC (and in src/system/smlnj/MLRISC)
3732 :     are now generated by perl scripts.
3733 :    
3734 :     ----------------------------------------------------------------------
3735 : blume 771 Name: Matthias Blume
3736 : blume 774 Date: 2001/01/10 21:55:00 JST
3737 :     Tag: blume-20010110-rcc
3738 :     Description:
3739 :    
3740 :     The RCC stuff now seems to work (but only on the x86).
3741 :     This required hacking of the c-calls interface (and -implementation) in
3742 :     MLRISC.
3743 :    
3744 :     Normal compiler users should be unaffected.
3745 :    
3746 :     ----------------------------------------------------------------------
3747 :     Name: Matthias Blume
3748 : blume 773 Date: 2001/01/09 01:20:00 JST
3749 :     Tag: blume-20010109-rcc
3750 :     Description:
3751 :    
3752 :     This is a fairly big patch, flushing out a large number of pending
3753 :     changes that I made to my development copy over the last couple of days.
3754 :    
3755 :     Of practical relevance at this moment is a workaround for a pickling
3756 :     bug that Allen ran into the other day. The cause of the bug itself is
3757 :     still unknown and it might be hard to fix it properly, but the
3758 :     workaround has some merits of its own (namely somewhat reducing pickling
3759 :     overhead for certain libraries). Therefore, I think this solution should
3760 :     be satisfactory at this time.
3761 :    
3762 :     The rest of the changes (i.e., the vast majority) has to do with my
3763 :     ongoing efforts of providing direct support for C function calls from
3764 :     ML. At the moment there is a new primop "RAW_CCALL", typing magic
3765 :     in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
3766 :     case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
3767 :     translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
3768 :     handle RCC.
3769 :    
3770 :     The last part (the changes to mlriscGen.sml) are still known to be
3771 :     wrong on the x86 and not implemented on all other architectures. But
3772 :     the infrastructure is in place. I had to change a few functor
3773 :     signatures in the backend to be able to route the CCalls interface
3774 :     from MLRISC there, and I had to specialize the mltree type (on the
3775 :     x86) to include the necessary extensions. (The extensions themselves
3776 :     were already there and redy to go in MLRISC/x86).
3777 :    
3778 :     Everything should be very happy as soon as someone helps me with
3779 :     mlriscGen.sml...
3780 :    
3781 :     In any case, nothing of this should matter to anyone as long as the
3782 :     new primop is not being used (which is going to be the case unless you
3783 :     find it where I hid it :). The rest of the compiler is completely
3784 :     unaffected.
3785 :    
3786 :     ----------------------------------------------------------------------
3787 :     Name: Matthias Blume
3788 : blume 772 Date: 2001/01/05 00:30:00 JST
3789 :     Tag: blume-20010105-primops
3790 :     Description:
3791 :    
3792 :     Added some experimental support for work that I am doing right now.
3793 :     These changes mostly concern added primops, but there is also a new
3794 :     experimental C library in the runtime system (but currently not enabled
3795 :     anywhere except on Linux/X86).
3796 :    
3797 :     In the course of adding primops (and playing with them), I discovered that
3798 :     Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
3799 :     broken. (Zhong was very right he labeled this stuff as "major gross hack".)
3800 :     To recover, I made type information in INL_PRIM mandatory and changed
3801 :     prim.sml as well as built-in.sml accordingly. The InLine structure now
3802 :     has complete, correct type information (i.e., no bottom types).
3803 :    
3804 :     Since all these changes mean that we need new binfiles, I also bumped the
3805 :     version number to 110.32.1.
3806 :    
3807 :     ----------------------------------------------------------------------
3808 :     Name: Matthias Blume
3809 : blume 771 Date: 2000/12/30 22:10:00 JST
3810 :     Tag: blume-20001230-various
3811 :     Description:
3812 :    
3813 :     Added proxy libraries for MLRISC and let MLRISC libraries refer
3814 :     to each other using path anchors. (See CM manual for explanation.)
3815 :    
3816 :     Updated CM documentation.
3817 :    
3818 :     Fixed some bugs in CM.
3819 :    
3820 :     Implemented "proxy" libraries (= syntactic sugar for CM).
3821 :    
3822 :     Added "-quiet" option to makeml and changed runtime system accordingly.
3823 :    
3824 :     Added cleanup handler for exportML to reset timers and compiler stats.
3825 :    
3826 :     ----------------------------------------------------------------------
3827 : george 761 Name: Lal George
3828 :     Date: 2000/12/22 22:22:58 EST 2000
3829 :     Tag: Release_110_32
3830 :     Description:
3831 :    
3832 :     Infinite precision used throughout MLRISC.
3833 :     see MLRISC/mltree/machine-int.sig
3834 :    
3835 :     ----------------------------------------------------------------------
3836 : blume 756 Name: Matthias Blume
3837 : blume 760 Date: 2000/12/22 23:16:00 JST
3838 :     Tag: blume-20001222-warn
3839 :     Description:
3840 : blume 759
3841 : blume 760 Corrected wording and formatting of some CM warning message which I
3842 :     broke in my previous patch.
3843 : blume 759
3844 :     ----------------------------------------------------------------------
3845 :     Name: Matthias Blume
3846 : blume 763 Date: 2000/12/22 21:20:00 JST
3847 :     Tag: blume-20001222-anchorenv
3848 :     Description:
3849 :    
3850 :     Fixed CM's handling of anchor environments in connection with CMB.make.
3851 :    
3852 :     ----------------------------------------------------------------------
3853 :     Name: Matthias Blume
3854 : blume 758 Date: 2000/12/22 13:15:00 JST
3855 :     Tag: blume-20001222-cleanup
3856 :     Description:
3857 :    
3858 :     Removed src/cm/ffi which does not (and did not) belong here.
3859 :    
3860 :     ----------------------------------------------------------------------
3861 :     Name: Matthias Blume
3862 : blume 757 Date: 2000/12/21 23:55:00 JST
3863 :     Tag: blume-20001221-exn
3864 :     Description:
3865 :    
3866 :     Probably most important: CM no longer silently swallows all exceptions
3867 :     in the compiler.
3868 :     Plus: some other minor CM changes. For example, CM now reports some
3869 :     sizes for generated binfiles (code, data, envpickle, lambdapickle).
3870 :    
3871 :     ----------------------------------------------------------------------
3872 :     Name: Matthias Blume
3873 : blume 756 Date: 2000/12/15 00:01:05 JST
3874 :     Tag: blume-20001215-dirtool
3875 :     Description:
3876 :    
3877 :     - "dir" tool added.
3878 :     - improvements and cleanup to Tools structure
3879 :     - documentation updates
3880 :    
3881 :     ----------------------------------------------------------------------
3882 : leunga 752 Name: Allen Leung
3883 : leunga 755 Date: Thu Dec 14 03:45:24 EST 2000
3884 :     Description:
3885 :     Tag: leunga-20001214-int-inf
3886 :     Description:
3887 :    
3888 :     In IntInf, added these standard functions, which are missing from our
3889 :     implementation:
3890 :    
3891 :     andb : int * int -> int
3892 :     xorb : int * int -> int
3893 :     orb : int * int -> int
3894 :     notb : int -> int
3895 :     << : int * word -> int
3896 :     ~>> : int * word -> int
3897 :    
3898 :     Not tested, I hope they are correct.
3899 :    
3900 :     ----------------------------------------------------------------------
3901 :     Name: Allen Leung
3902 : leunga 752 Date: Fri Dec 8 19:23:26 EST 2000
3903 :     Description:
3904 :     Tag: leunga-20001208-nowhere
3905 :     Description:
3906 :    
3907 :     Slight improvements to the 'nowhere' tool to handle OR-patterns,
3908 :     to generate better error messages etc. Plus a brief manual.
3909 :    
3910 :     ----------------------------------------------------------------------
3911 : george 749 Name: Lal George
3912 :     Date: 2000/12/08 09:54:02 EST 2000
3913 :     Tag: Release_110_31
3914 :     Description:
3915 : george 705
3916 : george 749 - Version 110.31
3917 : leunga 591 ----------------------------------------------------------------------
3918 : leunga 744 Name: Allen Leung
3919 :     Date: Thu Dec 7 22:01:04 EST 2000
3920 :     Tag: leunga-20001207-cell-monster-hack
3921 :     Description:
3922 :    
3923 :     Major MLRISC internal changes. Affect all clients.
3924 :     Summary:
3925 :    
3926 :     1. Type CELLS.cell = int is now replaced by a datatype.
3927 :     As a result, the old regmap is now gone. Almost all interfaces
3928 :     in MLRISC change as a consequence.
3929 :    
3930 :     2. A new brand version of machine description tool (v3.0) that generates
3931 :     modules expecting the new interface. The old version is removed.
3932 :    
3933 :     3. The RA interface has been further abstracted into two new functors.
3934 :     RISC_RA and X86RA. These functors have much simpler interfaces.
3935 :     [See also directory MLRISC/demo.]
3936 :    
3937 :     4. Some other new source->source code generation tools are available:
3938 :    
3939 :     a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
3940 :     b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
3941 :     I use this tool to generate the peephole optimizers---with the new
3942 :     cell type changes, peephole rules are becoming difficult to write
3943 :     without conditional pattern matching.
3944 :    
3945 :     5. More Intmap -> IntHashTable change. Previous changes by Matthias didn't
3946 :     cover the entire MLRISC source tree so many things broke.
3947 :    
3948 :     6. CM files have been moved to the subdirectory MLRISC/cm.
3949 :     They are moved because there are a lot of them and they clutter up the
3950 :     root dir.
3951 :    
3952 :     7. More detailed documentation to come...
3953 :    
3954 :     NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
3955 :     a makeml -rebuild first. This is because of other other
3956 :     changes that Matthias has made (see below).
3957 :    
3958 :    
3959 :     ----------------------------------------------------------------------
3960 : blume 733 Name: Matthias Blume
3961 : blume 742 Date: 2000/11/30 23:12:00 JST
3962 :     Tag: blume-20001130-filereorg
3963 :     Description:
3964 :    
3965 :     Some manual updates and some file reorganizations in CM.
3966 :    
3967 :     ----------------------------------------------------------------------
3968 :     Name: Matthias Blume
3969 : blume 737 Date: 2000/11/24 17:45:00 JST
3970 :     Tag: blume-20001124-link
3971 :     Description:
3972 :    
3973 :     Drastically improved link traversal code for the case that the dynamic
3974 :     value was already loaded at bootstrap time. As a result, CM and CMB
3975 :     now both load blazingly fast -- even on a very slow machine. Also,
3976 :     memory consumption has been further reduced by this.
3977 :    
3978 :     Warning: The format of the PIDMAP file has changed. THerefore, to
3979 :     bootstrap you have to do this:
3980 :    
3981 :     1. Run CMB.make
3982 :     2. Make a symbolic link for the boot directory:
3983 :     ln -s sml.boot.ARCH-OS xxx
3984 :     3. "Rebuild" the boot directory:
3985 :     ./makeml -boot xxx -rebuild sml ; rm xxx
3986 :     4. Boot normally:
3987 :     ./makeml
3988 :    
3989 :     ----------------------------------------------------------------------
3990 :     Name: Matthias Blume
3991 : blume 735 Date: 2000/11/21 21:20:00 JST
3992 :     Tag: blume-20001121-tools
3993 :     Description:
3994 :    
3995 :     Continued hacking on autoloading problem -- with success this time.
3996 :     Also changed tool-plugin mechanism. See new CM manual.
3997 :    
3998 :     ----------------------------------------------------------------------
3999 :     Name: Matthias Blume
4000 : blume 734 Date: 2000/11/19 14:30:00 JST
4001 :     Tag: blume-20001119-autoload
4002 :     Description:
4003 :    
4004 :     Some hacking to make autoloading faster. Success for CMB, no success
4005 :     so far for CM. There is a reduced structure CM' that autoloads faster.
4006 :     (This is a temporary, non-documented hack to be eliminated again when
4007 :     the general problem is solved.)
4008 :    
4009 :     ----------------------------------------------------------------------
4010 :     Name: Matthias Blume
4011 : blume 733 Date: 2000/11/17 14:10:00 JST
4012 :     Tag: blume-20001117-pickle-lib
4013 :     Description:
4014 :    
4015 :     1. Eliminated comp-lib.cm
4016 :     2. Made pickle-lib.cm
4017 :     3. Eliminated all uses of intset.sml (from comp-lib.cm)
4018 :     4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
4019 :     equivalent constructs from smlnj-lib.cm (INtHashTable).
4020 :     5. Point 4. also goes for those uses of intmap.* in MLRISC.
4021 :     Duplicated intmap modules thrown out.
4022 :     6. Hunted down all duplicated SCC code and replaced it with
4023 :     equivalent stuff (GraphSCCFn from smlnj-lib.cm).
4024 :     7. Rewrote Feedback module.
4025 :     8. Moved sortedlist.sml into viscomp-lib.cm. Eventually it
4026 :     should be thrown out and equivalent modules from smlnj-lib.cm
4027 :     should be used (IntRedBlackSet, IntListSet, ...).
4028 :    
4029 :     Confirmed that compiler compiles to fixpoint.
4030 :    
4031 :     ----------------------------------------------------------------------
4032 : leunga 731 Name: Allen Leung
4033 :     Date: 2000/11/10 18:00:00
4034 :     Tag: leunga-20001110-new-x86-fp
4035 :    
4036 :     A new x86 floating point code generator has been added.
4037 :     By default this is turned off. To turn this on, do:
4038 :    
4039 :     CM.autoload "$smlnj/compiler.cm";
4040 :     Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
4041 :    
4042 :     Changes:
4043 :    
4044 :     1. Changed FTAN to FPTAN so that the assembly output is correct.
4045 :     2. Changed the extension callback for FTANGENT to generate:
4046 :    
4047 :     fptan
4048 :     fstp %st(0)
4049 :     instead of
4050 :     fptan
4051 :     fstpl ftempmem
4052 :    
4053 :     3. Numerous assembly fixes for x86.
4054 :    
4055 :     5. Cleaned up the machine code output module x86/x86MC.sml and added
4056 :     support for a whole bunch of instructions and addressing modes:
4057 :    
4058 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st, %st(n)
4059 :     faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp %st, %st(n)
4060 :     fadd/fsub/fsubr/fmul/fdiv/fdivr %st(n), %st
4061 :     fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
4062 :     fxch %st(n)
4063 :     fld %st(n)
4064 :     fst %st(n)
4065 :     fst mem
4066 :     fstp %st(n)
4067 :     fucom %st(n)
4068 :     fucomp %st(n)
4069 :    
4070 :     All these are now generated when the fast fp mode is turned on.
4071 :    
4072 :     6. Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
4073 :    
4074 :     ----------------------------------------------------------------------
4075 : blume 729 Name: Matthias Blume
4076 :     Date: 2000/11/09 11:20:00 JST
4077 :     Tag: blume-20001109-scc
4078 :     Description:
4079 :    
4080 :     Eliminated some code duplication:
4081 :    
4082 :     1. Added "where" clause to GraphSCCFn in SML/NJ Library.
4083 :     (Otherwise the functor is useless.)
4084 :     2. Used GraphSCCFn where SCCUtilFun was used previously.
4085 :     3. Got rid of SCCUtilFun (in comp-lib.cm).
4086 :    
4087 :     ----------------------------------------------------------------------
4088 : george 721 Name: Lal George
4089 :     Date: 2000/11/06 09:02:21 EST 2000
4090 :     Tag: Release_110_30
4091 :     Description:
4092 :    
4093 :     - Version 110.30
4094 :     ----------------------------------------------------------------------
4095 : blume 715 Name: Matthias Blume
4096 : blume 716 Date: 2000/11/04 14:45:00
4097 :     Tag: blume-20001104-mlbuild
4098 :     Description:
4099 :    
4100 :     - Made ml-build faster on startup.
4101 :     - Documentation fixes.
4102 :    
4103 :     ----------------------------------------------------------------------
4104 :     Name: Matthias Blume
4105 : blume 715 Date: 2000/11/02 17:00:00 JST
4106 :     Tag: blume-20001102-condcomp
4107 :     Description:
4108 :    
4109 :     - Small tweaks to pickler -- new BOOTFILES!
4110 :     - Version bumped to 110.29.2.
4111 :     - Added conditional compilation facility to init.cmi (see comment there).
4112 :     ----------------------------------------------------------------------
4113 : leunga 713 Name: Allen Leung
4114 :     Date: 2000/10/23 19:31:00
4115 :     Tag: leunga-20001023-demo-ra
4116 :    
4117 :     1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
4118 :     2. Test programs for the graph library updated
4119 :     3. Some new MLRISC demo programs added
4120 :    
4121 :     ----------------------------------------------------------------------
4122 : blume 710 Name: Matthias Blume
4123 : blume 711 Date: 2000/08/31 22:15:00 JST
4124 :     Tag: blume-20001017-errmsg
4125 :     Description:
4126 :    
4127 :     More error message grief: Where there used to be no messages, there
4128 :     now were some that had bogus error regions. Fixed.
4129 :    
4130 :     ----------------------------------------------------------------------
4131 :     Name: Matthias Blume
4132 : blume 710 Date: 2000/08/31 17:30:00 JST
4133 :     Tag: blume-20001017-v110p29p1
4134 :     Description:
4135 :    
4136 :     I made a version 110.29.1 with new bootfiles.
4137 :    
4138 :     Changes: Modified pickler/unpickler for faster and leaner unpickling.
4139 :     CM documentation changes and a small bugfix in CM's error reporting.
4140 :    
4141 :     ----------------------------------------------------------------------
4142 : george 705 Name: Lal George
4143 :     Date: 2000/09/27 14:42:35 EDT
4144 :     Tag: george-20000927-nodestatus
4145 :     Description:
4146 :    
4147 :     Changed the type of the nodestatus, so that:
4148 :    
4149 :     SPILLED(~1) is now SPILLED
4150 :     SPILLED(m) where m>=0 is now MEMREG(m)
4151 :     SPILLED(s) where s<~1 is now SPILL_LOC(~s)
4152 :    
4153 :     ----------------------------------------------------------------------
4154 : blume 697 Name: Matthias Blume
4155 : blume 703 Date: 2000/09/07 14:45:00 JST
4156 :     Tag: blume-20000907-cmerrmsg
4157 :     Description:
4158 :    
4159 :     Small tweak to CM to avoid getting ML syntax error messages twice.
4160 :    
4161 :     ----------------------------------------------------------------------
4162 :     Name: Matthias Blume
4163 : blume 701 Date: 2000/08/31 18:00:00 JST
4164 :     Tag: blume-20000831-cvsbootfiles
4165 :     Description:
4166 :    
4167 :     New URL for boot files (because the 110.29 files on the BL server do
4168 :     now work correctly with my updated install scripts for yacc and lex).
4169 :    
4170 :     ----------------------------------------------------------------------
4171 :     Name: Matthias Blume
4172 : blume 697 Date: 2000/08/08 12:33:00 JST
4173 :     Tag: blume-20000808-manual
4174 :     Description:
4175 :    
4176 :     Tiny update to CM manual.
4177 :    
4178 :     ----------------------------------------------------------------------
4179 : leunga 695 Name: Allen Leung
4180 :     Date: 2000/08/7 19:31:00
4181 :     Tag: leunga-20000807-a-whole-bunch-of-stuff
4182 :    
4183 :     Moby, C--, SSA, x86, machine descriptions etc. Should only affect C--
4184 :     and Mobdy.
4185 :    
4186 :     1. x86
4187 :    
4188 :     a. Fixes to peephole module by John and Dan.
4189 :     b. Assembly fix to SETcc by Allen.
4190 :     c. Fix to c-call by John.
4191 :     d. Fix to spilling by John. (This one deals with the missing FSTPT case)
4192 :     e. Instruction selection optimization to SETcc as suggested by John.
4193 :    
4194 :     For example,
4195 :    
4196 :     MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
4197 :    
4198 :     should generate:
4199 :    
4200 :     MOVL a, x
4201 :     SUBL b, x
4202 :     SHRL 31, x
4203 :    
4204 :     2. IR stuff
4205 :    
4206 :     A bunch of new DJ-graph related algorithms added. These
4207 :     speed up SSA construction.
4208 :    
4209 :     3. SSA + Scheduling
4210 :    
4211 :     Added code for SSA and scheduling to the repository
4212 :    
4213 :     ----------------------------------------------------------------------
4214 : george 694 Name: Lal George
4215 :     Date: 2000/07/27 11:53:14 EDT
4216 :    
4217 :     Tag: lal-20000727-linux-ppc
4218 :     Description:
4219 :    
4220 :     Made changes to support Linux PPC.
4221 :     p.s. I have confirmation that the 110.29 boot files work fine.
4222 :    
4223 :     ----------------------------------------------------------------------
4224 : blume 684 Name: Matthias Blume
4225 : blume 692 Date: 2000/07/27 17:40:00 JST
4226 :     Tag: blume-20000727-scripts
4227 :     Description:
4228 :    
4229 :     !!!! WARNING !!!!
4230 :     You must recompile the runtime system!
4231 :     !!!! WARNING !!!!
4232 :    
4233 :     This is basically another round of script-enhancements:
4234 :    
4235 :     1. sml, ml-build, and ml-makedepend accept options -D and -U to define
4236 :     and undefine CM preprocessor symbols.
4237 :    
4238 :     2. ml-build avoids generating a new heap image if it finds that the
4239 :     existing one is still ok. (The condition is that no ML file had to
4240 :     be recompiled and all ML files are found to be older that the heap
4241 :     file.)
4242 :    
4243 :     To make this work smoothly, I also hacked the runtime system as
4244 :     well as SMLofNJ.SysInfo to get access to the heap image suffix
4245 :     (.sparc-solaris, ...) that is currently being used.
4246 :    
4247 :     Moreover, the signature of CM.mk_standalone has changed. See the
4248 :     CM manual.
4249 :    
4250 :     3. ml-makedepend accepts additional options -n, -a, and -o. (See the
4251 :     CM manual for details.)
4252 :    
4253 :     4. More CM manual updates:
4254 :     - all of the above has been documented.
4255 :     - there is now a section describing the (CM-related) command line
4256 :     arguments that are accepted by the "sml" command
4257 :    
4258 :     ----------------------------------------------------------------------
4259 :     Name: Matthias Blume
4260 : blume 691 Date: 2000/07/25 16:20:00 JST
4261 :     Tag: blume-20000725-makedepend
4262 :     Description:
4263 :    
4264 :     Added a script called ml-makedepend. This can be used in makefiles
4265 :     for Unix' make in a way very similar to the "makedepend" command for
4266 :     C.
4267 :    
4268 :     The script internally uses function CM.sources.
4269 :    
4270 :     Synopsis:
4271 :    
4272 :     ml-makedepend [-f makefile] cmfile targetname
4273 :    
4274 :     The default for the makefile is "makefile" (or "Makefile" should
4275 :     "makefile" not exist).
4276 :    
4277 :     ml-makedepend adds a cmfile/targetname-specific section to this
4278 :     makefile (after removing the previous version of this section). The
4279 :     section contains a single dependency specification with targetname on
4280 :     the LHS (targetname is an arbitrary name), and a list of files derived
4281 :     from the cmfile on the RHS. Some of the files on the RHS are
4282 :     ARCH/OPSYS-specific. Therefore, ml-makedepend inserts references to
4283 :     "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
4284 :     path names. The makefile writer is responsible for making sure that
4285 :     these variables have correct at the time "make" is invoked.
4286 :    
4287 :     ----------------------------------------------------------------------
4288 :     Name: Matthias Blume
4289 : blume 690 Date: 2000/07/22 23:30:00 JST
4290 :     Tag: blume-20000722-urlupdate
4291 :     Description:
4292 :    
4293 :     Changed BOOT and config/srcarchiveurl to point to BL server:
4294 :    
4295 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
4296 :    
4297 :     ----------------------------------------------------------------------
4298 :     Name: Matthias Blume
4299 : blume 686 Date: 2000/07/18 18:00:00 JST
4300 :     Tag: blume-20000718-Version_110_29
4301 :     Description:
4302 :